• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Curr Biol. Author manuscript; available in PMC Dec 6, 2012.
Published in final edited form as:
PMCID: PMC3237752

The P4-ATPase TAT-5 inhibits the budding of extracellular vesicles in C. elegans embryos



Cells release extracellular vesicles (ECVs) that can influence differentiation, modulate the immune response, promote coagulation, and induce metastasis. Many ECVs form by budding outwards from the plasma membrane, but the molecules that regulate budding are unknown. In ECVs, the outer leaflet of the membrane bilayer contains aminophospholipids that are normally sequestered to the inner leaflet of the plasma membrane, suggesting a role for lipid asymmetry in ECV budding.


We show that loss of the conserved P4-ATPase TAT-5 causes the large-scale shedding of ECVs and disrupts cell adhesion and morphogenesis in C. elegans embryos. TAT-5 localizes to the plasma membrane and its loss results in phosphatidylethanolamine exposure on cell surfaces. We show that RAB-11 and the ESCRT complex, which regulate the topologically analogous process of viral budding, are enriched at the plasma membrane in tat-5 embryos and are required for ECV production.


TAT-5 is the first protein identified to regulate ECV budding. TAT-5 provides a potential molecular link between loss of phosphatidylethanolamine asymmetry and the dynamic budding of vesicles from the plasma membrane, supporting the hypothesis that lipid asymmetry regulates budding. Our results also suggest that viral budding and ECV budding may share common molecular mechanisms.


Extracellular vesicles (ECV) can act as messengers between cells. For example, ECVs can influence developmental fates by delivering morphogens such as Hedgehog [1]. ECVs released by platelets induce coagulation and can activate an immune response [2, 3]. Tumor cells release ECVs that promote metastasis and infectious viral particles co-opt ECV biogenesis machinery to spread. ECVs form through two major routes. One pathway is via the direct budding of plasma membrane to form ECVs known as microparticles, microvesicles, or ectosomes. Alternatively, ECVs called exosomes form when multivesicular bodies fuse with the plasma membrane, releasing their intralumenal vesicles outside of the cell. While the mechanisms responsible for exosome production are better understood, the molecular mechanisms that lead to ectosome budding remain a mystery.

Because the outer surfaces of ectosomes expose atypically high levels of the aminophospholipids phosphatidylserine (PS) and phosphatidylethanolamine (PE), it has been proposed that lipid externalization could facilitate ectosome formation on the plasma membrane [4]. PS and PE are normally asymmetrically enriched in the inner leaflet of the membrane bilayer. These lipids have conical shapes that can bend membranes, and PS is an anionic lipid that contributes to the negative charge of the membrane [5]. Interestingly, activated platelets from Scott syndrome patients cannot externalize PS or PE and fail to release ECVs to induce coagulation [6], supporting a potential link between loss of lipid asymmetry and ECV budding.

Lipid asymmetry is mediated by transmembrane flippases from the P4 family of ATPases, which internalize phospholipids by translocating them from the outer to the inner leaflet of the bilayer [7]. PS translocase activity was definitively proven for yeast Drs2p and human ATP8B2 [8, 9]. Mutations in each of the five S. cerevisiae P4-ATPases cause defects in vesicle biogenesis or transport [10]. Humans have fourteen P4-ATPases, many of which are uncharacterized, but mutations in the PS flippase ATP8B1 cause intrahepatic cholestasis [7]. Hepatocytes progressively lose microvilli and ATP8B1 is implicated in maintaining membrane integrity [11]. Despite the potential connection between disrupted lipid asymmetry and ectosome production, no P4-ATPase has been shown to inhibit ectosome formation.

A link between lipid flipping and cellular phenotype has also been established in C. elegans, which contains six P4-ATPases (TAT-1 to TAT-6) [12]. The Drs2p homolog TAT-1 is required to prevent PS exposure and thereby regulates cell corpse engulfment [13, 14]. TAT-1 has also been implicated in lysosome biogenesis and endocytosis [15, 16]. TAT-2, TAT-3, and TAT-4 regulate sterol or branched chain fatty acid metabolism through mechanisms that have not yet been established [12, 17]. The Neo1p homolog TAT-5 is the sole essential P4-ATPase in C. elegans, but its role has not yet been determined.

Here, we identify TAT-5 for its requirement during morphogenesis in the C. elegans embryo. We show that TAT-5 localizes to the plasma membrane and is needed to prevent PE exposure. Loss of TAT-5 causes the large-scale budding of vesicles from the plasma membrane, compromising cell adhesion and leading to defects in cell shape. We show that the robust production of ECVs depends on RAB-11, a GTPase associated with recycling endosomes; and the ESCRT complex proteins, which form multivesicular bodies. These proteins are enriched at the plasma membrane in tat-5 embryos and are known to regulate the budding of viruses from the plasma membrane [18, 19]. Our findings identify the first P4-ATPase that inhibits ECV budding, providing a link between PE asymmetry, regulators of viral budding, and ECV biogenesis.


TAT-5 is needed for robust cell adhesion and dynamic changes in cell shape

We identified TAT-5 in an RNAi screen for essential genes important for morphogenesis in C. elegans embryos. Morphogenesis begins at gastrulation when the endodermal precursors flatten their apical surfaces and ingress into the center of the embryo (Figure 1A; Movie S1) [20]. In tat-5(RNAi) embryos, the endodermal precursors remained rounded and failed to ingress (Figure 1B; Movie S1), even though they expressed endoderm-specific markers (Figure 1C,D) and showed characteristic cell cycle lengths. Therefore, ingression defects in tat-5(RNAi) embryos did not result from a defect in specification.

Figure 1
TAT-5 is required for cell shape changes and gastrulation

To confirm the specificity of tat-5 RNAi, we examined embryos in two uncharacterized deletions that remove portions of tat-5 (tm1741 and tm1772). TAT-5 contains an ATPase domain and ten transmembrane domains. Both deletions remove part of the ATPase domain and the third transmembrane domain, and are predicted to create frameshifts that eliminate the remainder of the protein. Over 95% of tat-5 mutants were sterile (n=179), but rare worms produced a few embryos with rounded cells that invariably died (Figure S1B) like tat-5(RNAi) embryos. Because tat-5(RNAi) and tat-5 mutant embryos have indistinguishable phenotypes, hereafter we use RNAi for phenotypic analysis of tat-5. Mutations or RNAi that target the other P4-ATPases did not cause lethality [12] or disrupt cell shape (data not shown). Thus, TAT-5, but not other P4-ATPases, is required for the cell shape changes and movements that occur during gastrulation.

Because similar defects are observed in embryos lacking the adhesion proteins HMR-1/E-cadherin and SAX-7/L1CAM [21], we investigated whether tat-5 embryos have defects in cell adhesion. In wild-type 4-cell embryos, the ventral cell EMS extends an adhesive protrusion along the anterior cell ABa (Figure 1E; Movie S1). By contrast, EMS remained rounded and did not protrude in tat-5 embryos (Figure 1F). When we removed the confining eggshell, it was evident that all tat-5 cells were rounded in comparison to wild-type (Figure 1G,H; Movie S2). There was a significant increase in cell circularity in tat-5 embryos, and the length of cell-cell contacts was reduced (see Supplemental Experimental Procedures). Cells in tat-5 embryos were also easily dissociated during devitellinization (14/30 embryos), in contrast to wild-type (1/23 embryos). However, loss of TAT-5 does not disrupt morphogenesis by preventing the localization of HMR-1 and SAX-7 to cell contacts (Figure S1C–F). Thus, cell adhesion is weakened in tat-5 mutant embryos through a novel mechanism, and poor adhesion may explain the observed defects in cell shape and morphogenesis.

Extracellular tubulovesicular structures accumulate in tat-5 embryos

We next asked whether tat-5 embryos had defects in the organization of the plasma membrane that could disrupt cell adhesion. Using a reporter (PHPLC1[partial differential]1::mCherry) that localizes to the plasma membrane, we observed patches of increased labeling in tat-5 embryos (Figure 2B) compared to wild-type embryos (Figure 2A). Similarly increased membrane labeling was observed in embryos from tat-5 deletion mutants (Figure S2D), and also with transmembrane proteins such as HMR-1 and SAX-7 (Figure S1D,F). This phenotype suggests that tat-5 embryos accumulate extra membrane on the surfaces of cells, which could disrupt cell contacts.

Figure 2
Extracellular vesicles are present on cell surfaces and between cell-cell contacts in tat-5 embryos

To reveal the nature of the extra membrane in tat-5 embryos, we performed an ultrastructural analysis using transmission electron microscopy (TEM). Cell-cell contacts in control embryos were evident as a pair of 5 nm thick plasma membranes that were rarely separated (Figure 2C). In tat-5 embryos, the plasma membranes were separated by a gap up to 300 nm in thickness, which contained numerous extracellular vesicles or tubules (Figure 2D). Hereafter, we refer to these heterogeneous structures as ECVs. ECVs were also present on the surface of one-cell tat-5 embryos (Figure 2F), but were rare in wild-type embryos (Figure 2E). We conclude that the extra labeling with PHPLC1[partial differential]1::mCherry in tat-5 embryos corresponds to the ECVs we observe by TEM. We hypothesize that these membrane structures interrupt the contacts between cells and are responsible for the weakened cell adhesion we observe in tat-5 embryos.

To demonstrate that the ECVs were vesicles not villi, we used electron tomography to generate 3D reconstructions of the cell surface. In tat-5 embryos, we found spherical vesicles, cup-shaped vesicles, and large tubulovesicular structures (Figure 3A and Movie S3). Most ECVs within a 200 nm section did not connect to the plasma membrane and therefore are unlikely to be villi or intercellular bridges. To confirm that most ECVs were outside the cell, we asked whether proteins on the ECV membrane or their contents were protected from degradation. When proteins tagged with the PIE-1 ZF1 domain are expressed in the early embryo, they are rapidly degraded in somatic cells [22]. We tagged the integral plasma membrane protein SYX-4 with the ZF1 domain and GFP to assess its ability to be degraded in tat-5 embryos. In wild-type embryos, GFP::ZF1::SYX-4 initially localized to the plasma membrane, but was rapidly endocytosed and degraded as somatic cells formed (Figure 3C). In tat-5 embryos, GFP::ZF1::SYX-4 was internalized and degraded over much of the surface of somatic cells, but remained within thickened patches that co-stained with SAX-7 (Figure 3D). We observed a similar pattern of protection when ZF1-tagging the cortical proteins PAR-3 and PAR-6 in tat-5 embryos (data not shown). These findings suggest that ECVs protect ZF1-tagged proteins from degradation because they are not connected to the cell, supporting the tomography data.

Figure 3
Extracellular vesicles are generated by budding from the plasma membrane

Extracellular vesicles in tat-5 embryos form by plasma membrane budding

We next analyzed whether the ECVs in tat-5 embryos were exosomes or ectosomes. Exosomes are 40–100 nm diameter vesicles derived from the intralumenal vesicles of multivesicular bodies (MVBs) [3]. To determine if ECVs in tat-5 embryos are exosomes, we compared the diameter of ECVs and intralumenal MVB vesicles. The ECVs were significantly larger than intralumenal vesicles in tat-5 embryos (p<10−40), with little overlap (Figure 3B). Rare ECVs in wild-type embryos were also larger than wild-type intralumenal vesicles (ECV: 218 ± 62 nm, n=20 from 5 embryos; MVB: 57 ± 16 nm, n=106 from 3 embryos, p<10−9). Additionally, we never observed MVBs fusing with the plasma membrane in tomograms from tat-5 embryos (over 40 μm of cell contact examined). The simple shapes of exosomes [3] are also inconsistent with the complicated tubulovesicular structures we observed in tat-5 ECVs (Figure 3A). Finally, tat-5 ECVs did not label with LMP-1/LAMP1 (Figure S3), a marker of exosomes in other systems [2]. Thus, we can find no evidence that the ECVs derive from MVBs.

We next tested whether ECVs formed through the direct budding of the plasma membrane as ectosomes. Ectosomes are larger than exosomes (100–1000 nm), consistent with the observed size range of ECVs, and tend to have more irregular shapes [3]. We also observed 36 vesicles in tat-5 embryos (in 40 μm of cell contact) that were connected to the plasma membrane (Figure 3A'). We never observed a vesicle attached to the plasma membrane in tomograms from wild-type embryos (in 24 μm examined) and cannot conclude whether the ECVs from wild-type and tat-5 embryos are equivalent. The labeling of ECVs with plasma membrane markers such as PHPLC1[partial differential]1::mCherry, GFP::ZF1::SYX-4, HMR-1, and SAX-7 is also consistent with a cell surface origin (Figure 2B, ,3D,3D, S1D,F). Together, our observations suggest that most or all ECVs in tat-5 embryos are ectosomes that form by plasma membrane budding.

ECVs are formed throughout embryogenesis

To determine whether ECVs are made continuously, we used a reporter that would mark new but not pre-existing ECVs. The med-1::GFP::CAAX transgene is expressed in a subset of cells after the 7-cell stage and labels the cytoplasmic face of the membrane. In control embryos, GFP::CAAX labeled the plasma membrane and thin filopodia-like protrusions (Figure 3E). In tat-5 embryos, GFP::CAAX accumulated in large patches on the surface of cells (Figure 3F). These observations suggest that early embryonic cells produce ECVs continuously.

TAT-5 localizes to the plasma membrane

To determine if TAT-5 functions where ectosomes are produced, we expressed a GFP-tagged transgene. GFP::TAT-5 is functional and can rescue the sterility of tat-5 deletion alleles (>60% fertile, n=61 adults, 2 transgenes). In early embryos, GFP::TAT-5 was found primarily at the plasma membrane, but also labeled some cytoplasmic vesicles (Figure 4A). GFP::TAT-5 also rescued the extra membrane observed in tat-5 embryos (Figure 4C). The localization of GFP::TAT-5 to the plasma membrane is consistent with TAT-5 acting on the cell surface.

Figure 4
TAT-5 localizes to the plasma membrane and maintains PE asymmetry

The ATPase activity of P4-ATPases is required for their lipid flippase action [8]. To determine whether ATPase activity is required for TAT-5 to prevent plasma membrane budding, we mutated two sites essential for activity of P-type ATPases: the aspartyl phosphorylation site (D439) or the motif required for dephosphorylation of D439 [23]. GFP::TAT-5(D439E) did not localize to the plasma membrane (Figure S4A–D), while GFP::TAT-5(E246Q) localized to the plasma membrane (Figure 4B) but failed to rescue the sterility and lethality of tat-5 mutants (n=120 adults, 2 transgenes). GFP::TAT-5(E246Q) localized to membrane thickenings in embryos from tat-5 mutants (Figure 4D), demonstrating that TAT-5 ATPase activity is required to prevent ECV formation.

TAT-5 is required to prevent cell surface exposure of PE

Ectosomes have been associated with high levels of externalized PS and PE [4]. As a P4-ATPase, TAT-5 is predicted to regulate the bilayer asymmetry of one or more aminophospholipids. To establish whether PS asymmetry was altered in tat-5 embryos, we stained the surface of unpermeabilized embryos with Annexin V, a probe that specifically recognizes PS [24]. In control embryos, Annexin V labeled the plasma membrane and endocytosed puncta weakly (Figure 4E,H). In tat-1 mutants, embryonic cells stained brightly with Annexin V (Figure 4F,H), consistent with reports that PS was exposed on the surface of tat-1 germ cells [13]. In tat-5 embryos, we observed weak labeling similar to control embryos, even at regions where ectosomes accumulate (Figure 4G,H). We conclude that TAT-5 is not required to prevent cell surface exposure of PS (in contrast to TAT-1), and we find no evidence that ectosomes in tat-5 mutants externalized PS.

To analyze PE asymmetry we used duramycin, which specifically binds PE [25]. In cultured cells, PE is externalized during cytokinetic furrow ingression [26]. Consistent with these observations, duramycin stained a subset of cell contacts in unpermeabilized wild-type embryos (Figure 4I), and staining was similar in tat-1 embryos (Figure 4J). However, in tat-5 embryos, duramycin staining was more uniform around the plasma membrane and brightly labeled membrane thickenings (Figure 4K). The intensity of membrane staining, both at regions with and without patches of ectosomes, was significantly higher in tat-5 embryos than in control embryos (Figure 4L, p<0.01). Increased surface PE was not due to a generally higher level of PE in tat-5 mutants, since levels of PE and other phospholipids were equivalent in wild-type and tat-5 (Figure S4E–F). These findings indicate that PE is inappropriately externalized on tat-5 embryonic cells.

Clathrin, RAB-11, and the ESCRT complex are enriched at the plasma membrane in tat-5 embryos

Given the unusual vesicles and mislocalized lipids in tat-5 embryos, we examined whether endocytosis and secretion were disrupted. We found no evidence for a defect in secretion using transgenes expressing secreted proteins (Figure S5A–D), and markers for clathrin-dependent endocytosis were normal in tat-5 (Figure S5C–F). We also found no defect in the localization of early endosomal, late endosomal, Golgi, or ER markers in tat-5 embryos (Figure S5G–N), although lysosomes appeared enlarged (Figure S3A–B, S5O–P). Thus, loss of TAT-5 does not grossly disrupt membrane trafficking.

TAT-5 ortholog Neo1p has been linked to clathrin recruitment in yeast [27], so we examined clathrin localization using a tagged heavy chain subunit (CHC-1) [28]. In control embryos, GFP::CHC-1 was rapidly lost from the plasma membrane after fertilization, and was mostly cytoplasmic by the 4-cell stage (Figure 5A). In 4-cell tat-5 embryos, GFP::CHC-1 was still enriched at the plasma membrane (Figure 5B). However, given that clathrin-mediated endocytosis was not defective in tat-5 embryos (Figure S5E–F), the enrichment of clathrin at the plasma membrane suggests that clathrin recruitment or dissociation is aberrant in tat-5 embryos.

Figure 5
Clathrin, RAB-11, and the ESCRT complex are recruited to the plasma membrane in tat-5 embryos

We found a similar plasma membrane accumulation of proteins linked to viral shedding, which is provocative because viral particles can also form by plasma membrane budding. The small GTPase Rab11 is a regulator of recycling endosomes that is required for viral budding [18]. In control embryos, GFP::RAB-11 had a punctate distribution in the cytoplasm and on the cortex (Figure 5C). In tat-5 embryos, GFP::RAB-11 became continuous along the plasma membrane and had the same thickened appearance as other membrane markers (Figure 5D). The recycling endosome protein RME-1 showed a similar distribution (data not shown). Thus, recycling endosome markers are over-recruited to the plasma membrane or trapped in ECVs.

The ESCRT complex, which forms the intralumenal vesicles of MVBs, is also required for viral shedding [19]. Since ESCRT proteins do not stably associate with membranes unless the ESCRT recycling protein VPS-4 is inhibited [29], we examined the localization of early and late ESCRT proteins in vps-4(RNAi) embryos. Proteins from the ESCRT-I (MVB-12) and ESCRT-III (VPS-32) complexes stably labeled MVBs as well as a small number of puncta on the plasma membrane of vps-4 embryos (Figure 5E & 5G). In embryos from tat-5 + vps-4 dsRNA injections, we observed twice as many ESCRT-III puncta on the plasma membrane and ESCRT-I proteins were similarly enriched (Figure 5F & 5H). ESCRT-III proteins do not label the plasma membrane in tat-5 embryos when VPS-4 is present, suggesting that ESCRT-III is not trapped within ectosomes (data not shown). These findings demonstrate that the ESCRT complex is recruited to the plasma membrane in tat-5 embryos.

RAB-11 and the ESCRT complex are required for ectosome formation

Although externalized PE is predicted to promote outward curvature of the plasma membrane, PE exposure is unlikely to form complete vesicles without the assistance of proteins [5]. Given that viral budding regulators RAB-11 and the ESCRT complex were enriched at the plasma membrane in tat-5 embryos, we asked whether reducing their levels could suppress the thickened membrane phenotype (Figure 6A,B & Table S1). In embryos from tat-5 dsRNA injected adults, we found that 82% had severe membrane thickening (Figure 6D) compared to wild type (Figure 6C). However, when we co-injected rab-11 and tat-5 dsRNA, embryos more closely resembled rab-11 single injections with only 11% of embryos showing the severe thickening typical of tat-5 (Figure 6E). We verified that our double injections effectively targeted tat-5 by injecting the GFP::TAT-5 strain (Figure S6). In addition, RNAi targeting clathrin adaptors AP1 at the Golgi (aps-1) or AP2 at the plasma membrane (aps-2) did not suppress tat-5 membrane thickening (Figure 6A), demonstrating that clathrin is unlikely to have a role in budding ectosomes and that suppression by RAB-11 is specific. These findings suggest that ectosomes arise from a RAB-11-dependent membrane pool or through a RAB-11-dependent process (see Discussion).

Figure 6
RAB-11 and the ESCRT complex mediate extracellular vesicle production

Next, we examined a role for the ESCRT complex in ectosome formation. The ESCRT complex is broken into four subcomplexes: ESCRT-0 clusters cargo and recruits ESCRT-I; ESCRT-I and ESCRT-II form membrane buds; and ESCRT-III causes vesicle scission and release [19]. We found that dsRNA targeting ESCRT-0 or ESCRT-I complex members were able to partially suppress tat-5 membrane thickening (Figure 6B; Table S1). For example, only 33% of embryos from tat-5 + hgrs-1 (ESCRT-0) injections and 40% of embryos from tat-5 + vps-28 (ESCRT-I) injections showed severe membrane thickening (Figure 6F,G). Targeting ESCRT-II, ESCRT-III (Figure 6H), or associated factors VPS-4 and ALX-1 did not significantly suppress membrane thickenings, although these proteins could be involved in ectosome release rather than bud formation (see Discussion). Thus, these suppression and localization data suggest that ESCRT complex proteins are recruited to the plasma membrane in tat-5 embryos and are required to bud vesicles out of the cell.


Impact of cell surface organization on morphogenesis

The lethality of tat-5 embryos highlights the importance of proper surface membrane structure for cell shape changes and morphogenesis. We propose that the accumulation of intercellular ectosomes in tat-5 embryos compromises morphogenesis by disrupting adhesion. For example, anterior spreading of the EMS cell and endodermal cell ingression do not occur in tat-5 embryos, and both of these events involve the formation of adhesive cellular protrusions [30, 31]. Although similar morphogenetic defects can arise from improper cell specification, specification occurs normally in tat-5 early embryos.

Plasma membrane aminophospholipids and ectosome formation

PS and PE are externalized on ectosomes, and it was proposed that their externalization induces ectosome formation [4]. Our system allows us to test the potential contributions of PS and PE separately because of the specific effects tat-1 and tat-5 mutants have on lipid localization. PS asymmetry was not disrupted on the plasma membrane or on ectosomes in tat-5 embryos. Additionally, despite the dramatic externalization of PS in tat-1 mutants, they do not accumulate excess surface membrane (our unpublished data). Therefore, although PS externalization has important signaling roles [32], PS externalization is neither necessary nor sufficient for ectosome formation. Although TAT-5 could regulate additional lipids, the physical characteristics of PE (described below) support the hypothesis that the loss of PE asymmetry induces ectosome budding.

A model for ectosome formation

TAT-5 is the first protein shown to repress both PE externalization and ectosome production, and provides a potential link between these events. PE externalization occurs slowly unless induced by lipid scramblases, which abolish lipid asymmetry [32]. We predict that TAT-5 opposes the action of a scramblase to maintain PE asymmetry and inhibit ectosome formation (Figure 7). Patients with Scott syndrome, whose platelets fail to release ECVs [6], were found to carry mutations in a potential scramblase [33], further correlating loss of lipid asymmetry and ECV production. This raises the possibility that human P4-ATPases (such as TAT-5 homologs ATP9A/B) may regulate ECV formation and coagulation in platelets.

Figure 7
Two models for the role of TAT-5 and PE externalization in ectosome production

PE is an abundant, zwitterionic lipid with a small headgroup and conical shape [5], and we can envision both indirect and direct roles for PE in budding (Figure 7A–B). First, PE externalization could alter the local charge of the plasma membrane by increasing the density of sparser lipids in the inner leaflet, such as anionic PS and phosphatidylinositols (PI), resulting in the formation of anionic microdomains. Anionic microdomains could recruit or stabilize membrane-sculpting proteins that promote ectosome formation (Figure 7A). For example, mammalian Hrs (ESCRT-0) binds anionic lipids [19]. We observed that HGRS-1/Hrs was required for efficient ectosome production, and ESCRT proteins were enriched at the cell surface in tat-5 embryos. Clathrin, which was similarly enriched at the cell surface in tat-5 embryos, can also bind to anionic lipids or be recruited by Hrs [19, 34].

A second potential mechanism by which externalized PE could lead to ectosome budding is through the formation of PE-rich microdomains within the outer leaflet. Because of PE's conical shape [5], microdomains of PE could act as hinges to induce negative membrane curvature (Figure 7B). Indeed, PE-rich microdomains that contain membrane-localized ESCRT proteins have been implicated in HIV budding [35].

RAB-11 and the ESCRT complex in ectosome production

Our suppression studies indicate that RAB-11 has an important role in ectosome production. Both recycling endosome markers we examined showed plasma membrane accumulation in tat-5 embryos, raising the possibility that RAB-11-dependent recycling of lipids or proteins to the plasma membrane may be necessary for ectosome production. One possibility is that RAB-11 delivers the ESCRT complex to the plasma membrane. Alternatively, RAB-11 may have a direct role in promoting ectosome budding, as has been suggested for viral budding [18]. RAB-11 also promotes exosome production [3], suggesting that ectosomes and exosomes may be regulated by common pathways despite their distinct subcellular origins.

The ESCRT proteins are well known for their role in forming viral particles, but only recently have they been implicated in ectosome budding [36]. HIV-1 budding requires ESCRT-I, ESCRT-III, ALIX, and VPS4. The viral Gag protein recruits ESCRT-I in place of ESCRT-0 [19]. We found that inhibiting ESCRT-0 or ESCRT-I suppressed ECV formation in tat-5 embryos, but surprisingly, inhibiting ESCRT-III, ALX-1, or VPS-4 did not. However, we believe that ESCRT-III is likely to take part based on its recruitment to the plasma membrane in tat-5 embryos. Given that the ESCRT-III complex is required for scission, lack of suppression might reflect our inability to distinguish stalled buds from released ectosomes by light microscopy. By contrast, reducing ESCRT-0/I should prevent bud formation early, and therefore reduce buildup of excess membrane. Notably, inhibiting ESCRT-0/I proteins only partially suppressed ectosome production in tat-5 embryos. Although this may reflect partial knockdown, another possibility is that ESCRT proteins regulate the production of a subset of ectosomes. Indeed, a VPS4-independent mechanism for viral budding has recently been identified, and this budding is dependent on a Rab11-binding protein [37].

The cellular role of TAT-5 homologs

The role of the TAT-5 subfamily of P4-ATPases has not been examined in other multicellular organisms, but studies in yeast suggest that some aspects of TAT-5 function may be conserved. For example, although yeast Neo1p localizes predominantly to endomembranes and Golgi, neo1 mutants have ultrastructural phenotypes such as the formation of unusual tubular and multivesicular structures [38, 39]. Since the intralumenal vesicles of MVBs share the same topology as ectosomes, this phenotype could reflect an analogous role for Neo1p in inhibiting the outward budding of endosomal membranes. It will be important to determine if yeast and mammalian TAT-5 homologs have a similar role in maintaining PE asymmetry, and to test whether they can influence the release of ectosomes or the budding of viral particles from cells.

Experimental Procedures

Worms were grown at 25°C unless stated otherwise. Timelapse, DIC, and fluorescent images were acquired and processed as described [40, 41]. Extended Experimental Procedures are available in the Supplemental Information.

Supplementary Material




We thank the Skirball Image Core Facility and the New York Structural Biology Center for their assistance with TEM, especially KD Derr and Eric Roth. We thank Julie Ahringer, Jon Audhya, Andy Fire, Michael Glotzer, Renaud Legouis, Morris Maduro, Shohei Mitani, Karen Oegema, James Priess, Geraldine Seydoux, Phil Thorpe, Simon Tuck, John White, the Caenorhabditis Genetics Center, and the Developmental Studies Hybridoma Bank for sharing reagents. We thank members of the Skirball DG department and Ellie Heckscher for critical reading of this manuscript. This work was supported by American Cancer Society Postdoctoral Fellowship PF-08-157-01-DDC (AMW) and by NIH grants R01GM067237 (BDG) and R01GM078341 (JN).


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Supplemental Information Supplemental Information includes Extended Experimental Procedures, six figures, three movies and one table.


1. Tanaka Y, Okada Y, Hirokawa N. FGF-induced vesicular release of Sonic hedgehog and retinoic acid in leftward nodal flow is critical for left-right determination. Nature. 2005;435:172–177. [PubMed]
2. Gyorgy B, Szabo TG, Pasztoi M, Pal Z, Misjak P, Aradi B, Laszlo V, Pallinger E, Pap E, Kittel A, et al. Membrane vesicles, current state-of-the-art: emerging role of extracellular vesicles. Cell Mol Life Sci. 2011;68:2667–2688. [PMC free article] [PubMed]
3. Théry C, Ostrowski M, Segura E. Membrane vesicles as conveyors of immune responses. Nat Rev Immunol. 2009;9:581–593. [PubMed]
4. Basse F, Gaffet P, Rendu F, Bienvenue A. Translocation of spin-labeled phospholipids through plasma membrane during thrombin- and ionophore A23187-induced platelet activation. Biochemistry. 1993;32:2337–2344. [PubMed]
5. Morris RJ. Ionic control of the metastable inner leaflet of the plasma membrane: Fusions natural and artefactual. FEBS LETTERS. 2009;584:1665–1669. [PubMed]
6. Zwaal RF, Comfurius P, Bevers EM. Scott syndrome, a bleeding disorder caused by defective scrambling of membrane phospholipids. Biochim Biophys Acta. 2004;1636:119–128. [PubMed]
7. Folmer DE, Elferink RPJO, Paulusma CC. P4 ATPases - lipid flippases and their role in disease. Biochim Biophys Acta. 2009;1791:628–635. [PubMed]
8. Coleman JA, Kwok MCM, Molday RS. Localization, purification, and functional reconstitution of the P4-ATPase Atp8a2, a phosphatidylserine flippase in photoreceptor disc membranes. JBC. 2009;284:32670–32679. [PMC free article] [PubMed]
9. Zhou X, Graham TR. Reconstitution of phospholipid translocase activity with purified Drs2p, a type-IV P-type ATPase from budding yeast. PNAS. 2009;106:16586–16591. [PMC free article] [PubMed]
10. Muthusamy B-P, Natarajan P, Zhou X, Graham TR. Linking phospholipid flippases to vesicle-mediated protein transport. Biochim Biophys Acta. 2009;1791:612–619. [PMC free article] [PubMed]
11. Groen A, Romero MR, Kunne C, Hoosdally SJ, Dixon PH, Wooding C, Williamson C, Seppen J, van Den Oever K, Mok KS, et al. Complementary Functions of the Flippase ATP8B1 and the Floppase ABCB4 in Maintaining Canalicular Membrane Integrity. Gastroenterology. 2011;141:1927–1937. [PubMed]
12. Lyssenko NN, Miteva Y, Gilroy S, Hanna-Rose W, Schlegel RA. An unexpectedly high degree of specialization and a widespread involvement in sterol metabolism among the C. elegans putative aminophospholipid translocases. BMC Dev Biol. 2008;8:96. [PMC free article] [PubMed]
13. Darland-Ransom M, Wang X, Sun C-L, Mapes J, Gengyo-Ando K, Mitani S, Xue D. Role of C. elegans TAT-1 protein in maintaining plasma membrane phosphatidylserine asymmetry. Science. 2008;320:528–531. [PubMed]
14. Züllig S, Neukomm LJ, Jovanovic M, Charette SJ, Lyssenko NN, Halleck MS, Reutelingsperger CPM, Schlegel RA, Hengartner MO. Aminophospholipid translocase TAT-1 promotes phosphatidylserine exposure during C. elegans apoptosis. Curr Biol. 2007;17:994–999. [PubMed]
15. Chen B, Jiang Y, Zeng S, Yan J, Li X, Zhang Y, Zou W, Wang X. Endocytic Sorting and Recycling Require Membrane Phosphatidylserine Asymmetry Maintained by TAT-1/CHAT-1. PLoS Genet. 2010;6:e1001235. [PMC free article] [PubMed]
16. Ruaud A-F, Nilsson L, Richard F, Larsen MK, Bessereau J-L, Tuck S. The C. elegans P4-ATPase TAT-1 regulates lysosome biogenesis and endocytosis. Traffic. 2009;10:88–100. [PubMed]
17. Seamen E, Blanchette JM, Han M. P-type ATPase TAT-2 negatively regulates monomethyl branched-chain fatty acid mediated function in post-embryonic growth and development in C. elegans. PLoS Genet. 2009;5:e1000589. [PMC free article] [PubMed]
18. Bruce EA, Digard P, Stuart AD. The Rab11 pathway is required for influenza A virus budding and filament formation. J Virol. 2010;84:5848–5859. [PMC free article] [PubMed]
19. Henne WM, Buchkovich NJ, Emr SD. The ESCRT pathway. Dev Cell. 2011;21:77–91. [PubMed]
20. Rohrschneider MR, Nance J. Polarity and cell fate specification in the control of Caenorhabditis elegans gastrulation. Dev Dyn. 2009;238:789–796. [PMC free article] [PubMed]
21. Grana TM, Cox EA, Lynch AM, Hardin JD. SAX-7/L1CAM and HMR-1/cadherin function redundantly in blastomere compaction and non-muscle myosin accumulation during C. elegans gastrulation. Dev Biol. 2010;344:731–744. [PMC free article] [PubMed]
22. Reese KJ, Dunn MA, Waddle J, Seydoux G. Asymmetric segregation of PIE-1 in C. elegans is mediated by two complementary mechanisms that act through separate PIE-1 protein domains. Mol Cell. 2000;6:445–455. [PubMed]
23. Lenoir G, Williamson P, Puts CF, Holthuis JCM. Cdc50p plays a vital role in the ATPase reaction cycle of the putative aminophospholipid transporter Drs2p. J Biol Chem. 2009;284:17956–17967. [PMC free article] [PubMed]
24. Schlaepfer DD, Mehlman T, Burgess WH, Haigler HT. Structural and functional characterization of endonexin II, a calcium- and phospholipid-binding protein. PNAS. 1987;84:6078–6082. [PMC free article] [PubMed]
25. Zhao M. Lantibiotics as probes for phosphatidylethanolamine. Amino Acids. 2009 [PMC free article] [PubMed]
26. Emoto K, Kobayashi T, Yamaji A, Aizawa H, Yahara I, Inoue K, Umeda M. Redistribution of phosphatidylethanolamine at the cleavage furrow of dividing cells during cytokinesis. PNAS. 1996;93:12867–12872. [PMC free article] [PubMed]
27. Singer-Krüger B, Lasić M, Bürger A-M, Hausser A, Pipkorn R, Wang Y. Yeast and human Ysl2p/hMon2 interact with Gga adaptors and mediate their subcellular distribution. EMBO J. 2008;27:1423–1435. [PMC free article] [PubMed]
28. Sato K, Sato M, Audhya A, Oegema K, Schweinsberg PJ, Grant BD. Dynamic regulation of caveolin-1 trafficking in the germ line and embryo of Caenorhabditis elegans. Mol Biol Cell. 2006;17:3085–3094. [PMC free article] [PubMed]
29. Audhya A, McLeod IX, Yates JR, Oegema K. MVB-12, a fourth subunit of metazoan ESCRT-I, functions in receptor downregulation. PLoS ONE. 2007;2:e956. [PMC free article] [PubMed]
30. Nance J, Priess JR. Cell polarity and gastrulation in C. elegans. Development. 2002;129:387–397. [PubMed]
31. Pohl C, Bao Z. Chiral Forces Organize Left-Right Patterning in C. elegans by Uncoupling Midline and Anteroposterior Axis. Dev Cell. 2010;19:402–412. [PMC free article] [PubMed]
32. Bevers EM, Williamson PL. Phospholipid scramblase: An update. FEBS LETTERS. 2010;584:2724–2730. [PubMed]
33. Suzuki J, Umeda M, Sims PJ, Nagata S. Calcium-dependent phospholipid scrambling by TMEM16F. Nature. 2010;468:834–838. [PubMed]
34. Creutz CE, Snyder SL. Interactions of annexins with the mu subunits of the clathrin assembly proteins. Biochemistry. 2005;44:13795–13806. [PubMed]
35. Booth AM, Fang Y, Fallon JK, Yang J-M, Hildreth JEK, Gould SJ. Exosomes and HIV Gag bud from endosome-like domains of the T cell plasma membrane. JCB. 2006;172:923–935. [PMC free article] [PubMed]
36. Gan X, Gould SJ. Identification of an inhibitory budding signal that blocks the release of HIV particles and exosome/microvesicle proteins. Mol Biol Cell. 2011;22:817–830. [PMC free article] [PubMed]
37. Utley TJ, Ducharme NA, Varthakavi V, Shepherd BE, Santangelo PJ, Lindquist ME, Goldenring JR, Crowe JE. Respiratory syncytial virus uses a Vps4-independent budding mechanism controlled by Rab11-FIP2. PNAS. 2008;105:10209–10214. [PMC free article] [PubMed]
38. Hua Z, Graham TR. Requirement for neo1p in retrograde transport from the Golgi complex to the endoplasmic reticulum. Mol Biol Cell. 2003;14:4971–4983. [PMC free article] [PubMed]
39. Wicky S, Schwarz H, Singer-Krüger B. Molecular interactions of yeast Neo1p, an essential member of the Drs2 family of aminophospholipid translocases, and its role in membrane trafficking within the endomembrane system. Mol Cell Biol. 2004;24:7402–7418. [PMC free article] [PubMed]
40. Achilleos A, Wehman AM, Nance J. PAR-3 mediates the initial clustering and apical localization of junction and polarity proteins during C. elegans intestinal epithelial cell polarization. Development. 2010;137:1833–1842. [PMC free article] [PubMed]
41. Anderson DC, Gill JS, Cinalli RM, Nance J. Polarization of the C. elegans embryo by RhoGAP-mediated exclusion of PAR-6 from cell contacts. Science. 2008;320:1771–1774. [PMC free article] [PubMed]
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Gene
    Gene links
  • GEO Profiles
    GEO Profiles
    Related GEO records
  • HomoloGene
    HomoloGene links
  • MedGen
    Related information in MedGen
  • Pathways + GO
    Pathways + GO
    Pathways, annotations and biological systems (BioSystems) that cite the current article.
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem Substance links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...