ARF6-mediated endosomal transport of Telencephalin affects dendritic filopodia-to-spine maturation
2012; Springer Nature; Volume: 31; Issue: 15 Linguagem: Inglês
10.1038/emboj.2012.182
ISSN1460-2075
AutoresTim Raemaekers, Aleksandar Perić, Pieter Baatsen, Ragna Sannerud, Ilse Declerck, Veerle Baert, Christine Michiels, Wim Annaert,
Tópico(s)Cell death mechanisms and regulation
ResumoArticle10 July 2012Open Access ARF6-mediated endosomal transport of Telencephalin affects dendritic filopodia-to-spine maturation Tim Raemaekers Corresponding Author Tim Raemaekers Laboratory of Membrane Trafficking, Center for Human Genetics, KU Leuven and VIB Center for the Biology of Disease, Leuven, Belgium Search for more papers by this author Aleksandar Peric Aleksandar Peric Laboratory of Membrane Trafficking, Center for Human Genetics, KU Leuven and VIB Center for the Biology of Disease, Leuven, Belgium Search for more papers by this author Pieter Baatsen Pieter Baatsen Electron Microscopy Facility (EMCORF), VIB Center for the Biology of Disease, Leuven, Belgium Search for more papers by this author Ragna Sannerud Ragna Sannerud Laboratory of Membrane Trafficking, Center for Human Genetics, KU Leuven and VIB Center for the Biology of Disease, Leuven, Belgium Search for more papers by this author Ilse Declerck Ilse Declerck Laboratory of Membrane Trafficking, Center for Human Genetics, KU Leuven and VIB Center for the Biology of Disease, Leuven, Belgium Search for more papers by this author Veerle Baert Veerle Baert Laboratory of Membrane Trafficking, Center for Human Genetics, KU Leuven and VIB Center for the Biology of Disease, Leuven, Belgium Search for more papers by this author Christine Michiels Christine Michiels Laboratory of Membrane Trafficking, Center for Human Genetics, KU Leuven and VIB Center for the Biology of Disease, Leuven, Belgium Search for more papers by this author Wim Annaert Corresponding Author Wim Annaert Laboratory of Membrane Trafficking, Center for Human Genetics, KU Leuven and VIB Center for the Biology of Disease, Leuven, Belgium Search for more papers by this author Tim Raemaekers Corresponding Author Tim Raemaekers Laboratory of Membrane Trafficking, Center for Human Genetics, KU Leuven and VIB Center for the Biology of Disease, Leuven, Belgium Search for more papers by this author Aleksandar Peric Aleksandar Peric Laboratory of Membrane Trafficking, Center for Human Genetics, KU Leuven and VIB Center for the Biology of Disease, Leuven, Belgium Search for more papers by this author Pieter Baatsen Pieter Baatsen Electron Microscopy Facility (EMCORF), VIB Center for the Biology of Disease, Leuven, Belgium Search for more papers by this author Ragna Sannerud Ragna Sannerud Laboratory of Membrane Trafficking, Center for Human Genetics, KU Leuven and VIB Center for the Biology of Disease, Leuven, Belgium Search for more papers by this author Ilse Declerck Ilse Declerck Laboratory of Membrane Trafficking, Center for Human Genetics, KU Leuven and VIB Center for the Biology of Disease, Leuven, Belgium Search for more papers by this author Veerle Baert Veerle Baert Laboratory of Membrane Trafficking, Center for Human Genetics, KU Leuven and VIB Center for the Biology of Disease, Leuven, Belgium Search for more papers by this author Christine Michiels Christine Michiels Laboratory of Membrane Trafficking, Center for Human Genetics, KU Leuven and VIB Center for the Biology of Disease, Leuven, Belgium Search for more papers by this author Wim Annaert Corresponding Author Wim Annaert Laboratory of Membrane Trafficking, Center for Human Genetics, KU Leuven and VIB Center for the Biology of Disease, Leuven, Belgium Search for more papers by this author Author Information Tim Raemaekers 1,‡, Aleksandar Peric1,‡, Pieter Baatsen2, Ragna Sannerud1, Ilse Declerck1, Veerle Baert1, Christine Michiels1 and Wim Annaert 1 1Laboratory of Membrane Trafficking, Center for Human Genetics, KU Leuven and VIB Center for the Biology of Disease, Leuven, Belgium 2Electron Microscopy Facility (EMCORF), VIB Center for the Biology of Disease, Leuven, Belgium ‡These authors contributed equally to this work *Corresponding authors. Laboratory of Membrane Trafficking, Center for Human Genetics, KU Leuven and VIB Center for the Biology of Disease, Campus Gasthuisberg, POB 602, O&N4, Room 7.159, Leuven, Belgium. Tel.:+32 16 330520; Fax:+32 16 330939; E-mail: [email protected] or E-mail: [email protected] The EMBO Journal (2012)31:3252-3269https://doi.org/10.1038/emboj.2012.182 PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Dendritic filopodia are dynamic structures thought to be the precursors of spines during synapse development. Morphological maturation to spines is associated with the stabilization and strengthening of synapses, and can be altered in various neurological disorders. Telencephalin (TLN/intercellular adhesion molecule-5 (ICAM5)) localizes to dendritic filopodia, where it facilitates their formation/maintenance, thereby slowing spine morphogenesis. As spines are largely devoid of TLN, its exclusion from the filopodia surface appears to be required in this maturation process. Using HeLa cells and primary hippocampal neurons, we demonstrate that surface removal of TLN involves internalization events mediated by the small GTPase ADP-ribosylation factor 6 (ARF6), and its activator EFA6A. This endocytosis of TLN affects filopodia-to-spine transition, and requires Rac1-mediated dephosphorylation/release of actin-binding ERM proteins from TLN. At the somato-dendritic surface, TLN and EFA6A are confined to distinct, flotillin-positive membrane subdomains. The co-distribution of TLN with this lipid raft marker also persists during its endosomal targeting to CD63-positive late endosomes. This suggests a specific microenvironment facilitating ARF6-mediated mobilization of TLN that contributes to promotion of dendritic spine development. Introduction Dendritic filopodia are long, thin, actin-rich, and dynamic protrusions that are considered to be the precursors of mature, mushroom-shaped spines, both during early neural development and later into adulthood (Yuste and Bonhoeffer, 2004). This maturation process that generates post-synaptic sites of synapses is reversible, and reflects the plastic nature of synaptic connections (Matus, 2000). Alterations of the underlying remodelling machinery can result in abnormal spine structures, as seen in various neurological disorders including Fragile X syndrome and Alzheimer's disease (AD) (Kaufmann and Moser, 2000; Knafo et al, 2009). Since the mechanism driving spine morphogenesis is still unclear, gaining further insight herein remains an important challenge. Telencephalin (TLN), also known as intercellular adhesion molecule-5 (ICAM5), plays an important role in spine morphogenesis. It prominently localizes to dendritic filopodia where it facilitates their formation and maintenance in a process that requires actin-binding ERM (ezrin/radixin/moesin) proteins (Matsuno et al, 2006; Furutani et al, 2007). In contrast to other adhesion molecules, TLN slows maturation and stabilization of synapses. In agreement, TLN deficiency or knockdown of ERM proteins increases spine maturation, while overexpression of TLN or constitutively active ezrin favours more filopodia (Matsuno et al, 2006; Furutani et al, 2007). Although the exact mechanism governing filopodia to spine transition is still unclear, it may require exclusion of TLN from filopodia, as spines are largely devoid of it (Yoshihara et al, 2009). We hypothesize that this is most likely mediated by until now unexplored internalization events, aside of the already described proteolytic shedding of TLN (Tian et al, 2007). We previously demonstrated that TLN accumulates aberrantly and prominently in presenilin1 (PSEN1)-deficient hippocampal neurons, as opposed to culture-matched wild-type neurons (Annaert et al, 2001; Esselens et al, 2004). PSEN1 is a key component of the γ-secretase complex important in AD (De Strooper and Annaert, 2010). It interacts with TLN, but rather than cleaving it, PSEN1 modulates its trafficking in a γ-secretase-independent manner (Esselens et al, 2004). Importantly, intracellular entrapment of TLN may affect its normal surface levels/function. As aberrant endosomal trafficking and synaptic dysfunctions have both been described as early stage events in various neurodegenerative diseases, including AD (Pimplikar et al, 2010), we searched for regulators of internalization/trafficking of TLN, and explored how they affect its role in spine morphogenesis. The small GTPase, ADP-ribosylation factor 6 (ARF6), is known to regulate endosomal trafficking and actin dynamics (D'Souza-Schorey and Chavrier, 2006; Grant and Donaldson, 2009). Its role herein has been mostly studied in HeLa cells, while in neurons, ARF6 was shown to participate in spine morphogenesis, where its activation by the guanine nucleotide exchange factor (GEF), EFA6A (Sakagami, 2008), promoted formation and maintenance of spines in a Rac1-dependent manner (Choi et al, 2006). This clearly contrasts the aforementioned function of TLN in slowing spine maturation (Matsuno et al, 2006). In this study, we demonstrate that (i) TLN recruits EFA6A and (ii) requires ARF6-activation for internalization. We furthermore show that its endosomal targeting to CD63-positive multivesicular bodies (MVBs)/late endosomes involves the association with flotillin-positive microdomains. Collectively, our data support the ARF6-dependent mobilization of TLN from dendritic filopodia, which consequently contributes to their maturation into spines. Results TLN and the ARF6 activator, EFA6A, interact in vitro, and associate with flotillin-positive domains in neurons In order to identify novel regulators of TLN internalization, we focused on proteins involved in membrane trafficking and spine development. We opted to assess if the endogenous distribution pattern of such candidates changed as a consequence of an induced change in the somato-dendritic localization of TLN. We established previously that the addition of polystyrene microbeads to primary hippocampal neurons results in a marked re-distribution of TLN to the microbead attachment sites (Esselens et al, 2004). TLN appears to act here specifically as microbeads seem not to adhere to axons (Supplementary Figure S1), nor to young neurons (Supplementary Figure S2), both of which are devoid of TLN. The feature of TLN herein is further reinforced by the lack of prominent recruitment to such sites of another dendritic cell adhesion protein, N-Cadherin (Figure 1A). Figure 1.Telencephalin and EFA6A interact in vitro, and associate with flotillin-positive domains in neurons. (A) Top panel: Hippocampal neurons (DIV 14) were incubated overnight with polystyrene microbeads (1 μm) and stained for TLN (red), EFA6A (green) and N-Cadherin (blue). Arrowheads indicate microbead attachment sites, where TLN and EFA6A, but not N-Cadherin are strongly co-enriched. Bottom panel: as in the top panel but using HeLa cells co-expressing TLN–Cherry and EFA6A–GFP. (B) Hippocampal neurons (DIV 14) were stained for endogenous TLN (red) and EFA6A (green) with or without ARF6 (blue). Top panel: Dendrite showing the close association of TLN and EFA6A. Note that each TLN-positive filopodium also contains EFA6A. Middle and bottom panels: TLN and EFA6A are closely apposed to ARF6 within filopodia as well as at the cell body, where the staining pattern of the three proteins appears confined to distinct surface microdomains (arrowheads). (C) Hippocampal neurons (DIV 14) co-expressing TLN–Cherry and flotillin2–GFP. Both proteins show, as in (B), a similar distribution along dendrites and at the cell body. (D) Acquired images (see B, C) were analysed for overlap between TLN and EFA6A or flotillin2 at the surface microdomains. Data were collected and averaged from several independent neurons and depicted in graphs. The obtained percentages are for EFA6A: 92.5±0.4, and flotillin2: 92.1±1.2. (E) Cell lysates obtained from HeLa cells co-expressing either TLN–mRFP or a cytosolic-domain deletion mutant of TLN (ΔIC) and/or EFA6A–GFP, as indicated on the top, were subjected to co-immunoprecipitation experiments with anti-GFP Ab. Note that in contrast to full-length TLN, TLNΔIC did not co-immunoprecipitate with EFA6A. Bars: 10 μm. Download figure Download PowerPoint Interestingly, we showed that the cup-like dendritic protrusions at microbead attachment sites also contain phosphatidylinositol 4,5-biphosphate (PIP2; Esselens et al, 2004). PIP2 is a surface-associated phospholipid generated by phosphatidylinositol 4-phosphate 5-kinase (PIP5-kinase) upon its activation by ARF6 (Brown et al, 2001). Besides regulating endosomal transport and actin dynamics (D'Souza-Schorey and Chavrier, 2006), ARF6 also promotes spine development (Choi et al, 2006), making it a prime candidate regulator of TLN trafficking. As the ARF6-GEF, EFA6A, has a similar somato-dendritic localization (Sakagami et al, 2007; Sannerud et al, 2011) as TLN, we first tested if it co-enriched with TLN at microbead attachment sites in hippocampal neurons and (transfected) HeLa cells. HeLa cells do not normally express TLN, but they do offer the advantage of working with a well-established cell system for studying EFA6A/ARF6-related trafficking (Donaldson et al, 2009). For transfection experiments in HeLa cells and in neurons, we tagged TLN with an internal, transmembrane proximal mCherry/mRFP tag (Figure 4A), thereby keeping both protein termini free for physiologically relevant interactions (Tian et al, 2000; Furutani et al, 2007). This internal tag did not interfere with TLN's recruitment and attachment to microbeads (Figure 1A; Supplementary Figure S3), nor does it affect its somato-dendritic polarization during development (Nicolaï et al, 2010). In both differentiated hippocampal neurons and co-transfected HeLa cells, EFA6A was indeed co-recruited with TLN to microbeads (Figure 1A). Next, we analysed the distribution of endogenous EFA6A/ARF6 and TLN in 14 DIV hippocampal neurons and showed that as expected they have closely overlapping distribution along the dendritic shaft and within filopodia. At the basal surface of neuronal cell bodies, they also displayed a strikingly similar mesh-like pattern (Figure 1B and D), suggesting their mutual preference for a defined membrane (lipid/protein) microenvironment. This is further strengthened by the observation that N-Cadherin has a more homogeneous surface distribution (Supplementary Figure S4). Interestingly, the nature of these microdomains is distinct from that of focal adhesion molecules as both vimentin and α-actinin are excluded from TLN-positive domains in TLN-expressing HeLa cells, where a reminiscent surface distribution of TLN is observed (Supplementary Figure S4). On the other hand, the lipid raft-associated protein flotillin/reggie displays a similar distribution (Neumann-Giesen et al, 2004) and our co-expression analysis of tagged TLN and flotillin2 in HeLa cells (Supplementary Figure S5) and in hippocampal neurons (Figure 1C and D) also confirms this. Here, along dendrites, both proteins were closely apposed within filopodia and in the shaft (Figure 1C), where the patchy staining pattern often persisted. This congruent microdomain association of expressed TLN and flotillin is further supported through their mutual recruitment to microbeads added to HeLa cells (Supplementary Figure S6) and the ability to, unlike N-Cadherin, resist the detergent extraction in neurons (Supplementary Figure S7; Ledesma et al, 1998). As EFA6A co-distributes with TLN we next assessed their interaction biochemically. EFA6A was previously shown to interact with two neuronal potassium channels, both depending on ARF6 for their trafficking (Decressac et al, 2004; Gong et al, 2007). Using similar co-immunoprecipitation conditions, we show in a reciprocal set-up (Figure 1E; Supplementary Figure S8) that EFA6A may indeed interact (in)directly with TLN and that this interaction could depend on the presence of the intracellular domain of TLN. Taken together, these data suggest that TLN and EFA6A may associate with flotillin-positive membrane domains. The relevance of these observations is underscored by the fact that flotillin, like EFA6A, plays a role in protein trafficking (Stuermer, 2010), has synaptic functions (Swanwick et al, 2010b), and as TLN, affects filopodia formation (Neumann-Giesen et al, 2004). Trafficking of TLN is regulated by ARF6 and its internalization involves EFA6A-mediated activation of ARF6 ARF6 regulates the trafficking of specific plasma-membrane proteins by switching from a GDP- to an active GTP-bound form via an exchange activity mediated by EFA6A (Figure 2C; D'Souza-Schorey and Chavrier, 2006; Franco et al, 1999). As TLN can interact with EFA6A, we therefore studied how altering the activation state of ARF6 affects the localization of TLN. When expressed alone in HeLa cells, TLN is found almost exclusively at the cell surface, localizing to filopodia and microvilli-like structures, which are induced by its expression (Figure 2A). On the contrary, co-expression with EFA6A causes filopodia and microvilli-like structures to disappear and TLN to redistribute from the surface to a punctate vesicular localization. This likely reflects its enhanced internalization, as these punctae are not immunostained by a TLN ectodomain-specific antibody in non-permeabilized cells (Figure 2A and B; Supplementary Figure S9, and see later). This agrees with the role of EFA6A in co-ordinating endocytosis with cytoskeletal remodelling at the cell surface, both requiring the activity of ARF6 and its downstream target, the Rho family GTPase, Rac1 (Franco et al, 1999; Figure 2C). To demonstrate that EFA6A-mediated TLN internalization indeed occurs via ARF6 activation, we co-expressed TLN/EFA6A with ARF6T27N, a dominant mutant keeping ARF6 in its GDP-locked state. As expected, here we noticed little or no TLN internalization, while filopodia and microvilli-like structures remained preserved (Figure 2A and B). Figure 2.Trafficking of Telencephalin is regulated by ARF6 and its internalization involves EFA6A-mediated activation of ARF6. (A) Top panel: HeLa cell expressing TLN–Cherry. Insets highlight the presence of TLN at filopodia and microvilli-like structures evident at the apical surface. Middle panel: HeLa cells co-expressing TLN–Cherry and EFA6A–GFP. Arrowheads indicate internal vesicles containing both TLN and EFA6A. Lower panel: as in the middle panel, but co-expressed with ARF6T27N–HA (inactive ARF6), and stained for HA (blue). Note that TLN is hardly internalized and filopodia and microvilli-like structures remain intact. (B) List summarizing the presence/abundance of intracellular TLN-containing vesicles and filopodia/microvilli-like structures, when TLN is expressed alone, or in combination with EFA6A with or without ARF6T27N. More than 100 randomly selected cells from three independent experiments were analysed. (C) Literature-based scheme that depicts how inactive GDP-bound ARF6 gets activated to GTP-bound ARF6 by its GEF, EFA6A. The mutants used in the study (ARF6Q67L, ARF6T157A, ARF6T27N), and the activation state which they reflect, are also indicated. Note that ARF6-induced actin remodelling and membrane trafficking is dependent on Rac1. (D) Top panel: HeLa cells co-expressing TLN–Cherry and ARF6Q67L–HA were stained for HA (green) and the ARF6-cargo protein, MHCI (blue). All three proteins are clearly entrapped within the grape-like vacuoles. Middle panel: Hippocampal neurons (DIV 14) co-expressing TLN–Cherry and ARF6Q67L–HA. Arrowheads indicate TLN entrapment in vacuole-like structures that resemble those observed in HeLa cells. Bottom panel: Hippocampal neurons (DIV 14) expressing the ARF6Q67L–HA were stained for endogenous TLN (red) and HA (green). Note the clear entrapment of TLN in ARF6Q67L vacuoles. Bars: 10 μm. Download figure Download PowerPoint We next studied how ARF6 activation affects intracellular sorting of TLN using well-described ARF6 mutants, including ARF6T27N and the GTP-locked (constitutively active) ARF6Q67L. In HeLa cells, expression of ARF6Q67L results in the accumulation of grape-like vacuoles in which ARF6-cargo proteins like MHCI and CD59 are trapped (Naslavsky et al, 2004). Likewise, co-expressed TLN readily colocalized with MHCI in ARF6Q67L-positive endosomal structures (Figure 2D). ARF6 cargos, like MHCI and CD59, internalize in a clathrin-independent manner as opposed to, e.g., the transferrin receptor (TfR) that uses clathrin-mediated endocytosis (Naslavsky et al, 2004). As such, TfR, did not accumulate in ARF6Q67L vacuoles, thereby underscoring the selective ARF6-dependent endosomal sorting of TLN (Supplementary Figure S10). Importantly, in primary hippocampal neurons, endogenous as well as overexpressed TLN similarly accumulated in ARF6Q67L vacuoles (Figure 2D). On the other hand, the GDP-locked mutant ARF6T27N is known to block recycling of specific ARF6 cargos from a perinuclear recycling compartment to the cell surface (D'Souza-Schorey et al, 1998; Sannerud et al, 2011). Expectedly, upon co-expression with ARF6T27N in HeLa cells, TLN co-accumulated with ARF6T27N and endogenous CD59 in discrete perinuclear compartments (Supplementary Figure S11). Notably, colocalization of expressed TLN and ARF6T27N was also observed in hippocampal neurons (Supplementary Figure S12). Taken together, these data identify TLN as a novel cargo protein of ARF6-mediated internalization and endosomal sorting. Internalization of TLN involves Rac1-mediated actin remodelling, ERM dephosphorylation as well as its release from binding to ERM proteins As Rac1 activity is required for ARF6-mediated spine maturation (Choi et al, 2006), we tested how this downstream target of ARF6 (Figure 3D; Franco et al, 1999; Koo et al, 2007) affects TLN localization. In HeLa cells, co-expression of TLN with a dominant active Rac1 mutant (GTP-bound Rac1V12; Ridley et al, 1992) resulted in prominent dorsal ruffling and endocytosis of TLN (Figure 3A). Cortical actin polymerization/remodelling underlies membrane ruffling (Franco et al, 1999) and is usually accompanied by the uptake of large portions of extracellular fluid (macropinocytosis) (Donaldson et al, 2009). To assess if internalized TLN indeed arises from macropinocytic uptake, we incubated TLN/Rac1V12-expressing cells with 70 kDa dextran beads. Confocal analysis indeed revealed a clear colocalization of a part of TLN-positive endosomes with this macropinocytic marker (Figure 3B). Figure 3.Internalization of Telencephalin involves Rac1-mediated actin remodelling, ERM dephosphorylation as well as its release from binding to ERM proteins. (A) HeLa cells co-expressing TLN–Cherry and Rac1V12–GFP were stained for surface TLN (blue) in non-permeabilized cells using anti-Cherry antibody to show the internal nature of the TLN-containing vesicles (arrowheads). (B) HeLa cells co-expressing untagged TLN (colour-coded: red) and Rac1V12–GFP (colour coded: blue) were incubated with fluorescent dextran beads (70 kDa) (colour coded: green) to assay macropinocytosis. Arrowheads indicate overlap of beads with some of the TLN-positive endosomes. (C) HeLa cells co-expressing TLN–Cherry and Rac1V12–GFP were stained for phosphorylated ERM proteins (pERM; blue). Arrowheads indicate internalized TLN, while arrows point out a membrane ruffle where TLN and Rac1 are found colocalizing. Note that the transfected cell displays overall reduced pERM staining and is largely devoid of filopodia and microvilli-like structures. (D) Literature-based scheme depicting how GTP-bound ARF6 activates Rac1 by activating its GEF, thereby inducing actin remodelling and membrane trafficking. The active Rac1 mutant (Rac1V12) used in (A–C) is indicated. Also shown is the Rac1-mediated dephosphorylation/inactivation of ERM proteins, and their subsequent release from binding to plasma membrane proteins, like TLN. Together, these effects lead to the disassembly of filopodia and microvilli-like structures. (E) HeLa cell (top panel) and primary hippocampal neuron (DIV 14) (bottom panel) co-expressing TLN–Cherry and VSVG-tagged FERM domain were stained for VSVG (green). Arrowheads indicate TLN- and FERM-containing internal vesicles, while the arrow (bottom panel) indicates their association within a filopodium from where TLN likely internalizes. Bars: 10 μm. Download figure Download PowerPoint Figure 4.Deletion of an acidic cluster/ERM-binding domain of Telencephalin enhances its internalization. (A) Scheme depicting the signal peptide (SP), extra- and intracellular domains of TLN, as well as the position of the internal Cherry tag, proximal to the trans-membrane region (TM). Also shown is the amino-acid sequence of the intracellular domain of human TLN, and its alignment with those from other species. Boxed in blue is a juxta-membrane cluster of acidic residues that fall within a well-conserved region. The residues essential for binding to ERM proteins (blue arrowheads) and the alignment of the TLN acidic cluster with the acidic EX motif of the KAC motif, which is implicated in ARF6-dependent trafficking are also depicted. (B) HeLa cells expressing either TLN–Cherry (top panel) or its acidic cluster-deleted mutant (TLNΔAC) (bottom panel) were stained for pERM (green). Insets show that in contrast to full-length TLN, which localizes to pERM-positive filopodia and microvilli-like structures, TLNΔAC is found mostly in internal vesicles. Cells expressing TLNΔAC also have fewer filopodia and microvilli-like structures. (C) HeLa cells co-expressing either TLN–Cherry (top panel) or TLNΔAC–Cherry (bottom panel) with RAB5Q79L–Cerulean, fixed at the same time point after transfection. Insets show entrapment of TLN in enlarged RAB5 endosomes, which is more prominent for TLNΔAC than for full-length TLN. Multiple randomly selected cells were analysed and the accompanying quantification is shown in graphs. The obtained means±s.e.m. values were expressed relative to full-length TLN, which was defined as 100% (TLN: 100.0±12.9; TLNΔAC: 331.3±35.3). (D) As in (C) but studied in hippocampal neurons (DIV 14). The values are TLN: 100.0±6.4 and TLNΔAC: 176.5±22.4. Arrowheads indicate entrapment of TLN in Rab5Q79L endosomes. Note that, in contrast to TLNΔAC, full-length TLN is still evidently visible at the cell surface. P-values (Student's t-test): **P<0.01, ***P<0.001. Bars: 10 μm. Download figure Download PowerPoint As interaction of TLN with phosphorylated ERM (pERM) proteins is essential for its surface anchorage and function in filopodia formation/maintenance (Furutani et al, 2007), we next show that in addition to TLN internalization, Rac1 activation also causes a general reduction in staining intensity of pERMs (Figure 3C). This implies that the mechanism of TLN endocytosis could involve ERM dephosphorylation. Earlier studies in lymphocytes reported that Rac1-mediated dephosphorylation of ERMs translated into disassembly of microvilli via a loss of anchorage of membrane proteins to the underlying actin cytoskeleton (Faure et al, 2004; Nijhara et al, 2004; Figure 3D). Thus, we tested whether disrupting the TLN–ERM interaction is sufficient to induce TLN internalization (Figure 3E). Co-expression of TLN with the FERM domain of ERM proteins (which binds the cytosolic domain of target proteins including TLN, but not actin; Amieva et al, 1999; Allenspach et al, 2001), outcompeted binding of endogenous pERMs to TLN, resulting in the emergence of TLN-positive endosomal punctae, both in HeLa cells and in neurons (Figure 3E). Together, these data demonstrate the dependence of TLN internalization on Rac1-mediated cortical actin remodelling and ERM dephosphorylation. Deletion of an acidic cluster/ERM-binding domain of TLN enhances its internalization To elucidate further the endosomal route of TLN, we searched for sorting motifs within its cytosolic domain (Figure 4A). As sequence homology searches did not conclusively reveal any known sorting sequences, and in light of our findings (Figure 3), we hypothesized that the deletion of the ERM-binding domain of TLN on its own may be sufficient to facilitate its internalization. Previously, it has been suggested that ERM binding to TLN requires well-conserved amino-acid residues (Furutani et al, 2007), which we now show are situated within a conserved acidic cluster (AC) (Figure 4A), slightly resembling the EX motif of the KAC (potassium channel AC), a sorting sequence involved in ARF6-mediated trafficking (Gong et al, 2007). We therefore generated a mutant TLN in which the AC was deleted (TLNΔAC). Expression of TLNΔAC in HeLa cells resulted in a loss of staining for active pERM proteins and fewer filopodia/microvilli-like stuctures, as compared to full-length TLN. This was paralleled by a clear redistribution of TLNΔAC towards a more internal localization (Figure 4B). Moreover, the observation that TLNΔAC still got trapped in the ARF6Q67L compartment (Supplementary Figure S13) suggests that the AC of TLN does not affect its ARF6-dependent routing, but rather plays a role in its surface retention, likely by binding the ERM proteins. To quantify the differential internalization of TLNΔAC relative to full-length TLN, we co-expressed each construct together with the dominant active GTP-locked RAB5 mutant (RAB5Q79L; Stenmark et al, 1994). This mutant blocks sorting and recycling from early endosomes resulting in the accumulation of surface-internalized cargos in enlarged RAB5 endosomes. This approach allowed us to clearly discriminate between surface-associated and endosomal TLN. As expected, in both HeLa cells and primary neurons, TLNΔAC accumulated more prominently in RAB5Q
Referência(s)