JBC Anatrace, Inc.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M601021200 on March 9, 2006

J. Biol. Chem., Vol. 281, Issue 18, 12352-12361, May 5, 2006
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
281/18/12352    most recent
M601021200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Klein, S.
Right arrow Articles by Luton, F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Klein, S.
Right arrow Articles by Luton, F.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

Role of the Arf6 GDP/GTP Cycle and Arf6 GTPase-activating Proteins in Actin Remodeling and Intracellular Transport*

Stéphanie Klein1, Michel Franco, Pierre Chardin, and Frédéric Luton2

From the Institut de Pharmacologie Moléculaire et Cellulaire, CNRS UMR 6097, 660, route des Lucioles, 06560 Valbonne, France

Received for publication, February 2, 2006 , and in revised form, March 2, 2006.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
We have analyzed both biochemically and functionally a series of Arf6 mutants, providing new insights into the molecular mode of action of the small G protein Arf6. First, by comparing a fast-cycling mutant (Arf6(T157N)) and a GTPase-deficient mutant (Arf6(Q67L)), we established the necessity for completion of the Arf6 GDP/GTP cycle for recycling of major histocompatibility complex molecules to the plasma membrane. Second, we found that aluminum fluoride (AlF), known for inducing membrane protrusion in cells expressing exogenous wild-type Arf6, stabilized a functional wild-type Arf6·AlFx · GTPase-activating protein (GAP) complex in vitro and in vivo. We also found that the tandem mutation Q37E/S38I prevented the binding of two Arf GAPs, but not the effector ARHGAP10, and blocked the formation of membrane protrusion and actin reorganization. Together, our results with AlFx and Arf6(Q37E/S38I) demonstrate the critical role of the Arf6 GAPs as effectors for Arf6-regulated actin cytoskeleton remodeling. Finally, competition experiments conducted in vivo suggest the existence of a membrane receptor for GDP-bound Arf6.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
The small G protein Arf6 (ADP-ribosylation factor-6) is involved in a wide variety of cellular events, including cell adhesion and migration, endocytosis, secretion, phagocytosis, and formation of invadopodia (1–4). It is believed to accomplish these diverse functions by modulating plasma membrane rearrangement, actin cytoskeleton organization, and intracellular transport. Classically, small G proteins are seen as molecular switches that cycle between an active GTP-bound form and an inactive GDP-bound form. This implies that a GTPase-defective mutant locked in a GTP-bound form would reproduce the functions controlled by the corresponding activated small G protein. However, it has long been appreciated for many G proteins that GTP hydrolysis as well is necessary to elicit the full biological response, suggesting that the completion of the full GDP/GTP cycle is important. For instance, the transforming mutant Ras(F28L) was found to display a dramatic decreased affinity for both nucleotides (5). This mutant cycles at an enhanced rate between its GDP- and GTP-bound states and is called "fast-cycling." The analogous Cdc42(F28L), Rac1(F28L), and Rho(AF30L) mutants were also found to be fast-cycling proteins with tumorigenic activity (6, 7). It is interesting that, in contrast to its fast-cycling mutant, the GTPase-defective mutant of Cdc42 does not confer the transforming characteristics when expressed in cells. The nucleotide cycle of the small G proteins of the Rab and Ran families is also critical to ensure the vectorial transport of vesicles and the shuttling in and out the nucleus, respectively (1, 8). The requirement for the completion of the nucleotide cycle has also been documented for Arf1 involved at different stages in the formation of coat complex protein I vesicles (9). Therefore, it appears that, in many cases, to recapitulate the biological functions of small G proteins, mutants that cycle constitutively between their GDP- and GTP-bound forms are more representative of the small G protein natural activities.

Studies with fast-cycling mutants underline the importance of the GTPase cycle and indicate that the GTPase-activating proteins (GAPs)3 cannot be considered as mere terminators, but rather as essential effectors (5–7). In fact, there has been recent accumulating evidence of molecular functions and signaling pathways controlled by GAPs for Ras and Rho (10, 11). This is also the case for the Arf GAPs, which represent the largest group of Arf-GTP-binding partners identified to date. Over 15 different Arf GAPs found in the human genome have been classified into two groups: Arf GAP1 and AZAP (Arf GAP with ankyrin repeats and pleckstrin homology domains) (12). Arf GAP1 proteins comprise Arf GAP1/3, the Git proteins, and additional proteins that contain an Arf GAP-like domain. The AZAP subgroup is further subdivided into four groups: ASAPs containing SH3 domains, ACAPs with predicted coiled-coil domains, AGAPs with GTP-binding protein-like domains, and ARAPs containing Rho GAP domains. The specificity of these enzymes has yet to be conclusively solved and, more important, established in vivo. Arf6 seems to be the target of several Arf GAPs, including ACAP1 and ACAP2, Git1 and Git2, or ASAP1 and ASAP2 (12). The fact that more Arf GAPs than Arf proteins exist make them likely candidates as effectors. Furthermore, their multidomain organization argues in favor of their acting as scaffolding molecules capable of transducing the signal of an activated Arf protein toward one or several biological functions. For instance, they might facilitate the concerted rearrangement of the plasma membrane and actin cytoskeleton required for cell migration and phagocytosis downstream of Arf6 activation.

Arf6-regulated cellular functions are classically approached by using dominant-negative (Arf6(T27N)) and constitutively active (Arf6(Q67L)) mutants. In many cases, both mutants exhibit a similar phenotype, suggesting that the blocking of the GDP/GTP cycle affects Arf6 biological effects (13–18). In fact, a fast-cycling mutant of Arf6 (Arf6(T157A)) was used to show that the Arf6 GDP/GTP cycle might be important for Arf6 effects on actin cytoskeleton remodeling and cell migration (19).

Here, we investigated the role of the Arf6 GAPs in Arf6-mediated cortical actin cytoskeleton remodeling and studied the importance of the completion of the full GDP/GTP cycle of Arf6 for the post-endocytic transport of major histocompatibility complex class I (MHCI). Collectively, our data support a model whereby Arf6 conducts its biological functions by cycling through its GDP- and GTP-bound forms and suggest an important role for the Arf6 GAPs. We also contribute new evidence for a plasma membrane receptor for Arf6. Altogether, our results provide new insights into Arf6 mechanisms of action.


   EXPERIMENTAL PROCEDURES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
RESULTS
 DISCUSSION
 REFERENCES
 
Cell Culture and Antibodies—Baby hamster kidney (BHK) cells were grown in BHK-21 medium (Invitrogen) containing 5% fetal calf serum, 10% tryptose phosphate broth, 100 units/ml penicillin, 100 µg/ml streptomycin, and 2 mM L-glutamine. HeLa cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 µg/ml streptomycin. Transient transfections were performed using FuGENE 6 transfection reagent (Roche Diagnostics, Mannheim, Germany) according to manufacturer's instructions. Anti-hemagglutinin (HA) epitope tag (clones 3F10 and 12CA5; Roche Diagnostics), anti-Myc epitope tag (clone 9E10; Roche Diagnostics), anti-FLAG epitope tag (M2; Sigma), anti-Arf6 (clone 8A6-2) (20), and anti-MHCI (clone HC10) (59) monoclonal antibodies were used. Fluorescently conjugated phalloidin and secondary antibodies were from Molecular Probes (Eugene, OR).

Production and Purification of Recombinant Arf Proteins—For production of myristoylated Arf1 and Arf6 proteins, BL21(DE3) bacteria were cotransformed with pET21b-Arf and pBB131, encoding yeast N-myristoyltransferase (21). Recombinant myristoylated Arf6 with a C-terminal hexahistidine tag was separated from the contaminating non-myristoylated protein by precipitation in 35% saturated ammonium sulfate, followed by adsorption to His·Bind resin (nickel-nitrilotriacetic acid-agarose; Qiagen Inc.) and elution with 200 mM imidazole (pH 8) in 10% glycerol. After gel filtration on NAP-10 columns (Amersham Biosciences), the proteins were stored in 50 mM Tris-HCl (pH 8), 1 mM MgCl2, 1 mM dithiothreitol, 0.25 mM phenylmethylsulfonyl fluoride, and 5 µM GTP or GDP. Myr-Arf1 was purified as described previously (22).

Biochemical Properties of Recombinant Arf6 Proteins—The N-myristoylated and C-terminally hexahistidine-tagged wild-type (WT) Arf6, Arf6(T157N), Arf6(Q37E/S38I), and Arf6(Q37E/S38I/T157N) proteins purified from bacteria were loaded with GDP or GTP in solution. After ultracentrifugation, all of the proteins were found to be stable, as >80% were recovered in the supernatant. The stability of the association with the nucleotide was also assessed as described (23). After 1 h of incubation at 37 °C, >80% of the proteins had kept their nucleotide. Finally, we tested whether the introduction of the mutations affected Arf6 interaction with the membrane. Sedimentation experiments with azolectin vesicles established that all mutants were capable of switching from a GDP-bound soluble form to a GTP-bound membrane-associated form. Altogether, these results demonstrate that the Arf6 mutants are stable GTP- or GDP-bound proteins, that the myristate and N-terminal helix are properly exposed for interacting with the lipids, and that the introduction of the C-terminal hexahistidine tag did not affect the interaction of Arf6 with the membrane. Unlabeled nucleotides and azolectin were from Sigma.

Preparation of Phospholipid Vesicles—Except for experiments with PZA, large unilamellar vesicles from crude soybean lipids (type IIS azolectin) were prepared as described (24). For in vitro studies with PZA, sucrose-loaded liposomes of defined composition were prepared by extrusion through a 0.4-µm pore polycarbonate filter (IsoporeTM, Millipore Corp., Molsheim, France) (25). The composition was 50% egg phosphatidylcholine, 19% liver phosphatidylethanolamine, 5% brain phosphatidylserine, 7.5% liver phosphatidylinositol, 16% cholesterol, and 2.5% phosphatidylinositol bisphosphate (PIP2). Lipids were from Avanti%20Polar%20Lipids">Avanti Polar Lipids, Inc. (Birmingham, AL).

GDP or GTP{gamma}S Dissociation Assay—Myr-Arf6-His proteins (2 µM) were first loaded with [3H]GDP for 30 min at 30 °C or with [35S]GTP{gamma}S for 30 min at 37 °C in buffer containing either 40 µM [3H]GDP (1000 cpm/pmol; Amersham Biosciences) and 70 µM free Mg2+ or 20 µM [35S]GTP{gamma}S (1000 cpm pmol; PerkinElmer Life Sciences) and 1 µM free Mg2+. When maximal loading was reached, the free Mg2+ concentration was raised to 1.5 mM ([3H]GDP) or 1 mM ([35S]GTP{gamma}S) to stabilize the nucleotide. Spontaneous nucleotide dissociation was initiated by the addition of 1 mM unlabeled GDP or GTP{gamma}S, and exchange was catalyzed by the addition of EFA6 (20 and 100 nM) and 1 mM unlabeled GTP. After various incubation times, protein and bound nucleotide were isolated by filtration through nitrocellulose filters and quantitated by liquid scintillation counting.

Sedimentation Assay—Myr-Arf6-His proteins (3 µM) were loaded with GDP with and without AlFx (10 mM NaF and 100 µM AlCl3) for 20 min at 37 °C in the absence or presence of PIP2-enriched liposomes (2 mg/ml) prepared as described above. Where indicated, PZA or the Arf-GTP-binding domain of ARHGAP10 (2 µM) was added and incubated for 15 min at 25 °C. Liposomes were then separated by ultracentrifugation, and the amount of proteins bound to liposomes was quantified by SDS-PAGE, SYPRO orange (Bio-Rad) staining, and densitometric analysis (Fujifilm LAS-3000 fluorescence imaging system).

Fluorescence Measurements—Tryptophan fluorescence was used to follow Arf conformational change upon GTP hydrolysis. Two tryptophans act as intrinsic probes of the conformation of Arf1 or Arf6. Upon GTP hydrolysis, these two aromatic groups move closer to each other, inducing a decrease in the intrinsic fluorescence of Arf (26). All fluorescence measurements were performed with a Shimadzu RF-5000 fluorescence spectrophotometer. GTP-loaded myristoylated Arf1 or Arf6 proteins (0.5 µM) were injected into a fluorescence cuvette containing liposomes of defined composition (0.2 mg/ml, 0.4 µM) in 50 mM HEPES (pH 7.5), 100 mM KCl, 1 mM MgCl2, and 1 mM dithiothreitol. GTP hydrolysis was initiated by the addition of 10 nM ACAP1 or PZA. PZA is a fragment of the Arf GAP ASAP1 that includes a pleckstrin homology domain (which binds PIP2), a zinc finger GAP domain (homologous to the catalytic domain of Arf GAP1), and ankyrin repeats (27). The decrease in fluorescence could be fit to a single exponential. The excitation and emission wavelengths were 297.5 and 340 nm, respectively. An emission bandwidth of 5 nm and small excitation bandwidth of 3 nm were used to maximize the signal/noise ratio.

Pull-down Assay for Arf6-GTP—The activation levels of expressed Arf6 proteins were assayed as described previously (28, 29). 24 h after transfection, BHK cells were lysed at 4 °C in 1 ml of lysis buffer (1% Triton X-100, 50 mM Tris-HCl (pH 8), 100 mM NaCl, 10 mM MgCl2, 0.05% sodium cholate, 0.005% SDS, 10% glycerol, 2 mM dithiothreitol, and protease inhibitors). Lysates were clarified by centrifugation at 13,000 x g for 10 min and incubated with 0.5% bovine serum albumin and 40 µg of GST-ARHGAP10 Arf-binding domain bound to glutathione-Sepharose beads (Amersham Biosciences) for 40 min. The beads were washed three times with lysis buffer, and bound proteins were eluted in 30 µl of SDS sample buffer. The presence of Arf6-GTP was detected by immunoblotting using anti-HA antibody.

Triton X-100 Solubilization and Immunoblotting—BHK cells (10 x 106) were transiently transfected with plasmids encoding the indicated Arf6 proteins. 48 h post-transfection, the cells were lysed in 1% Triton X-100, 20 mM Tris-HCl (pH 7.5), 120 mM NaCl, 1 mM MgCl2, 2 mM EDTA, and protease inhibitors for 10 min at 4 °C. The lysates were centrifuged at 13,000 x g for 1 h at 4 °C to separate the Triton X-100-soluble and -insoluble fractions. Aliquots of the pellet and supernatant fractions were analyzed by SDS-PAGE and immunoblotting using anti-HA antibody to detect the Arf6 proteins.


Figure 1
View larger version (17K):
[in this window]
[in a new window]
  		FIGURE 1.
GDP and GTP dissociation rates of Arf6 proteins. A, myristoylated Arf6 proteins (Arf6WT (•), Arf6(T157N) ({blacksquare}), Arf6(Q37E/S38I) ({diamondsuit}), and Arf6(Q37E/S38I/T157N) ({blacktriangleup})) were loaded with 40 µM [3H]GDP in the presence of 70 µM free Mg2+ for 30 min at 30 °C. The free Mg2+ concentration was then raised to 1.5 mM, and [3H]GDP dissociation initiated by the addition of 1 mM unlabeled GDP. The proteins were isolated on nitrocellulose filters, and the amount of bound nucleotide was quantitated by liquid scintillation counting. The curves were best fit assuming first-order kinetics for nucleotide dissociation and represent the means ± S.E. of three independent experiments. B, the GDP dissociation rate (k exch) for each Arf6 protein is reported, with the -fold GDP dissociation compared with Arf6WT. C, myristoylated Arf6 proteins were loaded with 20 µM [35S]GTP{gamma}S([355S]GTPgS) in the presence of 1 µM free Mg2+ for 30 min at 37 °C. The free Mg2+ concentration was then raised to 1 mM, and [35S]GTP{gamma}S dissociation was initiated by the addition of 1 mM unlabeled GTP{gamma}S.

 
Confocal Immunofluorescence Microscopy—For morphological studies, the cells were exposed to 30 mM NaF and 50 µM AlCl3 for 30 min at 37 °C where indicated. At the end of the incubation period, the cells were fixed in 4% paraformaldehyde and processed for immunofluorescence analysis as described previously (30). For the endocytosis assay, anti-MHCI antibody (10 µg/ml) was prebound to the cell surface at 4 °C for 1 h. The unbound excess was eliminated by three quick washes with ice-cold phosphate-buffered saline (PBS), and the cells were then incubated at 37 °C in Dulbecco's modified Eagle's medium to allow for endocytosis and intracellular trafficking of the anti-MHCI antibody. For recycling, the cells were treated as described above; but after 6 h of internalization at 37 °C, the samples were cooled down, and the anti-MHCI antibody present on the cell surface was stripped by a 2-min acid wash at 4 °C with 150 mM NaCl and 100 mM glycine HCl (pH 2.6). After one quick wash with PBS, the cells were incubated for 1 h at 37°C to allow for recycling of internalized anti-MHCI antibody. At the end of the incubation period, the samples were treated as described above for immunofluorescence analysis. For both endocytosis and recycling assays, the transfected cells expressing exogenous Arf6 were detected with anti-HA antibody 3F10. Confocal microscopy analysis was carried out using a Leica TCS-SP microscope equipped with a mixed-gas argon/krypton laser.

Co-immunoprecipitation—The experiment was performed exactly as described previously (31). Briefly, BHK cells were transfected with the indicated plasmids using FuGENE 6. 24 h post-transfection, the cells were exposed or not to AlFx (10 mM NaF and 100 µM AlCl3) for 30 min at 37 °C and solubilized in Triton X-100 buffer containing 0.5 mM 3,3'-dithiobis(sulfosuccinimidyl propionate) (Perbio Science France SAS, Brebières, France). Lysates were centrifuged, precleared, and incubated overnight at 4 °C with anti-FLAG antibody-agarose beads (Sigma). The washed immunoprecipitates were resolved by SDS-PAGE, and the indicated proteins were detected after immunoblotting by chemiluminescence (ECLTM, Amersham Biosciences) and revealed using the Fujifilm LAS-3000 fluorescence imaging system.

Biotinylation Recycling Assay—HeLa cells grown on 35-mm dishes were washed three times with ice-cold PBS, and cell surface-biotinylated in PBS containing 1 mg/ml succinimidyl 2-(biotinamido)-ethyl-1,3'-dithiopropionate. The cells were then washed, and the biotin was quenched by three quick washes with minimum Eagle's medium/bovine serum albumin (minimum Eagle's medium, 0.6% bovine serum albumin, 20 mM HEPES (pH 7.4), and penicillin/streptomycin) supplemented with 50 mM NH4Cl. The cells were incubated for 6 h at 37 °C in minimum Eagle's medium/bovine serum albumin to allow for cell-surface receptor endocytosis, followed by three washes with ice-cold PBS. Except for control samples, cells were treated with 50 mM glutathione, 50 mM Tris (pH 8.8), 100 mM NaCl, and 0.2% bovine serum albumin (pH 8.6) for 45 min at 4 °C to remove the remaining biotin from the cell surface. The cells were then incubated for different periods of time at 37 °C to allow for cell-surface recycling of pre-internalized biotinylated membrane receptors. The cells were submitted to glutathione treatment to remove the biotin from recycled pre-internalized biotinylated cell-surface receptors. The cells were then solubilized in 1% Triton X-100, 20 mM HEPES (pH 7.4), 2 mM EDTA, 125 mM NaCl, 0.2 mM phenylmethylsulfonyl fluoride, and protease inhibitors for 20 min at 4 °C and spun at 15,000 x g for 30 min in a microcentrifuge; the supernatant was precleared on Sepharose beads; and the biotinylated proteins were recovered after a 60-min incubation with streptavidin-agarose beads. The bound proteins were eluted by vortexing the beads for 10 min at 60 °C in an equal volume of 100 mM dithiothreitol and prepared for SDS-PAGE and immunoblot analysis using anti-MHCI antibody HC10. The controls included a sample for total biotinylation processed for immunoprecipitation immediately after the biotinylation step and a sample to test for the efficiency of the glutathione treatment immediately after biotinylation. The proteins were revealed by chemiluminescence using horseradish peroxidase-coupled secondary antibodies. Quantitation was performed using the Fujifilm LAS-3000 fluorescence imaging system.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
DISCUSSION
 REFERENCES
 
In this study, we analyzed a series of Arf6 mutants in vivo and in vitro. To allow for easy comparison, all of the mutants were analyzed side by side, and the results are presented in the same figures.

Role of the Arf6 GDP/GTP Cycle—To address the importance of the Arf6 nucleotide cycle, we designed a new fast-cycling mutant that would cycle more rapidly than Arf6WT between its GDP- and GTP-bound states. Similar to Phe28 of Ras, Thr157 interacts with the guanine base, and its mutation was predicted to affect the affinity of both nucleotides (32). In a previous study (19), Thr157 was replaced with alanine (Arf6(T157A)) to eliminate the van der Waals interactions existing between the threonine and guanine base. Here, we replaced Thr157 with the large positively charged amino acid asparagine to maximally interfere with the binding of the nucleotide. Essential to the characterization of a fast-cycling mutant is the determination of the spontaneous GDP dissociation rate, which should be increased. We measured the GDP dissociation rate of Arf6(T157N) in the presence of lipids at physiological Mg2+ concentration by monitoring the dissociation of prebound [3H]GDP over time. An internal experimental control included Arf1WT, which is known to have a lower GDP off-rate compared with Arf6WT (32). We observed that the T157N mutation increased the GDP off-rate by nearly 5-fold compared with Arf6WT (Fig. 1, A, • (Arf6WT) and {blacksquare} (Arf6(T157N)); and B). The GTP off-rate was also determined by following the dissociation of prebound [35S]GTP{gamma}S. The experiment was controlled using Arf6WT samples incubated in low magnesium-containing medium or in the presence of the exchange factor EFA6 (Fig. 1C). The GTP{gamma}S off-rate of Arf6(T157N) was found to be similar to that of Arf6WT. However, because the reaction was very slow, a slight difference could not be totally excluded.


Figure 2
View larger version (31K):
[in this window]
[in a new window]
  		FIGURE 2.
Analysis of EFA6 exchange activity and PZA GAP activity on Arf6 proteins. A, myristoylated Arf6 proteins (Arf6WT (•), Arf6(T157N) ({blacksquare}), Arf6(Q37E/S38I) ({diamondsuit}), and Arf6(Q37E/S38I/T157N) ({blacktriangleup})) were loaded with 40 µM [3H]GDP in the presence of 70 µM free Mg2+ for 30 min at 30 °C. The free Mg2+ concentration was then raised to 1.5 mM, and the exchange reaction was initiated by the addition of 20 or 100 nM EFA6 and 1 mM unlabeled GTP. B, the curves were best fit assuming first-order kinetics for nucleotide dissociation and represent the means ± S.E. of three independent experiments. The deducted rates of nucleotide exchange (k exch) are reported. For each Arf6 protein, the -fold stimulation was calculated as the ratio of the EFA6-catalyzed to spontaneous exchange rates. C and E, myristoylated Arf6 mutants and Arf6WT were compared for their sensitivity to ACAP1 or PZA, an Arf GAP protein domain derived from ASAP1. GTP hydrolysis in Arf proteins was initiated by the addition of 10 nM ACAP1 or PZA (indicated by the arrow) and monitored by the correlated variation in intrinsic tryptophan fluorescence over time. D and F, the decrease in fluorescence was best fit to a single exponential. The deducted rates of GTP hydrolysis are reported. These results are representative of at least two independent experiments.

 
Next, we tested the sensitivity of Arf6(T157N) to an Arf6-specific guanine nucleotide exchange factor (GEF) and GAP regulatory factors. The sensitivity of Arf6(T157N) to the Arf6-specific exchange factor EFA6 was tested by following the exchange of preloaded [3H]GDP for nonradioactive GTP over time (Fig. 2A). We found that the mutant was sensitive to EFA6 to the same extent as Arf6WT, with a 10-fold stimulation over the spontaneous exchange rate at 100 nM EFA6 (Fig. 2B). We then tested the sensitivity of Arf6(T157N) to the Arf6 GAP ACAP1 (33). Tryptophan fluorescence was used to follow Arf6 conformational change upon GTP hydrolysis, which was monitored by a decrease in fluorescence (Fig. 2C). Arf proteins have been shown to have very low intrinsic GTPase activity (26), and we found that the T157N mutation did not increase the spontaneous GTPase activity of Arf6 (Fig. 2C, see curve before the addition of the GAP indicated by the arrow). The mutation of Thr157 did not affect sensitivity to ACAP1, as Arf6(T157N) and Arf6WT displayed the same rate of catalyzed GTP hydrolysis (~0.04 min–1) (Fig. 2, C and D). As a negative control for the GAP reaction, we showed that Arf6(Q67L) was completely insensitive to ACAP1. ASAP1, another Arf GAP that is more specific toward Arf1 in vitro (34) but has been shown to affect Arf6 functions in vivo (35), was also tested and gave similar results (Fig. 2, E and F).

Decreasing the affinity for the nucleotides has been shown to affect the stability of Arf6 (23). Thus, before analyzing the effects of the mutant in vivo, its stability was assessed by Triton X-100 extraction. As described previously (23), the Arf6WT and Arf6(Q67L) proteins were found in the Triton X-100-soluble fraction, whereas the unstable aggregated Arf6(T27N) mutant was found in the Triton X-100-insoluble fraction (Fig. 3A). The Arf6(T157N) mutant was also found in the Triton X-100-soluble fraction, suggesting that it is stable when expressed in vivo. In cytosol/membrane fractionation, Arf6(T157N) was recovered mostly in the membrane fraction, indicating that it is mostly a GTP-bound protein in vivo (data not shown). This was confirmed in a pull-down experiment using a GST fusion protein of ARHGAP10, a specific effector of Arf-GTP (29). The proportion of Arf6(T157N) associated with GTP was 10 times that of Arf6WT and 70% that of Arf6(Q67L) (Fig. 3B). Furthermore, quantitation indicated that most of the Arf6(T157N) (60%) was in a GTP-bound form.


Figure 3
View larger version (27K):
[in this window]
[in a new window]
  		FIGURE 3.
In vivo Triton X-100 solubility and nucleotide status of Arf6 proteins. A, BHK cells were transiently transfected with plasmids encoding the indicated Arf6 proteins. 48 h post-transfection, cells were lysed at 4 °C in Triton X-100 extraction buffer. After centrifugation, the soluble and insoluble fractions were analyzed by SDS-PAGE and immunoblotting. The percentages of Arf6 Triton X-100 (Tx-100)-soluble proteins from three independent experiments are reported. B, BHK cells were transfected with plasmids encoding the indicated Arf6 proteins. 24 h post-transfection, cells were solubilized and submitted to a pull-down assay on GST-ARHGAP10 beads. Isolated Arf6-GTP and total Arf6 were quantitated by SDS-PAGE and immunoblotting. The calculated -fold of each protein in a GTP-bound state normalized to Arf6WT is indicated. The results are representative of six independent experiments.

 


Figure 4
View larger version (47K):
[in this window]
[in a new window]
  		FIGURE 4.
GDP/GTP cycle of Arf6 is necessary for the post-endocytic transport of MHCI. A, HeLa cells were transfected or not (Unt.) with plasmids encoding the indicated Arf6 proteins. Cells were incubated with anti-MHCI antibody HC10 for 30 min at 4 °C and then incubated for 6 h at 37°C to allow for endocytosis and intracellular transport. The cell-surface periphery was stripped by an acid wash, and the cells were allowed to recycle pre-internalized MHCI molecules for 1 h. Arf6-positive cells shown at t = 0 or indicated with an arrow (t = 6 h and t = +1 h) were detected with rat anti-HA monoclonal antibody 3F10. Anti-HA and anti-MHCI antibodies were revealed using fluorescently coupled secondary antibodies, and the samples were processed for confocal immunofluorescence analysis. B, HeLa cells were transfected or not with plasmids encoding the indicated Arf6 proteins, and post-endocytic recycling of biotinylated MHCI molecules was analyzed as described under "Experimental Procedures. Immunoblots of a representative experiment in which all cells were analyzed within the same experiment are shown. Tot, total; Glu.; glutathione control. C, the results of the recycling biotinylation assay in three independent experiments were quantitated.

 
Arf6 has been implicated in various intracellular transport pathways, including endocytosis, exocytosis, and recycling (1). We investigated the recycling transport pathway of the molecules MHCI and Tac, which are internalized through a clathrin-independent pathway in HeLa cells. Expression of Arf6(Q67L), but not Arf6WT, has been shown to accumulate MHCI molecules in an intracellular compartment presumably by blocking fusion with the "classical" early endosomal compartment (36, 37). The results suggested that GTP hydrolysis is required to achieve the efficient recycling of MHCI molecules. Thus, we asked whether the fast-cycling mutant Arf6(T157N) would accelerate the recycling of MHCI molecules. We first studied MHCI molecule endocytosis from and recycling to the plasma membrane by immunofluorescence. To follow internalization, anti-MHCI antibody was first prebound to the cell surface for 1 h at 4 °C. The cells were then incubated for an additional 6 h at 37 °C to allow for endocytosis (t = 6 h), and the cell surface was acid-stripped to visualize only the intracellular signal (Fig. 4A). As a control, some samples were directly processed for immunofluorescence before endocytosis (t = 0). In untransfected or Arf6WT-transfected cells, after 6 h of endocytosis, the MHCI molecules were found throughout the endocytic/recycling system, and disperse punctate staining was observed. Consistent with a previous report on cells expressing Arf6(Q67L) (37), we also observed the intracellular accumulation of MHCI molecules in a pack of large vesicles close to the nucleus. In cells expressing Arf6(T157N), there was no evidence of intracellular accumulation. Instead, the cells often displayed a much weaker staining presumably due to increased recycling. Next, we looked at the plasma membrane recycling of pre-internalized anti-MHCI antibody after acid stripping of the cell surface (t =+1 h). In untransfected and Arf6WT-transfected cells, little and incomplete cell-surface staining was visible, with no detectable decrease in intracellular staining indicating limited recycling. Arf6(Q67L) showed a total absence of cell-surface staining, with persistent intracellular accumulation of MHCI molecules confirming its inhibitory effect. In contrast, the periphery of Arf6(T157N) cells was well decorated with MHCI staining, presumably reflecting faster recycling to the cell surface. Similar results were obtained with Tac molecules (data not shown). To quantitate the recycling rate of MHCI molecules, we set up a biotinylation assay using cleavable biotin (Fig. 4B). We found that Arf6WT expression had no detectable effect, whereas Arf6(Q67L) significantly slowed down MHCI recycling (Fig. 4C). On the other hand, as suspected based on immunofluorescence, Arf6(T157N) greatly accelerated the recycling of MHCI molecules (Fig. 4C). These results are all the more significant because, for any given experiment, an average of only 30% of the cells were transfected. Therefore, the quantification underestimated both the inhibitory and stimulatory effects of Arf6(Q67L) and Arf6(T157N), respectively. Altogether, our data show that the full GDP/GTP cycle of Arf6 is required for cell-surface recycling of MHCI molecules.


Figure 5
View larger version (29K):
[in this window]
[in a new window]
  		FIGURE 5.
Arf6WT, but not Arf6(Q37E/S38I), forms a stable complex with PZA and ACAP1 in the presence of AlFx. A, myristoylated Arf6WT was loaded with GDP for 20 min at 37 °C in the absence or presence of AlFx and incubated for 15 min at 25 °C with or without ACAP1 and PIP2-enriched liposomes. After centrifugation, the supernatant (S) and pellet (P) were collected, and the reported percentages of Arf6 and ACAP1 bound to the vesicles were quantified after SDS-PAGE and SYPRO orange staining by densitometric analysis. MW, molecular weight. B, BHK cells were transfected with plasmids encoding FLAG-ACAP1 and HA-Arf6. 24 h post-transfection, the cells were exposed or not to AlFx for 30 min and solubilized, and the lysates were submitted to immunoprecipitation (IP) with anti-FLAG or control antibody (Ab). Co-immunoprecipitation of Arf6 was revealed by immunoblotting with anti-HA antibody. Ig LC indicates the light chain of the immunoprecipitating antibody. Aliquots of total lysates were also analyzed separately to determine the total amounts of transfected FLAG-ACAP1 and HA-Arf6. C, myristoylated Arf6WT and Arf6(Q37E/S38I) were loaded with GDP for 20 min at 37 °C in the absence or presence of AlFx and incubated for 15 min at 25 °C with or without PZA and PIP2-enriched liposomes. After centrifugation, the supernatant and pellet were collected, and the reported percentages of Arf6 and PZA bound to the vesicles were quantified after SDS-PAGE and SYPRO orange staining by densitometric analysis. D, the Arf-binding domain of ARHGAP10 (2 µM) was incubated for 30 min at 25 °C with azolectin liposomes in the absence or presence of GTP-loaded myristoylated Arf6WT or Arf6(Q37E/S38I) (3µM). After centrifugation, the reported percentages of Arf6 and ARHGAP10 bound to the liposomes were quantified after SDS-PAGE and SYPRO orange staining by densitometric analysis. The results are representative of three independent experiments.

 
Role of Arf6 GAPs—That the nucleotide cycle is important implies a critical role for the GAP to achieve at least some of the biological functions of Arf6. Therefore, we decided to explore the role of Arf6 GAPs, which represent possibly the largest group of Arf6-binding partners identified to date (12). We looked for an Arf6 mutation that would disrupt binding to the GAP. Small G proteins interact with their specific partners through their two switch domains (38–41). The only two amino acids dissimilar in the switch I domain of Arf1 (Glu41 and Ile42 and Arf6 (Gln37 and Ser38) have been suspected to account for the different morphological effects produced by exogenous expression of Arf1 and Arf6. When these two residues are exchanged in Arf6, the resulting Arf6(Q37E/S38I) mutant is no longer able to induce the formation of membrane protrusions in the presence of AlFx (36). It was proposed that AlFx acts indirectly on Arf6 through the activation of G{alpha}q and that the double mutation Q37E/S38I prevents the binding of effector(s) to Arf6 (36). However, AlFx has been shown to mimic the {gamma}-phosphate of GTP in its binding site (42) and to stabilize complexes formed between Ras, Cdc42, or RhoA and its cognate GAP within a transition state of the GTP hydrolysis reaction (43–45). Therefore, we hypothesized 1) that AlFx is capable of stabilizing a complex between Arf6 and an Arf6 GAP in vivo and 2) that Arf6(Q37E/S38I) cannot form membrane protrusions in response to AlFx because it cannot bind anymore to the Arf6 GAP. However, we have shown that the single mutation of Ser38 (Arf6(S38I)) decreases the GDP dissociation rate (31). Thus, before testing the ability of Arf6(Q37E/S38I) to bind to Arf6 GAPs, we studied its biochemical properties. In agreement with our previous data on the single mutant Arf6(S38I), the GDP dissociation rate of Arf6(Q37E/S38I) was decreased by 3-fold (Fig. 1, A and B). In addition, the GTP dissociation rate appeared to be unchanged at a physiological concentration of magnesium (Fig. 1C). At 100 nM EFA6, we also observed a 10-fold stimulation of EFA6-induced exchange activity comparable with Arf6WT, suggesting that Arf6(Q37E/S38I) can be loaded in vivo with GTP by a GEF (Fig. 2, A and B).

To test our hypotheses, we first analyzed the sensitivity of Arf6(Q37E/S38I) to the Arf GAP ACAP1 by following changes in tryptophan fluorescence. We observed a significant decrease of ~4-fold in the Arf6(Q37E/S38I) response to ACAP1 (Fig. 2, C and D). Because the double mutation Q37E/S38I converts the switch I domain of Arf6 into an Arf1-like switch I domain, we also tested its response to the presumed more Arf1-specific Arf GAP ASAP1. Using its minimal active PZA fragment, we observed again a similar decrease in Arf6(Q37E/S38I)-catalyzed GTP hydrolysis (Fig. 2, E and F).

The lack of significant GAP activity on Arf6(Q37E/S38I) was a strong indication that the mutant protein could no longer properly bind to either Arf GAP. It is not possible to demonstrate a direct interaction between a small G protein and its specific GAP because the catalyzed hydrolysis of the nucleotide immediately liberates the two proteins. Thus, we thought of using AlFx 1) to determine whether it can stabilize the formation of a complex between Arf6 and its GAPs and 2) to test whether this complex is possible with Arf6(Q37E/S38I). We performed sedimentation experiments to measure the formation of a complex between Arf6 and ACAP1 at the surface of PIP2-enriched lipid vesicles that were harvested after centrifugation and analyzed by immunoblotting (Fig. 5A). As predicted, Arf6WT loaded with GDP was found in the soluble fraction (58%), and ACAP1 was also found to be rather soluble in the absence and presence of AlFx (54 and 55% respectively). However, when mixed together in the presence of AlFx, both ACAP1 (67%) and Arf6WT (64%) translocated into the pellet. These results demonstrate for the first time that AlFx can stabilize a complex between Arf6 and a GAP on a lipid membrane. To assess whether such a complex can form in vivo, we carried out a co-immunoprecipitation experiment (Fig. 5B). 24 h post-transfection, BHK cells expressing a tagged version of Arf6WT (HA-Arf6) and/or ACAP1 (FLAG-ACAP1) were exposed or not to AlFx for 30 min. After cell solubilization, the precleared lysates were submitted to immunoprecipitation with anti-FLAG antibody, and the coprecipitated Arf6 protein was detected by immunoblotting with anti-HA antibody. High amounts of Arf6 were co-immunoprecipitated with ACAP1 in cells treated with AlFx. However, when cells were not exposed to AlFx, no Arf6 was detected in the anti-FLAG-ACAP1 immunoprecipitate. As a control for AlFx, lysates of cells exposed to AlFx were immunoprecipitated with an irrelevant antibody and shown to contain no Arf6. We concluded that cells exposed to AlFx sustain an Arf6-GDP·AlFx·ACAP1 complex. Together with our in vitro experiments, this result demonstrates that AlFx efficiently stabilizes the interaction between Arf6 and the Arf GAP ACAP1.


Figure 6
View larger version (39K):
[in this window]
[in a new window]
  		FIGURE 6.
Formation of large membrane protrusions is stimulated by Arf6(T157N) and inhibited by the double mutation Q37E/S38I. A and B, HeLa cells were transiently transfected with plasmids encoding the indicated Arf6 proteins. 24 h post-transfection, cells were incubated in the absence or presence of AlFx for 30 min at 37 °C. Transfected Arf6 proteins and the actin cytoskeleton were labeled with anti-HA antibody and phalloidin, respectively. The samples were analyzed by confocal immunofluorescence. C, the graph reports the percentages of transfected cells exhibiting membrane protrusions. The quantification is representative of at least three independent experiments in which all cell types were analyzed in parallel. Unt., untransfected.

 


Figure 7
View larger version (36K):
[in this window]
[in a new window]
  		FIGURE 7.
Competition experiment between Arf6(Q37E/S38I) and Arf6WT suggests the existence of Arf6-specific attachment sites at the plasma membrane. HeLa cells were cotransfected with increasing amounts of plasmid encoding Myc-tagged Arf6WT together with a constant amount of plasmid encoding HA-tagged Arf6(Q37E/S38I) and incubated with AlFx for 30 min at 37 °C. A, the ratios of Arf6(Q37E/S38I) to Arf6WT were evaluated by immunoblotting. B, cells were stained with anti-Myc and anti-HA antibodies and phalloidin to detect Arf6WT, Arf6(Q37E/S38I), and polymerized actin, respectively. C, the percentages of total cotransfected cells (N) with well developed membrane protrusions are reported. The quantification is representative of three independent experiments giving comparable results.

 
Thus, we could now test directly the interaction between Arf6(Q37E/S38I) and an Arf GAP. Because the double mutation converts the switch I domain of Arf6 into an Arf1-like domain, we analyzed the binding to the presumed Arf1-specific Arf GAP ASAP1. First, we controlled that Arf6WT could bind to ASAP1. We used the PZA fragment and found that the recruitment of PZA by Arf6 onto the lipid membranes was efficient in the presence of AlFx. In the same experiment, when the double mutant was analyzed, neither Arf6(Q37E/S38I) (49%) nor PZA (39%) was efficiently redistributed in the pelleted vesicle fraction in the presence of AlFx (Fig. 5C). In agreement with our GAP activity assays, these results demonstrate that substitution of Gln37 and Ser38 disrupted the interaction with PZA. We addressed the specificity of this effect by analyzing the binding to the Arf-GTP effector ARHGAP10, which is not a GAP for the Arf proteins (Fig. 5D) (29). In contrast with the two Arf GAPs, ARHGAP10 associated equally well with Arf6WT and Arf6(Q37E/S38I). Our finding suggests that the tandem mutation Q37E/S38I selectively abolishes the interaction between Arf6 and its Arf GAPs, but not ARHGAP10. It will be important to confirm that the tandem mutation does not affect the association of other non-GAP Arf6 effectors.

Next, we asked whether Arf6(Q37E/S38I), which cannot bind to the Arf GAPs, would inhibit the protrusions induced by the fast-cycling mutant Arf6(T157N) (19). To address this question, we introduced the double mutation in the fast-cycling Arf6(T157N) mutant. The biochemical properties of this new protein were similar to those of the double mutant Arf6(Q37E/S38I) notably with respect to its sensitivity to the Arf GEF and Arf GAP proteins (see Figs. 1, 2, 3). In agreement with a previous study (19), Arf6(T157N) expression in HeLa cells induced the formation of large membrane protrusions. The phenotype observed was indistinguishable from the one obtained when Arf6WT-expressing cells were exposed to AlFx (Fig. 6A) in that the percentage of cells displaying protrusions and their numbers/cells were very similar (see quantification Fig. 6C). Expression of Arf6(Q37E/S38I) was incapable of inducing membrane protrusions by itself. Furthermore, as expected for an Arf GAP-deficient binding protein, Arf6(Q37E/S38I) was also incapable of responding to AlFx (Fig. 6B). Most interesting, analysis of the triple mutant Arf6(Q37E/S38I/T157N) revealed that the double mutation Q37E/S38I completely blocked the formation of membrane protrusions as well as any rearrangement of the actin cytoskeleton normally induced by the fast-cycling mutation T157N (Fig. 6, A and B). This result suggests that the double mutation Q37E/S38I prevents binding to the effector responsible for remodeling the cortical actin cytoskeleton and the plasma membrane. Because Arf6(Q37E/S38I) does not bind to the GAP and does not respond to AlFx and because the results with the fast-cycling mutant Arf6(T157N) indicate that the GDP/GTP cycle of Arf6 is critical, our data support the proposal that the GAP acts as an effector.

Arf6 Receptor at the Plasma Membrane—A chimeric protein comprising the switch I and II domains of Arf1 and the C-terminal end of Arf6 has been shown to inhibit the formation of membrane protrusions induced by Arf6WT in the presence of AlFx (36). It was proposed that it acts as a dominant-negative mutant by sequestering downstream effectors. However, because the chimeric molecule possesses the GEF, GAP, and effector interaction domains of Arf1, it is unlikely to inhibit Arf6 by binding to its specific GAPs, GEFs, or effectors. Instead, we suspected that the overexpressed chimeric protein competes with Arf6WT (through domains other than the switch regions) for specific sites of attachment to the plasma membrane where it would encounter its specific partners, including those responsible for the formation of membrane protrusions. This is suggested by the following observations. 1) We observed that AlFx reproducibly induced the formation of membrane protrusions in a small fraction of untransfected control cells presumably by stabilizing endogenous Arf6, with its cognate GAP acting as an effector as shown above (Fig. 6C). However, we found this effect to be totally absent in cells overexpressing Arf6(Q37E/S38I) (Fig. 6C). 2) In agreement, the chimeric protein and Arf6(Q37E/S38I) appeared to be targeted to the same domains of the plasma membrane as Arf6WT (this study and Ref. 36). 3) Finally, in cytosol/membrane fractionation experiments, overexpressed Arf6 proteins always displaced endogenous Arf6 from the membrane into the cytosol (data not shown). To test our hypothesis, we examined whether these sites of attachment at the plasma membrane are places where Arf6 is activated and binds its effectors. We performed competition experiments by cotransfecting increasing amounts of Arf6WT together with a constant amount of Arf6(Q37E/S38I) at 1:1, 1:2, and 1:10 ratios controlled by immunoblotting (Fig. 7A). We observed that, in the presence of AlFx, numerous cells expressing higher levels of Arf6WT were capable of forming protrusions, whereas only a few cells expressing lower levels of Arf6WT displayed protrusions (Fig. 7B). After quantification, we found a significant positive correlation between the number of cells with protrusions and the level of expression of exogenous Arf6WT (Fig. 7C). Altogether, our experiments indicate the existence of specific sites of attachment for Arf6-GDP at the plasma membrane where it binds to its GEF and to specific effector(s) responsible for the formation of membrane protrusions. More experiments, including the identification of the receptor, are needed not only to demonstrate its existence, but also to assess the importance of the Arf6-GDP receptor for Arf6 biological functions.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
REFERENCES
 
Over the past 10 years, Arf6 has been implicated in a large variety of cellular functions, including cell adhesion and migration and phagocytosis, by affecting general mechanisms such as cell-surface membrane trafficking, the endocytic/recycling system, and the actin cytoskeleton. In this work, the detailed biochemical and functional analyses of a series of Arf6 mutants and the use of AlFx to stabilize Arf6 in complex with its cognate GAPs provided new insights into Arf6 molecular modes of action. In particular, our study has demonstrated the importance of the completion of the GDP/GTP cycle for Arf6-regulated recycling transport and the role of Arf GAP(s) as effector(s) for Arf6-mediated plasma membrane rearrangement.

Biochemical Analyses of Arf6 Mutants—To analyze the importance of the nucleotide cycle, we designed a so-called fast-cycling mutant by homology to the equivalent Ras mutant. The primary property of a fast-cycling mutant is an increase in its spontaneous GDP dissociation rate, which facilitates the spontaneous cycling between the GDP and GTP conformations. Our Arf6(T157N) mutant displayed a 5-fold in its GDP off-rate. However, its GTP{gamma}S off-rate was not measurably affected and remained similar to that of Arf6WT. Although a slight difference undetectable under our experimental conditions cannot be completely excluded, this result indicates that the contribution of Thr157 is less important for the binding of GTP. Crystallographic analyses have shown that the guanine base-binding pocket is identical in the GDP- and GTP-bound Arf6 forms, suggesting that the {gamma}-phosphate accounts for most of the GTP affinity for Arf6 (32, 46). Our Arf6(T157N) mutant seems to be much more in a GTP-bound form in vivo than the previously published Arf6(T157A) mutant (19). The reasons for this difference are unknown, as the GDP dissociation rate and the sensitivity to the GEF and GAP regulatory proteins have not been analyzed in vitro for this mutant.

The biochemical characterization of Arf6(Q37E/S38I) demonstrated that the protein has a slightly reduced GDP dissociation rate. However, the sensitivity of the protein to the GEF EFA6 is unaffected, indicating that the protein can be normally activated in vivo. Most important, our biochemical experiments demonstrated that the response of this mutant to the Arf GAPs is severely compromised. Thus, Arf6(Q37E/S38I) allowed us to address the role of the Arf GAPs in Arf6-regulated functions. Note that the functional analysis of Arf6(Q37E/S38I) originally led to the proposal that this mutant could discriminate between effectors regulating transport and those involved in actin remodeling (36). However, an exhaustive analysis of the effects of this mutant on all Arf6-regulated pathways has yet to be performed. In fact, in a recent study, the double mutant Arf6(Q37E/S38I) was reported to block beta1-integrin recycling (47). The mechanism of inhibition is unknown, but this double mutation may prevent the binding of an effector directly involved in transport and not actin remodeling.

Completion of the Arf6 GDP/GTP Cycle Is Necessary for MHCI Recycling—Similar to what had been observed for other small G proteins, the GTPase-insensitive (Arf6(Q67L)) and dominant-negative (Arf6(T27N)) mutants often display equivalent modulating effects suggestive of the necessity for Arf6 to cycle from the GTP to the GDP conformation to achieve some of its functions (13–18). Also, Arf6(Q67L) has been shown to block the recycling of non-classically internalized receptors such as Tac, beta-integrin, and MHCI molecules, suggesting that the hydrolysis of the GTP is important for plasma membrane recycling of these molecules (14, 37). To test this hypothesis, we studied the effects of the fast-cycling mutant on the intracellular fate of cell-surface internalized MHCI and Tac molecules. After expression of the GTPase-insensitive Arf6(Q67L) mutant, the MHCI and Tac molecules were found accumulated in an intracellular compartment caused by a block in fusion of the endocytic vesicles with the classical early endosome compartment (this study and Ref. 37). Here, we found that Arf6(T157N) did not block, but instead facilitated the recycling of MHCI as well as Tac molecules (data not shown), thus indicating that the catalytic activity of an Arf6 GAP is necessary for the recycling of the MHCI molecules to the plasma membrane. This is in agreement with a recent report that indirectly suggested a role for ACAP1 in endosome-to-plasma membrane recycling (48). The definitive demonstration of the role of Arf6-GTP hydrolysis in MHCI molecule recycling now requires the identification of the Arf6 GAP involved. We have also concluded that Arf6(Q67L) does not merely mimic a constitutive active phenotype, but also creates a phenotype consistent with an abortive GDP/GTP cycle and therefore, at best, only partially recapitulates Arf6 activation.

Arf6 GAP ACAP1 Acts as an Arf6 Effector for Membrane Protrusion Formation—Arf6 is involved in the formation of membrane protrusions and the rearrangement of the cortical actin cytoskeleton, and several Arf GAPs have been shown to modulate cell morphology (12, 27, 33, 35, 49–52). Here, we propose that the best candidate as an Arf6 effector for actin and plasma membrane remodeling is an Arf6 GAP. At first, we established that AlFx stabilizes a complex between Arf6 and the Arf GAPs ASAP1 and ACAP1 in vitro and in vivo. This result is important, as it helps to explain the actin rearrangement observed when cells exogenously expressing Arf6WT are exposed to AlFx. It also suggests an important role for an Arf GAP in Arf6-regulated cortical actin and plasma membrane remodeling. This is supported by the observation that the phenotype obtained in cells expressing Arf6WT plus AlFx was indistinguishable from that in cells expressing the fast-cycling mutant Arf6(T157N). We propose that, in both cases, a large proportion of the Arf6 GAP is mobilized: in a fast turnover by the fast-cycling mutant Arf6(T157N) and within a stable complex by Arf6WT and AlFx. Furthermore, we have shown that the double mutation Q37E/S38I precluded the formation of an Arf6·Arf GAP complex and that, when introduced into Arf6(T157N), completely prevented the formation of membrane protrusions. Collectively, our observations strongly indicate that an Arf GAP is the effector for Arf6-regulated remodeling of the plasma membrane and actin cytoskeleton. Nevertheless, one cannot rule out that another effector bound simultaneously to Arf6-GTP could also participate. Although we emphasize that the Arf6 GAP acts as an effector, we do not exclude that, in the cell, the Arf6 GAP catalyzes also the hydrolysis of the GTP. However, as suggested by another study, the action of the Arf6 GAP as a catalyst of GTP hydrolysis could be delayed to allow the GAP to accomplish its effector function before GTP hydrolysis and release of the Arf6·Arf6 GAP complex occur (35).

Why does Arf6(Q67L) not recapitulate the effects observed with Arf6WT + AlFx or Arf6(T157N)? Several explanations are possible, and a conclusive answer to this question necessitates the identification of the effector involved in actin remodeling. Based on our results and what is known about the molecular mechanism of GTP hydrolysis, we favor the following interpretation. First, the introduction of AlFx enforces a constraint in the environment of the {gamma}-phosphate that greatly favors the association with the Arf GAP at the expense of the other effectors. Second, when loaded with GTP, the fast-cycling Arf6(T157N) mutant displays an intact GTP-bound state equivalent to that found in Arf6WT, thus allowing for the efficient recruitment and activation of the GAPs that can act as effectors. In contrast, Gln67 in the catalytic domain of the small G proteins has been shown to interact with the catalytic arginine residue provided by the GAP (53). Therefore, the mutation of this glutamine disturbs the environment in the {gamma}-phosphate-binding pocket, disadvantaging the efficient binding of GAPs. This is supported by our observation that Arf6(Q67L)-expressing cells are insensitive to AlFx. The deleterious effect of the glutamine mutation on the binding of the GAP was recently demonstrated for Rab5 (54). Furthermore, the interaction between Rab5 and its cognate GAP (RabGAP-5) is strongest when the glutamine residue of Rab5 and the catalytic arginine of RabGAP-5 are both mutated (54).

Arf6 Receptor at the Plasma Membrane—When cotransfected with Arf6WT, a chimeric Arf1-Arf6 molecule that includes the effector-binding domains of Arf1 (switches I and II) was shown to block the formation of membrane protrusion in response to AlFx (36). It was proposed that the chimera acts by sequestering an Arf6 effector. We find this unlikely, as the chimera possesses the effector-binding domains of Arf1 and not Arf6. Instead, we hypothesize that the protein acts by competing with the endogenous Arf6-GDP protein for a plasma membrane receptor. The results of our cotransfection experiments using increasing amounts of Arf6WT versus a constant amount of Arf6(Q37E/S38I) are consistent with a competition mechanism. Furthermore, the cytosol/membrane fractionation experiments support our model. Indeed, whereas myristoylated Arf6-GDP and Arf1-GDP have the same affinity for membrane lipids in vitro,4 Arf6-GDP is present in a higher proportion in the membrane fraction in vivo (23, 55, 56). In addition, when we performed competitive experiments by transient transfection, an increased proportion of endogenous Arf6 was displaced into the cytosolic fraction by overexpressed Arf6WT, whereas transfected Arf6WT was mostly soluble (data not shown), thus strongly suggesting that the docking site for Arf6-GDP at the membrane is saturable. Our proposal of a receptor for Arf6-GDP is also consistent with the recent description of a protein (TRE17) that binds to the dominant-negative Arf6(T27N) mutant, but not to the GTP-bound Arf6(Q67L) mutant. In cotransfection experiments, TRE17 promotes the translocation of Arf6(T27N) to the plasma membrane and facilitates the activation of exogenous Arf6WT by EFA6 (57). Although these results have yet to be confirmed with endogenous Arf6 and although direct binding has not been demonstrated, the existence of an Arf6-binding protein at the plasma membrane is indicated. Our results suggest the existence of specific sites of attachment for Arf6-GDP at the plasma membrane where it binds to its GEF and to specific effector(s) responsible for the formation of membrane protrusions. We have also shown that the Arf6 GAPs act as effectors for the formation of membrane protrusions. These results support a model in which Arf6, its GEFs, GAPS, and effectors could function within a large complex or in proximity at the plasma membrane. Based on the results with the chimeric protein and the crystallized structures of Arf6-GDP and Arf6-GTP (32, 46), the plasma membrane-binding domain of Arf6 would be within its C-terminal region. The {alpha}3 and/or {alpha}4 helix exposed outside of the molecule and weakly conserved in Arf1 represents the best candidate for an Arf6-binding domain for a plasma membrane receptor. While this manuscript was in preparation, Donaldson and co-workers (58) published a study reporting a sequence contained within the {alpha}3 helix of Arf1 that is required for its Golgi targeting. This sequence appears to be necessary for Arf1-GDP binding to the Golgi SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor) protein membrin, which acts as a membrane receptor for Arf1. Taken together, the results suggest that Arf1 and Arf6 specificity could be supported not only by their GEFs, GAPs, and effectors, but also by their membrane receptors, which specify the discrete intracellular location where Arf6 activation occurs.


   FOOTNOTES
 
* This work was supported in part by a grant from the Association pour la Recherche sur le Cancer (to F. L. and M. F.) and by Cancéropole Provence-Alp-Côte d'Azur Projet AxeIII (to P. C.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

1 Recipient of a fellowship from the Ministère de la Recherche et de l'Education. Back

2 To whom correspondence should be addressed. Tel.: 33-4-9395-7770; Fax: 33-4-9395-7710; E-mail: luton{at}ipmc.cnrs.fr.

3 The abbreviations used are: GAPs, GTPase-activating proteins; MHCI, major histocompatibility complex class I; BHK, baby hamster kidney; HA, hemagglutinin; WT, wild-type; GTP{gamma}S, guanosine 5'-O-(thiotriphosphate); PIP2, phosphatidylinositol bisphosphate; GST, glutathione S-transferase; PBS, phosphate-buffered saline; GEF, guanine nucleotide exchange factor. Back

4 M. Franco, unpublished data. Back


   ACKNOWLEDGMENTS
 
We are grateful to Drs. P. Randazzo and V. Braud for providing the purified Arf GAPs and anti-MHCI antibody HC10, respectively. We are indebted to Dr. K. Singer for stimulating discussions and critical review of the manuscript. We also thank Mariagrazzia Partisani for technical skills.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Chavrier, P., and Goud, B. (1999) Curr. Opin. Cell Biol. 11, 466–475[CrossRef][Medline] [Order article via Infotrieve]
  2. Donaldson, J. G. (2003) J. Biol. Chem. 278, 41573–41576[Free Full Text]
  3. Sabe, H. (2003) J. Biochem. (Tokyo) 134, 485–489[Abstract/Free Full Text]
  4. D'Souza-Schorey, C. (2005) Trends Cell Biol. 15, 19–26[CrossRef][Medline] [Order article via Infotrieve]
  5. Reinstein, J., Schlichting, I., Frech, M., Goody, R. S., and Wittinghofer, A. (1991) J. Biol. Chem. 266, 17700–17706[Abstract/Free Full Text]
  6. Lin, R., Bagrodia, S., Cerione, R., and Manor, D. (1997) Curr. Biol. 7, 794–797[CrossRef][Medline] [Order article via Infotrieve]
  7. Lin, R., Cerione, R. A., and Manor, D. (1999) J. Biol. Chem. 274, 23633–23641[Abstract/Free Full Text]
  8. Kuersten, S., Ohno, M., and Mattaj, I. W. (2001) Trends Cell Biol. 11, 497–503[CrossRef][Medline] [Order article via Infotrieve]
  9. Yang, J. S., Lee, S. Y., Gao, M., Bourgoin, S., Randazzo, P. A., Premont, R. T., and Hsu, V. W. (2002) J. Cell Biol. 159, 69–78[Abstract/Free Full Text]
  10. Tocque, B., Delumeau, I., Parker, F., Maurier, F., Multon, M. C., and Schweighoffer, F. (1997) Cell. Signal. 9, 153–158[CrossRef][Medline] [Order article via Infotrieve]
  11. Symons, M., and Settleman, J. (2000) Trends Cell Biol. 10, 415–419[CrossRef][Medline] [Order article via Infotrieve]
  12. Randazzo, P. A., and Hirsch, D. S. (2004) Cell. Signal. 16, 401–413[CrossRef][Medline] [Order article via Infotrieve]
  13. D'Souza-Schorey, C., Li, G., Colombo, M. I., and Stahl, P. D. (1995) Science 267, 1175–1178[Abstract/Free Full Text]
  14. Radhakrishna, H., and Donaldson, J. G. (1997) J. Cell Biol. 139, 49–61[Abstract/Free Full Text]
  15. Zhang, C. J., Cavenagh, M. M., and Kahn, R. A. (1998) J. Biol. Chem. 273, 19792–19796[Abstract/Free Full Text]
  16. Altschuler, Y., Liu, S.-H., Katz, L., Tang, K., Hardy, S., Brodsky, F., Apodaca, G., and Mostov, K. (1999) J. Cell Biol. 147, 7–12[Abstract/Free Full Text]
  17. Claing, A., Chen, W., Miller, W. E., Vitale, N., Moss, J., Premont, R. T., and Lefkowitz, R. J. (2001) J. Biol. Chem. 276, 42509–42513[Abstract/Free Full Text]
  18. Hashimoto, S., Onodera, Y., Hashimoto, A., Tanaka, M., Hamaguchi, M., Yamada, A., and Sabe, H. (2004) Proc. Natl. Acad. Sci. U. S. A. 101, 6647–6652[Abstract/Free Full Text]
  19. Santy, L. C. (2002) J. Biol. Chem. 277, 40185–40188[Abstract/Free Full Text]
  20. Marshansky, V., Bourgoin, S., Londoño, I., Bendayan, M., and Vinay, P. (1997) Electrophoresis 18, 538–547[CrossRef][Medline] [Order article via Infotrieve]
  21. Duronio, R. J., Jackson-Machelski, E., Heuckeroth, R. O., Olins, P. O., Devine, C. S., Yonemoto, W., Slice, L. W., Taylor, S. S., and Gordon, J. I. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 1506–1510[Abstract/Free Full Text]
  22. Franco, M., Chardin, P., Chabre, M., and Paris, S. (1995) J. Biol. Chem. 270, 1337–1341[Abstract/Free Full Text]
  23. Macia, E., Luton, F., Partisani, M., Cherfils, J., Chardin, P., and Franco, M. (2004) J. Cell Sci. 117, 2389–2398[Abstract/Free Full Text]
  24. Szoka, F., Jr., and Papahadjopoulos, D. (1978) Proc. Natl. Acad. Sci. U. S. A. 75, 4194–4198[Abstract/Free Full Text]
  25. Macia, E., Paris, S., and Chabre, M. (2000) Biochemistry 39, 5893–5901[CrossRef][Medline] [Order article via Infotrieve]
  26. Kahn, R. A., and Gilman, A. G. (1986) J. Biol. Chem. 261, 7906–7911[Abstract/Free Full Text]
  27. Randazzo, P. A., Andrade, J., Miura, K., Brown, M. T., Long, Y. Q., Stauffer, S., Roller, P., and Cooper, J. A. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 4011–4016[Abstract/Free Full Text]
  28. Niedergang, F., Colucci-Guyon, E., Dubois, T., Raposo, G., and Chavrier, P. (2003) J. Cell Biol. 161, 1143–1150[Abstract/Free Full Text]
  29. Dubois, T., Paleotti, O., Mironov, A. A., Fraisier, V., Stradal, T. E., De Matteis, M. A., Franco, M., and Chavrier, P. (2005) Nat. Cell Biol. 7, 353–364[CrossRef][Medline] [Order article via Infotrieve]
  30. Franco, M., Boretto, J., Robineau, S., Monier, S., Goud, B., Chardin, P., and Chavrier, P. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 9926–9931