Identification of a Plasma Membrane-associated Guanine Nucleotide Exchange Factor for ARF6 in Chromaffin Cells

ADP-ribosylation factors (ARFs) constitute a family of structurally related proteins that forms a subset of the Ras superfamily of regulatory GTP-binding proteins. Like other GTPases, activation of ARFs is facilitated by specific guanine nucleotide exchange factors (GEFs). In chromaffin cells, ARF6 is associated with the membrane of secretory granules. Stimulation of intact cells or direct elevation of cytosolic calcium in permeabilized cells triggers the rapid translocation of ARF6 to the plasma membrane and the concomitant activation of phospholipase D (PLD) in the plasma membrane. Both calcium-evoked PLD activation and catecholamine secretion in permeabilized cells are strongly inhibited by a synthetic peptide corresponding to the N-terminal domain of ARF6, suggesting that the ARF6-dependent PLD activation near the exocytotic sites represents a key event in the exocytotic reaction in chromaffin cells. In the present study, we demonstrate the occurrence of a brefeldin A-insensitive ARF6-GEF activity in the plasma membrane and in the cytosol of chromaffin cells. Furthermore, reverse transcriptase-polymerase chain reaction and immunoreplica analysis indicate that ARNO, a member of the brefeldin A-insensitive ARF-GEF family, is expressed and predominantly localized in the cytosol and in the plasma membrane of chromaffin cells. Using permeabilized chromaffin cells, we found that the introduction of anti-ARNO antibodies into the cytosol inhibits, in a dose-dependent manner, both PLD activation and catecholamine secretion in calcium-stimulated cells. Furthermore, co-expression in PC12 cells of a catalytically inactive ARNO mutant with human growth hormone as a marker of secretory granules in transfected cells resulted in a 50% inhibition of growth hormone secretion evoked by depolarization with high K+. The possibility that the plasma membrane-associated ARNO participates in the exocytotic pathway by activating ARF6 and downstream PLD is discussed.

related GTP-binding proteins that form a subset of the Ras superfamily. ARFs are ubiquitous in eukaryotic cells with an amino acid sequence that is highly conserved across diverse species, suggesting a fundamental role in cellular physiology. Initially discovered as cofactors of the cholera toxin-catalyzed ADP-ribosylation of G␣ s (1), ARFs appear to be critical to vesicular trafficking in various subcellular compartments of the cell (2). More recently, members of the ARF family have been shown to activate phospholipase D (PLD) in several cellular systems as well as in isolated membranes (3)(4)(5). The ARF family has been divided into three classes based on size, sequence homology, gene structure, and phylogenetic analysis. Class I ARFs (ARFs 1-3) were initially identified as components of vesicles that originate from the Golgi (6, 7) and the endoplasmic reticulum (8), whereas class III ARF6, which is the most structurally divergent member of the family, has more recently been implicated in the exocytotic and endocytotic pathways at the plasma membrane (9 -12). Little is known about class II ARF4 and ARF5.
Like other G proteins, ARFs cycle between a GDP-bound and a GTP-bound conformation. The GTP-induced conformational change is the "on" signal that permits the ARF proteins to bind to and activate specific protein effectors. Isolated ARFs have little detectable GTPase activity and exchange bound nucleotide very slowly. In cells, their GTPase cycle requires an interaction with GTPase-activating proteins and guanine-nucleotide exchange factors (GEFs) which catalyze the nucleotide exchange activity on ARF. The identification of ARF1 GEFs was facilitated by the discovery that the fungal metabolite brefeldin A (BFA) disrupts Golgi trafficking by inhibiting a Golgi-associated ARF1 exchange factor (13). Several GEF activities have been described, but the breakthrough toward the identification of GEFs acting on ARF proteins was the cloning of two related BFA-sensitive ARF1 GEFs-encoding genes in yeast Saccharomyces cerevisiae, Gea1 and Gea2 (14). This lead to the discovery of cytohesin-1 (15), ARNO (ARF nucleotidebinding-site opener, Ref. 16), and GRP1 (17) which promote guanine nucleotide exchange on ARF1 by a BFA-insensitive catalytic mechanism. Subsequent studies demonstrated that ARNO, GRP1, and cytohesin-1 can also promotes GDP/GTP exchange on ARF6 in both cell free and intact cell assays * This work was supported in part by Association de la Recherche sur le Cancer ARC Grant 9101. 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.
‡ Supported by a fellowship from La Ligue Nationale Contre le Cancer.
¶ To whom correspondence should be addressed. Tel.: 33-3-88-45-67-13; Fax: 33-3-88-60-08-06; E-mail: bader@neurochem.u-strasbg.fr. 1 The abbreviations used are: ARF, ADP ribosylation factor; PLD, phospholipase D; GEF, guanine nucleotide exchange factor; BFA, brefeldin A; ARNO, ARF nucleotide-binding site opener; GH, growth hormone, PBS, phosphate-buffered saline; PA, phosphatidic acid; PEt, 1-palmitoyl-2-oleoyl-sn-3-phosphoethanol; GTP␥S, guanosine 5Ј-O-(thiotriphosphate); PCR, polymerase chain reaction; myrARF6, myristoylated ARF6 protein; RT, reverse transcriptase; D␤H, dopamine-␤hydroxylase; bp, base pair(s). (18 -20). Recently, Franco and co-workers (21) isolated a novel protein, EFA6, which triggers rapid and efficient exchange on ARF6, and it was suggested that EFA6 represents the first identified ARF6-specific GEF. All known ARF GEFs share a conserved central region, the Sec7 domain, responsible for their exchange activity and displaying a high degree of homology with the yeast sec7p, a molecule required for membrane traffic from the yeast (22). Similarities between the ARNO-related GEFs comprise also a coiled coil N terminus and a C-terminal plekstrin homology domain required for phosphoinositide binding and phosphatidylinositol (4,5)P 2 -dependent activation of ARNO exchange activity (16,17). We recently described the presence of a secretory granuleassociated ARF6 protein in adrenal medullary chromaffin cells (11). Secretagogue-evoked stimulation of chromaffin cells triggers the rapid translocation of ARF6 from secretory granules to the plasma membrane and the concomitant activation of PLD in the plasma membrane (23). The observation that both calcium-evoked PLD activation and calcium-induced catecholamine secretion can be inhibited by a synthetic peptide corresponding to the N-terminal domain of myristoylated ARF6, led us to propose that ARF6 participates in the exocytotic reaction by controlling a plasma membrane-bound PLD and the generation of fusogenic lipids at the exocytotic sites (23). Here, we examine the intracellular localization and function of the nucleotide exchange factor for ARF6 in chromaffin cells in order to identify the partners of ARF6 in calciumevoked secretion. Our data reveal the presence of a BFA-insensitive ARNO-related GEF activity for ARF6 in both cytosolic and plasma membrane-bound fractions. Furthermore, the introduction of anti-ARNO antibodies into the cytosol of permeabilized chromaffin cells strongly inhibited calcium-evoked PLD activation and catecholamine secretion. In line with this observation, overexpression of a catalytically inactive ARNO mutant reduced to a similar extent stimulated exocytosis in PC12 cells. We propose that ARNO plays a critical role in the exocytotic pathway in chromaffin cells, by controlling the ARF6 GTPase cycle at the plasma membrane and thereby the ARF6dependent activation of PLD.

[ 3 H]Noradrenaline Release from Permeabilized Chromaffin Cells-
Chromaffin cells were isolated from fresh bovine adrenal glands and maintained in primary culture essentially as described previously (24). Cells were usually cultured as monolayers on either 24 multiple 16-mm Costar plates (Cambridge, MA) at a density of 2.5 ϫ 10 5 cells/well or 100-mm Costar plates at a density of 5 ϫ 10 6 cells/plate. Catecholamine stores were labeled by incubating 3-7-day-old cultured chromaffin cells with [ (23,24). [ 3 H]Noradrenaline release was determined by measuring the radioactivity present in the incubation medium and in cells after precipitation with 10% (w/v) trichloroacetic acid. Release of [ 3 H]noradrenaline is expressed as a percentage of total radioactivity present in the cells before Ca 2ϩ -induced stimulation. In the figures, data are given as the mean of triplicate determinations on three different cell preparations Ϯ S.E.
Transfection of PC12 Cells and Assay of Human Growth Hormone (GH) Release-PC12 cells were grown in Dulbecco's modified Eagle's medium supplemented with glucose (4500 mg/liter) and containing 30 mM NaHCO 3 , 5% fetal bovine serum, 10% horse serum, and 100 units/ml penicillin/streptomycin. Cells were grown in 6-well plastic dishes for release experiments or on poly-D-lysine-coated glass coverslips for immunocytochemistry. Expression plasmids with DNA encoding GH, ARNO, or the ARNO mutant protein were introduced into PC12 cells (80% confluent) using GenePorter (Gene Therapy Systems) according to the manufacturer's instruction. After 5 h of incubation at 37°C, 1 ml of culture medium with fetal bovine serum, horse serum, and antibiotics was added.
Expression of GH, ARNO, or ARNO mutant proteins was assessed after 48 h by immunocytochemistry. Therefore, PC12 cells were fixed for 15 min in 4% paraformaldehyde in 0.12 M sodium/phosphate, pH 7.0, and for a further 10 min in fixative containing 0.1% Triton X-100. Following several rinses with phosphate-buffered saline (PBS), cells were pretreated with 3% bovine serum albumin, 10% (v/v) normal goat serum in PBS to reduce nonspecific staining. Cells were then incubated for 2 h at 37°C with either rabbit polyclonal anti-human GH antibodies or with rabbit polyclonal anti-ARNO antibodies in PBS containing 3% bovine serum albumin, washed, and incubated for 1 h in PBS containing cyanine 2-labeled anti-rabbit secondary antibodies (Amersham Pharmacia Biotech, Les Ulis, France). Coverslips were extensively washed with PBS, rinsed with water, and mounted in Mowiol 4 -88 (Hoechst). Stained cells were monitored using the Zeiss laser scanning microscope as described (25). Transfection efficiency was estimated by counting the number of GH-positive cells among 500 cells in randomly selected areas of the coverslips.
GH release experiments were performed 48 h after transfection. PC12 cells were washed twice with Locke's solution and then stimulated for 10 min with Locke's solution (basal release) or with a high K ϩ solution (Locke's containing 59 mM KCl and 85 mM NaCl). The supernatant was collected and cells were scrapped in 10 mM phosphatebuffered saline. The amounts of the GH secreted into the medium and retained in the cells were measured using a radioimmunoassay kit (Nichols Institute, San Juan, Capistrano, CA). The amount of GH secretion is expressed as a percentage of total GH present in the cells before stimulation.
Assay for Phospholipase D Activity in Cultured Chromaffin Cells-Chromaffin cells were labeled with 1 Ci/ml [9,10-3 H]myristic acid for 24 h at 37°C. Labeled cells were then washed, permeabilized with streptolysin-O, and stimulated with 20 M free Ca 2ϩ in the presence of 1% ethanol. Cells were subsequently collected and lipids extracted and separated with CH 3 OH, CHCl 3 , 0.1 N HCl (1:1:1, v/v). The lower lipidcontaining phase was collected, spiked with a mixture of standard lipids containing L-␣-phosphatidic acid (1,2-diacyl-sn-glycero-3-phosphate) (PA) and 1-palmitoyl-2-oleoyl-sn-3-phosphoethanol (PEt), dried under vaccum, and redissolved in 20 l of CHCl 3 /CH 3 OH (2:1, v/v). Lipids were separated on one-dimensional TLC 0.25-mm oxalate-coated silica gel plates in a solvent system composed of CHCl 3 /CH 3 OH/CH 3  Subcellular Fractionation of Chromaffin Cells-Chromaffin cells were isolated from fresh bovine adrenal glands, homogenized in 0.32 M sucrose, 10 mM Tris, pH 7.4, and then centrifuged at 800 ϫ g for 15 min. After centrifugation at 20,000 ϫ g for 20 min, the pellet containing the crude membrane fraction was resuspended in 0.32 M sucrose (10 mM Tris, pH 7.4), layered on a continuous sucrose density gradient (sucrose 1-2.2 M, 10 mM Tris, pH 7, 4), and centrifuged for 90 min at 100,000 ϫ g to separate the plasma membrane from secretory granules. Twelve 1-ml fractions were collected from the top to the bottom and analyzed for protein content by the Bradford procedure. The distribution of dopamine-␤-hydroxylase activity (D␤H; chromaffin granule marker) and Na ϩ /K ϩ ATPase activity (plasma membrane marker) in the fractions of the gradient was routinely estimated as described previously (25). Fractions were diluted 10 times in 10 mM Tris, pH 7.4, 1 mM EDTA and membranes were collected by centrifugation for 30 min at 100,000 ϫ g. Plasma membranes were purified from fractions 2 and 3 containing the highest Na ϩ /K ϩ ATPase activity and secretory granule membranes from fractions 10 and 11 containing the highest D␤H activity (25). The cytosol and the Golgi membranes were recovered from the 20,000 ϫ g supernatant. After centrifugation at 100,000 ϫ g for 45 min, the supernatant (cytosol) and the pellet (enriched in Golgi membranes) were saved and protein content was determined.
Protein bound radioactivity was determined by nitrocellulose filter trapping (28). Nonspecific binding to nitrocellulose was estimated with 3 M [ 35 S]GTP␥S (ϳ5 ϫ 10 6 cpm) and this value was subtracted from all determinations. Specific binding to ARF6 was determined by measuring, in a parallel set of samples, the binding in the absence of ARF6 and this value was subtracted from the total binding measured in the presence of ARF6. Results shown are representative of three independent experiments.
RNA Isolation and Amplification of cDNA-Total RNA was isolated from cultured chromaffin cells (1.3 ϫ 10 6 cells/assay) according to Chomczynski and colleagues (29). Total RNA (1 g) or purified poly(A) ϩ RNA (50 ng), prepared with the Micro-FastTrack kit from Invitrogen, were transcribed into cDNA with Superscript RNase H Ϫ Moloney murine leukemia virus reverse transcriptase (Life Technologies, Inc.) and an adapter primer (see Table I) from Life Technologies, Inc. or 6-nucleotide random primers. An aliquot (2 l) of the cDNA produced was used as a template for PCR (total volume of 50 l) using Taq polymerase (MBI fermentas) and specific primers. Expression of ARNO cDNA was tested with primers arno. for and arno. rev ( Table I). Expression of EFA6 cDNA was tested with primers efa. for1 with efa. rev1 or 2, and efa. for2 with efa. rev2 (Table I), according to Franco et al. (21). After 5 min of denaturation at 90°C, 35 cycles of the following profile were done: 94°C for 1 min, 40°C or 60°C for 1 min, and 72°C for 2 min. Finally, an extension step of 10 min was added. The PCR products were purified using the Cleanmix Kit (Talent), cloned using the pMOSBlue T-vector Kit (Amersham Pharmacia Biotech), and transformed into Escherichia coli. The plasmids were isolated with Qiagen columns and sequenced on an automated system (Applied Biosystems).
cDNA Cloning and Preparation-The human ARNO cDNA was cloned in the cytomegalovirus-based mammalian expression vector pCB7 as described earlier (19). The E156K mutation in ARNOpCB7 was obtained with the QuickChange Site-directed mutagenesis kit (Stratagene, La Jolla, CA). Sequences of the two constructs were confirmed by restriction digests and DNA sequencing. The cDNA for human GH was provided by Dr. R. W. Holz (Department of Pharmacology, University of Michigan Medical School, Ann Arbor, MI). Plasmids were amplified from 200 ml of overnight growth with the Endofree Maxiprep kit (Qiagen).
Antibodies and Recombinant Proteins-Polyclonal anti-ARNO antibody was raised in rabbits with His 6 -tagged recombinant human ARNO used as antigen (18). Anti-ARNO Fab fragments were prepared by incubating 100 g of IgG with 1 g of papain in a buffer containing 100 mM sodium phosphate, pH 7.5, 2 mM EDTA, 10 mM cysteine for 20 h at 37°C as described (30). Fab fragments were separated from Fc fragments and intact IgGs using protein A-Sepharose (30). Rabbit polyclonal antibodies against dopamine-␤-hydroxylase (D␤H) were prepared in our laboratory and their specificity demonstrated (31). Rabbit polyclonal anti-N-CAM antibodies were kindly provided by Dr. Geneviève Rougon (Laboratoire de Génétique et Physiologie du Developpement, CNRS, Marseille-Luminy, France) and rabbit polyclonal anti-EFA6 antibody (21) was a generous gift from Dr. Philippe Chavrier (Center d'Immunologie INSERM-CNRS, Marseille, France). Rabbit polyclonal anti-human GH antibodies were generously provided by Dr. A. F. Parlow and the NIDDKs National Hormone and Pituitary Program (Torrance, CA).
Recombinant myristoylated ARF6 protein (myrARF6) was produced in a bacterial expression system containing the N-myristoyl transferase in the plasmid pACYC177/ET3d, and purified according to the procedure previously described (32). Recombinant ARNO was expressed as described (18). Recombinant ARNOSec7 domain was cloned in the plasmid pQE30, expressed, and purified as a His 6 fusion protein as described (33).
Protein Determination, Electrophoresis, and Immunoblotting-Protein concentration was routinely determined using the Bradford procedure with Bio-Rad Dye reagent and bovine serum albumin as standard. One-dimensional SDS-gel electrophoresis was performed on 12% acrylamide gels in Tris glycine buffer. Proteins were transferred to nitrocellulose sheets at a constant current of 25 mA for 1 h. Blots were developed with secondary antibodies coupled to horseradish peroxidase (Amersham Pharmacia Biotech) and immunoreactive bands were detected with the ECL system (Amersham Pharmacia Biotech).

Subcellular Localization of the Nucleotide Exchange Activity
for ARF6 in Chromaffin Cells-To determine the intracellular localization of the ARF6 GEF activity in chromaffin cells, we measured the ability of various subcellular fractions to catalyze the binding of guanosine 5Ј-(␥-thio)triphosphate ([ 35 S]GTP␥S) on recombinant myristoylated ARF6 (myrARF6). The common characteristic of the ARF exchange factors is the centrally located Sec7 domain essential for GEF activity. Fig. 1A illustrates the time course for nucleotide exchange on myrARF6 catalyzed by the recombinant ARNOSec7 domain. At a physiological concentration of Mg 2ϩ (5 mM), ARNOSec7 promoted a rapid binding of [ 35 S]GTP␥S to myrARF6 which reached a maximal rate within 10 min. Using these experimental conditions, we measured the nucleotide exchange activity present in the cytosol or associated with the plasma membrane or secretory granule membrane prepared by subcellular fractionation of cultured chromaffin cells. As shown in Fig. 1B, chromaffin granule membranes were unable to catalyze nucleotide exchange on myrARF6. In contrast, substantial GEF activity was found in both cytosolic and plasma membrane-bound fractions (Fig. 1B). Note that boiling completely abolished the ARF6 GEF activity detected in these two fractions (data not shown). These results suggest that the GDP-GTP exchange factor for ARF6 is localized both on the plasma membrane and in the cytosol of chromaffin cells.
Exchange factors for ARF proteins can be divided into two classes depending on their size and susceptibility to brefeldin A of their Sec7 domain: the large ϳ200-kDa GEFs (Sec7, Gea1, and Gea2) are inhibited by brefeldin A, whereas those of the cytohesin family (ϳ55 kDa) are brefeldin A-insensitive (34,35). To further characterize the type of ARF6 GEF activity present on the plasma membrane and in the cytosol in chromaffin cells, we examined the effect of brefeldin A. We found that the nucleotide exchange on myrARF6 catalyzed by either the plasma membrane or the cytosol was not significantly affected by the addition of 300 M brefeldin A (Fig. 1B). This observation indicates that the ARF6 GEF activity present in chromaffin cells is likely to belong to the cytohesin family.
Presence and Distribution of ARNO in Chromaffin Cells-The brefeldin A-insensitive GEFs include ARNO, cytohesin-1, GRP1/ARNO3, and the more recently identified, EFA6. Since both ARNO and EFA6 have been reported to catalyze nucleotide exchange on ARF6, we investigated the presence of their mRNA in chromaffin cells using RT-PCR. cDNA from bovine chromaffin cells were prepared from total RNA using random primers or the oligo(dT) adapter primer ( Table I). Expression of ARNO and EFA6 was then tested by PCR with the specific primers described in Table I. An intense band of 1200 bp, corresponding to the size of the ARNO cDNA ( Fig. 2A) was detected with an annealing temperature of 60°C in both cDNA preparations. Sequencing of this band confirmed that it corresponded to ARNO and not to cytohesin-1 or GRP1/ARNO3 (data not shown). The bovine ARNO cDNA sequence presents 96% identity with the published human ARNO sequence (16,18), indicating a high degree of conservation across species of the ARNO gene. In contrast, we were not able to detect any product corresponding to the entire EFA6 cDNA (expected size of 1970 bp) or to EFA6 fragments (5Ј-terminal 1250-bp or 3Ј-terminal 720-bp fragments) with the PCR conditions described above or when the annealing temperature was decreased to 40°C. Moreover, expression of EFA6 mRNA was not detected by RT-PCR when purified poly(A) ϩ RNA were used.
The intracellular localization of ARNO in various subcellular fractions of chromaffin cells was analyzed by immunodetection on nitrocellulose sheets (Fig. 2). We found that ARNO was predominantly localized in the cytosol and in the upper fractions of a sucrose density gradient layered with a crude membrane preparation. The presence of the peak of ATPase Na/K activity (25) and the cell surface adhesion protein N-CAM (Fig.  2D) indicated that these fractions were enriched in plasma membranes. Note that ARNO was not detected in the fractions containing secretory granules identified by the presence of dopamine-␤-hydroxylase (Fig. 2D). To evaluate the portion of ARNO present in the cytosol, cultured chromaffin cells were collected, and the content of ARNO was estimated in three fractions defined as the total homogenate, the cytosol and the membrane-bound compartment by immunoreplica analysis. As illustrated in Fig. 2B, we found that approximately 40% of the total ARNO was present in the cytoplasmic pool. This distribution of ARNO was not significantly modified when subcellular fractions were prepared from cells previously stimulated for 2 min with 10 M nicotine (data not shown). Furthermore, no immunoreactivity with EFA6 antibodies was detected in the cytosol or any of the membrane-bound fractions (Fig. 2C). Taken together, these findings suggest that ARNO represents most likely the nucleotide exchange factor for ARF6 detected in the plasma membrane and in the cytosol in chromaffin cells. Thus, ARNO may be responsible for ARF6 activation during regulated exocytosis.

Effect of Brefeldin A and Anti-ARNO Antibodies on Phospholipase D Activation and Catecholamine Secretion in Chromaffin
Cells-We previously reported that in chromaffin cells, catecholamine secretion requires the transient activation of an ARF6-regulated PLD associated with the plasma membrane. To probe the functional importance of ARF6 GEF in the exocytotic reaction, we first examined the effect of brefeldin A on Ca 2ϩ -evoked PLD activation and catecholamine secretion in permeabilized chromaffin cells. In the range of concentrations tested (0 -50 M), brefeldin A modified neither the basal nor the calcium-evoked [ 3 H]noradrenaline release (data not shown). It also had no effect on either basal or calcium-stimulated PLD activity as estimated by the formation of [ 3 H]PEt (data not shown). Thus, the exchange factor required in the exocytotic machinery for ARF6 activation and subsequent PLD stimulation is likely to belong to the brefeldin A-insensitive cytohesin family. This finding correlates well with the presence of ARNO at the plasma membrane in chromaffin cells.

FIG. 2. Expression and subcellular distribution of ARNO in chromaffin cells.
A, amplification by RT-PCR of ARNO cDNA from chromaffin cell total RNA using the primers arno. for and arno. rev (see Table I). 12 l of the total PCR products were applied on a 1.3% agarose gel. A single band of 1200 bp corresponding to the ARNO cDNA was amplified. B, cultured chromaffin cells were collected, homogenized, and processed to separate the cytosol from the crude chromaffin membrane pellet. Fractions were then subjected to protein determination and ARNO immunodetection on nitrocellulose sheets. Values obtained by autoradiography and scanning densitometry analysis are expressed relative to the total amount of ARNO detected in the cell homogenate. Similar results were obtained in three separate experiments. C, proteins (200 g) from cytosol, plasma membranes, and chromaffin granule membranes were separated by SDS-gel electrophoresis and analyzed by immunoblotting with anti-EFA6 antibodies. Recombinant EFA6 protein (20 ng) was loaded in an adjacent lane. D, fractions 3-12 (100 g of protein/fraction) collected from a continuous sucrose density gradient layered with the crude chromaffin membrane pellet were subjected to gel electrophoresis and immunodetection on nitrocellulose sheets using either anti-N-CAM antibodies to detect plasma membranes, or anti-D␤H antibodies to detect chromaffin granules or anti-ARNO antibodies. Recombinant His-ARNO (25 ng) was loaded in an adjacent lane for comparison. Note that ARNO is exclusively detected in the fractions containing plasma membranes. To further integrate ARNO in the exocytotic pathway, we investigated the effect of anti-ARNO antibodies on catecholamine secretion in chromaffin cells. When introduced into the cytosol of permeabilized cells, anti-ARNO antibodies induced a dose-dependent inhibition of the Ca 2ϩ -evoked secretory response without modifying basal noradrenaline release, estimated in the absence of calcium (Fig. 3A). Neither preimmune serum nor anti-ARNO antibodies preincubated with an excess of recombinant ARNO, significantly affected calcium-dependent catecholamine secretion (Fig. 3B). It is also interesting to note that the presence of anti-EFA6 antibodies in the cytosol did not reduce the calcium-stimulated secretory activity in streptolysin-O-permeabilized cells (Fig. 3B). To exclude a possible effect resulting from the bivalence of native immunoglobulins, Fab fragments were tested for their effect on secretory activity. As illustrated in Fig. 3C, the incubation of permeabilized chromaffin cells with increasing concentrations of anti-ARNO Fab fragments also resulted in inhibition of calcium-evoked catecholamine secretion, reaching a value comparable to that obtained with native immunoglobulins.
We then examined the effect of the anti-ARNO and anti-EFA6 antibodies on PLD activity in stimulated chromaffin cells. We found a close correlation between their effects on PLD activation and their ability to inhibit catecholamine secretion in stimulated cells. Indeed, only anti-ARNO antibodies were able to inhibit the formation of [ 3 H]PEt and [ 3 H]PA in permeabilized cells stimulated with 20 M free calcium (Fig. 4). In contrast, preimmune serum or anti-EFA6 antibodies had no significant effect on calcium-stimulated PLD activation. These results suggest that ARNO seems to be implicated in regulated secretion in chromaffin cells most likely by activating a pathway leading to the stimulation of PLD. Since both ARNO, ARF6, and the activated PLD are located at the plasma membrane following stimulation (Ref. 23; and the present report), it is tempting to propose that ARNO is the upstream regulator of the ARF6-dependent PLD required for exocytosis in chromaffin cells.
Co-expression of ARNO or a Catalytically Inactive ARNO Mutant with Human GH in PC12 Cells-To further probe the implication of ARNO in the molecular machinery underlying dense-core granule exocytosis, we performed transient transfection studies in PC12 cells. Co-transfection with a plasmid encoding GH has been used by several investigators to study proteins involved in regulated exocytosis (36 -38). GH following PC12 cell transfection is targeted to dense-core granules (Fig.  5A) and it has been established that secretagogue-evoked GH release shows all the characteristics expected for dense-core granule exocytosis (39). In addition, the assay offers the advantage to analyze secretion in only those cells that take up the plasmids. In the present experiments, we estimated by immunocytochemistry with anti-GH antibodies, that 5 to 15% of cells were transfected. Previous studies established that more than 90% of the cells expressing GH also express the second protein (36,37). We have confirmed this high efficiency of co-expression with a plasmid encoding the green-fluorescent protein. Essen- tially all cells expressing GH expressed also green-fluorescent protein as determined by fluorescence microscopy (data not shown). GH expression levels measured by radioimmunoassay were in the range of 5 ng/well. Transfection with plasmids encoding wild type ARNO or the catalytically inactive ARNO (E156K) mutant (19) did not modify total GH levels within cells 2 days after transfection (Fig. 5B). Moreover, overexpression of ARNO or ARNO (E156K) did not affect the intracellular distribution of GH in PC12 cells. GH was essentially stored within secretory granules as revealed by the punctuate immunofluorescent staining pattern observed in cell bodies and neurites in both control and ARNO-transfected PC12 cells (Fig. 5A). Thus, overexpression of ARNO did not apparently affect the synthesis and transport of GH through the early secretory pathway. As illustrated in Fig. 5C, overexpression of ARNO moderately increased the extent of GH release stimulated by a depolarizing concentration of potassium. However, the basal GH release was not affected and the increase of stimulated GH secretion was significant and reproducible in three separate transfection experiments. To examine whether the nucleotide exchange activity of ARNO was required for the exocytotic process, we used the catalytically inactive ARNO mutant previously described as being incapable of catalyzing nucleotide exchange on ARF6 (19). Indeed, we found that overexpression of ARNO (E156K) mutant in PC12 cells did not affect the basal GH release but clearly reduced the high K ϩ -induced GH secretion by around 50% (Fig. 5C). In line with the observation that anti-ARNO antibodies reduce the secretory activity in permeabilized cells, these findings reinforce the idea that ARNO's nucleotide exchange activity is required in the pathway leading to granule exocytosis in neuroendocrine cells. DISCUSSION ARF proteins have traditionally been thought to play a role in the regulation of intracellular membrane trafficking (40). ARF1, which is the best characterized of the six mammalian ARFs, is recruited from the cytosol to the Golgi complex where it mediates the binding of coat proteins and adaptins to Golgi membranes (2,12,41). It is clear from recent work that ARF6 serves distinct functions in eukaryotic cells. In contrast to ARF1 which is generally associated with the Golgi complex, ARF6 has been localized at the plasma membrane where it is likely to modulate some aspects of the vesicular trafficking to and from the cell surface (9,10,12,43). Several lines of evidence also support a unique role for ARF6 in regulating cortical actin structure and function (44,45). In view of the convergence in molecules and mechanisms underlying the multiple steps of intracellular vesicular transport, we previously investigated the possible function of members of the ARF family in calciumregulated exocytosis (11,23). Our findings suggested that a granule-bound ARF6 plays a direct role in the late stages of exocytosis, most likely by stimulating a PLD activity located at the plasma membrane.
Important to our understanding of ARF6 function in the exocytotic pathway will be the identification of the factors which catalyze nucleotide exchange (GEFs) and GTP hydrolysis (GTPase-activating proteins). In the present study, we report the presence of a GEF activity for ARF6 in the cytosol and at the plasma membrane of chromaffin cells. It is worth noting that no GEF activity could be detected on purified secretory granules. This observation suggests that ARF6 is in its inactive GDP-bound conformation when associated with the chromaffin granule membrane. Previous results obtained by chemical cross-linking and coimmunoprecipitation indicated that the membrane receptor that stabilizes the interaction of ARF6 with chromaffin granules is likely to be the G␤␥ subunits of trimeric G protein (11). In line with these results, it has been previously reported that the interaction between ARF1 and G␤␥ is favored when ARF1 is in its GDP-liganded form (46,47). Because ARF6 is apparently not cytosolic in chromaffin cells (11), it may be predicted that ARF6 undergoes its GTPase cycle and becomes activated at the plasma membrane where its GEF appears to be located. This idea correlates well with our previous observations showing the activation of an ARF6-regulated PLD at the plasma membrane in secretagogue-stimulated cells (23). Taken together, our results support a model in which docking of secretory granules to the exocytotic sites at the plasma membrane upon cell stimulation allows ARF6 to transiently switch from the granular G␤␥ complex to the plasma membrane-associated GEF. Formation of GTP-bound ARF6 subsequently activates downstream effector(s) in the exocytotic machinery, in particular the ARF6-regulated PLD present at the plasma membrane (23). In view of the strong inhibition induced by the synthetic N-terminal domain of ARF6 on both PLD activation and catecholamine secretion in permeabilized chromaffin cells (11,23), ARF6-dependent activation of the plasma membrane-associated PLD is likely to represent a key event in the exocytotic pathway.
We report here that the brefeldin A-insensitive GEF, ARNO, colocalizes with ARF6 at the plasma membrane in stimulated FIG. 5. Effect of the overexpression of wild-type ARNO or catalytically inactive ARNO(E156K) mutant on secretion of co-expressed GH from PC12 cells. PC12 cells were transfected (4 g/well of each plasmid) with pCB7 (Control), pCB7-ARNO, or pCB7-ARNO (E156K) plasmids along with plasmid (4 g/well) encoding GH. 48 h after transfection, cells were washed and subsequently stimulated for 10 min with a depolarizing concentration of K ϩ . Basal release was estimated by incubating cells for 10 min with Locke's solution. A, intracellular distribution of GH in transfected PC12 cells. Immunofluorescence confocal micrographs with anti-GH antibodies (diluted 1:100) detected with cyanine 2-conjugated anti-rabbit antibodies (diluted 1:2000). Optical sections were taken through the center of the nucleus. Note the punctuate staining pattern in the cell extensions consistent with the storage of GH within secretory granules. B, total GH content per well estimated by radioimmunoassay (n ϭ 6). C, GH secreted into the extracellular medium and expressed as a percentage of total GH level present in each well before stimulation. The K ϩ -evoked secretory response was obtained by subtracting the basal release from the release measured in the presence of 59 mM K ϩ . Control, ARNO-, and ARNO (E156K)-transfected PC12 cells incubated in Locke's solution released within 10 min 13.6 Ϯ 0.4, 13.3 Ϯ 0.3, and 13.2 Ϯ 0.3% of the total GH, respectively. Data are given as the mean values Ϯ S.E. (n ϭ 6). Similar results were obtained in three independent experiments done with different cell cultures. *, p Ͻ 0.001 when tested by Student's test. chromaffin cells. Indeed, both PLD activation and catecholamine secretion in permeabilized chromaffin cells are insensitive to brefeldin A. In agreement, recent work indicates that brefeldin A disrupts the Golgi membranes, but does not affect the pattern of catecholamine release monitored by carbon fiber amperometry nor does it modify the calcium sensitivity, granule mobilization, or initial rate of exocytosis in single rat chromaffin cells (48). ARNO-related proteins, first characterized as efficient exchange factors for ARF1 (16,49) have subsequently been shown to regulate ARF6 activity and functions in intact cells (19,20). In the present study, we examined whether an exchange factor specific for ARF6 might be expressed in chromaffin cells. EFA6, a protein mainly expressed in brain (50), promotes nucleotide exchange preferentially on ARF6 (21). EFA6 was an attractive candidate to play the role of an ARF6-specific GEF in the exocytotic machinery since studies in transfected HeLa cells indicate that the protein is preferentially localized in the subplasmalemmal region where it seems to regulate membrane trafficking and the organization of peripheral actin filaments (21). However, EFA6 was not detectable in chromaffin cells both in RT-PCR and Western blot expression studies. Furthermore, the introduction of anti-ARNO antibodies, but not anti-EFA6 antibodies, into the cytosol of permeabilized chromaffin cells, inhibited to a similar extent PLD activation and catecholamine secretion in calciumstimulated cells. Consistent with this result, overexpression of a catalytically inactive ARNO mutant reduced the secretory activity in transfected PC12 cells. Collectively, these findings strongly suggest that the plasma membrane-associated ARNO described here represents the endogenous GEF for ARF6 in the exocytotic pathway in chromaffin cells.
In contrast to a recent report describing ARNO mostly as a cytosolic protein in HeLa cells (51), we found here that a large portion of ARNO is associated with the plasma membrane in chromaffin cells. This localization of ARNO was apparently not modified in secretagogue-stimulated cells. An obvious candidate for targeting ARNO to the plasma membrane is its Cterminal plekstrin homology domain (16). Plekstrin homology domains are generally thought to mediate recruitment of proteins to membrane surfaces through their interaction with phosphoinositides (52). Plekstrin homology domains have been subdivided into classes based on their affinity for different polyphosphoinositide species (53). Interestingly, recent work indicates that ARNO exhibits a much higher affinity for phosphatidylinositol 3,4,5-trisphosphate than for either phosphatidylinositol 4,5-bisphosphate or phosphatidylinositol 3,4bisphosphate in vitro (54,55). In adipocytes, insulin-dependent translocation of ARNO to the plasma membrane requires phosphatidylinositol 3-kinase activity (55). These observations fit well with our previous findings that a phosphatidylinositol 3-kinase present in the subplasmalemmal region in chromaffin cells plays a critical function in the pathway leading to catecholamine secretion (56). Thus, it will be quite interesting to determine whether there is regulatory cross-talk between ARNO and the generation of 3-phosphorylated inositides in the plasma membrane during exocytosis.
The present observations that ARNO antibodies inhibited in parallel PLD activation and catecholamine secretion in permeabilized chromaffin cells are consistent with the predicted importance of the ARNO/ARF6/PLD cascade in the exocytotic machinery in neuroendocrine cells (23). PLD hydrolyzes phosphatidylcholine to generate phosphatidic acid and choline. The conversion of phosphatidylcholine to phosphatidic acid in the plasma membrane replaces a nonfusogenic phospholipid with a fusogenic one, a potentially positive lipid modification that may allow the membrane fusion machinery to function. On the other hand, the fact that an ARF6-dependent PLD activation accompanies the exocytotic reaction does not necessarily preclude the occurrence of some other downstream effector(s) for ARNO/ ARF6 in the regulated secretory pathway. ARF6 is unique in its ability to regulate the structure of the cortical actin cytoskeleton (44,45). Furthermore, overexpression of ARNO in HeLa cells promotes a protein kinase C-dependent remodeling of the cortical actin cytoskeleton, similar to those observed in cells expressing GTP-bound ARF6 (19). In secretory cells, actin filaments form a cortical barrier that excludes the majority of secretory granules from the subplasmalemmal zone. Secretagogue-evoked stimulation produces a specific reorganization of the actin network, which is a prerequisite for exocytosis enabling docking and fusion of secretory granules with the plasma membrane (31,57,58). Interestingly, activation of protein kinase C by phorbol ester treatment disrupts cortical F-actin and increases both the number of granules within the subplasmalemmal zone and the initial rate of stimulated catecholamine secretion in chromaffin cells (58). In view of their actin remodeling properties, ARNO/ARF6 represent attractive candidates capable of mediating actin rearrangements underlying membrane trafficking at the site of exocytosis. One molecule that has been shown to play a role in the ARF6-induced cytoskeletal reorganization is the Rac1-interacting protein POR1 (44). The occurrence of POR1 in secretory cells is currently unknown, but would represent an interesting subject of investigation in view of the coordinated control apparently exerted by ARF6 and Rho-related GTPases on regulated exocytosis (42, 59 -61).