ACAP3 (ArfGAP with coiled-coil, ankyrin repeat and pleckstrin homology domains 3) belongs to the ACAP family of GAPs (GTPase-activating proteins) for the small GTPase Arf (ADP-ribosylation factor). However, its specificity to Arf isoforms and physiological functions remain unclear. In the present study, we demonstrate that ACAP3 plays an important role in neurite outgrowth of mouse hippocampal neurons through its GAP activity specific to Arf6. In primary cultured mouse hippocampal neurons, knockdown of ACAP3 abrogated neurite outgrowth, which was rescued by ectopically expressed wild-type ACAP3, but not by its GAP activity-deficient mutant. Ectopically expressed ACAP3 in HEK (human embryonic kidney)-293T cells showed the GAP activity specific to Arf6. In support of this observation, the level of GTP-bound Arf6 was significantly increased by knockdown of ACAP3 in hippocampal neurons. In addition, knockdown and knockout of Arf6 in mouse hippocampal neurons suppressed neurite outgrowth. These results demonstrate that ACAP3 positively regulates neurite outgrowth through its GAP activity specific to Arf6. Furthermore, neurite outgrowth suppressed by ACAP3 knockdown was rescued by expression of a fast cycle mutant of Arf6 that spontaneously exchanges guanine nucleotides on Arf6, but not by that of wild-type, GTP- or GDP-locked mutant Arf6. Thus cycling between active and inactive forms of Arf6, which is precisely regulated by ACAP3 in concert with a guanine-nucleotide-exchange factor(s), seems to be required for neurite outgrowth of hippocampal neurons.
The small GTPases Arfs (ADP-ribosylation factors) function as a molecular switch in signal transduction systems by cycling between GDP-bound inactive and GTP-bound active forms. This cycle is precisely regulated by GEFs (guanine-nucleotide-exchange factors), which facilitate exchange of GDP for GTP on Arfs, and GAPs (GTPase-activating proteins), which stimulate GTPase activity of Arfs to hydrolyse GTP on Arfs to GDP . In general, Arfs exist as the GDP-bound inactive form in the non-stimulated resting state of cells. In response to agonist stimulation of the cell, Arfs are converted into the GTP-bound active form by the action of GEFs, then regulate activities or subcellular location of their effectors to transduce signals downstream. Thereafter, Arfs hydrolyse the bound GTP to GDP by its GTPase activity stimulated by GAPs, thereby becoming inactive.
The mammalian Arf family consists of six isoforms, Arf1–Arf6, which are classified into three classes based on their amino acid sequence similarity. Classes I and II of Arfs, which include Arf1–Arf3 and Arf4 and Arf5 respectively, localize at the endoplasmic reticulum and Golgi, and play an important role in vesicular trafficking between these intracellular organelles [1,2]. On the other hand, Arf6 which is the sole member of class III primarily localizes to the plasma membrane and endosomal compartments, and regulates membrane dynamics-based cellar events such as endocytosis, exocytosis, endosomal recycling, membrane ruffling and cytokinesis through reorganization of actin cytoskeleton [3,4].
We have previously demonstrated that the active form of Arf6 activates the lipid-metabolizing enzyme PIP5K (phosphatidylinositol 4-phosphate 5-kinase) to produce the versatile phospholipid PtdIns(4,5)P2, which in turn induces membrane ruffling in EGF (epidermal growth factor)-stimulated HeLa cells . PLD (phospholipase D), which produces the signalling lipid PA (phosphatidic acid), is also a downstream effector for the active form of Arf6 [6,7]. In a physiological setting, however, cycling between inactive and active forms of Arf6 seems to be required for appropriately regulating cellular events. This notion is supported by the report showing that the Arf6 GAP SMAP1 (small ArfGAP1) regulates clathrin-mediated endocytosis . In addition, it has been reported that ectopic expression of a constitutively active Arf6 mutant or overexpression of the Arf6 GEF EFA6 (exchange factor for Arf6) inhibits transferrin uptake/internalization [9,10], implying that hyperactivation of Arf6 could disrupt the Arf6-regulated cell events. Thus GAPs as well as GEFs could play an important role in the signal transduction to exert cell functions through Arf6.
To date, 31 Arf GAPs have been identified and classified into ten families based on their sequence similarity . One of the Arf GAP families, ACAP (ArfGAP with coiled-coil, ankyrin repeat and pleckstrin homology domains) family, includes three isoforms: ACAP1–ACAP3. ACAP1 and ACAP2 have been identified as Arf6-specific GAPs, and regulate reorganization of actin cytoskeleton . ACAP1 is involved in the recycling of β1 integrin and GLUT4 (glucose transporter 4) from the endosome to the plasma membrane [13–15]. ACAP2 is involved in FcγR-mediated phagocytosis by macrophages and in neurite outgrowth of PC12 cells and hippocampal neurons [16,17]. However, specificity of the ACAP3 GAP activity to Arf isoenzymes and its physiological functions at the cellular level remain unknown.
In the present study, we investigated physiological functions of ACAP3 and specificity of its GAP activity to Arf isoforms. The results obtained demonstrate that ACAP3 positively regulates neurite outgrowth through its GAP activity specific to Arf6 in mouse hippocampal neurons. In addition, our results suggest that cycling between active and inactive states of Arf6, which is regulated by ACAP3 and unidentified GEF(s), is required for promoting neurite outgrowth.
MATERIALS AND METHODS
cDNAs for ACAP2, ACAP3 and ArfGAP1 were amplified from total RNA prepared from immortalized mouse embryonic endothelial cells by RT (reverse transcription)–PCR, and inserted into the mammalian expression vectors pEGFP-C2 (Clontech) and pCAGGS (a gift from Dr J. Miyazaki, Osaka University, Osaka, Japan)  for transfection into HEK (human embryonic kidney)-293T cells and primary cultured hippocampal neurons respectively. cDNAs encoding mouse Arfs, which were gifts from Dr K. Nakayama (Kyoto University, Kyoto, Japan), were subcloned into pcDNA3 (Life Technologies) with a C-terminal HA (haemagglutinin) or Myc tag. GAP activity-deficient mutants of Arf GAPs (ACAP2 R442Q, ACAP3 R446Q and ArfGAP1 R50K), and active (Q67L), inactive (T44N) and fast cycle (T157A) mutants of Arf6 were generated by PCR-based site-directed mutagenesis. For construction of shRNA plasmids, we employed the expression vector containing dual promoters CMV (cytomegalovirus) and H1. cDNAs encoding GFP and shRNAs were inserted downstream of CMV and H1 respectively. Target sequences of shRNAs were 5′-GGTAGAAACAGATGTGGTTGA-3′ (#1) and 5′-GCACCAAGTGGTGTGGTAATG-3′ (#2) for ACAP3, and 5′-AGCTGCACCGCATTATCAA-3′ (#1)  and 5′-CCAGGAGCTGCACCGCATTAT-3′ (#2) for Arf6. The sequence 5′-CGAATCCTACAAAGCGCGC-3′ was used as a control shRNA. For the rescue experiments, shRNA-resistant ACAP3 cDNAs were produced by mutating four nucleotides in the target region of ACAP3 shRNA#1 without any changes in amino acid sequences.
Generation of Arf6-KO (knockout) mice was described previously . All experiments with mice were conducted according to the Guideline for Proper Conduct of Animal Experiments, Science Council of Japan. The protocols of the experiments were approved by the Animal Care and Use Committee, University of Tsukuba.
Tissue distribution of ACAPs
Tissue distributions of ACAP1–ACAP3 mRNAs and ACAP3 protein were analysed by semi-quantitative RT–PCR and Western blotting respectively.
For semi-quantitative RT–PCR of ACAP1–ACAP3 mRNAs, total RNA was extracted from tissues of C57BL/6J mice with TRIzol® reagent (Life Technologies) and reverse-transcribed with Superscript III (Life Technologies) according to the manufacturer's protocol. PCR primers were 5′-GCAAGTCATC-TGAGATGACGGTCAAGC-3′ and 5′-ACCTCTGCATTGTGG-GTAAGAGCAG-3′ for ACAP1, 5′-AACAGCACGAGGTTGA-AGAGGCTGC-3′ and 5′-CTGGAGAAATCCTTCTGCTGGAT-CG-3′ for ACAP2, 5′-TGGACAAGCTGGTCAAACTGTGCAG-3′ and 5′-TCACCAAGGACAACTCCATGTCCTC-3′ for ACA-P3, and 5′-GAGGGGCCATCCACAGTCTTC-3′ and 5′-CATCA-CCATCTTCCAGGAGCG-3′ for GAPDH (glyceraldehyde-3-phosphate dehydrogenase).
For Western blotting of ACAP3, tissues of C57BL/6J mice were homogenized in a buffer consisting of 20 mM Tris/HCl, pH 7.5, 1 mM EDTA, 1% SDS and protease inhibitor cocktail (Nacalai Tesque). Tissue lysates were sonicated on ice and centrifuged at 1000 g for 20 min at 4°C. Supernatants obtained were subjected to Western blotting.
Cell culture and transfection
To analyse expression levels of shRNA-resistant ACAP3 proteins and specificity of ACAP3 GAP activity to Arf isoenzymes, HEK-293T cells were employed. Cells were cultured in DMEM (Dulbecco's modified Eagle's medium) containing 4.5 g/ml glucose (Nacalai Tesque) supplemented with 10% (v/v) FBS (Life Technologies) and 1% penicillin/streptomycin (Nacalai Tesque) for 2 h, and transfected with plasmids for shRNAs and/or proteins using Lipofectamine™ 2000 (Life Technologies) according to the manufacturer's protocol. After incubation at 37°C for 1.5 days under 5% CO2, cells were subjected to analyses for protein expression and ACAP3 GAP activity.
For the experiments with hippocampal neurons, hippocampi were dissected from ICR WT (wild-type) mice and Arf6-KO mice back-crossed with C57BL/6J  at E (embryonic day) 17 or E18, and treated with 10 units/ml papain and 100 units/ml DNase in DMEM at 37°C for 20 min. The dissociated hippocampal neurons were plated on polyethyleneimine-coated dishes or glass coverslips, and cultured in Neurobasal medium (Life Technologies) supplemented with B-27 (Life Technologies), 0.5 mM L-glutamine and 1% penicillin/streptomycin (Nacalai Tesque). To assess the knockdown efficiencies for ACAP3 and Arf6 (see Figures 3A and 7A respectively), hippocampal neurons were transfected with plasmids encoding shRNAs for these proteins by a Nucleofector™ 2b Device (AAB-1001, Lonza) with Amaxa Mouse Neuron Nucleofector Kit (VPG-1001, Lonza) at DIV (days in vitro) 0, and expression levels of these proteins were determined by Western blotting at DIV5. Under these conditions, transfection efficiency of the plasmid was ∼60% of total cells, which was determined as the percentage of cells expressing GFP. In the experiments for the subcellular location of overexpressed ACAP3 and Arf6, primary cultured hippocampal neurons were transfected with plasmids for HA–ACAP3 and Arf6–Myc using the calcium phosphate method  at DIV4, and subjected to immunocytochemical analysis with anti-HA and -Myc antibodies at DIV6. The ratiometric image and the line profile for fluorescence intensity of HA–ACAP3 over that of co-expressed GFP were obtained with ImageJ software (NIH). For the assay of neurite outgrowth with hippocampal neurons prepared from ICR mice, shRNA plasmid for ACAP3 or Arf6 with or without plasmids encoding WT and mutant ACAP3 or Arf6 were transfected into hippocampal neurons using the calcium phosphate method at DIV0, 1, 4 and 9, and total neurite length and total neurite number were assessed at DIV3, 6, 9 and 12 respectively. In the experiment with hippocampal neurons prepared from WT and Arf6-KO mice, hippocampal neurons were cultured as described above, and neurite outgrowth was assessed at DIV6. In the experiment to investigate the effects of ACAP3 knockdown on the level of the active form of Arfs, cultured hippocampal neurons were co-transfected with plasmids for ACAP3 shRNA and for Arf6–HA using the calcium phosphate method at DIV4, and levels of the active form of Arf6–HA were assessed at DIV9.
Assay for active forms of Arfs
GTP-bound active forms of Arfs were assessed by the pull-down method with GST-conjugated GGA3 (Golgi-associated, γ-adaptin ear-containing, Arf-binding protein 3) (amino acids 1–226)  or LZII (leucine zipper region II) of JNK (c-Jun N-terminal kinase)-interacting leucine zipper protein (amino acids 389–455) , followed by Western blotting.
HEK-293T cells overexpressed with Arfs–HA and GFP–Arf GAPs were lysed in lysis buffer (50 mM Tris/HCl, pH 7.5, 100 mM NaCl, 2 mM MgCl2, 0.1% SDS, 0.5% sodium deoxycholate, 1% Triton X-100, 10% glycerol, 1 μg/ml aprotinin and 1 μg/ml leupeptin). After centrifugation, supernatants were incubated with COSMOGEL® GST-Accept (Nacalai Tesque) pre-conjugated with GST–GGA3 at 4°C for 30 min with rotation. The beads were washed five times with the washing buffer (50 mM Tris/HCl, pH 7.5, 100 mM NaCl, 2 mM MgCl2, 1% NP-40, 10% glycerol, 1 μg/ml aprotinin and 1 μg/ml leupeptin). The active forms of Arfs bound to beads were eluted with SDS/PAGE sample buffer and detected by Western blotting with anti-HA antibody.
For assay with cultured hippocampal neurons, cells transfected with plasmids for Arf6–HA and shRNAs were lysed with the lysis buffer (50 mM Tris/HCl, pH 7.5, 100 mM NaCl, 10 mM MgCl2, 0.5% sodium deoxycholate, 1% Triton X-100 and 10% glycerol) supplemented with protease inhibitor cocktail (Nacalai Tesque), and the active form of Arf6 was pulled down with GST–LZII-conjugated beads. After the beads were washed with the buffer consisting of 50 mM Tris/HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 5 mM MgCl2, 1 mM sodium deoxycholate and 10% glycerol, the active form of Arf6 was eluted with SDS/PAGE sample buffer and detected by Western blotting with anti-HA antibody.
Proteins of samples were separated by SDS/PAGE, and transferred on to PVDF membrane (Millipore). Membranes were blocked with Blocking-One P (Nacalai Tesque) at room temperature for 20 min, and incubated with primary antibody at room temperature for 1.5 h. After washing with 0.05% Tween 20 in PBS, the membranes were incubated with horseradish peroxidase-conjugated secondary antibody (Cell Signaling Technology) at room temperature for 1 h, and washed with 0.05% Tween 20 in PBS. Reacted proteins were visualized with the Chemi-Lumi One enhanced detection reagents (Nacalai Tesque), and signals were detected with a luminescent image analyser LAS-4000 mini (Fujifilm). Primary antibodies used were as follows: anti-HA antibody (3F10, Roche), anti-GFP antibody (598, MBL), anti-ACAP3 antibodies (17570-1-AP, Proteintech, and sc-160232, Santa Cruz Biotehnology), anti-GAPDH antibody (MAB374, Millipore), anti-PSD95 (postsynaptic density 95) antibody (MA1-045, Thermo Scientific) and anti-actin antibody (A2066, Sigma–Aldrich). Rabbit anti-Arf6 antibody was generated as described previously .
In situ hybridization
In situ hybridization was performed as described previously . Brains dissected from P (postnatal day) 14 mice were fixed with 4% (w/v) PFA (paraformaldehyde) in PBS, and embedded in OCT compound. After preparation of cryostat sections, they were fixed with 4% (w/v) PFA in PBS at room temperature for 10 min, washed with PBS and immersed in 0.1 M triethanolamine containing 0.25% acetic anhydride for 10 min. They were then washed three times with PBS and blocked in the pre-hybridization solution [50% formamide, 5× SSC (1× SSC is 0.15 M NaCl/0.015 M sodium citrate), 1× Denhardt's, 250 μg/ml tRNA and 500 μg/ml herring sperm DNA] at 4°C overnight. The cRNA probes were hybridized in the hybridization buffer (50% formamide, 300 mM NaCl, 20 mM Tris/HCl, pH 8.0, 5 mM EDTA, 10 mM Na2HPO4, 10% dextran sulfate, 1× Denhardt's, 500 μg/ml tRNA and 200 μg/ml sperm DNA) at 65°C overnight. The sections were washed four times with 0.2× SSC, pH 7.5 at 65°C for 30 min. After rinsing with 0.2× SSC and buffer A (0.1 M Tris/HCl, pH 7.5, and 0.15 M NaCl), the sections were blocked with 10% (v/v) normal sheep serum in buffer A at room temperature for 1 h, and incubated with the alkaline phosphatase-conjugated anti-DIG (digoxigenin) antibody (Roche) in buffer A supplemented with 1% (v/v) sheep serum at 4°C overnight. They were washed with buffer A supplemented with 0.1% Triton X-100 three times, and developed by incubating with NBT (Nitro Blue Tetrazolium)/BCIP (5-bromo-4-chloro-3-indolyl phosphate) (1:200 NBT/BCIP stock solution, Roche) in a buffer consisting of 0.1 M Tris/HCl, pH 9.5, 0.1 M NaCl and 50 mM MgCl2. Images were obtained using a Biozero BZ-8000 microscope (Keyence).
For the detection of overexpressed proteins in hippocampal neurons, cells were fixed with 4% (w/v) PFA in PBS for 20 min, permeabilized with 0.1% Triton X-100 and 0.1% Tween 20 for 10 min, blocked with 1% (w/v) BSA in PBS at room temperature for 1 h, and stained by sequential incubation with primary antibodies and Alexa Fluor® 488- or Alexa Fluor® 546-conjugated secondary antibodies (Life Technologies). Primary antibodies used were as follows: anti-HA antibody (16B12, Covance), anti-GFP antibody (598, MBL), anti-Myc antibody (562, MBL) and anti-tubulin antibody (T6199, Sigma–Aldrich).
For the detection of endogenous ACAP3 and Arf6 of hippocampal neurons, cells were fixed with 4% (w/v) PFA and 4% (w/v) sucrose in PBS for 20 min, and post-fixed with trichloroacetic acid on ice for 15 min. The fixed neurons were permeabilized with 0.1% Triton X-100 for 5 min, blocked with 2% (w/v) BSA in PBS at 4°C overnight, and stained by sequential incubation with primary antibodies diluted with Can Get Signal Solution B (NKB-601, Toyobo) at 4°C overnight and with Alexa Fluor® 488- or Alexa Fluor® 546-conjugated secondary antibodies (Life Technologies). Primary antibodies used were as follows: anti-ACAP3 antibody (17570-1-AP, Proteintech) and guinea pig anti-Arf6 antibody (a gift from Dr H. Sakagami, Kitasato University, Kanagawa, Japan).
Fluorescence images were acquired with the confocal laser-scanning microscope FV10i (Olympus), confocal microscope Leica TCS SP5 (Leica Microsystems) or BZ-X710 (Keyence). Total neurite length and total neurite number of 33–90 neurons and fluorescence intensities for ACAP3 and tubulin of 40–45 neurons were analysed using ImageJ software.
Each value shows the means±S.E.M. from more than three independent experiments. Statistical significance was calculated by Student's t test or one-way ANOVA with Tukey's HSD (honest significant difference) or Dunnett's multiple comparison tests.
Abundant expression of ACAP3 in the brain
To obtain the information for tissues or cells in which ACAP3 functions, tissue distribution of ACAP3 mRNA in P56 adult mice was analysed by semi-quantitative RT–PCR in comparison with that of ACAP1 and ACAP2 mRNAs (Figure 1A). Consistent with the previous study , ACAP1 mRNA was not detected in the cerebrum and cerebellum, but was abundantly expressed in the lung, spleen, adipose tissue and testis. ACAP2 and ACAP3 mRNAs were ubiquitously expressed. Notably, ACAP3 mRNA was abundant in the cerebrum, cerebellum, heart and stomach. When expression levels of the ACAP3 protein in P56 mouse tissues were analysed by Western blotting, however, extremely higher expression of ACAP3 was observed in the brain compared with other tissues (Figure 1B). Furthermore, it was found that the expression level of ACAP3 in the brain was dependent on the mouse development (Figure 1C): ACAP3 expression was detectable in the brain of E12–E14 mouse embryos and increased with development, reaching a plateau by E16–E18. In situ hybridization analysis in the brain of P14 mice revealed evident expression of ACAP3 mRNA in the cortex, hippocampus and cerebellum (Figure 1D); in particular, the signal in the hippocampus was very strong.
ACAP3 is abundantly expressed in the brain
ACAP3 is involved in neurite outgrowth of hippocampal neurons
Since ACAP3 mRNA was highly expressed in the hippocampus (Figure 1D), it is plausible that ACAP3 functions in hippocampal neurons. In primary cultured mouse hippocampal neurons, expression level of ACAP3 was detectable at DIV6, and increased with prolonged culture, reaching a plateau by DIV12, which was almost parallel to the expression pattern of the postsynaptic maker protein PSD95 (Figure 2A), suggesting that ACAP3 plays a role in the development of hippocampal neurons. When HA–ACAP3 was expressed in hippocampal neurons, it localized in the soma and neurites, especially at the tip of the growth cone as revealed by the ratiometric image and line profile (Figures 2B and 2C). Consistent with this result, endogenous ACAP3 was found to locate at the tips of the growth cone in DIV6–DIV12 hippocampal neurons (Figure 2D), although it was not detectable at DIV3 due to its lower expression levels as shown in Figure 2(A). These results strongly suggest the involvement of ACAP3 in axonal outgrowth.
Expression levels and subcellular localization of ACAP3 in hippocampal neurons
To address this possibility, ACAP3 in hippocampal neurons was efficiently knocked down with shRNAs (Figure 3A). Under these conditions, total neurite length was significantly decreased, whereas total neurite number was not affected in DIV3–DIV12 hippocampal neurons (Figures 3B–3D), supporting the possibility described above that ACAP3 is involved in neurite outgrowth throughout these developing stages, possibly axonal outgrowth, without any effects on initial formations of the neurite.
ACAP3 is involved in neurite outgrowth of hippocampal neurons
GAP activity of ACAP3 is required for regulation of neurite outgrowth
To investigate whether regulation of neurite outgrowth by ACAP3 requires its GAP activity, we conducted rescue experiments of neurite outgrowth inhibited by ACAP3 knockdown with shRNA-resistant WT (WTRES) and the GAP activity-deficient mutant R446Q of ACAP3 (R446QRES), in which the conserved arginine residue necessary for the GAP activity of ACAPs was substituted by glutamine . When these molecules as well as shRNA-non-resistant WT and the R446Q mutant of ACAP3 were expressed in HEK-293T cells, expression levels of WTRES and R446QRES were not significantly disturbed by ACAP3 shRNA, whereas expression of the shRNA-non-resistant molecules was completely suppressed (Figure 4A). Expression of shRNA-resistant molecules was confirmed in ACAP3-knocked-down hippocampal neurons (Supplementary Figure S1). Under these conditions, WTRES, but not R446QRES, rescued neurite outgrowth inhibited by ACAP3 knockdown (Figures 4B and 4C). Thus ACAP3 regulates neurite outgrowth through its GAP activity.
Regulation of neurite outgrowth by ACAP3 requires its GAP activity
GAP activity of ACAP3 is specific to Arf6
The result described above showing that GAP activity of ACAP3 is critical to regulate neurite outgrowth raises a possibility that regulation of neurite outgrowth by ACAP3 is mediated by an Arf isoenzyme(s). To clarify the specificity of ACAP3 GAP activity to Arf isoforms, effects of overexpressed WT and the GAP activity-deficient mutant of ACAP3 (ACAP3 WT and ACAP3 R446Q respectively) on the levels of GTP-bound active form of ectopically expressed Arfs–HA in HEK-293T cells were analysed (Figure 5). Wild-type ArfGAP1 and ACAP2 specific to Arf1 and Arf6 respectively and their GAP activity-deficient mutants (ArfGAP1 R50K and ACAP2 R442Q respectively) were also overexpressed as controls. As reported in [12,25], ArfGAP1 and ACAP2 had GAP activities specific to Arf1 and Arf6 respectively (Figures 5A and 5C). ACAP3 WT and ACAP3 R446Q did not affect the levels of the active form of Arf1 and Arf5, the representatives of classes I and II respectively (Figures 5A and 5B). In contrast, ACAP3 WT significantly decreased the level of the GTP-bound active form of Arf6, whereas ACAP3 R446Q did not (Figure 5C). These results demonstrate that ACAP3 functions as an Arf6-specific GAP.
GAP activity of ACAP3 is specific to Arf6
ACAP3 co-localizes with Arf6 and functions as Arf6 GAP in hippocampal neurons
An expression pattern of endogenous Arf6 during in vitro culture of mouse hippocampal neurons was similar to that of ACAP3: the expression levels of Arf6 increased with prolonged culture and reached a plateau at DIV9–DIV12 (Figure 6A). It has been reported that Arf6 localizes at the cell body, axons and dendrites of hippocampal neurons . Consistent with this earlier paper, Arf6–Myc expressed in hippocampal neurons localized at these compartments and co-localized with co-expressed HA–ACAP3 (Figure 6B). It is noteworthy that co-localization of these two molecules was observed at the neurite tip. This was also true for endogenous Arf6 and ACAP3 (Figure 6C): co-localization of Arf6 and ACAP3 at the neurite tip was obvious at DIV6 and DIV9, although it was not clear at DIV3 and DIV12. Furthermore, it was found that knockdown of ACAP3 in hippocampal neurons increased the level of the active form of Arf6–HA (Figure 6D), demonstrating that ACAP3 has the Arf6 GAP activity in hippocampal neurons. These results, taken together with the results shown in Figure 4, support the notion that ACAP3 regulates neurite outgrowth by controlling Arf6 activity through its Arf6-specific GAP activity.
ACAP3 co-localizes with Arf6 and functions as Arf6 GAP in hippocampal neurons
Regulation of neurite outgrowth by ACAP3 is mediated by Arf6
Finally, we investigated the involvement of Arf6 in neurite outgrowth and the functional relationship between ACAP3 and Arf6. When primary cultured hippocampal neurons were transfected with plasmids for Arf6 shRNAs, the levels of Arf6 were decreased by approximately 30–40% (Figure 7A). Under these conditions, total neurite length, but not total neurite number, was inhibited by ∼50% (Figures 7B–7D). Consistent with these results, total neurite length, but not total neurite number, of hippocampal neurons prepared from Arf6-KO embryos was significantly inhibited (Figures 7E–7H). Thus Arf6 also positively regulates neurite outgrowth without any effects on the number of neurites.
Involvement of Arf6 in neurite outgrowth
Furthermore, it was found that neurite outgrowth suppressed by ACAP3 knockdown was rescued by overexpression of the fast cycle mutant of Arf6, T157A, which spontaneously exchange nucleotides on the mutant T157A , but not by WT and GTP- and GDP-locked Arf6 mutants of Arf6, Q67L and T44N  respectively (Figure 8). These results indicate that Arf6 functions as a downstream effector of ACAP3 in the neurite outgrowth signalling and cycling between active and inactive states of Arf6 is required for neurite outgrowth.
Effect of overexpressed Arf6 mutants on neurite outgrowth in ACAP3-knocked-down hippocampal neurons
In the present study, we demonstrated that the ACAP family member of Arf GAP, ACAP3, has Arf6-specific GAP activity and regulates neurite outgrowth by effectively controlling the cycle of Arf6 between active and inactive forms through its GAP activity in mouse hippocampal neurons.
Although the molecular mechanism by which Arf6 cycling between the inactive and active states regulates neurite outgrowth remains to be elucidated, it is plausible that the PtdIns(4,5)P2-producing enzyme PIP5K functions downstream of ACAP3/Arf6 in neurite outgrowth. This idea is supported by the previous study demonstrating that Arf6 directly activates PIP5K to produce the versatile phospholipid PtdIns(4,5)P2 at the plasma membrane, thereby inducing dynamic actin cytoskeleton remodelling , which is a critical cellular event at the growth cone for neurite outgrowth . If this is the case, a PIP5K isoform regulating neurite outgrowth downstream of Arf6 might be PIP5Kγ687, since this PIP5K isoenzyme is specifically expressed in the brain and positively regulates neuronal development . Regulation of actin cytoskeleton remodelling by PtdIns(4,5)P2 is mediated by actin-modulating proteins such as α-actinin , gelsolin  and cofilin , which regulate polymerization/depolymerization or assembly/disassembly of actin filaments. In addition, PtdIns(4,5)P2 directly binds to and regulates the activity of proteins anchoring actin filaments to the plasma membrane, such as ERM (ezrin/radixin/moesin) proteins  and vinculin , to control the cell adhesion/detachment to/from extracellular matrix. Such a modulation of cell matrix adhesion is critical for regulating dynamic membrane structures such as filopodia and lamellipodia, which are also critical cellular events for neurite outgrowth. Since tandem repeat of adhesion/detachment of filopodia and lamellipodia to/from the extracellular matrix and of actin filament reorganization at growth cones is required for the extension of neurites , it is reasonable to speculate that oscillation of the PtdIns(4,5)P2 level is regulated though cycling Arf6 states by ACAP3 in concert with an unidentified Arf6 GEF(s). Alternatively, another small GTPase Rac1 may function downstream of Arf6 to regulate neurite outgrowth: Arf6 has been reported to activate Rac1 , which plays a pivotal role in actin polymerization and neurite outgrowth . The cycle of active and inactive Arf6 regulated by ACAP3 may lead to oscillation of the Rac1 activity, thereby contributing to the regulation of actin cytoskeleton reorganization at the growth cones.
Another cellular event essential for neurite outgrowth is membrane trafficking or recycling. During neurite extension, membrane proteins are destined to locate at the growth cone membrane and proximal axons, such as NGF (nerve growth factor) receptors [Trk (tropomyosin receptor kinase) receptors] and cell adhesion molecule L1/NgCAM (neuron-glia cell adhesion molecule), should be supplied from endosomes to the plasma membrane of these areas by membrane trafficking or recycling [37,38]. Since it has been reported that Arf6 is involved in membrane trafficking and recycling in many types of cells, Arf6 signalling would regulate neurite outgrowth by controlling these intracellular events. Consistent with this idea, it has been suggested that in PC12 cells, Arf6-mediated membrane trafficking is involved in neurite outgrowth stimulated by NGF: scission of vesicles from recycling endosomes by the key player of vesicle scission EHD1 (EH-domain-containing 1) , whose recruitment to the endosome is regulated by Arf6 , is critical for the neurite outgrowth . In the process of Arf6-dependent recruitment of EHD1, PtdIns4P functions as a key recruiter of EHD1 . Since Arf6 is the activator of the PtdIns(4,5)P2-producing enzyme PIP5K and an Arf6 GAP is involved in the recruitment of EHD1 , it is reasonable to expect that inactivation of Arf6 by Arf6 GAP at the recycling endosomes terminates the activation of PIP5K to decrease the PtdIns(4,5)P2 level and thereby increases PI4P level. Although it has been demonstrated that the Arf6-specific GAP ACAP2 functions in this process , it has not been examined whether ACAP3 is involved in this process. It is of interest to examine whether ACAP3 is also involved in the recruitment of EHD1 to the recycling endosome and regulates recycling of the cargo proteins essential for neurite outgrowth, such as Trk receptors and L1/NgCAM.
ACAP2 as well as ACAP3 is expressed in the brain, whereas ACAP1 is not expressed in the brain (Figure 1A). In addition, both ACAP2 and ACAP3 have an Arf6-specific GAP activity . These results and the report raise a possibility that ACAP2 also positively regulates neurite outgrowth by the same mechanism as ACAP3, consistent with the report that in PC12 cells and hippocampal neurons knockdown of ACAP2 suppresses neurite outgrowth , although we did not investigate the involvement of ACAP2 in neurite outgrowth in the present study. Should this be the case, a question arises as to how ACAP2 and ACAP3 share their functions to regulate neurite outgrowth. Recycling of endosomal proteins to the plasma membrane includes multiple steps, e.g. budding of vesicles from endosomes, transport of the vesicles to the plasma membrane and tethering/fusion of the vesicles to/with the plasma membrane. This fact led us to speculate that different populations of Arf6 regulate distinct steps of the recycling process in neurite outgrowth due to the spatiotemporal regulation of Arf6 by distinct Arf6 GAPs in concert with Arf6 GEFs. This speculation is consistent with the reports suggesting that Arf6 is implicated in several steps for recycling of GLUT4 in differentiated 3T3-L1 adipocytes  and of β1 integrin in vascular endothelial cells . Thus sequential regulation of Arf6 by distinct Arf6 GAPs including ACAP2 and ACAP3 may drive the entire process of membrane recycling. These points should be clarified by further investigation.
In the present study, we emphasize that cycling of Arf6 between active and inactive states is crucial for regulating neurite outgrowth. This resulted from our findings that depletion of Arf6 inhibited neurite outgrowth (Figure 7), ACAP3 positively regulated neurite outgrowth through its Arf6-specific GAP activity (Figures 3B–3D, 4B, 4C and 5) and neurite outgrowth inhibited by ACAP3 knockdown was rescued by expression of the fast cycle mutant of Arf6, but not by that of WT, constitutively active and dominant-negative mutants of Arf6 (Figure 8). This notion is also supported by the report suggesting that hyperactivation of Arf6 interferes with the clathrin-mediated endocytosis: overexpression of the Arf6 GEF EFA6 inhibits transferrin uptake/internalization . Finally, requirement of cycling between active and inactive states of Arf6 is suggested in the membrane trafficking of the membrane-associated protease MT1-MMP (membrane-type 1 matrix metalloproteinase) from the sorting endosome to invadopodia via the late endosome in the breast cancer cell line MDA-MB-231 cells, which is indispensable for extracellular matrix degradation by invadopodia of cancer cells [45,46]. Thus cycling of GTP/GDP-bound Arf6 seems to be important to regulate a wide variety of cell functions.
Yuki Miura, Tsunaki Hongu and Yasunori Kanaho conceived and initiated the project. Yuki Miura performed the majority of the experiments with data analyses, and contributed to writing the paper. Tsunaki Hongu carried out the initial experimental work and designed experimental strategy. Yohei Yamauchi, Yuji Funakoshi, Naohiro Katagiri and Norihiko Ohbayashi interpreted the experimental results, designed the experiments and contributed to the preparation of the paper. Yasunori Kanaho designed the research strategy, interpreted the experimental results and revised the paper.
We are grateful to Dr J. Miyazaki (Osaka University, Osaka, Japan), Dr K. Nakayama (Kyoto University, Kyoto, Japan) and Dr H. Sakagami (Kitasato University, Kanagawa, Japan) for kindly providing pCAGGS expression vector, cDNAs encoding mouse Arfs, and guinea pig anti-Arf6 antibody respectively.
This study is supported by the Ministry of Education, Culture, Sports, and Technology (MEXT) of Japan [grant numbers 15K07039 (to N.O.) and 15H02503 (to Y.K.)], the Kao Melanin Workshop (to N.O.), the Cosmetology Research Foundation (to N.O.) and the Program for Leading Graduate Schools of the Japan Society for the Promotion of Science (to Y.M.).
ArfGAP with coiled-coil, ankyrin repeat and pleckstrin homology domains
days in vitro
Dulbecco's modified Eagle's medium
exchange factor for Arf6
Golgi-associated, γ-adaptin ear-containing, Arf-binding protein 3
glucose transporter 4
human embryonic kidney
honest significant difference
leucine zipper region II
Nitro Blue Tetrazolium
neuroglia cell adhesion molecule
nerve growth factor
phosphatidylinositol 4-phosphate 5-kinase
postsynaptic density 95
These authors contributed equally to this work.