Physical and functional interaction of rabphilin-3A with alpha-actinin.

Rabphilin-3A is a downstream target molecule of Rab3A small GTP-binding protein and implicated in Ca2+-dependent neurotransmitter release. Here we have isolated a rabphilin-3A-interacting molecule from a human brain cDNA library by the yeast two-hybrid method and identified it to be α-actinin, known to cross-link actin filaments into a bundle. α-Actinin interacts with the N-terminal region of rabphilin-3A, with which GTP-Rab3A interacts, and this interaction stimulates the activity of α-actinin to cross-link actin filaments into a bundle. The interaction of rabphilin-3A with α-actinin is inhibited by guanosine 5′-(3-O-thio)triphosphate-Rab3A. These results suggest that the Rab3A-rabphilin-3A system regulates the α-actinin-regulated reorganization of actin filaments. It has been shown that reorganization of actin filaments is also involved in Ca2+-dependent exocytosis. Therefore, rabphilin-3A may serve as a linker for Rab3A and cytoskeleton.

The Rab small GTP-binding protein family, consisting of more than 30 members, is implicated in intracellular vesicle trafficking (for reviews, see Refs. [1][2][3]. The Rab3 subfamily consists of four members, Rab3A, -B, -C, and -D. Of these members, Rab3A and -C have been shown to be implicated in Ca 2ϩ -dependent exocytosis, particularly in neurotransmitter release (for a review, see Ref. 4). We have identified a downstream target molecule of the Rab3 subfamily members that specifically interacts with their GTP-bound form and named it rabphilin-3A (5,6). Structural and biochemical analyses indicate that rabphilin-3A has at least two functionally different domains: the N-terminal GTP-Rab3-binding domain and the C-terminal two C2-like domains that interact with Ca 2ϩ and phospholipid (6,7). Tissue and subcellular distribution analyses of rabphilin-3A indicate that it is specifically expressed in neuronal cells where it is highly concentrated on synaptic vesicles (6,8,9). Overexpression of the N-terminal or C-terminal fragment of rabphilin-3A in bovine adrenal chromaffin cells and PC12 cells (10,11) or microinjection of these fragments into the presynaptic nerve terminal of squid giant axon 1 has been shown to inhibit Ca 2ϩ -dependent exocytosis, suggesting that rabphilin-3A as well as Rab3A is involved in Ca 2ϩ -dependent exocytosis. However, it still remains to be clarified how Rab3A and rabphilin-3A regulate Ca 2ϩ -dependent exocytosis.
Synaptotagmin has been shown to have two C2-like domains that interact with Ca 2ϩ and phospholipid and to serve as a Ca 2ϩ sensor for neurotransmitter release (for reviews, see Refs. 12 and 13). The mode of action of this protein in neurotransmitter release has not definitely been clarified, but it has been shown to interact with syntaxin, one plasma membrane component of the SNARE 2 system. Recently, the binding of Ca 2ϩ to synaptotagmin has been shown to induce its tighter interaction with syntaxin (12). The assembly of the SNARE components is known to be followed by the assembly of the NSF/SNAP system, which finally causes the fusion of the vesicle with the plasma membrane (for a review, see Ref. 14). It has been suggested that synaptotagmin regulates these SNARE and NSF/SNAP system in response to Ca 2ϩ . Moreover, synaptotagmin has been shown to interact with many substances, such as AP2, neurexin, IP4, and ␤-SNAP (12,15,16). AP2 is an adaptor protein for clathrin, known to be involved in endocytosis (for a review, see Ref. 17); neurexin is the receptor for ␣-latrotoxin, known to induce massive neurotransmitter release (18); IP4 is a potential second messenger, known to regulate the intracellular Ca 2ϩ concentration and to be a blocker of neurotransmitter release (19); and ␤-SNAP is a brain-specific SNAP, but its role is still unclear (14).
In contrast to synaptotagmin, any protein directly interacting with rabphilin-3A, except the Rab3 subfamily members, has not been identified. In this study, we attempted to isolate a rabphilin-3A-interacting protein by the use of the yeast twohybrid system and identified it to be ␣-actinin, known to interact with actin filaments and to cross-link them into a bundle.

EXPERIMENTAL PROCEDURES
Materials and Chemicals-Recombinant rabphilin-3A and Rab3A were purified from the membrane fraction of overexpressing Spodoptera frugiperda cells (Sf9 insect cells) (20,21). Rab3A used for measuring low shear viscosity was further purified by hydroxyapatite column chromatography. GTP␥S-Rab3A was prepared as described (22). Chicken gizzard ␣-actinin was purified as described (23). Recombinant chicken lung type ␣-actinin was purified from overexpressing Escherichia coli. 3 An anti-rabphilin-3A polyclonal antibody and an anti-␣-actinin monoclonal antibody (CP2-1) were prepared * The work at Osaka University Medical School was supported by grants-in-aid for Scientific Research and for Cancer Research from the Ministry of Education, Science, Sports, and Culture, Japan (1996), by grants-in-aid for Abnormalities in Hormone Receptor Mechanisms and for Aging and Health from the Ministry of Health and Welfare, Japan (1996), and by grants from the Human Frontier Science Program (1996) and the Uehara Memorial Foundation (1996). 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.
§ Present address: The Third Dept. of Internal Medicine, Faculty of Medicine, Kyushu University, Fukuoka 812-82, Japan.
Screening for Rabphilin-3A-interacting Molecule by the Two-hybrid Method-A strain L40 was transformed with a derivative of pBTM116 bearing the N-terminal fragment (1-280 amino acids (aa)) of rabphilin-3A fused to the LexA DNA-binding domain: pBTM116-rabphilin-3A-N. A strain L40 carrying pBTM116-rabphilin-3A-N was transformed with the library DNA (MATCHMAKER human brain oligo(dT)primed library in pGAD10, Clontech). Approximately 2.4 ϫ 10 6 transformants were screened for the growth on SD plate media lacking tryptophan, leucine, and histidine, but containing 0.5 mM 3-amino-1,2,4-triazole, which is a specific inhibitor of the HIS3 gene product. His ϩ colonies were then placed on the nitrocellulose filter and stained with 5-bromo-4-chloro-3-indolyl-␤-D-galactopyranoside for ␤-galactosidase activity as described (27). From the 36 His3 ϩ and lacZ ϩ positive clones obtained with this screening, library plasmids were recovered through E. coli transformation. The recovered plasmids were transformed again into L40 containing pBTM116-rabphilin-3A-N, and it was found that 11 clones conferred the His ϩ and lacZ ϩ phenotypes on L40 containing pBTM116-rabphilin-3A-N. The nucleotide sequences of the insert DNA of these 11 clones were determined.
In Vitro Binding of Chicken ␣-Actinin to Recombinant Rabphilin-3A-Purified ␣-actinin (100 pmol) was incubated at 4°C for 90 min with HA-tagged full-length rabphilin-3A, the HA-tagged N-terminal fragment, or the HA-tagged C-terminal fragment (20 pmol each), which was immobilized on protein A-Sepharose (Pharmacia Biotech Inc.) through the anti-HA antibody, in 20 mM Tris/HCl at pH 8.0, 1 mM dithiothreitol, 1 mM EGTA, 0.26% CHAPS, and 150 mM NaCl. After washing three times with the same buffer, ␣-actinin associated with the beads was detected by SDS-PAGE followed by immunoblotting with the anti-␣actinin monoclonal antibody (CP2-1).
Immunoprecipitation of Rabphilin-3A from PC12 Cells-Confluent PC12 cells were washed with serum-free Dulbecco's modified Eagle's medium and lysed in 0.25 ml/dish of TNE buffer (10 mM Tris/HCl at pH 8.0, 1% Nonidet P-40, 150 mM NaCl, 1 mM EDTA, and 10 M (pamidinophenyl)methanesulfonyl fluoride). The extract was centrifuged at 15,000 ϫ g for 20 min, and the supernatant was incubated at 4°C for 90 min with either the anti-rabphilin-3A polyclonal antibody or rabbit IgG, which was immobilized on protein A-Sepharose (28). The immunoprecipitate was washed three times with TNE buffer and subjected to SDS-PAGE followed by immunoblotting with either the anti-rabphilin-3A polyclonal antibody or the anti-␣-actinin monoclonal antibody (CP2-1).
Low Shear Viscometry-Effect of rabphilin-3A on the actin filament bundling activity of recombinant chicken lung type ␣-actinin was examined by measuring low shear viscosity using a falling ball viscometer (29). Recombinant chicken lung type ␣-actinin (120 nM) was incubated at 4°C for 1 h with the indicated amounts of HA-tagged full-length rabphilin-3A, the HA-tagged N-terminal fragment, or the HA-tagged C-terminal fragment in 20 mM Tris/HCl at pH 7.2, 4 mM EDTA, 8 mM MgCl 2 , 0.12% CHAPS, 0.6 mM dithiothreitol, and 100 mM KCl. After the incubation, each sample was mixed with F-actin (1.0 mg/ml) in 20 mM Tris/HCl at pH 7.2, 2 mM EDTA, 6 mM MgCl 2 , 0.06% CHAPS, 100 mM KCl, 180 M ATP, and 0.65 mM EGTA, and the solution was sucked into a 0.1-ml micropipette. After incubation at 25°C for 1 h, the time for a stainless steel ball to fall a fixed distance in the pipette was measured and converted into viscosity in centipoise using various concentrations of sucrose solution at 20°C as a standard.
Electron Microscopy-A part of each sample prepared for measuring low shear viscosity was negatively stained with 2% uranyl acetate and viewed with a Hitachi electron microscope (model H-7100) (30,31).
Other Procedures-SDS-PAGE was performed as described (32). Rabphilin-3A and ␣-actinin transferred to nitrocellulose sheets were detected using the ECL immunoblotting detection system (Amersham Corp.). Protein concentrations were determined with bovine serum albumin as a standard protein (33).

RESULTS
Interaction of Rabphilin-3A with ␣-Actinin Estimated by the Yeast Two-hybrid System-We attempted here to isolate a rabphilin-3A-interacting protein by the use of the yeast two-hybrid system with the N-terminal fragment or the C-terminal fragment of rabphilin-3A as a bait from a human brain cDNA library. We did not isolate any clones with the C-terminal fragment as a bait, but with the N-terminal fragment as a bait, we isolated several clones and selected one clone encoding an amino acid position from 344 to 895 of ␣-actinin (EMBL accession number M86406), because ␣-actinin is known to cross-link actin filaments into a bundle (for reviews, see Refs. 34 and 35) and reorganization of actin filaments has been implicated in Ca 2ϩ -dependent exocytosis (for a review, see Ref. 36). The isolated clone indeed interacted with the N-terminal fragment of rabphilin-3A by the yeast two-hybrid method, and this interaction was comparable with the interaction of Rab3A with rabphilin-3A (Fig. 1A). This clone also interacted with fulllength rabphilin-3A, but not with its C-terminal fragment (Fig.  1B). The region of rabphilin-3A interacting with ␣-actinin was overlapped with that interacting with GTP-Rab3A (Fig. 1C).
Interaction of Recombinant Rabphilin-3A with Purified ␣-Actinin-The interaction of recombinant rabphilin-3A with ␣-actinin purified from chicken gizzard was then examined. The HA-tagged full-length rabphilin-3A, its N-terminal fragment, and its C-terminal fragment were separately immobilized on protein A-Sepharose, and the binding of ␣-actinin to these immobilized proteins were analyzed. ␣-Actinin indeed interacted with both full-length rabphilin-3A and its N-terminal fragment, but to little extent with its C-terminal fragment ( Fig.  2A). The stoichiometry of binding between ␣-actinin and fulllength rabphilin-3A was at least 0.1 under the assay conditions used here.
Inhibition by Rab3A of the Rabphilin-3A-␣-Actinin Interactions-Because ␣-actinin interacted with the N-terminal region of rabphilin-3A, with which GTP-Rab3A interacts (7,37), effect of GTP␥S-Rab3A on the interaction of recombinant rabphilin-3A with purified ␣-actinin was next examined. GTP␥S-Rab3A was first incubated with full-length rabphilin-3A or its N-terminal fragment immobilized on the beads. The interactions of ␣-actinin with the immobilized proteins were then analyzed. Only a small amount of ␣-actinin interacted with full-length rabphilin-3A and its N-terminal fragment in the presence of GTP␥S-Rab3A, indicating that Rab3A inhibits the interaction of rabphilin-3A with ␣-actinin (Fig. 2C).
Enhancement of the Actin Filament Bundling Activity of ␣-Actinin by Rabphilin-3A-Finally, effect of recombinant rabphilin-3A on the actin filament bundling activity of recombinant ␣-actinin was examined by measuring low shear viscosity using a falling ball viscometer (29). Both the full-length protein and the N-terminal fragment increased the viscosity in a dosedependent manner, whereas the C-terminal fragment did not affect the viscosity (Fig. 3). However, any of them did not affect the viscosity in the absence of ␣-actinin (data not shown). This activity of rabphilin-3A was confirmed by transmission electron microscopy of negatively stained specimens (30,31). ␣-Actinin caused actin filaments to associate into a bundle, but full-length rabphilin-3A alone did not show such an effect (Fig.  4, A-C). Full-length rabphilin-3A enhanced this activity of ␣-actinin (Fig. 4D). We examined the effect of GTP␥S-Rab3A on this activity of rabphilin-3A using a falling ball viscometer. However, a highly purified Rab3A sample interfered the measurement of low shear viscosity and by itself increased the viscosity in the absence of ␣-actinin (data not shown). The sample may be still contaminated by some unidentified interfering materials, because both GTP␥S-Rab3A, GDP-Rab3A, and their boiled samples showed the similar effect. FIG. 2. In vitro and in vivo binding of ␣-actinin to rabphilin-3A. A, interaction of recombinant rabphilin-3A with purified ␣-actinin. Purified ␣-actinin (100 pmol) was incubated at 4°C for 90 min with HA-tagged full-length rabphilin-3A, the HA-tagged N-terminal fragment, or the HA-tagged C-terminal fragment (20 pmol each), which was immobilized on protein A-Sepharose through the anti-HA antibody. After washing, ␣-actinin associated with the beads was detected by SDS-PAGE followed by immunoblotting with the anti-␣-actinin monoclonal antibody (CP2-1). Lane 1, protein A-Sepharose alone; lane 2, full-length rabphilin-3A; lane 3, the N-terminal fragment; lane 4, the C-terminal fragment. B, coimmunoprecipitation of ␣-actinin with rabphilin-3A from intact PC12 cells. Confluent PC12 cells were washed with serum-free Dulbecco's modified Eagle's medium and lysed in 0.25 ml/dish of TNE buffer. The lysate was incubated at 4°C for 90 min with either the anti-rabphilin-3A polyclonal antibody or rabbit IgG, which was immobilized on protein A-Sepharose. The immunoprecipitate was washed and subjected to SDS-PAGE followed by immunoblotting with either the anti-rabphilin-3A polyclonal antibody or the anti-␣-actinin monoclonal antibody (CP2-1). Panel a, with anti-rabphilin-3A antibody. Panel b, with anti-␣-actinin antibody. Lane 1, with control rabbit IgG; lane 2, with anti-rabphilin-3A antibody. Arrowhead, rabphilin-3A; arrow, ␣-actinin. C, inhibition by Rab3A of the rabphilin-3A-␣-actinin interactions. HA-tagged full-length rabphilin-3A or the HA-tagged Nterminal fragment, which was immobilized on protein A-Sepharose, was incubated in the presence or absence of 0.6 M GTP␥S-Rab3A. After washing, the beads containing GTP␥S-Rab3A bound rabphilin-3A were incubated with purified ␣-actinin, and ␣-actinin associated with the beads was detected by SDS-PAGE followed by immunoblotting with the anti-␣-actinin monoclonal antibody (CP2-1). Lane 1, full-length rabphilin-3A alone; lane 2, full-length rabphilin-3A with GTP␥S-Rab3A; lane 3, the N-terminal fragment alone; lane 4, the N-terminal fragment with GTP␥S-Rab3A. Recombinant chicken lung type ␣-actinin (120 nM) was incubated at 4°C for 1 h with the indicated amounts of HA-tagged full-length rabphilin-3A, the HA-tagged N-terminal fragment, or the HA-tagged Cterminal fragment. After the incubation, each sample was mixed with F-actin (1.0 mg/ml), and the solution was sucked into a 0.1-ml micropipette. After incubation at 25°C for 1 h, the time for a stainless steel ball to fall a fixed distance in the pipette was measured and converted into viscosity in centipoise using various concentrations of sucrose solution at 20°C as a standard. q, with full-length rabphilin-3A; å, with the N-terminal fragment; f, with the C-terminal fragment.

DISCUSSION
We have shown here that rabphilin-3A directly interacts with ␣-actinin and stimulates its actin filament bundling activity. We have moreover shown here that GTP-Rab3A inhibits this interaction of rabphilin-3A with ␣-actinin. Both Rab3A and rabphilin-3A have been shown to be associated with synaptic vesicles (8,9,22,38,39). It has not yet been reported that ␣-actinin is associated with synaptic vesicles, but ␣-actinin (40), Rab3A (41,42), and rabphilin-3A 4 have been shown to be associated with large dense core vesicles in chromaffin cells and PC12 cells. Present results together with these earlier observations suggest that rabphilin-3A physically and functionally interacts with ␣-actinin in intact cells.
Electron microscopic analysis on the spatial distribution of synaptic vesicles at active zones of the squid giant axon shows that microinjection of the N-terminal or C-terminal fragment of rabphilin-3A causes a selective accumulation of predocked vesicles 50 -100 nm away from the plasma membrane, 1 whereas the microinjection of the peptides from the C2-like domain of synaptotagmin, known to be associated with synaptic vesicles and to be involved in neurotransmitter release, has been shown to inhibit neurotransmitter release but to cause a selective accumulation of docked vesicles within 50 nm from the plasma membrane (43). These results suggest that rabphilin-3A as well as synaptotagmin is involved in Ca 2ϩ -dependent neurotransmitter release, but that rabphilin-3A is involved in the predocking or docking process, whereas synaptotagmin is involved in the fusion process that follows the docking process. On the other hand, there are several lines of evidence that reorganization of actin filaments is also involved in Ca 2ϩ -dependent exocytosis (36). One possible speculative role of the Rab3A-rabphilin-3A system in the regulation of the reorganization of actin filaments is that the conversion of GTP-Rab3A to GDP-Rab3A by the action of Rab3 GTPase-activating protein and the subsequent dissociation of GDP-Rab3A from rabphilin-3A initiates the interaction of rabphilin-3A with ␣-actinin, eventually stimulating the bundling of actin filaments, which may be involved in the docking and fusion processes. We have shown previously that rabphilin-3A interacts with ␤-adducin through the C2-like domain in the presence of Ca 2ϩ and phospholipid (44,45). ␤-Adducin has been implicated in the assembly of spectrin-actin complexes (for a review, see Ref. 46). The function of this interaction of rabphilin-3A with ␤-adducin remains to be clarified, but the present results together with these earlier observations suggest that rabphilin-3A may be an important linker for Rab3A and various cytoskeletal proteins.