Inhibitory role of endophilin 3 in receptor-mediated endocytosis.

Endophilin 1 (Endo1) participates in synaptic vesicle biogenesis through interactions of its Src homology 3 domain with the polyphosphoinositide phosphatase Synaptojanin and the GTPase Dynamin. Endo1 has also been reported to affect endocytosis by converting membrane curvature via its lysophosphatidic acid acyltransferase activity. Here we report that a closely related isoform of Endo1, Endo3, inhibits clathrin-mediated endocytosis. Mutational analyses showed that the variable region of Endo3 is important in regulating transferrin endocytosis. In the brain, Endo3 is co-localized with dopamine D2 receptor in olfactory nerve terminals and inhibits its clathrin-mediated endocytosis in COS-7 cells. Furthermore, overexpression of Endo3 in an olfactory epithelium-derived cell line suppressed dopamine D2 receptor-mediated endocytosis and therefore accelerated its dopamine-induced differentiation. These results indicate that Endo3 may act as a negative regulator of clathrin-mediated endocytosis in brain neurons.


Endophilin 1 (Endo1) participates in synaptic vesicle biogenesis through interactions of its Src homology 3 domain with the polyphosphoinositide phosphatase
Synaptojanin and the GTPase Dynamin. Endo1 has also been reported to affect endocytosis by converting membrane curvature via its lysophosphatidic acid acyltransferase activity. Here we report that a closely related isoform of Endo1, Endo3, inhibits clathrin-mediated endocytosis. Mutational analyses showed that the variable region of Endo3 is important in regulating transferrin endocytosis. In the brain, Endo3 is co-localized with dopamine D2 receptor in olfactory nerve terminals and inhibits its clathrin-mediated endocytosis in COS-7 cells. Furthermore, overexpression of Endo3 in an olfactory epithelium-derived cell line suppressed dopamine D2 receptor-mediated endocytosis and therefore accelerated its dopamine-induced differentiation. These results indicate that Endo3 may act as a negative regulator of clathrin-mediated endocytosis in brain neurons.
In addition to the activity of its SH3 domain, the N-terminal domain of Endo1 has been shown to have lysophosphatidic acid acyltransferase (LPA-AT) activity; this enzyme converts lysophosphatidic acid (LPA) and acyl-CoA to phosphatidic acid (PA) (11). Since the cone shape of PA is more compatible with a high curvature than the inverted cone shape of LPA, Endo1 may promote negative membrane curvature by altering the composition of the donor membrane at the invaginating bud, thereby promoting the conversion of a shallow pit (11). Moreover, the binding of Endo1 to liposomes is sufficient to deform them into narrow tubules (18). These observations provide direct evidence to support the hypothesis that Endo1 generates membrane curvature during synaptic vesicle endocytosis (19).
In contrast to the neural expression of Endo1, the closely related protein Endophilin 2 (Endo2) is ubiquitously distributed throughout many tissue types (2). Endo2 has been reported to interact with Endo1 via its coiled-coil domain, thus linking a number of neuronal targets to the endocytic machinery (20). In this study, we assayed the involvement of a third mammalian Endophilin isoform, Endophilin 3 (Endo3), in receptor-mediated endocytosis. We show that, while structurally similar to Endo1 and -2, Endo3 is functionally distinct in its ability to inhibit clathrin-mediated transferrin endocytosis. We also show that Endo3 is expressed in the brain, including the olfactory nerve (ON) terminals, where presynaptic D2Rs are co-localized. We therefore investigated the ability of Endo3 to regulate D2R-mediated endocytosis in COS-7, HEK293T, and olfactory epithelium-derived cell lines.

EXPERIMENTAL PROCEDURES
Dopamine was purchased from Sigma; (Ϫ)-[methoxy-3 H]sulpiride (specific activity, 73.7 Ci/mmol) from PerkinElmer Life Sciences; and restriction enzymes, DNA ligase, and KOD Dash DNA polymerase from Toyobo Co. (Osaka, Japan). FuGENE 6, alkaline phosphatase, and monoclonal antibody against human myc (9E10) were obtained from Roche Diagnostics (Basel, Switzerland) and pBluescript II SK was purchased from Stratagene (La Jolla, CA). EGFP-N1 and DsRed2-N1 were purchased from Clontech. Monoclonal antibody against human dopamine D2 receptor was obtained from Santa Cruz Biotechnology (Santa Cruz, CA); monoclonal antibody against Dynamin-1 (Hudy-1) was purchased from Upstate Biotechnology (Lake Placid, NY); anti-Synaptojanin antibody was purchased from Medical and Biological * This work was supported by grants from the program grants-in-aid for Scientific Research of the Japan Society for the Promotion of Science (to H. S. and K. Y.); the Ministry of Education, Science, Sports, and Culture of Japan; the Japan Epilepsy Research Foundation; and the Pharmacopsychiatry Research Foundation (to K. Y.). 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.
Laboratories (Nagoya, Japan); and Cy3-conjugated donkey anti-mouse IgG was obtained from Jackson ImmunoResearch (West Grove, PA). Alexa Fluor(R) 488-conjugated goat anti-rabbit and anti-mouse IgGs and Alexa Fluor(R) 568-conjugated goat anti-rabbit IgG were purchased from Molecular Probes (Eugene, OR). G protein-coupled receptor kinase 2 (GRK2) cDNA was the kind gift of Dr. R. J. Lefkovitz; cDNA encoding the short form of D2R was the kind gift of Dr. D. K. Grandy; the mammalian expression vector pEF-BOS was the gift of Dr. S. Nagata; and HEK293T cells were the gift of Dr. R. Takahashi.
Cloning of Rat Endophilin cDNAs-Rat endophilin cDNAs were PCR amplified from a rat hippocampus cDNA library (21) using specific primers for endo1 (5Ј-ATGTCGGTGGCMGGSCTSAAGAA-3Ј and 5Ј-CTAATGGGGCAGRGCAACCAGAA-3Ј), endo2 (5Ј-ATGTCGGTGG-CGGGGCTGAAGA-3Ј and 5Ј-TCACTGAGGCAGAGGCACCAGCA-3Ј), and endo3 (5Ј-ATGTCGGTGGCYGGGCTSAAGAA-3Ј and 5Ј-AGGATG-TCACAGAACACCRCACTG-3Ј) (22). The amplified DNA fragments were directly sequenced in both directions. A rat testis cDNA library (Stratagene) was screened with the rat endo3 cDNA probe according to standard procedures (23). The expression plasmid used in this study was pEF-BOS (24). Deletion constructs and chimeras between endo1 and 3 were made by two-step amplification using the overlap extension method (25). The internal oligomer pairs were designed such that the 5Ј appendage of one primer was complementary to the other oligomer, and the two DNA fragments resulting from the first PCR amplification were annealed prior to the second round of amplification. Site-directed mutations were created using the QuikChange system (Stratagene).
Yeast Two-hybrid System Screening-Yeast two-hybrid screening was performed as described previously (21). Rat hippocampal cDNAs were subcloned in the SalI/NotI sites of the pPC86 vector, which contains the GAL4 activation domain, and the variable region and SH3 domain of Endo3 were subcloned in-frame in the SalI/NotI sites of the pPC97 vector, which contains the GAL4 DNA binding domain. The plasmids were used to transform PCY2 cells, and positive clones were selected on double-minus plates (Leu Ϫ, Trp Ϫ) and assayed for ␤-galactosidase activity.
Cell Culture and Transient Transfection-COS-7, HEK293T, and 13.S.1.24 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum at 37°C in 5% CO 2 . Cells grown to 50 -70% confluence in 60-mm dishes were transfected with plasmids using FuGENE 6 or TransFast (Promega, Madison, WI) according to the manufacturer's instructions. For the internalization assay, cells were transfected with pEF-BOS-D2R along with plasmids encoding GRK2, endo3, endo1, or their mutants. In each transfection, the total amount of plasmids was adjusted to 5 g by adding pEF-BOS vector. Dopamine-induced sequestration of human c-myc-tagged D2R was visualized by immunofluorescence confocal microscopy as described previously (26). Cells were treated with 10 M dopamine at 37°C for 120 min, fixed with 4% paraformaldehyde, and incubated overnight at 4°C with anti-myc monoclonal antibody (9E10). The cells were washed with phosphate-buffered saline (PBS(Ϫ)) and incubated with Cy3-conjugated anti-mouse secondary antibody for 120 min. Prior to drug treatment of the 13.S.1.24 cells, they were transfected with EGFP-N1, EGFP-endo1, or EGFP-endo3, and apomorphine was added after 2 days. The populations of differentiated bipolar cells were quantified by hand scoring. Statistical analysis was performed using the unpaired Student's t test.
Transferrin Uptake-COS-7 fibroblasts were transfected using TransFast for 36 h, after which the cells were transferred to serum-free medium and incubated overnight. The cells were subsequently incubated for 30 min at 37°C with 25 g/ml biotinylated human transferrin (Sigma) and fixed in 4% paraformaldehyde. Transferrin uptake was visualized with Alexa488-conjugated streptavidin (green) using a Leica scanning confocal microscope (TCS-SP). Endo1 and Endo3 were detected using anti-Endophilin polyclonal antibodies followed by Al-exa568-conjugated anti-rabbit antibody (Molecular Probes). Inhibition of endocytosis in transfected cells was defined as uptake of less than 5% of the transferrin taken up by untransfected cells in the same field, which ensured scoring of only strongly inhibited cells.
Assay of Agonist-dependent Internalization-The assay for agonistdependent internalization has been described previously (26 -28). In brief, 1 day after transfection, cells in two 60-mm dishes were trypsinized and resuspended in 15 ml of medium, and 1.5 ml of suspension was dispensed into each well of a 12-well dish. The following day, 44 -48 h after transfection, 10 M dopamine was added, and the cells were incubated for various times. The reaction was stopped by placing the dishes on ice and washing the cells three times with cold PBS(Ϫ). The amount of D2R on the cell surface was determined by the binding of the hydrophilic, membrane-impermeable ligand, [ 3 H]sulpiride (28).
The cells were therefore incubated with 2.8 nM [ 3 H]sulpiride in 1 ml of HEPES-buffered saline (pH 7.4) at 4°C for 150 min, washed twice with ice-cold PBS(Ϫ), and treated with 1% Triton X-100 solution. To each sample was added 3.5 ml scintillation mixture (Aquasol-2, PerkinElmer Life Sciences), and the amount of [ 3 H]sulpiride remaining was quantified with a scintillation counter (Tri-Carb, PerkinElmer Life Sciences). Nonspecific binding of [ 3 H]sulpiride was determined by incubation in the presence of 10 M haloperidol, a D2R ligand, and was found to be a maximum of 5% of the total binding. Each sample was assayed in triplicate.
Production of Recombinant Proteins and Antibodies-cDNA encoding full-length Endo3 (aa 1-347) was subcloned into pGEX4T-3 (Amersham Biosciences, Uppsala, Sweden), and the construct was verified by sequencing. The host strain BL21 (DE3) was transformed with the resulting expression vector, and the recombinant protein was purified on glutathione-Sepharose (Amersham Biosciences) using standard methods. New Zealand White rabbits were immunized with GST-Endo3 fusion protein, and the antisera were affinity-purified by chromatography on GST-Endo3 immobilized on Affi-Gel (Bio-Rad). The antibodies were further purified by passing them through glutathione-Sepharose columns coupled with GST-Endo1 and GST-Endo2.
Western Blot Analysis-Western blot analyses of rat brain and transfected cell lysates with anti-Endo3 or anti-pan-Endophilin antibody were performed as described previously (29).
Immunoprecipitation-An olfactory bulb was extracted in 1.2 ml of TNE buffer (10 mM Tris-HCl (pH 7.6), 1% Nonidet P-40, 0.15 M NaCl, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride). The extract was precleared by incubation for 1 h at 4°C with 20 l of protein A-Sepharose and divided into two parts. 2 l of affinity-purified anti-Endo3 antibody was added to one part, while 2 l of preimmune serum was added to the other part, and the lysates were incubated for 1 h at 4°C. The immune complexes were precipitated with 10 l of protein A-Sepharose (Amersham Biosciences) and eluted into 20 l of SDS-PAGE sample buffer. Following electrophoresis on 7.5% SDS-PAGE, the proteins were transferred and incubated with monoclonal antibody to Dynamin-1 or Synaptojanin-1.
Immunohistochemistry-Adult female Wistar rats (8 weeks old, 200 g) were perfused with 2% paraformaldehyde fixative for 20 min. The brain was postfixed in the same fixative for 2 h, cryoprotected in 30% sucrose in PBS(Ϫ) solution and frozen in Tissue-Tek OCT Compound (Miles, Elkhart, IN) at Ϫ80°C. All experiments were conducted in accordance with the Guideline for the Care and Use of Animals (Tokyo Metropolitan Institute for Neuroscience, 2000). Frozen olfactory bulbs were cut into 20-m-thick sections with a cryostat. The sections were thaw-mounted onto silanized slides, air-dried, frozen, and stored at Ϫ80°C until used. The brain sections were incubated with 2.5% bovine serum albumin and 10% Block Ace (Snow Brand Milk, Sapporo, Japan) for 1 h and then with rabbit anti-Endo3 (1:100 dilution) and mouse anti-D2R (1:500 dilution) antibodies at 4°C for 4 days. Endo3 was visualized with the Alexa568-labeled secondary antibody. D2R signals were amplified with a Renaissance TSA Amplification kit (PerkinElmer Life Sciences). For signal amplification, the sections were incubated with biotinylated secondary antibody, followed by biotinyltyramide and horseradish peroxidase, and visualized using Alexa 488conjugated streptavidin.

Inhibition of Transferrin Endocytosis by Endophilin 3-To
clarify the role of Endo3 in clathrin-mediated endocytosis, we assayed the effect of Endo3 on transferrin uptake (31) by transient expression of endo3 in COS-7 cells, with an empty vector or endo1 as controls. In contrast to control cells, which showed marked transferrin uptake, we unexpectedly observed an inhibition of transferrin uptake in endo3-expressing cells (Fig. 1b).
To confirm this inhibition and to determine the Endo3 domain responsible, we interchanged the variable and SH3 regions of Endo1 and Endo3. Replacement of Endo1 domains with those of Endo3 (Endo1-3) resulted in the Endo3-phenotype for transferrin uptake; conversely, replacement of Endo3 domains with those of Endo1 (Endo3-1) resulted in the Endo1 phenotype (data not shown). The same pattern was observed in the absence of the Endo3 variable domain (Endo3(⌬V)) ( Fig. 1, a-c). When we interchanged the variable regions of Endo1 and Endo3, we found that Endo1(V3) showed robust inhibition, whereas Endo3(V1) stimulated transferrin uptake (Fig. 1, a-c). These results strongly indicate that the variable domain of Endophilin is responsible for the regulation of endocytosis.
To test the role of the Endophilin variable region in the regulation of endocytosis, we introduced additional structural perturbations by site-directed mutagenesis (Fig. 1a). Secondary structure analyses of the Endophilins revealed that both Endo1 and 2 have short N-terminal ␣-helices in their variable regions, whereas the corresponding region of Endo3 has a ␤-turn. We therefore inserted the ␣-helical structure of Endo1 (MSLEF, corresponding to amino acids 261-265) into the N terminus of the variable region of Endo3 (Endo3(ϩH)), and we found that this fragment stimulated transferrin uptake (Fig. 1,  b and c). Conversely, introduction of a point mutation into the middle of the ␣-helix of the variable region of Endo1 (Endo1L263P) resulted in marked inhibition of transferrin uptake ( Fig. 1, a-c). These data are consistent with the notion that the ␣-helix in the variable region is important for the regulation of transferrin endocytosis.
Expression of Endo3 in D2R-containing Olfactory Glomeruli-To address the functional role of Endo3 in neurons, we searched brain neurons, which express endo3 but not endo1, by in situ hybridization (30) for optimal expression of endo3. We found that olfactory mucosae were suitable for determining the function of Endo3. In situ hybridization revealed that endo3 mRNA was strongly expressed in the middle two-thirds of the  (c, d). e, immunoblotting of brain extracts with anti-Endo3 antibody. OB, olfactory bulb; Cc, cerebral cortex; Hi, hippocampus; St, striatum; Cb, cerebellum; Te, testis; Endo3, endo3-expressing COS-7 cell extract. f, Dynamin-1 (Dyn1) was co-immunoprecipitated from the olfactory bulb with anti-Endo3 antibody. g, the same membrane was reprobed with anti-Synaptojanin 1 (SJ1). h-j, co-localization of Endo3 (h) and D2R (i) in olfactory glomeruli and the merged image (j) are shown. ONL, olfactory nerve layer; GL, glomerular layer; EPL, external plexiform layer. Scale bar, 100 m. olfactory epithelium (Fig. 2, a and c), corresponding to the location of the primary ON perikarya (30). In contrast, the expression of endo1 mRNA in olfactory mucosae was much weaker than that of endo3 (Fig. 2, b and d). In addition, little signal was detected in sections hybridized with the endo3-sense control (data not shown). To determine the distribution of Endo3 protein in olfactory neurons, we generated an anti-Endo3 antibody and performed Western blotting, and we found that Endo3 protein was expressed in all brain regions, including the olfactory bulb (Fig. 2e).
Among the G protein-coupled receptors expressed in the olfactory bulb, D2R is highly and selectively expressed on the ON terminals (32). To determine whether Endo3 is co-expressed with presynaptic D2R in this region, we incubated olfactory bulbs with anti-Endo3 and anti-D2R antibodies, and we found that D2Rs were co-localized with Endo3 at the same glomerular synapses (Fig. 2, h-j). Immunoelectron microscopy showed that Endo3 was highly enriched in the presynaptic ON terminals, which are characterized by the presence of numerous synaptic vesicles (data not shown). These data strongly suggest that Endo3 is co-localized with presynaptic D2Rs on the ON terminals and may be involved in D2R-mediated endocytosis.
Identification of Dynamin-1 and Synaptojanin-1 as Endo3binding Proteins in the Brain-Since Endo3 has been reported to be testis-specific and to associate with testis-specific Dynamin-3 (2), we used the variable and SH3 domains of endo3 cDNA as a bait in a yeast two-hybrid screening system to identify the brain proteins that interact with it (21). This screening led to the isolation of dynamin-1 and -3. To confirm that Endo3 is associated with Dynamin-1 in vivo, we immunoprecipitated Endo3 from an extract of olfactory bulbs and found that Dynamin-1 was co-precipitated (Fig. 2f). In addition, Synaptojanin-1 was co-precipitated with Endo3 (Fig. 2g), indicating that Endo3 forms a stable complex with Dynamin-1 and Synaptojanin-1 in olfactory bulbs.
Effects of Endo3 Proteins on Presynaptic D2R-mediated Endocytosis-To verify the involvement of Endo3 in presynaptic D2R-mediated endocytosis, we reconstituted the assay system of D2R-mediated internalization after dopamine stimulation (26 -28). D2Rs are reported to be internalized in an agonistdependent manner when GRK2 is co-expressed (26). We assayed the effect of Endo3 on D2R internalization, which was detected as a decrease in [ 3 H]sulpiride binding activity in cells treated with dopamine (Fig. 3a). Control cells, which were transfected with the expression vector alone, showed robust D2R internalization (41.9 Ϯ 0.5%, n ϭ 10) after 120 min of dopamine treatment. In contrast, there was little decrease in [ 3 H]sulpiride binding (6.5 Ϯ 0.7%, n ϭ 10) when endo3-expressing cells were treated with dopamine (Fig. 3, a and b). Moreover, this inhibition was much stronger than that observed in cells expressing the SH3 domain of Endo3 (Endo3D in Fig. 3b, 26.3 Ϯ 1.0%). Overexpression of endo1 led to an increase in the basal level of D2R endocytosis even before dopamine treatment (Fig. 3g), and dopamine-induced internalization of D2R in endo1-expressing cells was slightly reduced compared with that in cells expressing vector (Fig. 3b).
When we examined D2R internalization in transiently infected cells by confocal microscopy, we found that Endo3 suppressed dopamine-induced internalization of the c-myc-tagged D2Rs. In cells expressing D2R and GRK2, D2R was located predominantly at the cell surface in the absence of dopamine (Fig. 3c). When these cells were treated with dopamine, the receptors were distributed in the intracellular endosomes (Fig.  3d). In endo3-expressing cells, however, there was almost a complete absence of D2R redistribution into the endosomes (Fig. 3f). Since Endo2 is the primary Endophilin isoform in COS-7 cells (Fig. 3i), our results suggest that exogenous Endo3 may inhibit D2R endocytosis by disrupting the interaction of Dynamin with endogenous Endo2.
Promotion of Dopamine-induced Differentiation of an Olfactory Epithelial Cell Line Expressing Endo3-To confirm that Endo3 inhibits D2R endocytosis in olfactory neurons, we used a neuronal cell line generated by transfection of rat olfactory epithelium with the immortalizing recombinant oncogene E1A of adenovirus-2 (33). This 13.S.1.24 cell line has the phenotype of olfactory neuronal progenitors, and addition of dopamine has been shown to induce their differentiation into bipolar olfactory neurons. When we expressed the EGFP-Endo3 fusion protein in this cell line and treated it with the dopamine agonist apomorphine (10 M), we observed that a large proportion of epitheloid cells underwent morphological differentiation into bipolar cells (Fig. 4, e and h). In contrast, apomorphine treatment of cells expressing EGFP or EGFP-Endo1 fusion protein resulted in the differentiation of a significantly smaller proportion of cells (Fig. 4, f and g). In addition, when we co-transfected myc-tagged D2R and the pEF-endo3 into these 13.S.1.24 cells line and immunostained them with anti-myc antibody, we observed inhibition of D2R-mediated endocytosis; in contrast, endocytosis was stimulated in cells co-transfected with myctagged D2R and pEF-endo1 or pEF-BOS plasmids (Fig. 4, a-d).
These results suggest that Endo3 inhibits D2R endocytosis, resulting in the differentiation of epithelial cells into bipolar neurons. DISCUSSION We have shown here that Endo3 is an inhibitor of receptormediated endocytosis. While overexpression of endo3 in COS-7 cells was associated with the arrest of transferrin endocytosis, this inhibition of transferrin uptake was abolished by replacing the variable region of Endo3 with that of Endo1. In olfactory neurons, only Endo3, not Endo1, was expressed. Endo3 inhibited the endocytosis of dopamine D2 receptors localized on olfactory nerve terminals, and this inhibition facilitated the dopamine-induced differentiation of a cell line derived from olfactory epithelium.
Endo1 is required for the transition from early to late stages of clathrin-mediated endocytosis, as shown by the microinjection of anti-Endophilin antibodies at the giant synapse of the lamprey, which resulted in the accumulation of clathrin-coated pits at the presynaptic membrane (10). Furthermore, mutant Drosophila larvae lacking endophilin failed to take up FM1-43 dye in synaptic boutons, indicating an inability to retrieve synaptic membrane (13,14). Based on these findings and the similarity of Endo3 to Endo1 and Endo2, we initially expected that Endo3 would activate the endocytosis of synaptic vesicles. Unexpectedly, however, our results showed that expression of Endo3 almost completely inhibited receptor-mediated endocytosis. The variable domain of Endo3 is composed of a ␤-turn, whereas that of Endo1 has an additional ␣-helix at its aminoterminal end. As the crystal structures of the Endophilins are not yet known, the sites responsible for inhibition of endocytosis were identified by mutagenesis. For example, when we inserted five residues (MSLEF aa261-265) from Endo1, which have a high potential to form an ␣-helix, into the corresponding portion of the variable region of Endo3, we observed reduced inhibition of transferrin endocytosis compared with wild type Endo3. Conversely, when we replaced a leucine in the ␣-helix of Endo1 with proline, we observed an inhibition of transferrin uptake equal to that seen with Endo3. The proline residue introduced a kink into the ␣-helix, which bends away from the side of the helix continuing the proline (34). Taken together, these results suggest that an amino-terminal ␣-helical structure in the variable region of the Endophilins regulates transferrin endocytosis. Similar findings have been described for Endo2, in that the presence of glutamate at amino acid 264 was shown to be required for its binding to voltage-gated calcium channels (35). A point mutation at this amino acid was found to eliminate this calcium-dependent interaction, as well as impairing synaptic vesicle endocytosis, indicating that this region is necessary for facilitation of endocytosis. Although there is as yet no direct evidence that this glutamate binds Ca 2ϩ , the replacement of glutamate with serine at the corresponding position of Endo3 suggests that it may work as a dominant negative regulator for other Endophilins. NH 2 -terminal LPA-AT activity was also shown to be involved in clathrin-mediated endocytosis. When we compared LPA-AT activity of the three Endophilins, we found that, although Endo3 had some LPA-AT activity, it was significantly lower than that of Endo2 (data not shown). Thus, overexpression of Endo3 protein in COS-7 cells may act as a dominant negative regulator by replacing endogenous Endo2, which induces positive-negative lipid membrane curvature along with conversion of the inverted cone-shaped LPA into cone-shaped PA. This N-terminal low LPA-AT model does not necessarily exclude inhibition via a variable region model. Either or both models may explain the ability of Endo3 to inhibit receptor-mediated endocytosis.
The in vivo functional role of Endo3 in the ON terminals is of great interest. Axons from olfactory epithelium form the ON, which innervates the olfactory bulb (36). ON terminals make synapses with dendrites of both mitral and tufted and periglomerular cells, and periglomerular neurons can form reciprocal dendro-dendritic synapses with mitral and tufted cells (37). Dopamine and D2R agonists are known to block synaptic transmission from the olfactory nerve to mitral and tufted cells (38). In other words, the activation of D2Rs presynaptically inhibits olfactory nerve terminals and decreases the ability to detect odors. We therefore hypothesize that dopamine-induced D2R internalization is suppressed by co-expression of Endo3 in ON terminals. For example, if an animal is tonically exposed to a strong odor, the glomeruli activated by this odor might release more dopamine, resulting in greater presynaptic inhibition. In the absence of Endo3 from the glomeruli, however, D2R would be internalized rapidly after odor stimulation. Under these conditions, feedback inhibition would not work satisfactorily, leading to sensing of the offensive smells. Thus, presynaptic inhibition of ON terminals via D2R is a potential mechanism for increasing the range of adaptation to strong odors.
We have demonstrated here that Endo3 inhibits transferrin uptake and that its variable region is critical for the regulation of endocytosis. In addition, we have shown that suppression of D2R-mediated endocytosis by Endo3 accelerates dopamine-induced differentiation of olfactory epithelium-derived cells. These findings indicate that Endo3 acts as a negative regulator of clathrin-mediated endocytosis, suggesting a new role of Endophilin in mediating physiological sensations.