Rab5 small GTPase is a famous regulator of endocytic vesicular transport from plasma membrane to early endosomes. In neurons, Rab5 is found not only on endocytic vesicles in cell bodies but also on synaptic vesicles in nerve terminals. However, the function of Rab5 on synaptic vesicles remains unclear. Here, we elucidate the function of Rab5 on synaptic vesicles with in vivo and in vitro experiments using Drosophila photoreceptor cells. Functional inhibition of Rab5 with Rab5N142I, a dominant negative version of Drosophila Rab5, induced enlargement of synaptic vesicles. This enlargement was, however, suppressed by enhancing synaptic vesicle recycling under light illumination. In addition, synaptic vesicles prepared from Rab5N142I-expressing flies exhibited homotypic fusion in vitro. These results indicate that Rab5 functions to keep the size of synaptic vesicles uniform by preventing their homotypic fusion. By contrast, Rab5 was not involved in the endocytic reformation of synaptic vesicles, contrary to expectation from its conventional function. Furthermore, we electrophysiologically and behaviourally showed that the function of Rab5 is essential for efficient signal transmission across synapses.

Introduction

Synaptic transmission intensity is primarily determined by the amount of neurotransmitter released from presynaptic neurons. The transmitter is packaged in synaptic vesicles in the nerve terminal(Williams, 1997); the release is then quantitatively controlled by the number of synaptic vesicles fused to presynaptic membrane (Bajjalieh,2001; Brunger,2001). Therefore, for quantitative release of neurotransmitters,each vesicle must be uniform in size. At least two independent mechanisms must be present to assure size uniformity of synaptic vesicles: one to generate uniformly sized vesicles; and one to prevent them from homotypic fusion. The former mechanism has been studied eagerly and it has become clear that clathrin, adaptor proteins and dynamin are involved in the machinery for making vesicles uniform in size (Zhang et al., 1999). By contrast, almost no attention has been paid to the mechanism that prevents synaptic vesicles from homotypic fusion. The simplest interpretation is that synaptic vesicles lack molecular machinery required for homotypic fusion. However, considering that secretory vesicles in many other secretory cells possess machinery for homotypic fusion(Cochilla et al., 2000; Coorssen et al., 1998), it would be dangerous to believe this interpretation without careful examination. Another possibility is that synaptic vesicles have a potential mechanism for homotypic fusion like vesicles in other secretory cells but that this potential is blocked to assure quantitative release of neurotransmitter. In this case, some additional molecule that blocks homotypic fusion would be required.

Rab proteins, a family of small GTPases, play an essential role in the regulation of intracellular vesicle transport(Novick and Zerial, 1997). Each Rab protein is thought to work in a particular transport pathway through identifying the correct pairing of the transport vesicle and the target membrane. Rab5 is one of the well-studied Rab proteins, and has been shown to be a key regulator of the early endocytic process from plasma membrane to early endosome (Bucci et al.,1992; Chavrier et al.,1990; Gorvel et al.,1991; Stenmark et al.,1994). Besides its use as an early-endosome marker, Rab5 and its interacting proteins have been studied to reveal the pathways of endocytic transport and their molecular mechanisms. It is now believed that Rab5 is required for all the endocytic pathways known so far, not only in common endocytic processes including fluid phase endocytosis and receptor mediated endocytosis but also in specialized endocytic processes such as phagocytosis(Alvarez-Dominguez et al.,1996) and apical endocytosis in polarized cells(Bucci et al., 1994). The previous studies have also demonstrated that Rab5 in mammalian cells is localized on synaptic vesicles (de Hoop et al., 1994; Fischer von Mollard et al., 1994), suggesting a role in endocytic reformation of synaptic vesicles. However, no distinct evidence proving that Rab5 actually functions in the endocytosis of synaptic vesicles has been demonstrated. In addition, in recent model of synaptic vesicle reformation, synaptic vesicles are reformed directly from the endocytic vesicles without endosomal intermediates for sorting (Cremona and De Camilli, 1997; Sudhof,2000), and it would be difficult to interpret the role of Rab5 in synaptic vesicle recycling in this model based on our present knowledge of Rab5.

In the present study, we first investigated the physiological role of Rab5 on synaptic vesicles and found that Rab5 functions in keeping the size of synaptic vesicles uniform, probably by preventing homotypic fusion of the vesicles.

Materials and Methods

Fly strains and antibodies

Fly strains carrying hs-GAL4 and Rh1-GAL4 (F2-1) were donated by K. Sawamoto and C. Hama, respectively. The strains carrying shi1, w1118, Canton-S, B-#1104(w*;P{w+mC=GAL4-ninaE.GMR}12) and B-#1610(w*;Dr1/TMS, P{ry+t7.2=Delta2-3}99B) were supplied by Bloomington Drosophila Stock Center.

To raise antiserum against Drosophila Rab5, 6×His-Rab5 fusion protein was synthesized in Escherichia coli, purified on a polyhistidine resin (Qiagen, Germany) and injected into mice. The antiserum was used at a dilution of 1:300. In order to examine the antigen specificity of the antiserum, Rab proteins (Rab2, Rab4, Rab5, Rab6: with 6×His-tag,Rab3: without a tag) were expressed in E. coli cells(Satoh et al., 1997a), and the reactivity of anti-Rab5 antiserum against these proteins was tested using immunoblot analysis. Rabbit anti-synaptotagmin I (DSYT-2, 1:1,000)(Littleton et al., 1993) was provided by H. Bellen. Mouse anti-syntaxin (8C3, 1:20)(Fujita et al., 1982)developed by S. Benzer was supplied by the Developmental Studies Hybridoma Bank.

Construction of transgenic flies

Rab5 cDNA was cloned as a 1.5 kbp fragment from a Drosophila head cDNA library (Satoh et al.,1997b). Enhanced green fluorescent protein (EGFP) cDNA was amplified from a pEGFP plasmid DNA (Clontech, USA) by PCR, so that the Rab5 sequence could be fused in frame to the C terminus of EGFP. cDNA for Rab5N142I and Rab5Q88L were made by directed mutagenesis of Rab5 cDNA, and each was inserted into a pUAST vector. Wild-type Rab5 cDNA was inserted in reverse,downstream of the upstream activation sequence (UAS) for antisense nucleotide expression. The resulting constructs were injected into the eggs of B-#1610 flies. After crossing with w1118 flies, several heterozygous (balanced over SMCy or TM6B) or homozygous insertion lines were isolated, each containing a single copy of UAS-Rab5. Prior to use for experiments, males of each strain were crossed with females containing the Gal4 gene under the control of heat-shock protein (hsp70) promoter, rh1 opsin promoter or the GMR regulatory element. To express the Rab5N142I gene in shi1 flies, double mutant (Rh1-Gal4 UAS-Rab5N142I/TM6B) males were crossed with female shi1. Expressions of the mutant Rab5 proteins were confirmed by immunoblotting.

Electron microscopy and GFP observation

Conventional electron microscopy was carried out as described(Satoh et al., 1997a). Briefly, the heads were bisected and incubated in pre-fixative (2%paraformaldehyde, 2% glutaraldehyde, 0.1 M cacodylate buffer, pH 7.4) for 2 hours at 4°C, followed by postfixation with 2% OsO4 in 0.1 M cacodylate buffer, pH 7.4. For immunoelectron microscopy, the heads were bisected and incubated in fixative (4% paraformaldehyde, 0.1 M cacodylate buffer, pH 7.4) for 2 hours at 4°C, dehydrated in alcohol and embedded in LR-White (Nisshin EM, Japan). For immunogold labelling, ultrathin sections were incubated in anti-GFP rabbit IgG (1:20, Clontech USA) and then in mouse anti-rabbit IgG-gold conjugate (1:30, British Bio Cell International). For GFP observation, fly heads were bisected and incubated in fixative (4%paraformaldehyde, 0.1 M phosphate buffer, pH 7.4) for 2 hours at 4°C. The tissues were placed into OCT compound (Sakura, Japan) and frozen in isopentane at its melting point. Serial 15-μm sections were mounted on glass slides and observed using a confocal laser scanning microscope (MRC-1024, Bio-Rad,USA).

Subcellular fractionation and in vitro fusion assay

Heads from 3 g Canton-S flies were homogenized in 400 ml buffer A(10 mM HEPES, 1 mM EGTA, 0.1 mM MgCl2, 1 mM PMSF, pH 7.4)(van de Goor et al., 1995). The resulting homogenate was centrifuged at 1000 g for 10 minutes and the supernatant (∼200 μl) was loaded onto a 2-ml 10-30%sucrose gradient over a 200 μl 50% sucrose pad. The gradient was centrifuged at 100,000 g for 1 hour, and twelve 200 μl fractions were collected from top of the gradient. For purification of synaptic vesicles, the homogenate was centrifuged at 18,000 gfor 10 minutes and the supernatant (100 μl) was loaded onto a stepwise sucrose gradient (100 μl 12%, 500 μl 18% and 300 μl 50%). The gradient was centrifuged at 25,000 g for 30 minutes and the top 200 μl fraction containing cytosol and synaptic vesicles was collected. The in vitro reconstitution assay of homotypic fusion between synaptic vesicles was carried out by incubating the purified synaptic vesicles and cytosol with 10 mM ATP, 0.2 mM GTP and 2 mM DTT at 37°C for 15 minutes. For velocity sedimentation analysis, samples were loaded onto a 1.4-ml 10-13%linear sucrose gradient formed on a stepwise sucrose gradient (200 μl 14%,200 μl 18%, 200 μl 28% and 200 μl 50%), and centrifuged at 100,000 g for 2 hours.

Electrophysiology and behaviour assay

Electroretinogram (ERG) recordings were performed on living individuals as described (Larrivee et al.,1981). Flies were light adapted under bright white light for 3 minutes and dark-adapted in complete darkness for 1 minute. To evoke ERG,flies were stimulated with weak orange (540 nm) light. Behaviour assay was carried out under dim-red background light. Ten dark-adapted flies were put into a transparent plastic vial in which flies could walk freely. After 10 seconds of illumination of the vial with diffuse white light, the light was abruptly turned off. The walking of the flies was continuously recorded by a highly sensitive CCD camera, and the number of flies that stopped walking when the lights were turned off was counted. Ten trials were carried out on each strain.

Image analysis

Six eyes were used for quantification of multivesicular endosomes and Golgi apparatus. From each eye, several cross-sections containing photoreceptor nuclei were prepared. From the sections, 18 photoreceptor cells were randomly selected and photographed using an electron microscope (total 108 cells). Organelles were identified by their morphological features, and manually counted. For the histograms showing the size distributions of synaptic vesicles, randomly selected terminals from three eyes were photographed, and the diameters of 1000 vesicles were manually measured. For quantification of the in vitro fusion assay, membrane pellets from four experiments were observed on an electron microscope and the diameters of 250 vesicles were measured. Five eyes were used for quantification of the diameter of the nerve terminals. From each eye, five cartridges (each including six photoreceptor terminals) were photographed and their diameters were manually measured.

Results

Localization of Rab5 in the Drosophila photoreceptor cell

In order to examine whether Rab5 is localized on synaptic vesicles in Drosophila photoreceptor cells, like in mammalian neurons, we expressed EGFP-tagged Drosophila Rab5 in photoreceptor cells. Fig. 1A is a confocal fluorescence micrograph of a longitudinal section of Drosophilaphotoreceptor cells in the retina and lamina layers. The micrograph indicates that strong EGFP fluorescence is observed in the lamina region, where the synaptic terminals of R1-R6 photoreceptor cells are integrated. For precise identification of the Rab5-carrying organelles, we performed immunoelectron microscopy using an anti-GFP antibody. The results indicated that EGFP-tagged Rab5 was localized to endosomes referred to as multivesicular bodies (MVBs)(Fig. 1B) and to the invaginations of the plasma membrane (Fig. 1C) in the photoreceptor cell body. By contrast, gold particles were specifically observed on synaptic vesicles in the nerve terminals(Fig. 1D). No detectable signals were found on the plasma membrane, the capitate projection or mitochondria (Fig. 1D). We further examined the localization of Rab5 using subcellular fractionation of a homogenate of Drosophila heads. The result showed that distribution of Rab5 revealed by the use of Rab5-specific antiserum(Fig. 1E) coincided with that of synaptotagmin, a marker protein of synaptic vesicles(Littleton et al., 1993)(Fig. 1F). These data indicate that Rab5 is abundantly localized on synaptic vesicles in Drosophilaphotoreceptor cells.

Fig. 1.

Localization of Rab5 and its functional inhibition with Rab5N142I. (A) EGFP fluorescence was observed in longitudinal section of the eye from Rh1-GAL4/UAS-EGFP Rab5 fly. R, retina; L, lamina. (B-D)Immunoelectron microscopy using an anti-GFP antibody identified subcellular localization of EGFP-Rab5 in the cell body (B,C) and the nerve terminal (D) of photoreceptor cells. Scale bars, 100 μm (A), 200 nm (B,C), 400 nm (D). (E)Antigen specificity of anti-Rab5 antiserum. (Top) CBB staining of Drosophila Rab2, Rab3, Rab4, Rab5, Rab6 expressed in E. coli. (Bottom) Immunoblot analysis using the anti-Rab5 antiserum against the above Rab proteins. (F) Heads homogenate was fractionated by velocity sedimentation. Fractions containing synaptotagmin (Syt), syntaxin (Syx) or Rab5 were identified by immunoblotting. (G) The numbers of MVBs and Golgi bodies were counted on electron micrographs of photoreceptor cells from wild-type (open bars) or hs-GAL4/UASRab5N142I mutant (shaded bars)flies.

Fig. 1.

Localization of Rab5 and its functional inhibition with Rab5N142I. (A) EGFP fluorescence was observed in longitudinal section of the eye from Rh1-GAL4/UAS-EGFP Rab5 fly. R, retina; L, lamina. (B-D)Immunoelectron microscopy using an anti-GFP antibody identified subcellular localization of EGFP-Rab5 in the cell body (B,C) and the nerve terminal (D) of photoreceptor cells. Scale bars, 100 μm (A), 200 nm (B,C), 400 nm (D). (E)Antigen specificity of anti-Rab5 antiserum. (Top) CBB staining of Drosophila Rab2, Rab3, Rab4, Rab5, Rab6 expressed in E. coli. (Bottom) Immunoblot analysis using the anti-Rab5 antiserum against the above Rab proteins. (F) Heads homogenate was fractionated by velocity sedimentation. Fractions containing synaptotagmin (Syt), syntaxin (Syx) or Rab5 were identified by immunoblotting. (G) The numbers of MVBs and Golgi bodies were counted on electron micrographs of photoreceptor cells from wild-type (open bars) or hs-GAL4/UASRab5N142I mutant (shaded bars)flies.

Inhibition of endosome formation with Rab5N142I in the photoreceptor cell body

To investigate the function of Rab5 on synaptic vesicles, we expressed Rab5N142I, a mutant protein of Drosophila Rab5 equivalent to mammalian Rab5N133I. The mutant protein has a much lower affinity to guanine nucleotides than wild-type protein, and inhibits the function of endogenous Rab5. It has been demonstrated that overexpression of Rab5N133I inhibits endocytosis in a dominant negative manner in mammalian cells(Bucci et al., 1992; Gorvel et al., 1991). To confirm the inhibitory effect of Rab5N142I in Drosophila, we expressed the mutant protein under control of the heat-shock protein (hsp70)promoter and examined by conventional electron microscopy its effect on the formation of MVBs in the somatic domain of photoreceptor cells. After the expression of Rab5N142I, the number of MVBs decreased to approximately a quarter of that in the wild-type fly, whereas the number of Golgi complexes was not affected (Fig. 1F). This result indicates that Rab5N142I works as a specific inhibitor of endogenous Rab5.

Enlargement of synaptic vesicles in vivo with Rab5N142I

We next investigated the effect of Rab5N142I on the synaptic vesicles. In order to achieve sufficient concentration of Rab5N142I in the nerve terminal,the mutant protein was expressed under the control of the Drosophilamajor opsin (rh1) promoter. In contrast to the uniformly sized vesicles in the wild-type fly (Fig. 2A), a number of enlarged synaptic vesicles were distributed throughout the nerve terminal in the mutant fly, some of which were bound to synaptic ribbons (Fig. 2B). These structures were mostly spherical but some had narrow tubular or tubulo-vesicular structures (Fig. 2C). Although the results strongly suggested that Rab5N142I induced enlargement of the synaptic vesicles through the specific inhibition of native Rab5 function, there still remained the possibility that protein overexpression non-specifically caused the enlargement of the synaptic vesicles. To exclude this possibility, we examined the effects of the antisense oligonucleotide of Rab5. As shown in Fig. 2D, expression of the antisense oligonucleotide significantly reduced the amount of endogenous Rab5(inset) and increased the size of the synaptic vesicles in the terminal. These results clearly indicate that Rab5 is essential for creating or maintaining a uniformly sized population of synaptic vesicles. Interestingly, many enlarged and tubular synaptic vesicles also appeared in cells expressing Rab5Q88L, a GTP-binding form of Rab5 (Fig. 2E). This action of Rab5Q88L on synaptic vesicles is very similar to that of Rab5N142I (Fig. 2B,C) (inhibitory to native Rab5). By contrast, Rab5Q88L facilitated the formation of MVBs in the photoreceptor cell body (H. Shimizu et al., unpublished data), supporting the function of Rab5 on synaptic vesicles distinct from that on conventional endocytic vesicles. We further examined whether the synaptic vesicle size is affected by the increase of wild-type Rab5 protein using the fly overexpressing wild-type Rab5-GFP fusion protein. It has been shown in the mammalian cells that the fusion protein normally functions like wild-type Rab5 protein(Nielsen et al., 1999; Roberts et al., 1999). Although the fusion proteins were certainly transported to the photoreceptor nerve terminal (Fig. 1A), they did not cause any defects in synaptic vesicles(Fig. 2F). This result suggests that the mutation of Rab5 in the guanine nucleotide-binding, not the increase of Rab5 protein, might cause the enlargement of synaptic vesicles.

Fig. 2.

Enlargement of synaptic vesicles caused by dysfunction of Rab5. (A-C) Nerve terminals of the photoreceptor cells from wild-type (A) and Rh1-GAL4/UAS-Rab5N142I mutant (B,C) flies. (D) A nerve terminal of the photoreceptor cell expressing antisense Rab5 RNA(GMR-GAL4/UAS-antisenseRab5). Inset, immunoblot analysis of the expression of Rab5 in the eyes of the wild-type (WT) and the GMR-GAL4/UAS-antisenseRab5 (AS) flies. (E) A typical nerve terminal of the Rh1-GAL4/UASRab5Q88L mutant. (F) A nerve terminal of Rh1-GAL4/UAS-EGFP-Rab5. The arrowheads and arrows indicate enlarged and tubular synaptic vesicles, respectively. Scale bar, 250 nm.

Fig. 2.

Enlargement of synaptic vesicles caused by dysfunction of Rab5. (A-C) Nerve terminals of the photoreceptor cells from wild-type (A) and Rh1-GAL4/UAS-Rab5N142I mutant (B,C) flies. (D) A nerve terminal of the photoreceptor cell expressing antisense Rab5 RNA(GMR-GAL4/UAS-antisenseRab5). Inset, immunoblot analysis of the expression of Rab5 in the eyes of the wild-type (WT) and the GMR-GAL4/UAS-antisenseRab5 (AS) flies. (E) A typical nerve terminal of the Rh1-GAL4/UASRab5Q88L mutant. (F) A nerve terminal of Rh1-GAL4/UAS-EGFP-Rab5. The arrowheads and arrows indicate enlarged and tubular synaptic vesicles, respectively. Scale bar, 250 nm.

Suppression of Rab5N142I-induced vesicle enlargement by enhanced recycling of synaptic vesicles

Enlarged synaptic vesicles can be generated either when the budding process from the presynaptic plasma membrane is perturbed(Fergestad et al., 1999; Zhang et al., 1998; Zhang et al., 1999) or when homotypic fusion of small vesicles is abnormally induced. To discover the process in which Rab5 functions, we investigated the influence of Rab5N142I in different conditions of vesicle recycling activity by keeping flies in light(high recycling rate) or dark (low recycling rate). If Rab5N142I influences the homotypic fusion of synaptic vesicles, it would be more effective when vesicles are recycled slowly (i.e. in the dark) and have more chance to fuse with each other. By contrast, if Rab5 functions in the budding step of synaptic vesicles, Rab5N142I would influence the vesicle size, irrespective of the rate of vesicle recycling. Figure 3 shows histograms showing the distributions of synaptic vesicle diameters in wild-type and Rab5N142I-expressing flies. The results demonstrate that enlargement of synaptic vesicles with Rab5N142I was markedly enhanced when vesicles were slowly recycled in the dark, but such enhancement was not observed when the wild-type fly was kept in the dark. This supports the concept in which Rab5N142I causes homotypic fusion of synaptic vesicles and suggests that Rab5 on synaptic vesicles contributes to the prevention of homotypic fusion between the vesicles.

Fig. 3.

Enlargement of the synaptic vesicles in the dark incubation. Histograms indicate the size distributions of synaptic vesicles in the wild-type flies under light illumination (open bars) or dark condition (hatched bars), and in the Rh1-GAL4/UAS-Rab5N142I flies under light illumination (shaded bars) or in the dark (closed bars).

Fig. 3.

Enlargement of the synaptic vesicles in the dark incubation. Histograms indicate the size distributions of synaptic vesicles in the wild-type flies under light illumination (open bars) or dark condition (hatched bars), and in the Rh1-GAL4/UAS-Rab5N142I flies under light illumination (shaded bars) or in the dark (closed bars).

Homotypic fusion of synaptic vesicles in vitro

To verify this idea, we investigated whether homotypic fusion can be reconstituted in vitro with purified synaptic vesicles. Synaptic vesicles and cytosol were prepared from fly heads expressing low amounts of Rab5N142I under hsp70 promoter. The average size of the synaptic vesicles in the mutant fly was slightly larger than that in the wild-type(Fig. 4A,C). After in vitro incubation of synaptic vesicles in cytosol at 37°C for 15 minutes, vesicle size dramatically increased (Fig. 4B,C), demonstrating that synaptic vesicles could be homotypically fused in the presence of Rab5N142I protein. Velocity sedimentation analysis on sucrose gradient also showed that the incubation shifted the synaptic vesicles from lower to upper fractions (Fig. 4D), the latter containing vesicles of larger diameters(Fig. 4E). Neither alteration in vesicle size nor sedimentation velocity shift occurred in the wild-type fly(Fig. 4C,D).

Fig. 4.

Reconstitution of homotypic fusion of synaptic vesicles in vitro. (A,B)Synaptic vesicles prepared from the hs-GAL4/UASRab5N142I mutant flies were electron microscopically observed before (A) and after (B) fusion reaction. Scale bar, 200 nm. (C) The averaged diameters of synaptic vesicles prepared from the wild-type and mutant flies were measured on electron micrographs before (open bars) and after (shaded bars) fusion reaction. (D)Synaptic vesicles prepared from the wild-type and mutant flies were subjected to velocity sedimentation before (0 min) and after (15 min) the fusion reaction, and detected by immunoblotting using anti-synaptotagmin antibodies.(E) Synaptic vesicles from fractions (3+4) and (5+6) of the mutant flies were separately collected and their averaged diameters were measured on the electron micrographs.

Fig. 4.

Reconstitution of homotypic fusion of synaptic vesicles in vitro. (A,B)Synaptic vesicles prepared from the hs-GAL4/UASRab5N142I mutant flies were electron microscopically observed before (A) and after (B) fusion reaction. Scale bar, 200 nm. (C) The averaged diameters of synaptic vesicles prepared from the wild-type and mutant flies were measured on electron micrographs before (open bars) and after (shaded bars) fusion reaction. (D)Synaptic vesicles prepared from the wild-type and mutant flies were subjected to velocity sedimentation before (0 min) and after (15 min) the fusion reaction, and detected by immunoblotting using anti-synaptotagmin antibodies.(E) Synaptic vesicles from fractions (3+4) and (5+6) of the mutant flies were separately collected and their averaged diameters were measured on the electron micrographs.

Exocytosis and endocytosis of synaptic vesicles in the Rab5N142I mutant

Finally, we investigated whether any physiological defects arose from the enlargement of synaptic vesicles caused by the functional inhibition of Rab5. We first examined exocytosis and endocytosis of the enlarged synaptic vesicles using a temperature sensitive shibire mutant(shi1). The shibire gene encodes dynamin, which functions in the 'pinch-off' step of endocytic vesicles from plasma membrane(Chen et al., 1991; van der Bliek and Meyerowitz,1991). In the mutant, endocytosis was completely blocked at 30°C (restrictive temperature) by the dysfunction of dynamin, which (under light stimulus) resulted in the depletion of synaptic vesicles and a concomitant increase in the surface area of the nerve terminal (approximately double the diameter; Fig. 5D). When flies were returned to 20°C (permissive temperature), the nerve terminal again filled with synaptic vesicles through a rapid endocytic reformation, and the diameter of the nerve terminal also recovered(Koenig and Ikeda, 1996). As shown in Fig. 5A, Rab5N142I expressed in the shi1 mutant also induced enlargement of vesicles. These vesicles then disappeared when the flies were kept at the restrictive temperature (Fig. 5B). This result indicated that the enlarged vesicles could fuse directly with the presynaptic plasma membrane like normal synaptic vesicles. When flies were returned to the permissive temperature, not only the normal synaptic vesicles but also the enlarged vesicles were re-formed in the terminal (Fig. 5C). We also estimated the amount of exocytotic release and endocytic reformation of synaptic vesicles in this condition by measuring the diameters of the nerve terminals at each temperature, but there was no significant difference between wild type and Rab5 mutant fly (Fig. 5D). These results indicate that Rab5 is not required for exocytotic fusion or endocytic reformation of synaptic vesicles. Furthermore,it is also suggested that the enlarged synaptic vesicles induced by the interference of Rab5 function can undergo exocytotic fusion.

Fig. 5.

Membrane cycle of enlarged synaptic vesicles induced by Rab5N142I. (A-C)Cross sections of photoreceptor nerve terminals indicate the exo- and endocycle of synaptic vesicles in the shi1;;Rh1-GAL4 UAS-Rab5N142I/+ double mutant. After keeping the mutant flies at 20°C(A), they were put at 30°C for 1 minute under light illumination (B)followed by the incubation at 20°C for 30 minutes (C). Scale bar, 500 nm.(D) Changes in the diameter of the R1-6 photoreceptor nerve terminals of shi1 (open bars) and shi1;;Rh1-GAL4 UAS-Rab5N142I/+ (shaded bars) mutants caused by the above temperature shift.

Fig. 5.

Membrane cycle of enlarged synaptic vesicles induced by Rab5N142I. (A-C)Cross sections of photoreceptor nerve terminals indicate the exo- and endocycle of synaptic vesicles in the shi1;;Rh1-GAL4 UAS-Rab5N142I/+ double mutant. After keeping the mutant flies at 20°C(A), they were put at 30°C for 1 minute under light illumination (B)followed by the incubation at 20°C for 30 minutes (C). Scale bar, 500 nm.(D) Changes in the diameter of the R1-6 photoreceptor nerve terminals of shi1 (open bars) and shi1;;Rh1-GAL4 UAS-Rab5N142I/+ (shaded bars) mutants caused by the above temperature shift.

Impaired synaptic transmission in the Rab5N142I mutant

To examine whether such exocytotic fusion of enlarged synaptic vesicles is functional in synaptic transmission, we next performed ERG recording, an extracellularly recorded, light-evoked mass response of the eye. In the ERG,on and off transients are generated by neurons postsynaptic to R1-6 photoreceptor cells, and represent synaptic transmission between photoreceptor cells and lamina neurons (Kelly,1983). As shown in Fig. 6A, the ERGs from the wild-type fly demonstrated both on- and off-transient responses. By contrast, ERG from the mutant expressing dominant negative Rab5 protein contained no on-transient response. Furthermore, ERG of the light-adapted mutant fly showed small off-transient response, which then disappeared when the fly was dark-adapted(Fig. 6A). The electron microscopy observation demonstrated that the mutant flies contained enlarged synaptic vesicles in light, which further grew during dark adaptation(Fig. 3). Therefore, the above results indicated that the enlargement of synaptic vesicles greatly reduced the efficiency of synaptic transmission, although the vesicles were still able to fuse to presynaptic membrane (Fig. 5). We further examined synaptic transmission between photoreceptor cells and lamina neurons by measuring the stop-walk response of flies. It has been reported that wild-type flies suddenly stop walking and exhibit an unusual jump response to a light-off stimulus. Such behaviour is closely associated with the amplitude of the off-transient response of ERG(Kelly, 1983). In the dominant negative mutant of Rab5, the number of flies exhibiting the stop-walk response dramatically reduced (Fig. 6B),indicating again that the enlargement of synaptic vesicles lowered the efficiency of synaptic transmission.

Fig. 6.

Impaired synaptic transmission in Rab5N142I-expressing fly. (A) ERG recordings from the light-adapted (L) and dark-adapted (D) wild-type and Rh1-GAL4/UAS-Rab5N142I flies. Arrowheads indicate the off-transient responses. (B) The proportion of wild-type and Rh1-GAL4/UASRab5N142Iflies exhibiting stop-walk responses against light-off stimuli.

Fig. 6.

Impaired synaptic transmission in Rab5N142I-expressing fly. (A) ERG recordings from the light-adapted (L) and dark-adapted (D) wild-type and Rh1-GAL4/UAS-Rab5N142I flies. Arrowheads indicate the off-transient responses. (B) The proportion of wild-type and Rh1-GAL4/UASRab5N142Iflies exhibiting stop-walk responses against light-off stimuli.

Discussion

In the present study, we found that the dysfunction of Rab5 induced the enlargement of synaptic vesicles in vivo. Moreover, homotypic fusion occurred in vitro on the synaptic vesicles prepared from flies expressing dominant negative mutant Rab5. This is the first report demonstrating that synaptic vesicle can potentially fuse with each other (homotypic fusion). Nevertheless,no such fusion or enlargement of synaptic vesicles occurs in the nerve terminal of the wild-type fly. This finding therefore indicates that homotypic fusion of synaptic vesicles is effectively blocked by Rab5 in vivo. Based on these findings, we now propose a novel role for Rab5 on synaptic vesicles as an essential factor preventative of their homotypic fusion.

Neurons have to release neurotransmitter quantitatively, corresponding to the extent of their excitation. They control the released amount of transmitter by packing it into small, uniformly sized vesicles and limiting the number of vesicles fusing with the presynaptic plasma membrane. To generate uniformly sized vesicles, clathrin and adaptor proteins (APs)assemble a cage with hexagonal and pentagonal faces, which recruits a constant amount of presynaptic membrane (Zhang et al., 1999). In fact, it has been shown in Drosophila and Caenorhabditis elegans that mutations in AP180 (lap/unc-11) and stoned protein increase the size of the synaptic vesicles(Fergestad et al., 1999; Nonet et al., 1999; Zhang et al., 1998). In the present study, we demonstrated that synaptic vesicles possess potential ability to fuse homotypically. Unregulated homotypic fusion, however,decimates the vesicle-size uniformity generated by clathrin-adaptor protein system and thus perturbs quantitative control of neurotransmitter release. The role of Rab5 that underlies the prevention of homotypic fusion of synaptic vesicles is therefore essential to regulate synaptic transmission quantitatively.

We observed here that many enlarged and tubular synaptic vesicles appeared in cells expressing Rab5Q88L, a GTP-binding form of Rab5(Fig. 2F). This phenotype suggests that GTP-binding Rab5 promotes the synaptic vesicle fusion. In addition, expression of Rab5N142I, a GTP/GDP-free form of Rab5, did not function for preventing homotypic fusion of the vesicles but rather induced the vesicle fusion like Rab5Q88L. Superficially, this effect of Rab5N142I on synaptic vesicles seems to contradict previous work on the role of Rab5 in conventional endosome fusion: in this case, the dominant negative mutant decreases vesicle fusion. However, this discrepancy can be interpreted by postulating that Rab5N142I stochastically takes both conformations mimicking GTP-binding and GDP-binding forms of Rab5. On endosome fusion, GTP-binding form of Rab5 (active Rab5) mediates vesicle fusion and overexpression of Rab5N142I supplies inactive form of Rab5 abundantly, which then inhibits the function of active Rab5. By contrast, Rab5 on synaptic vesicles predominantly binds GDP (Stahl et al.,1994). Overexpression of Rab5N142I might provide a significant amount of Rab5 mimicking the GTP-binding form, which then induces vesicle fusion between synaptic vesicles. Although the precise molecular mechanism for the dominant negative action of Rab5N142I has not been elucidated, our present result that overexpression of the wild-type Rab5, which probably takes a GDP-binding form in the nerve terminal, does not induce homotypic fusion supports above interpretation. Because the decrease of intrinsic Rab5 with the expression of antisense Rab5 mRNA also induced the homotypic fusion of synaptic vesicles, the presence of the GDP-binding Rab5 or sufficient Rab5 normally regulated between GDP- and GTP-binding forms might be essential to prevent the homotypic fusion of synaptic vesicles. According to this hypothesis, it is also suggested that Rab5 might control the homotypic fusion of synaptic vesicles through the GTP-GDP exchange. Recently,electrophysiological studies have suggested that the homotypic fusion of synaptic vesicles possibly occurs when neurotransmitter is massively released from the nerve terminal (reviewed in Parsons and Sterling, 2003). In addition to the prevention of homotypic fusion, Rab5 might be involved in the neurotransmitter release.

After exocytotic release of neurotransmitters, synaptic vesicles must be re-formed by a rapid endocytic process in order to sustain the population of synaptic vesicles during the period of synaptic activation. Previous studies have proposed two different models: endosomal recycling and direct recycling(Cremona and De Camilli, 1997; Mundigl and De Camilli, 1994; Sudhof, 2000). The endosomal-recycling model, in which the vesicles recycle via an early endosome, was first proposed after the observation that endosome-like cisternae appear after extensive stimulation of frog nerve terminals(Heuser and Reese, 1973). Thereafter, localization of Rab5 on synaptic vesicles and the formation of large vacuoles after the overexpression of Rab5Q79L in primary culture of hippocampal neurons have been demonstrated(de Hoop et al., 1994; Fischer von Mollard et al.,1994) and used to support this model. By contrast, vesicles recycle directly after endocytosis without any endosomal intermediate in the direct-recycling model. This model was originally proposed in the early 1970s(Ceccarelli et al., 1973) and was recently resurrected by several lines of evidence(Murthy and Stevens, 1998; Schmidt et al., 1997; Takei et al., 1996). Although these two pathways are quite different from each other, recent studies suggest that they are not mutually exclusive and that multiple pathways, and thus multiple kinds of vesicle pools, could participate in synaptic vesicle recycling (Kuromi and Kidokoro,1998; Richards et al.,2000; Shi et al.,1998). In the present study, we have demonstrated that the functional inhibition of Rab5 has absolutely no effect on the endocytic reformation of synaptic vesicles in the Drosophila photoreceptor nerve terminal. In this analysis, we used the shibire mutant, whose defective gene encodes dynamin. When mutant flies were kept at the restricted temperature, their photoreceptor nerve terminals are almost completely depleted of synaptic vesicles. After a shift to the permissive temperature,vesicles are immediately recovered not only in the active zone but also in other regions of the terminal, even when Rab5 inhibitor (Rab5N142I) is adequately expressed. These results thus indicate that, even if multiple pathways for synaptic vesicle recycling are present in the Drosophilanerve terminal (Koenig and Ikeda,1996; Kuromi and Kidokoro,1998), none of them requires Rab5 for vesicle reformation.

In electrophysiological and behavioural analyses of the Rab5N142I-expressing mutant, we found that ERG of the mutant was lacking on- and off-transient responses. Because these transient responses represent synaptic transmission between photoreceptor cells and lamina neurons, this finding indicated that the synaptic transmission was impaired by the expression of Rab5N142I. When Rab5N142I was expressed in the shi1 mutant, enlarged synaptic vesicles and normal vesicles were cleared from nerve terminal by light stimulus. This result demonstrated that enlarged vesicles were able to fuse with presynaptic plasma membrane like normal vesicles. Furthermore, off-transients in the Rab5N142I-expressing fly were recovered when fly was adapted in the light. In electron microscopic study, we demonstrated that enlargement of synaptic vesicles was suppressed by enhanced recycling of the synaptic vesicles by light illumination. Therefore, the loss of the off-transient is probably due to enlargement of synaptic vesicles. These results strongly suggest that fusion of enlarged synaptic vesicles with presynaptic membrane is not efficient enough to give transient signals in ERG, although the vesicles still keep the competence of exocytotic fusion. Correct sizing of synaptic vesicles must be essential not only for quantitatively controlled but also for highly facilitated release of neurotransmitter.

Acknowledgements

We thank H. J. Bellen for providing anti-synaptotagmin I antibody and are grateful to A. K. Satoh for her helpful discussion and technical advice. We also thank K. Sawamoto and C. Hama for donating fly strains. This work was partly supported by a Grant-in-Aid for Scientific Research to K.O., and the JSPS Research for the Future Program to S.K. H.S. was supported by the JSPS Research Fellowships for Young Scientists.

References

Alvarez-Dominguez, C., Barbieri, A. M., Beron, W.,Wandinger-Ness, A. and Stahl, P. D. (
1996
). Phagocytosed live Listeria monocytogenes influences Rab5-regulated in vitro phagosome-endosome fusion.
J. Biol. Chem.
271
,
13834
-13843.
Bajjalieh, S. (
2001
). SNAREs take the stage: a prime time to trigger neurotransmitter secretion.
Trends Neurosci.
24
,
678
-680.
Brunger, A. T. (
2001
). Structural insights into the molecular mechanism of calcium-dependent vesicle-membrane fusion.
Curr. Opin. Struct. Biol.
11
,
163
-173.
Bucci, C., Parton, R. G., Mather, I. H., Stunnenberg, H.,Simons, K., Hoflack, B. and Zerial, M. (
1992
). The small GTPase Rab5 functions as a regulatory factor in the early endocytic pathway.
Cell
70
,
715
-728.
Bucci, C., Wandinger-Ness, A., Lutcke, A., Chiariello, M.,Bruni, C. B. and Zerial, M. (
1994
). Rab5a is a common component of the apical and basolateral endocytic machinery in polarized epithelial cells.
Proc. Natl. Acad. Sci. USA
91
,
5061
-5065.
Ceccarelli, B., Hurlbut, W. P. and Mauro, A.(
1973
). Turnover of transmitter and synaptic vesicles at the frog neuromuscular junction.
J. Cell Biol.
57
,
499
-524.
Chavrier, P., Parton, R. G., Hauri, H. P., Simons, K. and Zerial, M. (
1990
). Localization of low molecular weight GTP binding proteins to exocytic and endocytic compartments.
Cell
62
,
317
-329.
Chen, M. S., Obar, R. A., Schroeder, C. C., Austin, T. W.,Poodry, C. A., Wadsworth, S. C. and Vallee, R. B. (
1991
). Multiple forms of dynamin are encoded by shibire, a Drosophila gene involved in endocytosis.
Nature
351
,
583
-586.
Cochilla, A. J., Angleson, J. K. and Betz, W. J.(
2000
). Differential regulation of granule-to-granule and granule-to-plasma membrane fusion during secretion from rat pituitary lactotrophs.
J. Cell Biol.
150
,
839
-848.
Coorssen, J. R., Blank, P. S., Tahara, M. and Zimmerberg, J.(
1998
). Biochemical and functional studies of cortical vesicle fusion: the SNARE complex and Ca2+ sensitivity.
J. Cell Biol.
143
,
1845
-1857.
Cremona, O. and De Camilli, P. (
1997
). Synaptic vesicle endocytosis.
Curr. Opin. Neurobiol.
7
,
323
-330.
de Hoop, M. J., Huber, L. A., Stenmark, H., Williamson, E.,Zerial, M., Parton, R. G. and Dotti, C. G. (
1994
). The involvement of the small GTP-binding protein Rab5a in neuronal endocytosis.
Neuron
13
,
11
-22.
Fergestad, T., Davis, W. S. and Broadie, K.(
1999
). The stoned proteins regulate synaptic vesicle recycling in the presynaptic terminal.
J. Neurosci.
19
,
5847
-5860.
Fischer von Mollard, G., Stahl, B., Walch-Solimena, C., Takei,K., Daniels, L., Khoklatchev, A., De Camilli, P., Sudhof, T. C. and Jahn,R. (
1994
). Localization of Rab5 to synaptic vesicles identifies endosomal intermediate in synaptic vesicle recycling pathway.
Eur. J. Cell Biol.
65
,
319
-326.
Fujita, S. C., Zipursky, S. L., Benzer, S., Ferrus, A. and Shotwell, S. L. (
1982
). Monoclonal antibodies against the Drosophila nervous system.
Proc. Natl. Acad. Sci. USA
79
,
7929
-7933.
Gorvel, J. P., Chavrier, P., Zerial, M. and Gruenberg, J.(
1991
). Rab5 controls early endosome fusion in vitro.
Cell
64
,
915
-925.
Heuser, J. E. and Reese, T. S. (
1973
). Evidence for recycling of synaptic vesicle membrane during transmitter release at the frog neuromuscular junction.
J. Cell Biol.
57
,
315
-344.
Kelly, L. E. (
1983
). An altered electroretinogram transient associated with an unusual jump response in a mutant of Drosophila.
Cell Mol. Neurobiol.
3
,
143
-149.
Koenig, J. H. and Ikeda, K. (
1996
). Synaptic vesicles have two distinct recycling pathways.
J. Cell Biol.
135
,
797
-808.
Kuromi, H. and Kidokoro, Y. (
1998
). Two distinct pools of synaptic vesicles in single presynaptic boutons in a temperature-sensitive Drosophila mutant, shibire.
Neuron
20
,
917
-925.
Larrivee, D. C., Conrad, S. K., Stephenson, R. S. and Pak, W. L. (
1981
). Mutation that selectively affects rhodopsin concentration in the peripheral photoreceptors of Drosophilamelanogaster.
J. Gen. Physiol.
78
,
521
-545.
Littleton, J. T., Bellen, H. J. and Perin, M. S.(
1993
). Expression of synaptotagmin in Drosophilareveals transport and localization of synaptic vesicles to the synapse.
Development
118
,
1077
-1088.
Mundigl, O. and De Camilli, P. (
1994
). Formation of synaptic vesicles.
Curr. Opin. Cell Biol.
6
,
561
-567.
Murthy, V. N. and Stevens, C. F. (
1998
). Synaptic vesicles retain their identity through the endocytic cycle.
Nature
392
,
497
-501.
Nielsen, E., Severin, F., Backer, J. M., Hyman, A. A. and Zerial, M. (
1999
). Rab5 regulates motility of early endosomes on microtubules.
Nat. Cell Biol.
1
,
376
-382.
Nonet, M. L., Holgado, A. M., Brewer, F., Serpe, C. J., Norbeck,B. A., Holleran, J., Wei, L., Hartwieg, E., Jorgensen, E. M. and Alfonso,A. (
1999
). UNC-11, a Caenorhabditis elegans AP180 homologue, regulates the size and protein composition of synaptic vesicles.
Mol. Biol. Cell
10
,
2343
-2360.
Novick, P. and Zerial, M. (
1997
). The diversity of Rab proteins in vesicle transport.
Curr. Opin. Cell Biol.
9
,
496
-504.
Parsons, T. D. and Sterling, P. (
2003
). Synaptic ribbon. Conveyor belt or safety belt?
Neuron
37
,
379
-382.
Richards, D. A., Guatimosim, C. and Betz, W. J.(
2000
). Two endocytic recycling routes selectively fill two vesicle pools in frog motor nerve terminals.
Neuron
27
,
551
-559.
Roberts, R. L., Barbieri, M. A., Pryse, K. M., Chua, M.,Morisaki, J. H. and Stahl, P. D. (
1999
). Endosome fusion in living cells overexpressing GFP-Rab5.
J. Cell Sci.
112
,
3667
-3675.
Satoh, A., Tokunaga, F., Kawamura, S. and Ozaki, K.(
1997a
). In situ inhibition of vesicle transport and protein processing in the dominant negative Rab1 mutant of Drosophila.
J. Cell Sci.
110
,
2943
-2953.
Satoh, A. K., Tokunaga, F. and Ozaki, K.(
1997b
). Rab proteins of Drosophila melanogaster: novel members of the Rab-protein family.
FEBS Lett.
404
,
65
-69.
Schmidt, A., Hannah, M. J. and Huttner, W. B.(
1997
). Synaptic-like microvesicles of neuroendocrine cells originate from a novel compartment that is continuous with the plasma membrane and devoid of transferrin receptor.
J. Cell Biol.
137
,
445
-458.
Shi, G., Faundez, V., Roos, J., Dell'Angelica, E. C. and Kelly,R. B. (
1998
). Neuroendocrine synaptic vesicles are formed in vitro by both clathrin-dependent and clathrin-independent pathways.
J. Cell Biol.
143
,
947
-955.
Stahl, B., von Mollard, G. F., Walch-Solimena, C. and Jahn,R. (
1994
). GTP cleavage by the small GTP-binding protein Rab3A is associated with exocytosis of synaptic vesicles induced by alpha-latrotoxin.
J. Biol. Chem.
269
,
24770
-24776.
Stenmark, H., Parton, R. G., Steele-Mortimer, O., Lutcke, A.,Gruenberg, J. and Zerial, M. (
1994
). Inhibition of Rab5 GTPase activity stimulates membrane fusion in endocytosis.
EMBO J.
13
,
1287
-1296.
Sudhof, T. C. (
2000
). The synaptic vesicle cycle revisited.
Neuron
28
,
317
-320.
Takei, K., Mundigl, O., Daniell, L. and De Camilli, P.(
1996
). The synaptic vesicle cycle: a single vesicle budding step involving clathrin and dynamin.
J. Cell Biol.
133
,
1237
-1250.
van de Goor, J., Ramaswami, M. and Kelly, R.(
1995
). Redistribution of synaptic vesicles and their proteins in temperature-sensitive shibirets1 mutant
Drosophila. Proc. Natl. Acad. Sci. USA
92
,
5739
-5743.
van der Bliek, A. M. and Meyerowitz, E. M.(
1991
). Dynamin-like protein encoded by the Drosophila shibire gene associated with vesicular traffic.
Nature
351
,
411
-414.
Williams, J. (
1997
). How does a vesicle know it is full?
Neuron
18
,
683
-686.
Zhang, B., Koh, Y. H., Beckstead, R. B., Budnik, V., Ganetzky,B. and Bellen, H. J. (
1998
). Synaptic vesicle size and number are regulated by a clathrin adaptor protein required for endocytosis.
Neuron
21
,
1465
-1475.
Zhang, B., Ganetzky, B., Bellen, H. J. and Murthy, V. N.(
1999
). Tailoring uniform coats for synaptic vesicles during endocytosis.
Neuron
23
,
419
-422.