γ-Amino butyric acid (GABA) release in the ciliated protozoon Paramecium occurs by neuronal-like exocytosis

Eukaryotic cells release molecules by fusing secretory vesicles with the plasma membrane, and, in fact, all cell types appear to have a constitutive pathway for secretion. Some specialized cells, such as neurons, also maintain a separate reservoir of vesicles that secrete their content only in response to specific stimuli, a phenomenon called regulated exocytosis (Kelly, 1985).

Vesicular exocytosis is the physiological mechanism of neurotransmitter release from axon terminals. Classical exocytosis occurs when the terminal plasma membrane depolarizes and voltage-sensitive calcium channels (VSCCs) open. This influx of Ca²⁺ causes previously docked vesicles to undergo Ca²⁺-dependent fusion and to release their messenger content directly into the extracellular space (Schweizer et al., 1995; Südhof, 1995). There is convincing evidence that exocytosis involves the assembly of core complexes from the so-called SNARE proteins. Core complexes are composed of the plasma membrane target proteins syntaxin and SNAP-25 (synaptosome-associated protein of 25 kDa) and different isoforms of the synaptic vesicle protein VAMP (vesicle-associated membrane protein) (Schweizer et al., 1995; Südhof, 1995). The core complex formation allows vesicles to dock at the nerve terminal plasma membrane, the first step of regulated exocytosis. Disassembly of the SNARE complex also involves other proteins that prime vesicles and make them competent for exocytosis (Fasshauer et al., 1997; Fukuda et al., 2000; Katz and Brennwald, 2000; Weber et al., 1998). These proteins include homohexameric ATPase, NSF and its adaptor protein α-soluble NSF attachment protein (α-SNAP). Additional proteins, such as the Ca²⁺ sensor synaptotagmin, facilitate fusion with the plasma membrane.

In the motile, ciliated, protozoon Paramecium, constitutive and stimulated exocytosis of dense-core vesicles (trichocysts) is a thoroughly studied process (Plattner et al., 1991; Vayssié et al., 2000). Trichocysts are defensive organelles that are secreted through a pathway of regulated exocytosis in response to extracellular stimulation (Plattner, 2002; Plattner and Kissmehl, 2003; Vayssié et al., 2000). Evidence originating from biochemical, electrophysiological and calcium-imaging experiments shows that a transient subcortical increase of intracellular free Ca²⁺ is necessary for exocytotic membrane fusion in Paramecium. This process involves a stimulus-dependent influx of Ca²⁺ from the external medium and a stimulus-dependent release of Ca²⁺ from internal stores (Erxleben et al., 1997; Kerboeuf and Cohen, 1990; Klauke and Plattner, 1997; Stelly et al., 1995).

We have previously reported that GABA_A and GABA_B receptors are expressed in Paramecium, where they regulate motor behavior and membrane potential (Bucci et al., 2005; Ramoino et al., 2003). In the present study, we investigated the presence and release of γ-amino butyric acid (GABA), the natural agonist of the above receptors. We demonstrated that this amino acid and the machinery for its synthesis, storage and release are present in Paramecium. In addition, we showed that GABA is released into the environment in both a spontaneous and stimulus-dependent manner. This release may exert GABA's autocrine regulatory functions. Finally, we found that GABA release depends on the presence of extracellular calcium and is blocked by bafilomycin and botulinum toxin C1 serotype (BoNT/C1).

Experiments were carried out on Paramecium primaurelia Sonneborn, stock 90, cultured at 25°C in lettuce medium (pH 6.9) inoculated with Enterobacter aerogenes (Sonneborn, 1970). Cells were harvested in the late log phase of growth.

Monoclonal antibodies anti-sintaxin-1 and anti-SNAP25 as well as polyclonal rabbit anti-VAMP-3 (cellulobrevin) were purchased from SynapticSystem (Goettingen, Germany); the polyclonal guinea pig anti-vGAT (vesicular GABA transporter), rabbit anti-GAD_65/67 (glutamate decarboxylase) and rabbit anti-GABA were obtained from Chemicon International (Temecula, CA, USA); the monoclonal antibody anti-GABA was obtained from Sigma Chemical Co. (St Louis, MO, USA). The secondary antibodies anti-mouse Alexa Fluor 488, anti-rabbit Alexa Fluor 488, anti-guinea pig Alexa Fluor 488, anti-mouse Alexa Fluor 594 and anti-rabbit Alexa Fluor 594 were obtained from Molecular Probes, Invitrogen (Carlsbad, CA, USA).

Pre-stained molecular mass markers were obtained from Amersham (Buckinghamshire, England); BoNT/C1 was purchased from Wako (Osaka, Japan); remaining chemicals were purchased from Sigma unless otherwise specified in the text.

Cells were fixed in 4% paraformaldehyde in phosphate buffered saline (PBS) buffer (0.01 mol l⁻¹, pH 7.4) for 30 min, washed three times with PBS and incubated for 60 min with 3% bovine serum albumin (BSA) in PBS plus 1% Triton X-100. The blocking permeabilizing buffer was removed, and the cells were incubated overnight at 4°C with one or two of the following antibodies: mouse anti-GABA (1:1000), mouse anti-syntaxin-1 (1:100), mouse anti-SNAP25 (1:500), rabbit anti-VAMP3 (1:100), rabbit anti-GABA (1:400), rabbit anti-GAD_65/67 (1:400) or guinea pig anti-vGAT (1:2000). After three washes in 1% BSA in PBS plus 0.1% Triton X-100 for 10 min each, the appropriate secondary antibody conjugated to Alexa Fluor 488 (dilution 1:300) was applied for 1 h at 37°C. After extensive washing in PBS, cells were mounted in glycerol/buffer. In control experiments, the absence of cross-reactivity between the secondary antibodies was verified by omitting one of the primary antibodies during incubation. In addition, for every combination involving double labeling, singly labeled vesicles were always observed in the cells. The specificity of primary antibodies was tested by preabsorbion with the appropriate immunogen for GABA, GAD and vGAT (Chemicon) for 2 h at room temperature before processing Paramecium cells. No staining was detected in the negative control experiments.

Images (1024×1024×8 bit) were acquired with a Leica TCS SP2 Confocal Scanner (Mannheim, Germany) equipped with an Ar (457–476–488–514 nm) 100 mW laser and a HeNe (543 nm) 1.5 mW laser using a one Airy disk unit pinhole diameter and an oil immersion HCX PL APO ×100/1.4 objective (Diaspro et al., 2006). Alexa Fluor 488 was excited with the 488 nm line of the Ar laser, and its fluorescence was collected in a spectral window of 500–530 nm. Alexa 594 fluorescence was collected using the 543 nm excitation wavelength and a 590–620 nm spectral window emission. Spectral windows were selected by the acousto-optic beam splitter of the Leica SP2 scanning head. Serial optical sections were taken through the cell at a z-step of 75 nm. The image acquisition through green and red channels was performed according to a sequential acquisition protocol to reject possible cross-talk artefacts.

The Leica Confocal Software programme was used for image acquisition, storage and analysis. Labeling experiments were repeated 4–5 times, and images are representative of observations of an average of 30 cells in each sample.

Cells were centrifuged at 600 g to a density of 200×10³ cells ml⁻¹ and re-suspended in water containing a protease inhibitor cocktail. Samples were homogenized, sonicated and subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE) followed by western blotting. 100 μg of Paramecium protein and 10 μg of rat cerebral cortex protein per lane were used. The appropriate pre-stained molecular mass markers were run concomitantly. Transfer to nitrocellulose (NC) membranes was performed electrophoretically at 100 V for 75 min. After saturation with 5% non-fat dry milk, the NC membranes were incubated for 1 h with the following primary antibodies: mouse anti-syntaxin-1 (1:5000), mouse anti-SNAP25 (1:1000), rabbit anti-VAMP3 (1:10,000), rabbit anti-GAD_65/67 (1:1000) and guinea pig anti-vGAT (1:1000). The membrane was incubated with the appropriate horseradish peroxidase-conjugated (sheep) secondary antibody and coated using the ECL western blotting detection system for 1 min. The membrane was immediately exposed to Amersham autoradiography film at room temperature for various periods (from 5 s to 60 min) in a film cassette.

BoNT/C1 was activated by reduction in 200 mmol l⁻¹ Tris-HCl (pH 8.0), 500 mmol l⁻¹ NaCl and 50 μmol l⁻¹ ZnCl₂ with 5 mmol l⁻¹ dithiothreitol for 30 min at 37°C to release the light chain (Schilde et al., 2008). Crude cell lysates from P. primaurelia were incubated with BoNT/C1 (final concentration 500 nmol l⁻¹) for 1 h at 37°C. The protein was methanol precipitated and analyzed by SDS-PAGE. Syntaxin was detected on western blots with an anti-rat syntaxin-1 monoclonal antibody.

Cells (100×10³ ml⁻¹) were adapted in the adaptation solution (CaCl₂ 1 mmol l⁻¹, Hepes 1 mmol l⁻¹, pH 7.4) for 10 min and were then treated with 20, 30 or 40 mmol l⁻¹ KCl (final concentration) for 10 min. After treatment, the free-cell medium was analyzed by high performance liquid chromatography (HPLC). GABA release was studied in standard or Ca²⁺-free medium. In some experiments, Paramecium cells were pre-treated with bafilomycin (500 μmol l⁻¹; 60 min) or with BoNT/C1 (40 nmol l⁻¹; 24 h). The control release samples were cells that had been adapted in medium for 10 min without depolarization.

The amount of intracellular and released GABA was determined by HPLC. The analytical method involved pre-column derivatization with o-phthalaldehyde followed by separation on a C₁₈ reverse-phase chromatography column (Chrompack, Middleburg, The Netherlands; 10 mm × 4.6 mm, 3 μm; 30°C) coupled with fluorometric detection (excitation wavelength 350 nm; emission wavelength 450 nm) (Raiteri et al., 2000). The amount of GABA in the biological samples was determined by interpolating the peak area of the Paramecium sample using a curve obtained with different standard GABA solutions. Areas have been normalized using homoserine as an internal standard. The sensitivity of the assay amounted to about 0.1 pmol GABA.

Differences between means (±s.e.m.) were determined using the Student's t-test or analysis of variance (ANOVA) with Dunnett's or Newman–Keuls post-hoc tests (GraphPad Prism, GraphPad, San Diego, CA, USA). Tests were repeated on four different occasions over several weeks.

The localization and distribution of GABA in P. primaurelia were examined using a polyclonal antibody and laser confocal microscopy. Labeling was detected primarily at the cell surface and in the food vacuoles of cells in the late log phase of growth (Fig. 1A). Fig. 1B documents staining of dot-like structures located on the outlines of the regularly arranged surface fields (kinetics) that are characteristic of Paramecium cells.

The presence of the biosynthetic enzyme GAD and vGAT was determined by both immunofluorescence and immunoblotting. Using confocal microscopy, GAD- and vGAT-like immunoreactivity was detected at the cell surface and inside the cytoplasm (Fig. 1C,D). No immunostaining was observed in the negative controls when the primary antibody was omitted or when the primary antibody was pre-absorbed by GABA or the specific GAD or vGAT immunogen peptides.

Immunoblots showed two bands with an apparent molecular weight of about 67 kDa and 57 kDa, using antibodies raised against mammalian GAD and vGAT, respectively (Fig. 1E, lane 2). These molecular mass values were consistent with those obtained in synaptosomes prepared from rat cerebral cortex under the same experimental conditions (Fig. 1E, lane 1).

Transmitter release at the nerve terminal involves assembly of the SNARE complex proteins. VAMPs, in conjunction with syntaxins and SNAP-25, are thought to play a role in docking synaptic vesicles to the presynaptic plasma membrane (Rothman and Wieland, 1996; Sudhof, 1995). Evidence from both confocal microscopy and immunoblot indicated the presence in Paramecium of putative VAMP, syntaxin and SNAP proteins, which were identified using antibodies raised against mammalian proteins (monoclonal anti-syntaxin-1 and anti-SNAP-25 antibodies and the polyclonal antibody anti-cellubrevin/VAMP3 that cross-reacts with both synaptobrevins VAMP1 and VAMP2). By confocal microscopy, immunoreactivity appeared localized as a punctate pattern on the cell surface (Fig. 2A–C). Negative controls, obtained by omitting the primary antibody, showed no immunostaining.

Western blots of proteins derived from Paramecium cells, performed using anti-syntaxin-1 antibody, revealed two bands with estimated molecular masses of about 32 kDa and 36 kDa (Fig. 2D, line 2). Bands with an apparent molecular mass of about 20 kDa and 25 kDa were detected using a broad-spectrum anti-VAMP antibody and an anti-SNAP-25 antibody, respectively (Fig. 2D, line 2). These values are consistent with those obtained for rat cerebral cortex synaptosomes (Fig. 2D, line 1).

Double-labeling experiments in Paramecium not only demonstrated the relationship between GABA and the putative vGAT, which transports GABA into vesicles, but also showed the relationship between GABA and vesicular VAMP. GABA was stained with Alexa Fluor 594 (red fluorescence) whereas vGAT and VAMP were stained with Alexa Fluor 488 (green fluorescence). GABA was partially co-localized (yellow fluorescence) with vGAT (Fig. 3A) and VAMP (Fig. 3B), a co-localization that may occur at the vesicular level. Co-localization along the z-axes was demonstrated by the similarity of the green and red z-profiles of the fluorescence intensity of three double-stained vesicles collected at the top (Fig. 3Ai,Bi), middle (Fig. 3Aii,Bii) and bottom (Fig. 3Aiii,Biii) planes of a 2-μm thick z-stack.

We used KCl depolarization (from 0 mmol l⁻¹ up to 40 mmol l⁻¹ KCl; 10 min exposure) in the presence of Ca²⁺ to induce GABA release in Paramecium. The amino acid was spontaneously released into the environment (control condition, 0 mmol l⁻¹ KCl) but the amount of released amino acid increased after K⁺ depolarization in a concentration-dependent manner (Fig. 4A). The GABA released in the 10-min fraction collected before the onset of K⁺ stimulation amounted to 32±3.6 pmol 10⁴ cells⁻¹. The K⁺-evoked overflow was 65±7.0, 120±7.8 and 160±10.3 pmol 10⁴ cells⁻¹ at 20, 30 and 40 mmol l⁻¹ KCl, respectively. In these sets of experiments, the total GABA content in the paramecia amounted to about 2450±6.0 pmol 10⁴ cells⁻¹.

A lack of extracellular Ca²⁺ abolished the depolarization-induced GABA release (Fig. 4B), suggesting the involvement of regulated exocytosis. The involvement of vesicular GABA in amino acid release is supported by experiments with bafilomycin-A1, which was expected to prevent the accumulation of GABA into vesicles (Moriyama and Futai, 1990; Roseth et al., 1995). Cells were first incubated with bafilomycin-A1 (0.5 μmol l⁻¹, 60 min) and then exposed to 30 mmol l⁻¹ KCl. Pre-incubation with bafilomycin A1 greatly reduced the depolarization-evoked GABA release (Fig. 4B). As revealed by examination of GABA-labeled vesicle abundance using confocal microscopy, there was a substantial decrease in the number of labeled vesicles in Paramecium cells exposed to bafilomycin A1 (Fig. 5A).

We also tested the ability of BoNT/C1, which hampers neurotransmitter exocytosis by cleaving syntaxin-1 and SNAP-25 in mammals (Foran et al., 1996; Schiavo et al., 1995), to block GABA release. Incubation of cells for 24 h in the presence of the activated toxin (40 nmol l⁻¹) significantly inhibited GABA release (Fig. 4B). Activity of botulinum toxin was demonstrated in western blot experiments by directly measuring the amount of Paramecium syntaxin present after exposure of cell extracts to the botulinum toxin. Syntaxin, detected with the anti-rat syntaxin-1 monoclonal antibody used above, was totally cleaved by BoNT/C1 (Fig. 5B).

To identify putative cleavage sites for BoNT/C1 in the Paramecium syntaxin (PtSyx) sequences, a multiple sequence alignment was performed using the Rattus norvegicus syntaxin as consensus sequences (Fig. 6). Sequences were aligned using the ClustalW program in the Molecular Evolutionary Genetics Analysis (MEGA) software package (Kumar et al., 2004) (http://www.megasoftware.net/). Domain predictions (SNARE and transmembrane domains) were determined by PROSITE and TMHMM version 2.0 (http://www.cbs.dtu.dk/services/TMHMM/). Our analysis showed that some Paramecium syntaxin homologs (PtSyx 3-1/3-2), as well as some ciliate-specific syntaxins (PtSyx4-1 and PtSyx 11-1), possess a single Lys–Ala peptide bond in the same location as rat syntaxins. The classical functional domains described in mammal syntaxins, such as the single transmembrane domain at the C-terminus and the coiled-coil region before the transmembrane domain (SNARE domain), are both present in all Paramecium syntaxins (PtSyxs).

In this study, the presence of GABA synthesis and transport machinery in the ciliated protozoan P. primaurelia was demonstrated by western blotting and confocal microscopy. GAD is the main enzyme involved in the synthesis of GABA. Two GAD isoforms in the human brain, GAD₆₅ and GAD₆₇, have well-characterized molecular masses (Erlander and Tobin, 1991). GAD₆₅ and GAD₆₇ are present as homodimers or heterodimers in both soluble GAD (sGAD) and membrane-associated GAD (mGAD) pools (Sheikh and Martin, 1996). GABA synthesized by mGAD is preferentially transported into the vesicles by the vGAT, a 10-transmembrane protein (McIntire et al., 1997). The vGAT forms a protein complex with GAD on the vesicle and ensures an efficient transition between the synthesis of GABA and its packaging into the synaptic vesicle (Jin et al., 2003). In P. primaurelia cells in the late log phase of growth, both GAD-like and vGAT-like immunoreactivity was detected on the cell surface and inside the cytoplasm. Furthermore, immunoblot analysis showed that the molecular masses of GAD and vGAT proteins derived from Paramecium cells were consistent with corresponding proteins in mammalian cells. In a further indication of the above proteins in Paramecium, we not only detected GABA immunoreactivity but also found a co-localization of GABA with the vesicular transporter vGAT. This co-localization suggests that the amino acid is packaged into vesicles, an assumption supported by the blockade of GABA accumulation and release mediated by bafilomicyn A1 (see below).

GABA is the principal inhibitory neurotransmitter of the mammalian central nervous system, a function shared by a number of invertebrate systems. Immunocytochemical studies have localized GABA throughout the echinoderm nervous and muscular systems (Florey et al., 1975; Newman and Thorndyke, 1994), suggesting that GABA may play a role in modulating motor activities (Newman and Thorndyke, 1994). GABA has been detected in various digestive structures of Asterias and acts as a classic inhibitory neurotransmitter in both the holothurian cloaca (Hill, 1970) and body wall longitudinal muscles, where it causes muscle relaxation, suppresses spontaneous rhythmicity and inhibits cholinergic contractions (Devlin and Schlosser, 1999). The GABA system is involved in coordinating certain bilateral central pattern generator systems related to feeding and locomotion in the gastropod mollusk Aplysia californica (Diaz-Rios et al., 1999), and it controls foraging, regulates defecation and causes muscle relaxation during locomotion in the nematode Caenorhabditis elegans (White et al., 1986). Furthermore, GABA receptors modulate feeding response in the cnidarian Hydra vulgaris, one of the most primitive organisms bearing a neuronal system (Concas et al., 1998). These receptors are also found in sponges (Ramoino et al., 2007). In these organisms, which lack both nerves and muscles, GABA receptors are involved in the regulation of body contraction (Ellwanger and Nickel, 2006; Ellwanger et al., 2007).

In P. primaurelia, the unicellular organism chosen for this study, GABA_B receptors modulate swimming behavior by inhibiting dihydropyridine-sensitive Ca²⁺ channels via G-proteins (Ramoino et al., 2003). In addition, prolonged occupancy of GABA_B receptors by baclofen decreases GABA_B receptor function. This decrease is due not only to receptor internalization via a clathrin-dependent and -independent mechanism but also by the partial degradation via lysosomes (Ramoino et al., 2005; Ramoino et al., 2006). Conversely, GABA_A receptor activation induces membrane depolarization caused by a gradient-dependent efflux of chloride ions through GABA_A-associated chloride channels (Bucci et al., 2005). GABA_B receptors are also involved in Paramecium food uptake; administration of baclofen produces a dose-related increase in particle ingestion (P.R., A.D. and M.F., unpublished data). Furthermore, GABA_B receptors are involved in cell–cell interactions (Delmonte Corrado et al., 2002). In a related function, GABA_A receptors play a role in sexual processes involving mating-pair formation, and, in fact, the GABA_A receptor antagonists bicuculline and picrotoxin both prevent cell pairing. In Paramecium, GABA modulates a number of GABA receptor-mediated, physiologically relevant activities, creating the possibility of autocrine and/or paracrine control through functional GABA release.

In the present study, we demonstrated that P. primaurelia releases GABA into the medium and that acceleration of this release can be triggered by increasing the extracellular K⁺ concentration to 30 mmol l⁻¹. Physiologically, the resting membrane potential in Paramecium is about −25 mV and displays random fluctuations of small amplitude (Majima, 1980; Moolenaar et al., 1976). Larger fluctuations generate a more extensive, spike-like depolarization that is accompanied by both the opening of Ca²⁺ channels localized in the ciliary membrane (Dunlap, 1977; Ogura and Takahashi, 1976) and by a simultaneous influx of Ca²⁺. As a consequence of this depolarization and Ca²⁺ influx, a graded response is produced that reorients the cilia and reverses the swimming direction. In Paramecium, the intracellular K⁺ concentration ranges between 17 mmol l⁻¹ (Oertel et al., 1978) and 34 mmol l⁻¹ (Ogura and Machemer, 1980). It thus seems reasonable to assume that the membrane depolarizes in the presence of 30 mmol l⁻¹ of external K⁺ and that its potential can reach values moderately below 0 mV. Therefore, we propose that 30 mmol l⁻¹ KCl-induced depolarization provokes GABA release because it allows Ca²⁺ to enter the cell and trigger exocytosis of the neurotransmitter. Interestingly, exocytosis has already been reported to be triggered by Ca²⁺ influx in Paramecium (Erxleben et al., 1997; Klauke and Plattner, 1997). Further confirming the role of Ca²⁺ in exocytosis, the omission of Ca²⁺ from the medium abolished the stimulus-induced release of GABA under our experimental conditions. Moreover, depolarization-induced GABA release was inhibited by incubating Paramecium cells with bafilomicyn A1 or BoNT/C1 prior to the release experiments. Bafilomicyn A1 is a vesicle membrane V-type ATPase inhibitor (Bowman et al., 1988; Floor et al., 1990) that dissipates the proton gradient and prevents the vGAT-mediated accumulation of GABA into acidic vesicles (Moriyama and Futai, 1990; Roseth et al., 1995). BoNT/C1 is a Clostridium neurotoxin that cleaves specific aminoacid bonds of mammalian syntaxin-1 and SNAP-25, impeding SNARE complex formation and the consequent vesicle fusion to the cell membrane (Foran et al., 1996; Schiavo et al., 1995).

In Paramecium, V-ATPase is involved in osmoregulation, phagocytosis, endocytosis and dense-core secretory vesicle (tricocyst) biogenesis (Wassmer et al., 2005; Wassmer et al., 2006; Wassmer et al., 2009). It was found that the contractile vacuole complex V-ATPase in Paramecium is only sensitive to concanamycin and not to bafilomicyn (Fok et al., 1995). This lack of response to bafilomicyn even occurred when the c subunit of the enzyme included the Thr32, Phe136 and Tyr143 residues, which are reportedly involved in bafilomycin binding (Wassmer et al., 2005). The role of V-ATPase in tricocyst membrane fusion is not yet known (Wassmer et al., 2005; Wassmer et al., 2009). The V-ATPase was shown to be essential for the biogenesis of these secretory organelles (Wassmer et al., 2005), although they have been described as not acidic (Garreau de Loubresse et al., 1994; Lumpert et al., 1992). Our release and confocal microscopy experiments, performed with bafilomycin A1-pretreated Paramecium cells, not only showed a significant reduction of GABA release induced by depolarization but also a less efficient storage of GABA in the vesicular compartment. This finding indicates that bafilomycin A1 prevents GABA uptake into vesicles in Paramecium and that the Ca²⁺-dependent amino acid release is likely to be of vesicular origin.

In the release experiments, we used confocal microscopy to show that the release-machinery proteins syntaxin, SNAP and VAMP, which aggregate in mammals to form the SNARE complex, occur in this ciliated protozoon. Because GABA release is strongly reduced by BoNT/C1, these proteins probably also interact in Paramecium to form the SNARE complex. Co-localization of GABA and VAMP supports this view and suggests that GABA-containing vesicles can undergo docking and priming processes.

The presence of the SNARE-forming proteins syntaxin, VAMP and SNAP, involved in synaptic vesicle exocytosis, and of NSF, the SNARE-specific chaperone, has been recently demonstrated in Paramecium (Froissard et al., 2002; Kissmehl et al., 2002; Kissmehl et al., 2007; Schilde et al., 2006; Schilde et al., 2008). Seven subfamilies of synaptobrevins encoded by 12 genes, as well as at least 26 syntaxins grouped into 15 subfamilies, were identified (Kissmehl et al., 2007; Schilde et al., 2006). It is well known that, in mammals, syntaxins and SNAP25 are cleaved by BoNT/C1 at distinct sites close to the C-terminus. Specifically, the single Lys(K)–Ala(A) and Arg(R)–Ala(A) peptide bonds found in syntaxin1A/1B and SNAP25 are recognized by BoNT/C1 (Blasi et al., 1993; Schiavo et al., 1995). The sequence analysis that we performed suggests that this cleavage site for BoNT/C1 is conserved in P. tetraurelia syntaxin (Fig. 6). In addition, our experiments used crude cell lysates from P. primaurelia incubated with BoNT/C1 to show that the ciliated syntaxin is hydrolyzed by the botulinum toxin, explaining why the toxin also reduces GABA release.

To conclude, here we produced further evidence that a synaptic vesicle-like exocytotic pathway occurs in P. primaurelia. We also showed for the first time that the ciliated protozoon utilizes this process to release the messenger molecule GABA. In fact, our data suggest that (1) GABA newly synthesized by GAD is transported into acidic vesicles by vGAT, (2) that GABA-containing vesicles are competent for fusion to the plasma membrane upon depolarization, and (3) that GABA release involves a fusion-protein complex and occurs by means of a neuronal-like mechanism.

We thank Dr S. Maccione (DIP.TE.RIS., University of Genoa) for skilled technical aid and University of Genoa for financial support.

 
• ANOVA

  analysis of variance

 
• BoNT/C1

  botulinum toxin C1

 
• BSA

  bovine serum albumin

 
• GABA

  γ-amino butyric acid

 
• GAD

  glutamate decarboxylase

 
• HPLC

  high performance liquid chromatography

 
• mGAD

  membrane-associated GAD

 
• NC

  nitrocellulose

 
• PBS

  phosphate buffered saline

 
• sGAD

  soluble GAD

 
• VAMP

  vesicle-associated membrane proteins

 
• vGAT

  vesicular GABA transporter

 
• VSCC

  voltage-sensitive calcium channels