Synaptotagmin V is targeted to dense-core vesicles that undergo calcium-dependent exocytosis in PC12 cells.

Synaptotagmins (Syts) III, V, VI, and X are classified as a subclass of Syt, based on their sequence similarities and biochemical properties (Ibata, K., Fukuda, M., and Mikoshiba, K. (1998) J. Biol. Chem. 273, 12267-12273; Fukuda, M., Kanno, E., and Mikoshiba, K. (1999) J. Biol. Chem. 274, 31421-31427). Although they have been suggested to be involved in vesicular trafficking, as in the role of the Syt I isoform in synaptic vesicle exocytosis, their exact functions remain to be clarified, and even their precise subcellular localization is still a matter of controversy. In this study, we established rat pheochromocytoma (PC12) cell lines that stably express Syts III-, V-, VI-, and X-GFP (green fluorescence protein) fusion proteins, respectively, to determine their precise subcellular localizations. Surprisingly, Syts III-, V-, VI-, and X-GFP proteins were found to be targeted to specific organelles: Syt III-GFP to near the plasma membrane, Syt V-GFP to dense-core vesicles, Syt VI-GFP to endoplasmic reticulum-like structures, and Syt X-GFP to vesicles (other than dense-core vesicles) present in cytoplasm. We showed that Syt V-containing vesicles at the neurites of PC12 cells were processed to exocytosis in a Ca2+-dependent manner. Immunohistochemical analysis further showed that endogenous Syt V was also localized on dense-core vesicles in the mouse brain and specifically expressed in glucagon-positive alpha-cells in mouse pancreatic islets, but not in beta- or delta-cells. Based on these results, we propose that Syt V is a dense-core vesicle-specific Syt isoform that controls a specific type of Ca2+-regulated secretion.

Regulated vesicle trafficking is utilized for diverse cellular processes, including secretion of peptide hormones, neurotransmitter release, outgrowth of neurites in neurons, egg fertilization, and plasma membrane repair, and it is often triggered by Ca 2ϩ ions. Synaptic vesicle exocytosis, one of the well characterized forms of regulated vesicle trafficking, is strictly regulated by a rapid increase in Ca 2ϩ ions (10 -200 M) entering through voltage-gated Ca 2ϩ channels (1)(2)(3)(4). Although the precise mechanism by which synaptic vesicles sense such rapid increases in Ca 2ϩ ions remains unclear, synaptotagmin (Syt) 1 I, a Ca 2ϩ -and phospholipid-binding protein in synaptic vesicles, has been shown to regulate Ca 2ϩ -dependent neurotransmitter release (for review, see Refs. [5][6][7][8] and has therefore been proposed to be a major "Ca 2ϩ sensor" for neurotransmitter release (for review, see Refs. 9 and 10).
Syts III, V, VI, and X belong to the same branch in the phylogenetic tree and are classified as a subclass of Syt characterized by having almost no ability to bind inositol 1,3,4,5tetrakisphosphate (17) and an N-terminal Cys cluster responsible for dimer formation via disulfide bonding (12,19). However, the exact functions and subcellular localization of these Syts are still a matter of controversy. For instance, Sü dhof and co-workers (8) recently proposed that Syts III and VI function as plasma membrane Ca 2ϩ sensors, whereas others have shown that Syt III is localized on insulin-containing vesicles (34 -36) and that Syt VI is attached to various membrane fractions (23,24). Therefore, it is quite important to determine the exact localization of Syts to understand the functional diversity or functional redundancy of these Syt isoforms (III, V, VI, and X).
In this study, we investigated the subcellular distributions of the subclass of Syt (III, V, VI, and X), focusing especially on Syt V, which has never been characterized. Two different methods (stable expression of green fluorescence protein (GFP)-tagged Syt in PC12 cells and subcellular fractionation of brain) demonstrated that Syt V is localized predominantly on dense-core vesicles but that others are not enriched in these vesicles. In addition, Syt V was found to be specifically expressed in pancreatic ␣-cells. Based on our findings, we propose that Syt V is a cell type-specific regulator for Ca 2ϩ -dependent dense-core vesicle exocytosis.

Construction of the Expression Vectors-
The cDNA fragments encoding Syts III, V, VI, and X were prepared as described previously (12,14). Construction of the expression vector encoding FLAG-Syt-GFP (named pShooter-FLAG-Syt-GFP) (Invitrogen, Carlsbad, CA) was essentially carried out by PCR as described elsewhere (see Fig. 1A) (12,37). Briefly, the 3Ј-end of each Syt cDNA was ligated to the 5Ј-end of GFP cDNA to encode a fusion protein of Syt-GFP, and the sequences encoding the FLAG tag were inserted into the 5Ј-end of each Syt cDNA. A glycine linker was inserted between Syt and GFP to minimize interactions affecting the proper folding of the C2B domain (38). The cDNA fragments encoding these fusion proteins were subcloned into the NotI site of the modified pShooter vector (37) and verified by DNA sequencing as described previously (12).
Screening of Cell Lines Stably Expressing FLAG-Syt-GFP Protein-PC12 cells (2 ϫ 10 6 cells, the day before transfection) were cultured on 10-cm dishes in Dulbecco's modified Eagle's medium containing 10% horse serum and 10% fetal bovine serum at 37°C under 5% CO 2 . The expression vectors encoding the Syt-GFP fusion protein were transfected into PC12 cells with LipofectAMINE Plus reagent (Invitrogen) according to the manufacturer's instructions. The transfected cells expressing neomycin-resistant gene were selected using Geneticin (Invitrogen) at a concentration of 400 g/ml. Production of Syt-GFP fusion protein in each established cell line was verified by both immunoblot and immunocytochemical analyses.
Immunocytochemistry-The cloned PC12 cells (5 ϫ 10 4 cells/35-mm dish) stably expressing each Syt-GFP fusion protein were cultured with or without 100 ng/ml ␤-nerve growth factor (NGF) (Merck KGaA, Darmstadt, Germany) for 3 days. After washing twice with phosphatebuffered saline (PBS), the cells were fixed in 4% paraformaldehyde in 0.1 M sodium phosphate buffer for 20 min at room temperature, and the fixation was stopped with 0.1 M glycine. The cells were then permeabilized with 0.3% Triton X-100 in PBS for 2 min at room temperature. After blocking with 1% bovine serum albumin and 0.1% Triton X-100 in PBS for 1 h at room temperature, the cells were reacted with primary antibody (Ab), anti-Syt I (SYA-148, 1/250 dilution; StressGen Biotechnologies Corp., Victoria, BC, Canada), for 1 h at room temperature and then incubated with Alexa 568-conjugated anti-mouse IgG (1/5,000 dilution, Molecular Probes, Eugene, OR) for 1 h at room temperature. Immunoreactivity was analyzed with a confocal microscope (Fluoview, OLYMPUS, Tokyo, Japan) and Adobe Photoshop software (version 5.0).
Antibody Production-Anti-Syt III-C rabbit Ab was produced against the synthetic peptide corresponding to the C terminus of Syt III with an artificial Cys residue (CFTKGGKGLSEKENSE), as described previously (39). Anti-Syt V-C2A and anti-Syt V⌬C2AB Abs were raised against the C2A domain (amino acids 217-343) (17) and the cytoplasmic region between the transmembrane and the C2 domain (amino acids 71-216), respectively. Both antigens were produced as a fusion protein with glutathione S-transferase (GST) (40). Anti-Syt VI-C2A antibody was prepared as described previously (24). Anti-Syt X-C2A antibody was raised against the C2A domain of Syt X as the fusion protein with GST (17). The procedure for immunization and purification of IgG has been described previously (24). The specificity of the Abs was confirmed by immunoblot analysis with recombinant T7-tagged Syts I-XIII expressed in COS-7 cells as the loading samples (12,24,25,41). Except for anti-Syt V-C2A Ab, all of the Abs specifically recognized each antigen and did not cross-react with other Syt isoforms under our experimental conditions (data not shown).
Immunoblot Analysis-Three independent stable cell lines for each Syt-GFP construct were homogenized in 1% SDS with a 27-gauge syringe. After addition of SDS sample buffer, they were resolved on 10% polyacrylamide gel and transferred to polyvinylidene difluoride membranes. The blots were incubated with blocking buffer containing 1% skim milk and 0.1% Tween 20 in PBS for 1 h at room temperature and then with horseradish peroxidase-conjugated anti-FLAG M2 antibody (100 ng/ml, Sigma Chemical Co., St. Louis, MO) for 1 h at room temperature. The specific Abs described above (anti-Syt III-C rabbit antiserum (1/125 dilution), anti-Syt V⌬C2AB rabbit Ab, anti-Syt VI-C2A rabbit Ab, anti-Syt X-C2A rabbit Ab (2 g/ml)) were used to detect the endogenous expression of Syts III, V, VI, and X (24,41). Because the anti-Syt VI-C2A rabbit Ab weakly recognizes Syt I, the cross-reactive component was removed by preincubation with GST-Syt I-C2A recombinant proteins (24). The signals were detected by means of enhanced chemiluminescence systems (Amersham Biosciences). Subcellular Fractionation-The PC12 stable cell lines (5 ϫ 10 6 cells in a 10-cm dish) were cultured with NGF for 3 days. Cells from two confluent plates were harvested and homogenized in 1 ml of 0.3 M sucrose, 5 mM EDTA, and 5 mM HEPES-KOH, pH 7.2, in a glass-Teflon Potter homogenizer with 10 strokes at 1,000 rpm. The homogenate was centrifuged at 10,000 ϫ g for 10 min at 4°C to remove debris, and the supernatant was loaded on a linear sucrose gradient (0.6 -1.8 M) and spun at 100,000 ϫ g for 6 h in a SW41Ti rotor (Beckman Instruments, Fullerton, CA). 14 fractions of 0.75 ml each were collected from the top of the gradient and analyzed by immunoblotting. The following Abs were used as organelle markers: anti-secretogranin II (1 g/ml, Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and anti-synaptophysin (1 g/ml, Sigma). SYA-148 (shown above) or unpurified anti-Syt IX rabbit Ab, which can also react with Syt I (41), was used to detect Syt I. Subcellular fractionation of adult mouse whole brain was performed in a similar manner.
Uptake of Anti-FLAG Antibody into Stable PC12 Cell Lines-To investigate whether the vesicles containing Syt V-GFP fusion protein are involved in Ca 2ϩ -regulated exocytosis, anti-FLAG M2 antibody uptake experiments were performed as described previously (13,41). Briefly, PC12 stable cells cultured for 3 days in the presence of NGF were washed with low KCl buffer (5.6 mM KCl, 145 mM NaCl, 2.2 mM CaCl 2 , 5.6 mM glucose, 0.5 mM MgCl 2 and 15 mM HEPES-KOH, pH 7.4) twice and then incubated with anti-FLAG M2 antibody at a concentration of 1.1 g/ml in either low or high KCl buffer for 15 min at 37°C. High KCl buffer consists of 56 mM KCl, 95 mM NaCl, 2.2 mM CaCl 2 , 5.6 mM glucose, 0.5 mM MgCl 2 , and 15 mM HEPES-KOH, pH 7.4. They were then washed twice with PBS, fixed with 4% paraformaldehyde for 20 min, treated with 0.1 M glycine, and permeabilized with 0.3% Triton X-100 in PBS for 2 min at room temperature. After washing and incubating with blocking buffer containing 1% bovine serum albumin and 0.1% Triton X-100 in PBS, incorporated anti-FLAG M2 antibody was visualized by reacting with Alexa 594-conjugated anti-mouse IgG at a concentration of 1/5,000 dilution for 1 h at room temperature. Finally, they were washed with blocking buffer and processed for confocal microscopy as described above.
Immunoprecipitation-Whole brain or pancreas from an adult ICR female mouse was homogenized in 4 ml of 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% SDS, 1% Nonidet P-40, 0.5% deoxycholate, 10 M leupeptin, 10 M pepstatin A, and 0.1 mM phenylmethylsulfonyl fluoride, and the homogenate was centrifuged at 14,000 rpm for 10 min at 4°C. The supernatant was incubated with either 1 g/ml anti-Syt V-C2A Ab or normal rabbit IgG as a control for 1 h at 4°C and then with protein A-Sepharose (Amersham Biosciences) for 2 h at 4°C. After washing the beads three times with the homogenizing buffer, the immunoprecipitates were subjected to 7.5% SDS-PAGE followed by immunoblotting with anti-Syt V⌬C2AB Ab.
Immunohistochemistry-An ICR adult mouse was perfused with 4% paraformaldehyde, and the pancreas was excised and postfixed in the same fixative. It was cryosectioned into 10-m thick sections, dried, rehydrated in PBS, and fixed in 4% paraformaldehyde. After permeabilization in methanol for 30 min at Ϫ20°C, the sections were incubated in 2% skim milk, 0.1% Triton X-100, and 1% donkey serum in PBS for 1 h at room temperature and then with the following primary Abs: 5 g/ml anti-Syt V⌬C2AB Ab (described above), 10 g/ml anti-glucagon goat Ab (Santa Cruz Biotechnology, Inc.), anti-insulin guinea pig Ab (1/1 dilution; Zymed Laboratories, Inc.), or 2 g/ml anti-somatostatin goat Ab (Santa Cruz Biotechnology, Inc.). Positive immunoreactivity was visualized with the following secondary Abs: Alexa 488-conjugated anti-rabbit IgG, Alexa 594-conjugated anti-goat IgG, or Alexa 568conjugated anti-guinea pig IgG (1/1,000 dilution, Molecular Probes).

RESULTS
To investigate the subcellular localization of the subclass of Syts (Syts III, V, VI, and X) (12,17), we constructed expression vectors encoding each Syt protein with the C-terminal GFP tag (Fig. 1A) to detect its expression by fluorescence because Nterminal GFP tags have sometimes resulted in incorrect localization of certain isoforms. 2 Because PC12 cells are often used to study Ca 2ϩ -regulated exocytosis (41)(42)(43)(44)(45)(46) and are known to express Syts I, III, IV, and IX (25, 41-43, 47, 48), we selected PC12 cells for our initial study. Although the Syt family lacks a signal peptide sequence and should contain a specific mem-brane targeting signal itself (49), transient overexpression of some Syt-GFP proteins (especially Syt V) in PC12 cells was found to induce abnormal membranous structures (aggregation of GFP fluorescence) or abnormal cell shapes (data not shown). These effects are unlikely to be attributable to an artifact induced by fusion of GFP because transient overexpression of FLAG-Syt alone produced the same effects (data not shown). To eliminate the artifacts induced by forced overexpression, PC12 cell lines stably expressing each Syt-GFP or GFP alone, as a control, were established (Fig. 1, B and C, and Fig. 2). At least three independent cell lines for each construct were cloned, and the expression of FLAG-Syt-GFP fusion proteins was confirmed by immunoblotting with the anti-FLAG M2 Ab (Fig.  1B). The molecular weight of the fusion proteins was almost identical to their calculated molecular weight, and their amount was almost the same in all three lines but much lower than expression of GFP alone. Expression of FLAG-Syt-GFP in stable cells was further confirmed by specific Abs against each isoform (Fig. 1C, right lanes). Consistent with the previous report (41), however, we could not detect any endogenous Syt signals in control PC12 cells stably expressing FLAG-GFP (Fig.  1C, left lanes), indicating that the endogenous levels of expression of Syt III, V, VI, or X in normal PC12 cells are very low, if they are expressed at all.
We then attempted to determine the organelles to which the Syt-GFP fusion proteins are targeted by confocal microscopy (Fig. 2). The stable cell lines were also immunostained with anti-Syt I Ab, a marker for dense-core vesicles and synapticlike microvesicles in PC12 cells (41,42,50) to compare the distribution pattern of Syt-GFP with that of Syt I. To our surprise, each Syt-GFP had a distinct distribution, despite the well conserved amino acid sequences in the subclass (12). The Syt III-GFP protein was localized preferentially near the plasma membrane ( Fig. 2A) and barely colocalized with Syt I (Fig. 2, B and C), consistent with the previous reports that Syt III is localized at the plasma membrane, not at secretory granules (23,51). Interestingly, the cell bodies of all the Syt III-GFP stable cell lines were much smaller than those of other Syt-GFP stable cell lines (round shape), and they hardly extended their neurites in response to NGF (Fig. 2, A-C). At this stage, we could not determine whether these phenotypes were involved in certain biological events or were just expression artifacts. The Syt V-GFP protein had accumulated at the neurites (Fig.  2D, arrows) and was closely colocalized with Syt I (Fig. 2, E and The blots were then incubated with anti-actin Ab as an internal control to confirm that equivalent amounts of total proteins were loaded on each lane (bottom panels). The signals for each FLAG-Syt-GFP fusion protein were detected at the positions expected from their calculated molecular weight, and the amounts of each fusion protein did not vary much among three independent cell lines. Stable cell lines expressing FLAG-GFP alone were also established as a negative control and examined in the same way as the others. C, endogenous expression levels of Syts III, V, VI, and X in PC12 cells. Expression of endogenous Syts was examined by immunoblotting with the specific Abs for each isoform. The signal for each Syt was detected only in specific stable cell lines and not in control cell lines stably expressing GFP. The positions of the molecular weight markers (ϫ10 Ϫ3 ) are shown on the right. F), indicating that Syt V-GFP protein is localized on dense-core vesicles. The Syt VI-GFP fusion protein was observed mainly at the endoplasmic reticulum-like structures (Fig. 2G) as described previously (24) and not colocalized with Syt I (Fig. 2, H  and I). The Syt X-GFP protein was detected diffusely throughout the intracellular region of the cell body (Fig. 2J), presumably in the small vesicles. The Syt X-GFP fusion protein was not colocalized well with Syt I (Fig. 2, K and L). These patterns of different distributions of Syt-GFP shown here were identical in all three independent stable cell lines (data not shown). The results strongly suggested that only Syt V, not Syt III, VI, or X, is targeted to dense-core vesicles and that the four isoforms may have different roles in vesicular trafficking.
Localization of Syt V-GFP on Dense-core Vesicles by Sucrose Gradient Fractionation-Sucrose gradient fractionation was performed to confirm further that Syt V-GFP is present on dense-core vesicles in PC12 cells. Stable Syt-GFP cell lines were cultured for 3 days in the presence of NGF, and the homogenates of each cell line were separated in a 0.6 -1.8 M linear sucrose gradient and analyzed by immunoblotting with horseradish peroxidase-conjugated anti-FLAG M2 antibody (Fig. 3) (for details, see "Experimental Procedures"). The Syt V-GFP protein was enriched predominantly in fractions 8 -11 (Fig. 3, top panel), where secretogranin II, a marker for densecore vesicles, was also enriched (Fig. 3, third panel). It should be noted that only a small amount of Syt V-GFP protein was found in fractions 3-6, where synaptophysin, a marker for synaptic-like microvesicles, was enriched (Fig. 3, bottom panel), whereas Syt I was enriched in both dense-core vesicles and synaptic-like microvesicles (Fig. 3, second panel). By contrast, Syts III-, VI-and X-GFP proteins were distributed in much lighter fractions (data not shown), making them unlikely to be present at dense-core vesicles, consistent with our immunocytochemical analysis. These results strongly suggested that Syt V-GFP is a dense-core vesicle-specific isoform in the subclass of Syts III, V, VI, and X.

Syt V-GFP Is Localized on Secretory Vesicles That Undergo
Ca 2ϩ -dependent Exocytosis-Because PC12 cells contain various types of secretory vesicle (52), we next investigated whether Syt V-GFP is localized on Ca 2ϩ -regulated secretory vesicles. A positive result would be important evidence in support of Syt V involvement in Ca 2ϩ -regulated exocytosis in which Syts I and IX play a role (41)(42)(43)(44)(45). To do so, Ca 2ϩ -dependent uptake of Ab against the Syt N terminus (anti-FLAG Ab) was performed as described previously (41,(53)(54)(55), and we investigated whether Syt V-GFP-containing vesicles undergo exocytosis in response to Ca 2ϩ stimulation. If FLAG-Syt V-GFP-containing vesicles undergo exocytosis in response to Ca 2ϩ stimulation, the N terminus of FLAG-Syt V-GFP would be accessible on the outside surface of the cell membrane and should be recognized by anti-FLAG M2 Ab in the culture medium. The complex of FLAG-Syt V-GFP and anti-FLAG M2 Ab would then be incorporated into the cell by endocytosis. As expected, anti-FLAG M2 Ab was incorporated efficiently into the neurites after high KCl stimulation, and its signals were almost completely colocalized with GFP fluorescence (FLAG-Syt V-GFP) (Fig. 4, A-C). The uptake of anti-FLAG M2 Ab was Ca 2ϩ -dependent because neither in high KCl buffer containing 5 mM EGTA (Fig. 4, D-F) nor in low KCl buffer (Fig. 4, G-I), could we observe any anti-FLAG M2 signals in the neurites, and the weak dotted signals were only present at the edge of the neurites (Fig. 4, E and H). Although this might have resulted from nonspecific binding of Ab to the plasma membrane or the presence of a small amount of FLAG-Syt V-GFP at the plasma membrane, the former possibility is unlikely because no uptake of anti-FLAG M2 Ab was observed in control GFP stable cells even in the presence of high KCl buffer (data not shown). We therefore concluded that FLAG-Syt V-GFP is localized on dense-core vesicles that undergo Ca 2ϩ -dependent exocytosis.
Expression Patterns of Syt V in Vivo-Finally, we performed immunoprecipitation and subcellular fractionation studies to investigate whether endogenous Syt V is indeed targeted to dense-core vesicles in brain or endocrine tissues (e.g. pancreas). A single immunoreactive band with an apparent molecular weight of 65,000 was detected by anti-Syt V⌬C2AB Ab in total homogenates of both pancreas and brain (lanes 1 and 2 in  5A). The immunoreactive bands observed in brain and pancreas were specifically concentrated by immunoprecipitation with anti-Syt V-C2A Ab, but not with control rabbit IgG (lanes 3 and 4, and data not shown). In addition, the band was negative with anti-Syt I, III, VI, or X Ab, indicating that the signals observed in our experiments represented only Syt V (data not shown).
This specific antibody was used to investigate the localiza-tion of Syt V in dense-core vesicles by subcellular fractionation of the homogenate from whole brain. As shown in Fig. 5, B and C, Syt I peaked at around fractions 2 and 7. The lighter peak of Syt I (fraction 2) corresponds to synaptic vesicle-enriched fractions because this peak overlapped that of synaptophysin (data not shown). The denser peak of Syt I (fraction 7) corresponds to dense-core vesicles. Consistent with the results of recombinant Syt V-GFP in PC12 cells, brain Syt V had accumulated in fraction 7, whereas there was only a small amount of Syt V in fraction 2. These results suggested that Syt V is expressed in endocrine and neuroendocrine tissues and is targeted to dense-core vesicles.
To determine further whether Syt V is expressed uniformly (i.e. all dense-core vesicles) or in a cell type-specific manner (i.e. specific population of dense-core vesicles), immunohistochemical analysis was performed with specific anti-Syt V⌬C2AB Ab. In mouse brain, positive Syt V immunoreactivity was detected mainly in cells around the interpeduncular nucleus and essentially not at all in other brain regions, suggesting cell typespecific expression of Syt V (data not shown). Although this nucleus has been reported to be immunoreactive to Abs against opioid receptors and suggested to be involved in enkephalin metabolism (56,57), it is difficult to assign a function of Syt V in the interpeduncular nucleus at present because its neuroendocrine and peptide secretion properties have not been well characterized. We then focused on the expression patterns of Syt V in pancreatic islet cells because islet cells contain three functionally distinct cells (␣, ␤, and ␦) that store different peptide hormones in dense-core vesicles (i.e. glucagon, insulin, and somatostatin are specifically produced in pancreatic ␣-, ␤-, and ␦-cells, respectively). Interestingly, Syt V was expressed specifically in glucagon-containing ␣-cells in mouse pancreas (Fig. 6, A-C), but not in either insulin- (Fig. 6D) or somatostatin-containing cells (Fig. 6E). We therefore concluded that Syt V is expressed specifically in pancreatic ␣-cells and may be involved in glucagon secretion. DISCUSSION We previously identified a subclass of Syt (Syts III, V, VI, and X) which shows relatively high sequence similarity (12) and similar biochemical properties (12,17,19). One of the characteristic features of this class is the N-terminal cysteine residues that are responsible for dimer formation via disulfide bonding (12), suggesting a similar or redundant function of this class of Syt in vesicle trafficking. Surprisingly, however, the results of the present study showed that Syts III, V, VI, and X exhibit a distinct subcellular localization when stably expressed in PC12 cells (Figs. 2 and 3).
Syt III was first proposed as a candidate for the Ca 2ϩ sensor for dense-core vesicle exocytosis in insulin-secreting cells (34 -36) because Ab against the recombinant GST-Syt III (34,36) or C-terminal peptide of Syt III (35) recognized Syt molecules on insulin-containing vesicles. However, these studies did not provide sufficient evidence of the specificity of their antibody, and it is possible that their antibody cross-reacts with other Syt isoforms. By contrast, Gut and co-workers (51) recently showed that Syt III is most unlikely to be expressed on insulin-containing vesicles and that Syt III is targeted to the plasma membrane of primary islet cells when exogenous Syt III is transiently expressed. Our present study suggests that Syt III-GFP is localized near the plasma membrane of PC12 cells, consistent with the latter finding. In addition, Butz et al. (23) reported that Syt III is not concentrated in synaptic vesicles in the nervous system but enriched in synaptic plasma membranes. Therefore, Syt III is most likely present at the plasma membrane and may function as a plasma membrane Ca 2ϩ sensor (58) or be involved in neurite formation because PC12 cells stably or transiently expressing Syt III-GFP showed abnormal neurite formation ( Fig. 2A). 2 Almost all of the Syt VI proteins in brain are expressed as a transmembrane-deficient type (Syt VI⌬TM) (24,59), and they are attached to various membrane fractions as well as soluble fractions (23,24). Recently, full-length Syt VI having a transmembrane domain was demonstrated to be localized on the outer acrosomal membrane and to be involved in the Ca 2ϩtriggered acrosomal exocytosis of human spermatozoa (29). By contrast, our present study showed that Syt VI-GFP is distributed mainly in endoplasmic reticulum-like structures in PC12 cells (24). These discrepancies may be the result of the difference in cell types; the subcellular localization of Syt VI may be cell type-specific.
The syt X gene was originally isolated from the brain as a seizure-induced gene, suggesting that Syt X protein is involved in neuronal plasticity (60). However, the precise subcellular localization of Syt X has never been elucidated. Syt X-GFP showed a vesicular-like distribution (other than dense-core vesicles) in this study (Fig. 2J), but we could not determine the types of vesicle that contained the Syt X-GFP. Further analysis is required for a full understanding of the subcellular localization and functions of Syt X.
The most important and surprising finding was that Syt V is targeted to dense-core vesicles in both brain and neuroendocrine cell and essentially not to synaptic-like microvesicles (Figs. [2][3][4][5] because other members of this subclass (Syts III-, VI-, and X-GFP in PC12 cells) did not accumulate in dense-core vesicles. Based on this finding, together with the fact that the C2 domains of Syt V bind phospholipids or syntaxin in a Ca 2ϩdependent manner (58,61), we propose that Syt V is a densecore vesicle-specific isoform that regulates Ca 2ϩ -dependent exocytosis in brain and endocrine tissues (e.g. glucagon release in pancreatic ␣-cells). The possible involvement of Syt V in dense-core vesicle exocytosis in PC12 cells is being investigated in our laboratory by using recombinant fragments of C2 domains (so-called dominant negative approach) (42,43,45,46,58) or a functional block antibody specific to Syt V (41). In our preliminary experiments, however, the former strategy seemed unreliable for the following two reasons. First, the effect of the Syt V fragments on Ca 2ϩ -dependent neuropeptide Y release by Syt V-GFP-stable cells and control cells in which Syt V is not endogenously expressed was indistinguishable. Second, the inhibitory potential of the fragments of the Syt family did not reflect the endogenous expression levels of each Syt isoform (41,46,58). For instance, although Syts III and VII are not expressed well in PC12 cells (37,41), the effect of recombinant Syts III and VII was stronger than that of Syt I, the most abundant Syt isoform in normal PC12 cells (41), suggesting that the blocking site is non-Syt proteins (e.g. soluble NSF attachment protein receptor (SNARE) proteins) rather than endogenous Syt proteins (62). The second approach was also very difficult because we could not obtain highly specific Ab to the C2A domain of Syt V. 2 We generated four different anti-Syt V-C2A Abs, but all of them cross-reacted with Syts I and IX, both of which regulate dense-core vesicle exocytosis in PC12 cells (41), even after preincubation with an excess amount of GST-Syt I-C2A or -Syt IX-C2A. Thus, additional specific tool(s) are required to elucidate whether Syt V is involved in densecore vesicle exocytosis.
Another important finding was that the distribution pattern of Syt-GFP fusion proteins (especially Syts V and VII) expressed transiently was absolutely different from the pattern observed in the stable cell lines. For instance, transiently expressed Syt V-GFP fusion protein was often aggregated in cell bodies, whereas the stably expressed Syt V-GFP was localized predominantly on dense-core vesicles in PC12 cells. Similarly, transiently expressed Syt VII-GFP was localized on the plasma membrane (46), whereas the stably expressed Syt VII-GFP was enriched at the tips of the neurites as well as trans-Golgi network-like structures (37). These discrepancies may be explained as follows. The amount of the exogenous Syt-GFP fusion proteins expressed transiently is often too much, and such an excess protein overflows the normal sorting machinery. We therefore emphasize the advantage of using stable cell lines to evaluate the involvement of Syt in physiological events in cells. Of course, we cannot exclude the possibility that the sorting pathway of Syt-GFP fusion protein is different from that of endogenous Syt protein.
In conclusion, we have demonstrated distinct subcellular localizations of Syts III-, V-, VI-, and X-GFP in stable PC12 cells even though they belong to the same subclass of Syt. The results also suggest that Syt V, which is distributed predominantly in dense-core vesicles, is involved in Ca 2ϩ -mediated vesicle trafficking.