Synaptotagmin VII is targeted to dense-core vesicles and regulates their Ca2+ -dependent exocytosis in PC12 cells.

It has recently been proposed that synaptotagmin (Syt) VII functions as a plasma membrane Ca2+ sensor for dense-core vesicle exocytosis in PC12 cells based on the results of transient overexpression studies using green fluorescent protein (GFP)-tagged Syt VII; however, the precise subcellular localization of Syt VII is still a matter of controversy (plasma membrane versus secretory granules). In this study we established a PC12 cell line "stably expressing" the Syt VII-GFP molecule and demonstrated by immunocytochemical and immunoelectron microscopic analyses that the Syt VII-GFP protein is localized on dense-core vesicles as well as in other intracellular membranous structures, such as the trans-Golgi network and lysosomes. Syt VII-GFP forms a complex with endogenous Syts I and IX, but not with Syt IV, and it colocalize well with Syts I and IX in the cellular processes (where dense-core vesicles are accumulated) in the PC12 cell line. We further demonstrated by an N-terminal antibody-uptake experiment that Syt VII-GFP-containing dense-core vesicles undergo Ca2+ -dependent exocytosis, the same as endogenous Syt IX-containing vesicles. Moreover, silencing of Syt VII-GFP with specific small interfering RNA dramatically reduced high KCl-dependent neuropeptide Y secretion from the stable PC12 cell line (approximately 60% of the control cells), whereas the same small interfering RNA had little effect on neuropeptide Y secretion from the wild-type PC12 cells (approximately 85-90% of the control cells), indicating that the level of endogenous expression of Syt VII molecules must be low. Our results indicate that the targeting of Syt VII-GFP molecules to specific membrane compartment(s) is affected by the transfection method (transient expression versus stable expression) and suggested that Syt VII molecule on dense-core vesicles functions as a vesicular Ca2+ sensor for exocytosis in endocrine cells.

Although Syt VII has also been suggested as a regulator of secretory vesicle exocytosis in some endocrine cells and in neurons, the precise subcellular localization, expression level, and function of Syt VII are still matters of controversy (36 -43). Sü dhof and colleagues (38,43) have reported the finding that transient overexpressed Syt VII molecules with a C-terminal green fluorescence protein (GFP) tag were targeted to the plasma membrane in PC12 cells (or non-neuronal cells) and not endocytosed from the membrane, unlike endogenous Syt I. They also showed that Syt VII is abundantly and endogenously expressed in PC12 cells (43). Contradictory results, however, have been reported by other groups. First, there is a report that the level of endogenous expression of Syt VII in PC12 cells must be very low (less than 3% of Syt I) based on a quantitative immunoblot analysis (13,26), and, second, stably expressed Syt VII-GFP molecules have been found to be mainly targeted to the intracellular compartments of PC12 cells rather than to their plasma membrane (40). Such discrepancies are most likely attributable to the transfection methods used (transient expression versus stable expression), and transient overexpression of recombinant Syt proteins actually often causes mislocalization or aggregation of Syt proteins in living cells (40,44,45).
In this study we established a PC12 cell line that stably expresses Syt VII-GFP molecules and analyzed its precise subcellular localization by immunocytochemistry and immunoelectron microscopy. Unlike the previous transient overexpression studies, the Syt VII-GFP molecules were mainly targeted to dense-core vesicles that undergo Ca 2ϩ -dependent exocytosis and to other intracellular compartments such as the trans-Golgi network (TGN) and lysosomes. We also found that the selective knockdown of Syt VII-GFP by means of recently developed RNAi (double-stranded RNA-mediated interference) technology (reviewed in Refs. 46 and 47) causes inhibition of dense-core vesicle exocytosis in the stable PC12 cell line. Based on our findings, we propose that the Syt VII molecule is mainly targeted to dense-core vesicles and functions as a "vesicular Ca 2ϩ sensor" in endocrine cells.
Immunoprecipitation and Immunoblotting-Wild-type or Syt VII-GFP-expressing PC12 cells (two confluent 10-cm dishes) were homogenized in a buffer containing 1 l of 50 mM HEPES-KOH, pH 7.2, 150 mM NaCl, and protease inhibitors (0.1 mM phenylmethylsulfonyl fluoride, 10 M leupeptin, and 10 M pepstatin A) in a glass-Teflon Potter homogenizer with 10 strokes at 900 -1000 rpm, and the proteins were solubilized with 1% Triton X-100 at 4°C for 1 h. After removing the insoluble material by centrifugation at 15,000 rpm for 10 min, the supernatant (400 l) was incubated with anti-FLAG M2 affinity gel in the presence of 1 mM CaCl 2 (or 2 mM EGTA) for 1 h at 4°C. The beads were then washed five times with 10 mM HEPES-KOH, pH 7.2, 150 mM NaCl, 0.2% Triton X-100, 1 mM CaCl 2 (or 2 mM EGTA), and protease inhibitors, and the immunoprecipitates were subjected to a 10% SDS-PAGE followed by immunoblotting with anti-Syt I (1/250 dilution), anti-Syt IV (0.9 g/ml), anti-Syt IX (1/250 dilution), or horseradish peroxidase-conjugated anti-FLAG tag mouse monoclonal antibody (1/ 10,000 dilution) as described previously (53). Immunoreactive bands were visualized with enhanced chemiluminescence (Amersham Biosciences). The blots shown in this paper are representative of three independent experiments.
Construction of pSilencer-Syt VII Vectors for Syt VII Silencing and Neuropeptide Y (NPY) Release Assay-To silence rat and mouse Syt VII, a 19-nucleotide sequence (TCATCACCGTCAGCCTTAG) was selected according to the manufacturer's notes (Ambion, Inc., Austin, TX). The specificity of the sequence was verified by a BLAST search of the public data bases. pSilencer TM 1.0-U6 expression vectors (Ambion) that produce small interfering RNAs (siRNAs) targeted against Syt VII (named pSilencer-Syt VII) were also prepared according to the manufacturer's notes (29). In brief, two oligonucleotides (sense, 5Ј-TCATCACCGTCA-GCCTTAGTTCAAGAGACTAAGGCTGACGGTGATGATTTTTT-3Ј and Note that the Syt VII-GFP signals at the cell periphery colocalized well with the dense-core vesicle markers (Syt I, Syt IX, and Rab3A; yellow in C, F, and L, respectively), especially at the tips of the cellular processes, consistent with the results of our previous subcellular fractionation study (40). Because the Syt IV protein in undifferentiated PC12 cells is mainly localized in the Golgi and immature secretory vesicles (49,50,67), no colocalization between Syt VII-GFP and Syt IV was observed even in the tips of the cellular processes (I, inset, arrow) or in the perinuclear region. Confocal images were taken to highlight the localization of Syt VII-GFP in the tips of the cellular processes (i.e. densecore vesicle markers are accumulated), and as a result perinuclear localization of Syt VII-GFP is not so evident in Fig. 1 (see Fig. 4). Scale bar ϭ 20 m.
antisense, 5Ј-AATTAAAAAATCATCACCGTCAGCCTTAGTCTCTTGA-ACTAAGGCTGACGGTGATGAGGCC-3Ј; the underlined sequences contribute to forming small hairpin RNAs) were synthesized chemically, and the annealed oligonucleotides encoding short hairpin RNA were then subcloned into the ApaI/EcoRI site of the pSilencer TM 1.0-U6 vector. The efficiency and specificity of the siRNA targeted against Syt VII were assessed by transfection of pSilencer vector into the PC12 cell line stably expressing Syt VII-GFP molecule (29, 54) (see Fig. 6A).

Synaptotagmin VII-GFP in PC12 cells Is Predominantly
Localized on the Dense-core Vesicles-Although several independent groups have recently proposed that Syt VII regulates Ca 2ϩdependent vesicle exocytosis, its precise subcellular localization (plasma membrane, secretory granules, or lysosomes) is a matter of controversy (32,36,38,40,41,43). Because most previous studies on the function of Syt VII used transiently overexpressed Syt VII molecules, it is highly possible that excess recombinant Syt VII protein is mistargeted to other intracellular compartments by overflowing the normal transport pathway. To investigate this possibility, we established a PC12 cell line that stably expresses FLAG-Syt VII-GFP molecules (40). We then selected three independent cell lines and found that they exhibited the same intracellular distribution of Syt VII-GFP ( Figs. 1 and 4, A). The GFP fluo-rescence was found to be located in at least three distinct intracellular compartments of the PC12 cell line, the tips of the cellular processes (where dense-core vesicles are accumulated), the perinuclear regions, and adjacent to the plasma membrane. No GFP fluorescence was found around the plasma membrane when the stable cells were first established (40), but it became evident after several passages of the cells. In the first set of experiments, we investigated whether Syt VII-GFP was localized on the dense-core vesicles in the PC12 cell line by making comparisons with several dense-core vesicle markers at the immunocytochemical level. As shown in Fig. 1, Syt VII-GFP overlapped well with three dense-core vesicle markers (Fig. 1, C, F, and L, yellow, Syt I, Syt IX, and Rab3A, respectively) in the cellular processes, but not with Syt IV (Fig. 1I), which is not present on dense-core vesicles in undifferentiated PC12 cells (50,57). Interestingly, Syt VII-GFP was clearly associated with both Syts I and IX, but not with Syt IV, and the interaction between Syt VII-GFP and Syt IX was enhanced by Ca 2ϩ (Fig. 2,  compare lanes 3 and 4, third panel), consistent with the results of previous in vitro binding experiments (55, 58 -62). To confirm the dense-core vesicle localization of Syt VII-GFP molecules, we performed an immunoelectron microscopic analysis with anti-GFP antibody. As anticipated, immunogold signals were clearly associated with almost all of the dense-core vesicles in the neurites of the PC12 cell line (Fig. 3A), although some Syt VII-GFP signals were also found in other intracellular membrane structures. We also discovered that few Syt VII-GFP signals were associated with the plasma membrane  1 and 2). The asterisks indicate the heavy chain of IgG used for immunoprecipitation. Note that Syts I and IX, but not Syt IV, were coimmunoprecipitated with Syt VII-GFP irrespective of the presence of Ca 2ϩ (lanes 3 and 4), although the interaction between Syt VII-GFP and Syt IX was promoted by Ca 2ϩ (lane 4 in the third panel) (58,59). By contrast, neither the Syt I nor the Syt IX in the wild-type PC12 cells was trapped with the anti-FLAG M2 affinity gel alone (lanes 5 and 6). The positions of the molecular mass markers (ϫ10 Ϫ3 ) are shown on the left. (Fig. 3A, arrowheads). An N-terminal antibody-uptake experiment in the absence of high KCl stimulation (or low KCl stimulation) also supported this observation (see also Fig. 5E below). We therefore concluded that stably expressed Syt VII-GFP molecules are mainly targeted to the dense-core vesicles of PC12 cells rather than to their plasma membrane. Although the Syt VII-GFP fluorescence was partially overlapped with dextran (a marker for lysosomes; Fig. 4, P-R), consistent with a previous report (32), we found that Syt VII-GFP protein also partially overlapped with TGN38, a marker for the TGN (Fig.  4, D-F) (40), but not with GM130 (a marker for the cis-Golgi, Fig. 4, A-C), transferrin receptor (a marker for endosomes, Fig.  4, G-I), EEA1 (a marker for early endosomes, Fig. 4, J-L), or BiP (a marker for the endoplasmic reticulum, Fig. 4, M-O). The TGN localization of Syt VII-GFP molecules was also confirmed by the immunoelectron microscopic analysis. The immunogold signals for Syt VII-GFP were associated with the membranes of the TGN (Fig. 3B, shaded arrowheads), rather than those with the cis-Golgi (black arrowheads) or the trans-Golgi (black arrows). The localization of Syt VII-GFP in the TGN is not surprising because dense-core vesicles are thought to be formed from the TGN (i.e. budding from the TGN).
Synaptotagmin VII-GFP Is Localized on Mature Dense-core Vesicles That Undergo Ca 2ϩ -dependent Exocytosis in PC12 Cells-Because transiently overexpressed recombinant Syt VII molecules have been shown to remain in the plasma membrane FIG. 3. Localization of Syt VII-GFP protein in the PC12 cell line by immunoelectron microscopy. The panels show representative views of the dense-core vesicle localization (A) and TGN localization (B) of Syt VII-GFP stained with anti-GFP antibody followed by silver enhancement. Note that the Syt VII-GFP signals (i.e. gold particles) were clearly associated with dense-core vesicles as well as the plasma membrane in the neurites (A, arrowheads). Syt VII-GFP signals were also found at the membranes of the TGN (gray arrows) rather than the Golgi. The black arrowheads, black arrows, and gray arrows indicate the positions of cis-and trans-Golgi and TGN, respectively. Scale bar ϭ 500 nm. N, nucleus.  A and  B, D and E, G and H, J and K, M and N, and P and Q, respectively. Note that the majority of Syt VII-GFP signals in the perinuclear region colocalized well with TGN38 (inset in F), but not with GM130 (inset in C), and that some signals colocalized with dextran (inset in R), indicating that Syt VII-GFP is present in the TGN and lysosomes (32,40). Scale bar ϭ 20 m.
during dense-core vesicle exocytosis and endocytosis (42,43), we next sought to determine whether Syt VII-GFP-containing dense-core vesicles are fully mature enough to undergo regulated exocytosis in response to Ca 2ϩ -stimulation. To do so, we performed an N-terminal antibody-uptake experiment as described previously (25,45,50). As anticipated, both anti-FLAG tag antibody (Fig. 5, red signals) and anti-Syt IX-N antibody (blue signals) were incorporated into the cell bodies and neurites of the stable PC12 cell line in a high KCl-dependent manner (Fig. 5, A, B, I and J), although faint FLAG signals (i.e. Syt VII-GFP) were detected around the edge of the cells even under low KCl conditions (Fig. 5, E and M). The uptake of the antibodies against the extracellular domain of Syt was Ca 2ϩdependent, but not membrane depolarization-dependent, because antibody uptake did not occur in response to high KCl stimulation in the extracellular presence of 5 mM EGTA (45). Because the anti-FLAG tag antibody was unable to attach to the plasma membrane of the wild-type PC12 cells (45,63), a small population of Syt VII-GFP molecules is present at the plasma membrane (i.e. the extracellular domain of the Syt VII-GFP molecule is exposed to the extracellular space), consistent with the results of the immunoelectron microscopic analysis described above (Fig. 3A, arrowheads). Interestingly, anti-FLAG tag antibody uptake sometimes occurred independently of anti-Syt IX-N antibody uptake (Fig. 5, I-K, arrowheads), although anti-FLAG tag and anti-Syt IX-N antibodies were almost always incorporated into the same sites.
Synaptotagmin VII-GFP Is Involved in Ca 2ϩ -regulted Exocytosis of Dense-core Vesicles in PC12 Cells-Finally, we investigated whether exogenously expressed Syt VII-GFP molecules are involved in dense-core vesicle exocytosis by means of the NPY cotransfection assay (52, 55). As shown in Fig. 6B, three independent Syt VII-GFP-expressing cell lines exhibited increased high KCl-dependent NPY secretion activity compared with the wild-type PC12 cells (an ϳ4-fold greater release of NPY), although the level of expression of NPY did not differ between the wild-type and stable PC12 cells (data not shown).
Low KCl-dependent (or basal) NPY secretion from the stable cell lines was also increased compared with the wild-type PC12 cells, but the amount of basal NPY secretion from Syt VII-GFP-FIG. 6. Syt VII silencing caused a dramatic reduction of dense-core vesicle exocytosis in PC12 cells. A, specific reduction of Syt VII-GFP protein by the siRNA targeted against Syt VII. Three days after cotransfection of pSilencer-Syt VII with pShooter-NPY-T7-GST into the stable cell line, total cell lysates were subjected to 10% SDS-PAGE and immunoblotted with anti-Syt I, anti-Syt IV, anti-Syt IX, anti-T7 tag (left four panels), and anti-FLAG tag antibodies (right three panels). Note that the Syt VII-GFP expression level was specifically reduced by the siRNA without affecting the level of expression of other Syt isoforms or NPY. The positions of the molecular mass markers (ϫ10 Ϫ3 ) are shown on the left. B, Syt VII-GFP-expressing cell lines secrete NPY more efficiently than the wild-type PC12 cells. The NPY-T7-GST secretion assay was performed as described previously (52,55). The results are expressed as percentages of high KCl-dependent NPY secretion from the wild-type PC12 cells. L, low KCl-dependent NPY secretion (open bars); H, high KCl-dependent secretion (closed bars). C, effect of pSilencer-Syt VII on high KCl-dependent NPY secretion from the wild-type PC12 cells and Syt VII-GFP-expressing PC12 cell lines. The results are normalized by high KCl-dependent NPY secretion in control samples (closed bars). Bars indicate the means Ϯ S.E. of three determinations. The results shown are representative of at least three independent experiments. Note that Syt VII-GFP silencing resulted in a dramatic reduction in high KCl-dependent NPY secretion from the stable cell line (shaded bars), whereas the same siRNA slightly reduced high KCl-dependent NPY secretion from the wild-type PC12 cells (hatched bars). **, p Ͻ 0.001; *, p Ͻ 0.01, Student's unpaired t test. 1/2/3 cell lines was ϳ10 -15% of the high KCl-dependent NPY secretion, which is almost the same as the basal level secretion from the wild-type PC12 cells (52) (Fig. 6C).
To further determine whether the enhanced secretion ability of the Syt VII-expressing PC12 cell line is directly related to the Syt VII molecule and not an artifact caused during the screening of stable cell lines, we selectively knocked down Syt VII-GFP molecules by means of a recently developed RNAi technology. To knock down all of the Syt VII splicing variants (␣, ␤, and ␥ isoforms) (38,40), a 19-nucleotide sequence in the regions common to all Syt VII splicing isoforms, i.e. the transmembrane region, was selected. When the specific siRNA targeted against Syt VII was expressed in the stable PC12 cell lines, the level of expression of Syt VII-GFP protein was dramatically reduced without affecting the levels of expression of endogenous Syts I, IV, and IX proteins (Fig. 6A) or other proteins (e.g. SNAREs, Rabs, and Sec1/Munc18 proteins) that are involved in dense-core vesicle exocytosis (29) (data not shown). When NPY was coexpressed with Syt VII siRNA, high KCl-dependent NPY secretion from the stable PC12 cell lines was dramatically reduced (ϳ60% of the siRNA-untreated control cells, which is approximately twice as high as the high KCl-dependent NPY secretion from the wild-type PC12 cells; Fig. 6C, shaded bars). Residual NPY release activity may be mediated by residual Syt VII-GFP molecules that were not knocked down by the siRNA treatment or other Syt isoforms (e.g. Syt IX) (29). It should be noted that the Syt VII siRNA slightly reduced high KCl-dependent NPY secretion form the wild-type PC12 cells (ϳ85-90% of the control cells; this effect was not always statistically significant under our experimental conditions), which strongly supports the previous finding that the level of endogenous expression of Syt VII protein is not so high (less than 3% of that of Syt I) (13,26). By contrast, the siRNA targeted against Syt VII had virtually no effect on low KCl-dependent NPY secretion from either the wild-type or the stable PC12 cell lines (Fig. 6C, open bars).

DISCUSSION
Because phospholipid binding to the C2 domains of Syt VII is activated by lower concentrations of Ca 2ϩ (EC 50 value ϭ 1-2 M Ca 2ϩ ) than is required to activate Syt I, and the recombinant Syt VII transiently overexpressed in PC12 cells is mainly localized at the plasma membrane, it has been proposed that Syt VII functions as a plasma membrane Ca 2ϩ sensor for exocytosis (38,39,41). Recently, however, the plasma membrane localization of Syt VII and its abundance in PC12 cells have been questioned by other groups (13,26,40). First, the endogenous expression level of Syt VII has been found to be low (less than 3% of Syt I) (13,26), and there is no direct evidence that endogenous Syt VII molecules are present at the plasma membrane (i.e. the plasma membrane localization of Syt VII molecules has been shown by transient overexpression of recombinant Syt VII-GFP not by immunostaining of the endogenous Syt VII protein in PC12 cells). Actually, the anti-Syt VII antibody from Ref. 38 stained the cytoplasm of PC12 cells rather than the plasma membrane, and a 40-kDa immunoreactive band detected by this antibody in PC12 cells was unaffected by the Syt VII siRNA treatment (supplemental Fig. 1) indicating that the 40-kDa immunoreactive band is most unlikely to be an endogenous Syt VII␣ molecule. Second, transient overexpression of Syt isoforms often causes mislocalization of the recombinant proteins in PC12 cells, in contrast to the stable expression of Syt isoforms (45). For instance, recombinant Syt IX (or I) transiently overexpressed is often targeted to the plasma membrane of PC12 cells (supplemental Fig. 2), whereas endogenous Syt IX (or I) is clearly localized on dense-core vesicles, but not at the plasma membrane (25,64). Consistent with these observations, the present study on the PC12 cell line stably expressing Syt VII-GFP clearly demonstrated that the targeting of Syt VII molecules to specific membrane compartment(s) and the dynamics of Syt VII molecule during densecore vesicle exocytosis depend on the transfection method (transient expression versus stable expression). The results of the immunocytochemical (Figs. 1 and 4) and the immunoelectron microscopic analyses (Fig. 3) clearly indicated that stably expressing Syt VII molecules in PC12 cells are actually present on dense-core vesicles that undergo Ca 2ϩ -dependent exocytosis. In addition, because some proportions of Syt VII-GFP molecules are also present in lysosomes, Syt VII is likely to be involved in the control of lysosomal exocytosis in PC12 cells as has been shown in fibroblasts (32).
Because Syt VII is a Ca 2ϩ /phospholipid or SNARE-binding protein (14,38,39,41,65,66) like neuronal Ca 2ϩ sensor Syt I and Syt VII-GFP forms a hetero-oligomer with endogenous Syts I and IX (Fig. 2B), stably expressed Syt VII-GFP molecules are likely to be involved in Ca 2ϩ -dependent dense-core vesicle exocytosis in the stable PC12 cell line. Actually, all three independent cell lines showed increased high KCl-dependent NPY secretion compared with the wild-type cells (an ϳ4-fold increase), and selective knockdown of Syt VII-GFP with RNAi dramatically reduced high KCl-dependent NPY secretion but had no effect no low KCl-dependent NPY secretion (Fig. 6). Because the Syt VII siRNA slightly but always reduced high KCl-dependent NPY secretion from the wild-type PC12 cells (ϳ85-90% of the control cells; 1 case of 4 was statistically significant), we speculated that endogenous Syt VII molecules are also involved in the control of Ca 2ϩ -regulated dense-core vesicle exocytosis in the wild-type PC12 cells.
In summary, we have demonstrated that the Syt VII molecules in the stable PC12 cell line are actually present on densecore vesicles and that Syt VII-containing vesicles move dynamically during the exocytosis and endocytosis cycle in response to Ca 2ϩ stimulation. Because stable expression of Syt VII-GFP dramatically enhances stimulated NPY secretion and two C2 domains of Syt VII bind Ca 2ϩ in vitro, we proposed that Syt VII is involved in the control of dense-core vesicle exocytosis, possibly functioning as a Ca 2ϩ senosr on dense-core vesicles. In addition, it is also possible that Syt VII acts as a plasma membrane Ca 2ϩ -senosr because some Syt VII-GFP molecules are actually present at the plasma membrane of the PC12 cell line. Our findings also strongly indicate that the results of transient overexpression of recombinant Syt proteins should be carefully evaluated to avoid misunderstanding the function and localization of Syt isoforms.