Mechanisms Underlying the Neuronal Calcium Sensor-1-evoked Enhancement of Exocytosis in PC12 Cells*

Neuronal calcium sensor-1 (NCS-1) or the originally identified homologue frequenin belongs to a superfamily of EF-hand calcium binding proteins. Although NCS-1 is thought to enhance synaptic efficacy or exocytosis mainly by activating ion channel function, the detailed molecular basis for the enhancement is still a matter of debate. Here, mechanisms underlying the NCS-1-evoked enhancement of exocytosis were investigated using PC12 cells overexpressing NCS-1. NCS-1 was found to have a broad distribution in the cells being partially distributed in the cytosol and associated to vesicles and tubular-like structures. Biochemical and immunohistochemical studies indicated that NCS-1 partially colocalized with the light synaptic vesicle marker synaptophysin. When stimulated with UTP or bradykinin, agonists to phospholipase C-linked receptors, NCS-1 enhanced the agonist-mediated elementary and global Ca2+ signaling and increased the levels of downstream signals of phosphatidylinositol 4-kinase. NCS-1 enhanced the UTP-evoked exocytosis but not the depolarization-evoked Ca2+ responses or exocytosis, suggesting that the enhancement by NCS-1 should involve phospholipase C-linked receptor-mediated signals rather than the Ca2+ channels or exocytotic machinery per se. Taken together, NCS-1 enhances phosphoinositide turnover, resulting in enhancement of Ca2+ signaling and exocytosis. This is a novel regulatory mechanism of exocytosis that might involve the activation of phosphatidylinositol 4-kinase.

Frequenin markedly enhances synaptic efficacy by increasing neurotransmitter release in Drosophila (3) and Xenopus (4) through unknown mechanisms. Furthermore, the expression of NCS-1 (the mammalian homologue of frequenin) in other secretory cell types such as chromaffin cells suggests a more generalized role in exocytosis (5,6). In a model of synaptic transmission in PC12 cells, NCS-1 potentiated ATP-mediated exocytosis of dense core granules but did not affect Ca 2ϩ -induced exocytosis in permeabilized PC12 cells, suggesting that regulation by NCS-1 may be at the level of signal transduction rather than through direct effects on the exocytotic machinery (5,6). Although the precise mechanisms for such enhancement remain unknown, recent accumulating evidence suggests that NCS-1 or frequenin could modulate ion channels such as non-L-type (7,8) and P/Q-type (9) voltage-gated Ca 2ϩ channels and A-type K ϩ channel (10), which may explain such an increased exocytosis.
In addition to ion channels, NCS-1 and/or frequenin binds to multiple proteins such as G-protein receptor kinase 1 (11,12) to regulate exocytosis and to substitute for calmodulin (13). Recently, the yeast frequenin, but not vertebrate NCS-1 (14), has been demonstrated to bind to and activate phosphatidylinositol 4-OH kinase, a phosphatidylinositol 4-kinase (PtdIns4K) isoform (1). This class of enzyme is key to the supply of the phospholipase C substrate phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P 2 ) and therefore has the potential to regulate phosphoinositide and Ca 2ϩ signaling mediated by G-protein-coupled receptors (GPCRs) and receptor tyrosine kinases. The potentiation of this signaling pathway may explain the enhancement of GPCR-mediated vesicle exocytosis by NCS-1.
In the current study, we sought to test this hypothesis by examining the subcellular distribution of NCS-1 and its influence on GPCR-mediated phosphoinositide, elementary and/or global Ca 2ϩ signaling, and exocytosis in PC12 cells. The immunolocalization of NCS-1 is consistent with a role for this protein in exocytosis. Further, we demonstrate that NCS-1 enhances UTP-and bradykinin-evoked exocytosis, and this is associated with enhanced agonist-mediated phosphoinositide signaling and potentiated elementary/global Ca 2ϩ signaling events. These data suggest that the regulation of a PtdIns4K activity by NCS-1 (1) occurs in mammalian cells and could be the mechanism by which this protein enhances exocytotic responses.

EXPERIMENTAL PROCEDURES
Cell Culture and Generation of NCS-1-overexpressing PC12 Cells-PC12 cells were grown as described previously (15). To establish stable NCS-1 clones, PC12 cells were transfected with the rat ncs-1 cDNA (GenBank TM accession number L27421) in pClNeo (kindly provided by Jan Eggermont) and selected for neomycin resistance following standard protocols. Stable clones were further characterized by Western blotting and immunofluorescence as described below.
Subcellular Fractionation-NCS-1 cells were subjected to subcellular fractionation as previously described (16,17) with minor modifications. Briefly, cells were homogenized in buffer containing 0.25 M sucrose, 1 mM (CH 3 COO) 2 Mg, 2 mM EDTA, 10 mM HEPES, pH 7.4, with NaOH and supplemented with protease inhibitors (10 mg/ml aprotinin, 2 mg/ml pepstatin A). To separate the membranes from cytosol, the postnuclear supernatant (PNS) was centrifuged at 80,000 ϫ g for 20 min in a TLA 100.3 rotor using a Beckman Optima TL ultracentrifuge. The pellets were resuspended in homogenization buffer, loaded on the top of sucrose gradients (0.3-1.2 M sucrose), and then centrifuged at 110,000 ϫ g for 30 min (velocity gradient). Fractions of 1 ml were collected from the top of the velocity gradients. 300 l of each fraction from 4 to 8 were pooled and loaded on second sucrose gradients (1.1-2.0 M) and centrifuged at 110,000 ϫ g for 18 h (equilibrium gradients). Equal aliquots (250 l) of both gradient fractions and equal amounts of proteins of PNS prepared from wild type and NCS-1 cells were analyzed by Western blotting as described (18). Briefly, after electrophoresis, samples were electroblotted to nitrocellulose filters (Schleicher & Schuell) and probed with antibodies. The blots were then washed and incubated with the appropriate secondary antibodies conjugated to peroxidase. The peroxidase was detected using chemiluminescent substrate (Pierce). The levels of NCS-1 in the PNS were quantified by measuring the intensity of the bands in the autoradiograms using the analysis program Image 1.61 (National Technical Information Service, Springfield, VA).
Immunocytochemistry-Double immunofluorescence and confocal analysis on fixed cells were carried out as described (18). Briefly, the fixed cells were permeabilized with Triton X-100 and then incubated with the polyclonal anti-NCS-1 antibody and monoclonal antibodies against either synaptophysin or chromogranin B. After washing, the cells were incubated with the appropriate secondary antibodies conjugated to TRITC or fluorescein isothiocyanate, washed again, and mounted on glass coverslips with Mowiol. Images were collected in the MRC-1024 laser-scanning microscope (Bio-Rad) with ϫ40 or ϫ60 objective lenses. For comparison of double-stained patterns, images from the fluorescein isothiocyanate or TRITC channels were acquired independently from the same area of sample and then superimposed. Images were processed using Photoshop 4 (Adobe Systems, Mountain View, CA). Immunolabeling of frozen sections prepared from fixed and sucrose-embedded NCS-1 cells was performed as described (19). Monoclonal and polyclonal antibodies against the soluble granular proteins, chromogranin B and secretogranin II (SgII), were as previously described (19,20). Anti-synaptophysin antibodies were obtained from Roche Molecular Biochemicals.
Ca 2ϩ Imaging in Single Cells-Changes in [Ca 2ϩ ] i in single cells were measured by the fura-2 method as described by Grynkiewicz et al. (21) with minor modification (22). In brief, the culture medium was replaced with balanced salt solution of the following composition: 150 mM NaCl, 5.0 mM KCl, 1.8 mM CaCl 2 , 1.2 mM MgCl 2 , 25 mM HEPES, and 10 mM D-glucose, pH 7.4. Cells were loaded with fura-2 by incubation with 5 M fura-2 acetoxymethylester (fura-2/AM) at room temperature (20 -22°C) in balanced salt solution for 30 min, followed by a balanced salt solution wash and a further 30-min incubation to allow de-esterification of the loaded dye. Then the coverslips were mounted on a microscope (TMD-300, Nikon, Tokyo, Japan) equipped with a xenon lamp and band pass filters of 340-and 360-nm wavelength. Image data, recorded by a high sensitivity silicon intensifier target camera (C-2741-08, Hamamatsu Photonics, Hamamatsu, Japan) were regulated by a Ca 2ϩ analyzing system (Furusawa Laboratory Appliances).
Confocal Ca 2ϩ Imaging-For recording subcellular Ca 2ϩ signals, the cells were loaded with 5 M fluo-3/AM for 30 -40 min at room temperature. After a further 20 -30 min for de-esterification of the acetoxymethyl ester, the coverslips were mounted on a microscope (E-600; Nikon, Tokyo, Japan) equipped with a CSU-10 laser-scanning unit (Yokogawa, Tokyo, Japan) and ICCD camera (C2400-87V; Hamamatsu Photonics). Fluo-3 was excited with the 488-nm line of an argon ion laser, and the emitted fluorescence was collected at wavelengths of Ͼ515 nm). A Nikon CFI Fluor ϫ60 water immersion objective (1.0 numerical aperture) was used. In order to compensate for uneven distribution of the fluo-3, self ratios were calculated (R s ϭ F/F 0 ), which were subsequently converted into Ca 2ϩ concentration using the following equation.
The K d value of fluo-3 for PC12 cells was taken to be 707 nM, which was determined by an in vivo calibration method (23). Dopamine Release-The procedure for the measurement of released dopamine was the same as that described by Koizumi et al. (22). The  (24). Cell residues were extracted with acidified chloroform/methanol to recover the total lipid fraction, and deacylation and separation of glycerophosphoinositol (phosphates) was performed (25). Experiments to measure changes in Ins(1,4,5)P 3 mass were performed as described above except that cells were not prelabeled with myo-[ 3 H]inositol, and LiCl was omitted from the medium (26).

Subcellular Localization of Endogenous and Recombinant
NCS-1 in PC12 Cells-To determine the level of the expression and subcellular distribution of NCS-1, wild type or NCS-1transfected PC12 cells were initially analyzed by immunoblotting and subcellular fractionation. Using a specific antibody, the levels of NCS-1 were analyzed in two different PC12 cell clones (clones 251 and N2) and in a number of stable transfected clones obtained from N2 cells (Fig. 1A). Whereas the level of endogenous NCS-1 was similar in the wild type clones, transfection with NCS-1 resulted in markedly enhanced levels of expression in a number of selected clones. The clone selected for further investigation (from here on referred to as NCS-1 cells) expressed ϳ6 times more NCS-1 than the wild type N2 cells. Data from subcellular fractionation indicated a broad distribution of NCS-1 in transfected cells. As reported (5), a certain portion of NCS-1 was soluble in the cytosol. To study the distribution of the portion of protein associated with cellular organelles, the PNS were subjected to high speed centrifugation, and the pellets were resuspended in homogenization buffer. These were then analyzed by subcellular fractionation using two-step gradients. Analysis of the fractions obtained from the first (velocity) gradients, which separates organelles according to their size, revealed a broad distribution of NCS-1. In these velocity gradients, two peaks of NCS-1 were observed: one in the lighter fractions and the second in fractions 7-9 (Fig.  1B). This distribution is similar to that of synaptophysin, a protein present in the membrane of vesicles highly related to synaptic vesicles (referred to as synaptic-like microvesicles) (27) and known to recycle through the early endosomal compartment in PC12 cells (28,29). A well known marker of the secretory granules, SgII (30), peaked in fractions 4 -8, where some synaptophysin and NCS-1 were also present. In order to further analyze whether a portion of NCS-1 could be present in the membranes of secretory granules, aliquots of fractions 4 -8 were collected and loaded on the equilibrium gradients in order to achieve separation of organelles according to their density. Western blotting analysis of the fractions from these gradients revealed the presence of NCS-1 mainly in the earlier fractions. This distribution was again very similar to that of synaptophysin. There was no colocalization of NCS-1 with the marker for secretory granules, SgII, which was in the denser fractions.
The cellular localization of NCS-1 in stably transfected PC12 cells was also studied by immunofluorescence and immunocytochemistry. As shown in Fig. 2, A and B, when NCS-1 cells were labeled with NCS-1 antibodies, intense immunofluorescence was seen in a large number of cells (about 60%). The immunostaining for NCS-1 was partially diffuse in the positive cells in agreement with its partial distribution in the cytosol (Fig. 2, A and B) (5). The anti-NCS-1 antibody also gave a punctate staining and showed partial colocalization with the immunostaining pattern of synaptophysin (see Fig. 2, A-AЉ). In contrast, very little colocalization was observed with the immunofluorescent pattern of chromogranin B, a well known marker for the secretory granules of PC12 cells (Fig. 2, B-BЉ). In addition, NCS-1 immunostaining was observed at the periphery of the positive cells, raising the possibility that a certain amount of the protein could be bound to the plasma membrane. To confirm this possibility, ultrathin frozen sections prepared from NCS-1 cells were immunolabeled with anti-NCS-1 antibodies and secondarily labeled with anti-rabbit IgG conjugated to 5-nm gold particles. As shown in Fig. 2, C and D, this revealed intracellular NCS-1 immunoreactivity, with gold particles distributed both in the cytoplasm and on small clear vesicles and tubules. In line with the immunofluorescence data, some gold particles were also found to decorate the plasma membrane, thus confirming the presence of NCS-1 at the cell surface in NCS-1 cells.
Enhancement by NCS-1 of Ca 2ϩ Responses to UTP or Bradykinin-Stimulation of PC12 cells with the P2Y 2 -selective agonist UTP results in the activation of phospholipase C (PLC) and the subsequent generation of Ins(1,4,5)P 3 . Indeed, U73122 (10 M), an inhibitor of PLC, substantially decreased the UTPevoked elevation in intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ) in the cells. In the presence of U73122, the UTP-evoked [Ca 2ϩ ] i was decreased by 93.3 Ϯ 0.5% (n ϭ 53 cells from two independent experiments). In PC12 cells, the normal consequence of UTP stimulation is the rapid release of Ca 2ϩ from intracellular Ins(1,4,5)P 3 -sensitive stores, followed by a more sustained Ca 2ϩ entry across plasma membrane in response to store depletion (capacitative Ca 2ϩ entry (CCE)) (31). In the current study, a brief UTP stimulation caused a transient pattern of [Ca 2ϩ ] i elevation consistent with Ca 2ϩ mobilization and CCE in bothvector-transfectedcontrolandNCS-1cellsinaconcentrationdependent manner (Fig. 3A). However, UTP was significantly more potent in NCS-1 cells than in control cells (EC 50 ϭ 19 and 81 M, respectively). Furthermore, the maximal [Ca 2ϩ ] i responses evoked by 300 M UTP were significantly greater in NCS-1 cells than control cells (363 Ϯ 52 nM versus 269 Ϯ 31 nM, p Ͻ 0.05). The enhancement of Ca 2ϩ responses was not a clonal artifact, because in another NCS-1-expressing clonal cell line (clone 3) whose NCS-1 levels are greater than control but less than that of NCS-1 cells, UTP was more potent than in control cells (Fig. 3A, c). Bradykinin also elevates [Ca 2ϩ ] i in PC12 cells via the PLC-coupled bradykinin B 2 receptor (32). Similar to the UTP-mediated responses, bradykinin elevated [Ca 2ϩ ] i in NCS-1 cells more potently than in control cells (EC 50 values; 6.0 and 8.5 nM, respectively), and the maximal response was again enhanced (315 Ϯ 34 nM versus 232 Ϯ 21 nM, p Ͻ 0.05; 100 nM bradykinin) (Fig. 3B). This enhancement by NCS-1 of both UTP-and bradykinin-evoked Ca 2ϩ signaling in NCS-1 cells was also apparent in the absence of extracellular Ca 2ϩ (Ca 2ϩ o ) (Fig. 3, A (b) and B (b)), indicating that the Ca 2ϩ release process had itself been potentiated. U73122 (10 M) also inhibited the Ca 2ϩ responses to UTP in NCS-1 cells ([Ca 2ϩ ] i elevation was 5.3 Ϯ 0.62% of that with UTP alone; n ϭ 52 cells from two independent experiments). In contrast to these differences in UTP-and bradykinin-evoked responses between NCS-1 cells and control cells, the elevations in [Ca 2ϩ ] i evoked by depolarization with high K ϩ (25-100 mM) were almost identical in control and NCS-1 cells (Fig. 3B, a).
To ascertain whether the CCE process per se was affected by NCS-1, we determined CCE directly in control and NCS-1 cells. CCE was activated by inhibiting the Ca 2ϩ -ATPase of the endoplasmic reticulum using cyclopiazonic acid (30 M). The elevations of [Ca 2ϩ ] i induced by cyclopiazonic acid were not different in control and NCS-1 cells (Fig. 3C, a and d; 117.2 ϩ 16.5% of control, p Ͼ 0.05, n ϭ 32 cells), indicating that CCE per se was unaffected by overexpression of NCS-1. We next monitored Ca 2ϩ release and CCE separately to investigate the relationship between UTP-evoked Ca 2ϩ release and Ca 2ϩ entry. In the absence of Ca 2ϩ o , UTP produced concentration-dependent elevations of [Ca 2ϩ ] i in both control and NCS-1 cells (Fig. 3C, b and e). After the removal of UTP, readdition of Ca 2ϩ o produced a sustained rise in [Ca 2ϩ ] i , which slowly returned to basal levels after ϳ10 min. This rise was a consequence of CCE and was dependent upon the initial UTP concentration (Fig.  3C, b, c, e, and f). In other words, the extent of CCE is related to the extent of the UTP-mediated store depletion.
In addition to Ins(1,4,5)P 3 receptors, PC12 cells express functional type 2 ryanodine receptors on intracellular Ca 2ϩ stores (33,34). Activation of these Ca 2ϩ release channels with 40  Elementary Ca 2ϩ Release Events in NCS-1 Cells-Elementary Ca 2ϩ release signals, such as puffs and sparks, are the building blocks for global Ca 2ϩ signals (38,39). Such elementary events have been observed in a wide variety of cell types including PC12 cells (35). Thus, using confocal microscopy, we investigated whether overexpression of NCS-1 affects elementary Ca 2ϩ release events. Elementary Ca 2ϩ signals were observed spontaneously (Fig. 4B, left traces), and these were facilitated upon stimulation with UTP concentrations (0.1 or 0.3 M) that were subthreshold for global Ca 2ϩ signaling. These elementary Ca 2ϩ events were observed even in the absence of Ca 2ϩ o , suggesting that they were mainly due to Ca 2ϩ release from Ca 2ϩ stores. Low concentration (0.3 M) of UTP increased both the number of responding cells and the number of active elementary Ca 2ϩ release sites (Fig. 4C), and enhanced the event amplitude and frequency (Fig. 4D). When NCS-1 cells were stimulated with higher (e.g. 3 M) concentrations of UTP, these elementary events were followed by global Ca 2ϩ signals (Fig. 4, E and F). In control cells, spontaneous elementary Ca 2ϩ events also occurred, but there were no significant differences with NCS-1 cells in either the proportion of cells in which events were observed or the event amplitude. However, in control cells, 0.3 M UTP activated fewer additional elementary Ca 2ϩ release sites in fewer cells. Furthermore, the frequency of UTP-stimulated elementary events was lower than that seen in NCS-1 cells. These data demonstrate that overexpression of NCS-1 enhanced elementary Ca 2ϩ release events mediated by PLC-linked GPCRs, and as a consequence the NCS-1-expressing cells displayed global Ca 2ϩ signals at low levels of stimulation.
NCS-1 Enhances Secretion Evoked by UTP but Not by High K ϩ -UTP evokes the release of dopamine from PC12 cells (31), and the effects of NCS-1 overexpression on UTP-evoked secre-tory responses were therefore investigated. The UTP-evoked release of dopamine into the superfusate was determined by HPLC. UTP caused the release of dopamine in a concentrationdependent manner in control and NCS-1 cells. At each UTP concentration, there was a significantly greater release of dopamine in NCS-1 cells compared with control cells (Fig. 5). We also examined the UTP-evoked release of dopamine in clone 3 cells (see Fig. 3A, c). The dopamine release by clone 3 cells was greater than control but less than that of NCS-1 cells. This UTP-evoked release of dopamine was dependent on Ca 2ϩ o in all cell lines (Fig. 5, triangles). There was no significant difference in spontaneous or high K ϩ (25-100 mM)-evoked dopamine release between control, NCS-1 and clone 3 cells. In addition, no significant difference was observed in total cellular dopamine content between the clones (control, 287.8 Ϯ 22.6 (n ϭ 9); NCS-1 cell, 291.6 Ϯ 30.7 (n ϭ 9) ng/dish; p Ͼ 0.05).
Enhancement of Phosphoinositide Signaling by NCS-1-Finally, the levels of various phosphoinositides in both clones were investigated (Fig. 6). Under basal conditions, there was no significant difference in [ 3 H]PtdIns levels between the two clones ( Fig. 6A, a). In control cells, [ 3 H]phosphatidylinositol 4-phosphate (PtdIns(4)P) and [ 3 H]PtdIns(4,5)P 2 represented 1.0 and 2.2% of the total phosphoinositide pool, respectively. In NCS-1 cells, the levels of both PtdIns(4)P and PtdIns(4,5)P 2 were significantly increased to 2.2 and 3.7% of the total phosphoinositide pool, respectively (Fig. 6A, a-c). Stimulation of both clones with UTP (300 M) resulted in accumulation of [ 3 H]InsP x against a Li ϩ block of inositol monophosphatase (Fig.  6A, d). The response to UTP was significantly enhanced in NCS-1 cells such that accumulation was greater (198% over basal at 20 min compared with 83% in vector controls) (Fig. 6A,  d). UTP caused an increase in Ins(1,4,5)P 3 mass in control cells (Fig. 6B). There was a marked potentiation in the magnitude (228% over basal level at 20 s compared with 146% in vector controls) and the duration of the Ins(1,4,5)P 3 response to UTP in NCS-1 cells (Fig. 6B).

Subcellular Localization of NCS-1 and Implications for Its
Functional Role-NCS-1 showed a broad subcellular localization in PC12 cells, being partially distributed in the cytoplasm and associated with vesicles and tubular-like structures. The morphological and biochemical results suggest that some NCS-1 may be associated with the membrane of the synapticlike microvesicles. In contrast to previous observations (5), our results demonstrate that NCS-1 is mainly not associated with the membranes of secretory granules. Interestingly, NCS-1 is also associated with the plasma membrane. This localization may be consistent with the possible role of NCS-1 in the modulation of the function of plasma membrane channels and receptors (40), especially of voltage-gated Ca 2ϩ channels (7)(8)(9). On the other hand, NCS-1 at the plasma membrane and on vesicle membranes may play an important role in the metabolism of phosphoinositides. This phospholipid family is involved in both signal transduction and also in vesicle formation from the plasma membrane, trans-Golgi network and endosomes (41). In yeast, the homologue of NCS-1 has been demonstrated to exist in a complex with and to activate the yeast PtdIns4K, phosphatidylinositol 4-OH kinase (1), and very recent data have shown a direct interaction of NCS-1 with PtdIns4K in eukaryotic cells as well (42). The membranous distribution of NCS-1 in the present study raises the possibility that NCS-1 may contribute to the activation and recruitment of PtdIns4K to the plasma membrane (or other organelles), allowing an increase of PtdIns metabolism and synthesis of the signaling molecule Ins(1,4,5)P 3 , which therefore would affect subsequent intracellular responses such as Ca 2ϩ signaling and exocytosis (see Fig. 7). (43) and ligand-gated channels (44) through which Ca 2ϩ entry is able to evoke exocytosis. Of particular relevance to the current study, PC12 cells express P2X 2 nonselective cation channels (45). Although ATP effectively activates this receptor, UTP evokes little (0 -6% of the maximal ATP response) (46) or no (45) (34,48) further support the role of CCE in exocytosis.

Overexpression of NCS-1 Enhances Elementary and Global Ca 2ϩ Signaling and Secretory Responses-PC12 cells express both voltage-
NCS-1 did not influence depolarization (high K ϩ )-evoked exocytosis, supporting a previous observation that Ca 2ϩ sensitivity of the exocytotic process is not affected by NCS-1 overexpression (5). In addition, the effect of NCS-1 on voltage-gated Ca 2ϩ channels in PC12 cells appears to be negligible. The ability of NCS-1 to enhance exocytosis in PC12 cells therefore implies that this protein enhances CCE. However, CCE resulting from store depletion by inhibition of the endoplasmic reticulum Ca 2ϩ -ATPase with cyclopiazonic acid was unaffected. Furthermore, there was no difference in stored Ca 2ϩ contents or in responses to caffeine between the clones (see "Results"). These data demonstrate that the CCE mechanism per se was not potentiated by NCS-1 overexpression and indicate that NCS-1 exerts its effects on secretion through enhanced GPCRmediated Ca 2ϩ release from intracellular stores. Our data are fully in accord with this hypothesis and demonstrate that overexpression of NCS-1 enhances elementary and global Ca 2ϩ signaling in response to GPCR activation. Although the amplitudes of the initial, transient agonist-mediated global Ca 2ϩ signaling events were in excess of the more sustained elevation due to CCE, the former were unable to evoke exocytosis. This is in agreement with previous studies demonstrating that CCE in response to depletion of either Ins(1,4,5)P 3 -or ryanodine-sensitive stores triggers exocytosis more efficiently than the Ca 2ϩ release process (31,34,48,49). The reason for this is unclear. The exocytotic Ca 2ϩ sensors appear to possess a low affinity for Ca 2ϩ , requiring 10 -100 M for activation (50,51). Although global measurements following depolarization reveal [Ca 2ϩ ] i on the order of 1 M in many cell types, it is believed that much higher local [Ca 2ϩ ] i is achieved in subplasmalemmal microdomains due to the spatial packing of voltage-gated Ca 2ϩ channels (52,53). It is likely, therefore, that the spatial arrangement of Ca 2ϩ channels (e.g. CCE channels and Ins(1,4,5)P 3 receptors) and the exocytotic machinery dictate whether Ca 2ϩ fluxes are able to regulate vesicular release.
Overexpression of NCS-1 also enhanced the elementary Ca 2ϩ release events. Such elementary Ca 2ϩ signals were brief (lifetime ϳ1 s) [Ca 2ϩ ] i increases with a spatial spread of ϳ3-6 m. In control cells, the amplitudes of the elementary events typically ranged from 80 to 100 nM, and the frequency was generally less than 0.1 Hz (0.06 -0.1 Hz). All of these characteristic features are well in accord with elementary Ca 2ϩ events previously described in nerve growth factor-differentiated PC12 cells (35). Without stimulation (ϪUTP), there was little difference in the characteristics of elementary Ca 2ϩ release events between control and NCS-1 cells. However, upon stimulation with UTP, the number of cells showing elementary Ca 2ϩ signals increased, as did the number of active elementary Ca 2ϩ release sites within individual cells. In addition, the frequency of the events was dramatically enhanced. Global Ca 2ϩ signals arise via the coordinated recruitment of elementary Ca 2ϩ release sites (35,39,54). The potentiation of elementary Ca 2ϩ signaling in NCS-1 cells therefore underlies their propensity to show global Ca 2ϩ signals with low UTP concentrations.
Overexpression of NCS-1 Increases the Level of Polyphosphoinositides and Facilitates Agonist-mediated Ins (1,4,5)P 3 Production-Varieties of mechanisms have been proposed to account for the role of NCS-1 in synaptic enhancement and in exocytosis. Probable explanations for the NCS-1-evoked enhancement may be due to modulation of ion channels such as voltage-gated Ca 2ϩ channels (7-9) and A-type K ϩ channels (10). Recently, NCS-1 has been identified as an upstream activator of a specific isoform of PtdIns4K, phosphatidylinositol 4-OH kinase, in yeast (1). Here we demonstrated that in PC12 cells vertebrate NCS-1 significantly increased the levels of both PtdIns(4)P and PtdIns(4,5)P 2 without affecting those of PtdIns and facilitated Ins(1,4,5)P 3 production in response to UTP, an agonist to PLC␤-linked GPCRs. We therefore propose the following model for mechanisms that may underlie the enhancement by NCS-1 of exocytosis (see Fig. 7). First, NCS-1 binds to PtdIns4K located on the plasma membrane (or synaptic vesicles) and up-regulates its activity. This is followed by an increase in the local production of PtdIns(4)P and PtdIns(4,5)P 2 , which serves as substrate for PLC, resulting in an increase in the production of Ins(1,4,5)P 3 and diacylglycerol in response to UTP (or bradykinin). This local increase in Ins(1,4,5)P 3 evokes elementary and global Ca 2ϩ responses including both Ca 2ϩ release via the Ins(1,4,5)P 3 receptor and CCE, leading to an enhancement of the secretory response. The dynamics of the interaction between PtdIns4K and NCS-1 remain to be investigated in detail. The fact that the interaction and activation of phosphatidylinositol 4-OH kinase/PtdIns4K is dependent on the myristoylation of NCS-1 suggests that the interaction should occur in a membranous environment. NCS-1, which is also present on the synaptic vesicle fraction, could translocate to the plasma membrane upon stimulation and thereby either recruit PtdIns4K to the plasmamembrane or activate preexisting PtdIns4K (55). The stable PC12 cells now allow us to address these issues.
In summary, we have for the first time demonstrated a role of mammalian NCS-1, as an important member of this family of EF-hand of Ca 2ϩ -binding proteins, in phosphoinositide turnover, Ca 2ϩ signaling, and exocytosis. We have presented evidence that NCS-1 enhances exocytosis by increasing Ca 2ϩ signaling, which involves enhanced generation of Ins(1,4,5)P 3 . This is a novel regulatory mechanism of exocytosis that might involve the activation of PtdInd4K. The revealed regulatory mechanisms would provide the basis to study their importance in other neurosecretory cells such as chromaffin cells or neurons.