Synaptotagmin VII Is Targeted to Secretory Organelles in PC12 Cells, Where It Functions as a High-Affinity Calcium Sensor

Abstract

Synaptotagmin (syt) I is thought to act as a Ca²⁺ sensor that regulates neuronal exocytosis. Fifteen additional isoforms of syt have been identified, but their functions are less well understood. Here, we used PC12 cells to test the idea that different isoforms of syt impart cells with distinct metal (i.e., Ca²⁺, Ba²⁺, and Sr²⁺) requirements for secretion. These cells express syt's I and IX (syt IX sometimes referred to as syt V), which have low apparent metal affinities, at much higher levels than syt VII, which we show has a relatively high apparent affinity for metals. We found that syt I and VII partially colocalize on large dense core vesicles and that upregulation of syt VII produces a concomitant increase in the divalent cation sensitivity of catecholamine release from PC12 cells. Furthermore, RNA interference-mediated knockdown of endogenous syt VII reduced the metal sensitivity of release. These data support the hypothesis that the complement of syt's expressed by a cell, in conjunction with their metal affinity, determines the divalent cation sensitivity of exocytosis.

A complex composed of SNARE (for soluble NSF attachment protein receptor) proteins, and the Ca²⁺-binding protein synaptotagmin (syt) I, is thought to form the minimal Ca²⁺-triggered membrane fusion machine in neurons (38, 44). Fifteen additional isoforms of syt have been identified (7), suggesting that members of this family of proteins have diverged to fulfill distinct functions. One idea is that the wide-ranging Ca²⁺ requirements for membrane fusion observed in different cell types (e.g., references 4, 20, and 33) are due to the expression of distinct syt isoforms with different metal affinities.

The apparent Ca²⁺ affinities of a number of syt isoforms have been explored by measuring the [Ca²⁺]_1/2 for binding to liposomes composed of phosphatidylserine (PS) and phosphatidylcholine (PC). In a number of studies it was shown that all syt isoforms tested exhibited roughly the same apparent affinity for Ca²⁺. In particular, syt III and VII were shown to have virtually the same [Ca²⁺]_1/2 values as syt I (26, 27, 39). However, more recent studies suggested that syt III and VII may bind PS/PC at much lower [Ca²⁺] (36). Thus, whether different isoforms of syt have distinct affinities for Ca²⁺ remains an open issue. Moreover, there are no functional data to address the idea that syt's couple distinct ranges of metals to membrane fusion.

If syt's have in fact diverged to sense different ranges of Ca²⁺ and if high- and low-affinity isoforms of syt are both targeted to the same secretory organelle, it follows that the ratio of these isoforms might determine the metal (i.e., Ca²⁺, Ba²⁺, and Sr²⁺) requirements for exocytosis. At present, there is little consensus regarding the localization of syt isoforms other than syt I. For example, it has been reported that syt VII is targeted to intracellular structures, including the Golgi (16) and large dense core vesicles (LDCV) in PC12 cells (14). Other studies concluded that syt VII is not targeted to intracellular organelles in PC12 cells but rather is selectively localized to the plasma membrane in this cell type, as well as in neurons, and in non-neuronal cells (19, 35). Finally, Andrews and coworkers reported that syt VII is targeted to lysosomes in non-neuronal cells (28, 32).

Here, we focus on syt VII and PC12 cells to test the hypothesis that syt's have diverged to sense different ranges of [Ca²⁺] and that more than one isoform of syt is targeted to secretory organelles. Our data indicate that the ratio of syt's with high (e.g., syt VII) and low apparent affinities for Ca²⁺ (e.g., syt I and IX) can fine-tune the metal requirements for exocytosis. We note that syt IX (15, 22, 37, 48) is sometimes referred to as syt V (8, 21).

MATERIALS AND METHODS

Constructs and recombinant proteins.

cDNA encoding rat syt I ( (31a) and mouse syt VII (27) was provided by T.C. Südhof (University of Texas Southwestern Medical Institute, Dallas, TX) and M. Fukuda (RIKEN, Saitama, Japan), respectively. The D374 mutation in the syt I cDNA was corrected by replacement of the D with G (10). syt fragments encoding the cytoplasmic domain (C2A-C2B) of syt I (amino acids [aa] 96 to 421) and VII (aa 134 to 403) were subcloned into pGEX (Amersham Biosciences) vectors (6). Glutathione S-transferase (GST)-tagged proteins were purified as described previously (6), except that a high-salt buffer that contained DNase and RNase was used to remove bacterial contaminants from the GST-fused proteins (45). As a standard for the blots in Fig. 2, we also generated a GST fusion protein using residues 1 to 260 of syt VII as described previously (37); the fragment contains the epitope used to generate the syt VII antibody (37).

FIG. 2.

Upregulation of syt VII in PC12 cells. Quantitative immunoblots were carried out as described previously (37). The indicated amounts of GST-syt VII (aa 1 to 260) served as standards, and 20 μg of membrane preparations from wild-type PC12 cells, control cell lines (H1, B2, and E1), or syt VII⁺ cell lines (A1, G10, and G3) were subjected to SDS-polyacrylamide gel electrophoresis and immunoblot analysis. As a control for loading, we also blotted for syb II, syt I, and syt IX. syt VII was quantified by densitometry. The amount of syt VII in each cell line was determined with GST fusion protein as a standard. The abundance of syt VII was then plotted as percentage of total membrane protein (lower panel).

To build the syt VII overexpression construct, the single EcoRI site in the full-length mouse syt VII cDNA (accession number NM_018801) was removed by changing a nucleotide for amino acid N51 from AAT to AAC (AAT and AAC both encode N). The resulting cDNA was then subcloned into the pTRE vector (Clontech) using EcoRI and BamHI.

Oligonucleotides encoding shRNAs were obtained from QIAGEN and were phosphorylated using T4 polynucleotide kinase (Invitrogen). The resulting inserts were ligated into pSHAG-1 vector (provided by G. Hannon, Cold Spring Harbor, NY) via BseRI and BamHI sites. Three constructs were tested for their ability to knockdown exogenous and endogenous syt VII. The construct that exhibited the most effective knockdown of syt VII targets an mRNA sequence that corresponds to the following coding sequence: 1002 to 1020. The oligonucleotide sequence was 5′-TCCAGTTCAGTGTTGGCTAGAAGCTTGTAGCCAACACTGAACTGGATTTTTT-3′ (sense) and 5′-GATCAAAAAATCCAGTTCAGTGTTGGCTACAAGCTTCTAGCCAACACTGAACTGGACG-3′ (antisense). This construct was designated pSHAG-1-syt VII and was used to knock-down syt VII in the voltammetry assays shown in Fig. 7. All constructs were confirmed by DNA sequencing.

FIG. 7.

Knockdown of syt VII decreases the Ca²⁺, Sr²⁺, and Ba²⁺ sensitivity of exocytosis from PC12 cells. (A) Predicted structure of the shRNA transcript used to down-regulate syt VII mRNA (upper left panel). The DNA sequence that encodes this shRNA was inserted into the pSHAG-1 vector; the approximate position that the shRNA targets the syt VII mRNA is shown in the lower left panel. Transfection with the shRNA-encoding construct (syt VII⁻) effectively knocked down syt VII expression but had no effect on syt I and syt IX levels, as shown by the immunoblots in the right panel. (B to D) Cells transiently transfected with either pSHAG-1 alone (control) or the pSHAG-1-syt VII construct (syt VII⁻) were permeabilized and analyzed by using RDE voltammetry. Release was triggered by Ca²⁺, Sr²⁺, or Ba²⁺. Traces were collected and analyzed as described in Fig. 6, except that both normalized amplitude (amp) and rate of release (1/τ) are plotted as a function of [divalent cation]. [divalent cation]_1/2 values from the plots in panels B to D are listed in Table 3.

Lipids and liposome-binding assays.

Synthetic 1,2-dioleoyl-sn-glycero-3-[phospho-l-serine] (PS) and 1,2-dioleoyl-sn-glycero-3-phosphocholine (PC) were purchased from Avanti-Polar Lipids. l-3-Phosphatidyl[N-methyl-³H]choline-1,2-dipalmitoyl (³[H]PC) was purchased from Amersham Biosciences. Liposomes were prepared and ³H-labeled liposome binding assays were carried out as described previously (43). The data were fitted with a sigmoidal dose-response curve (variable slope) using Prism 2.0 software. We note that the [Ca²⁺]_1/2 value reported here (14 μM) is somewhat lower than the values reported previously (e.g., ∼20 μM [9]). This slight difference is because in the current study the free [Ca²⁺] was calculated, whereas in previous studies we determined the [Ca²⁺] by using a Ca²⁺ electrode (9). Using calculated values simplified comparisons between the relative abilities of Ca²⁺, Sr²⁺, and Ba²⁺ to trigger binding. Binding assays were carried out in triplicate on three independent occasions. The results in all cases were similar; representative data from one experiment are shown in Fig. 1, where we plot the mean ± the standard error of the mean (SEM).

FIG. 1.

The cytoplasmic domain of syt VII exhibits a higher apparent affinity for divalent cations than syt I. GST-immobilized cytoplasmic domains of syt I and VII were assayed for binding ³H-labeled liposomes composed of 25% PS and 75% PC. Counts-per-minute (CPM) values were plotted against [divalent cation], and the data were fitted with sigmoidal dose-response curves (Prism 2.0 software) to yield EC₅₀ values and Hill coefficients; data are expressed as means ± the SEM. (A) The [Ca²⁺]_1/2 values for syt I and syt VII were 14 ± 0.3 and 0.86 ± 0.2 μM, and the Hill coefficients were 3.5 ± 0.2 and 3.6 ± 0.2, respectively. (B) The [Sr²⁺]_1/2 values for syt I and syt VII were 26 ± 1.2 and 8.0 ± 0.7 μM, and the Hill coefficients were 1.5 ± 0.2 and 1.5 ± 0.2, respectively. (C) The [Ba²⁺]_1/2 values for syt I and syt VII were 51 ± 1.2 and 5.8 ± 1.3 μM, and the Hill coefficients were 1.4 ± 0.2 and 1.3 ± 0.1, respectively.

Cell culture, transfection, and immunoblot analysis.

PC12 cells were cultured according to standard procedures and were transfected by electroporation. Briefly, cells were harvested in Hanks balanced salt solution plus 1 mM EGTA and centrifuged at 200 × g for 3 min. Cell pellets were washed once with 10 ml of Cytomix (120 mM KCl, 0.15 mM CaCl₂, 10 mM KH₂PO₄, 25 mM HEPES, 2 mM EGTA, 5 mM MgCl₂ [pH 7.6]). Cells were then resuspended in 500 μl of Cytomix/plate and transferred to an electroporation cuvette with a 0.4-cm electrode gap (Bio-Rad). A total of 100 μg of the pSHAG-1 construct was added to the cuvette, mixed with cells, and immediately pulsed (290 V; 1,000 μF) by using a Gene Pulser II (Bio-Rad). Cells were then transferred into fresh normal media supplemented with 10% fetal bovine serum.

For immunoblot analysis of RNA interference (RNAi) knockdown cells, transiently transfected cells were harvested and lysed by using lysis buffer (0.5% Triton X-100, 0.05% sodium dodecyl sulfate [SDS], and 5 mM phenylmethylsulfonyl fluoride in phosphate-buffered saline) 48 to 72 h after transfection. Samples were centrifuged at 21,000 × g for 10 min, and the protein concentration of the supernatant was quantified with a BCA assay (Pierce). To detect syt VII expression, 200 μg of supernatant from each sample was subjected to SDS-polyacrylamide gel electrophoresis and immunoblot analysis. syt VII and IX were detected with polyclonal antibodies described in reference 37; syt I was detected with a monoclonal antibody (41.1; provided by R. Jahn, Gottingen, Germany). For detection of syt I and syt IX, 20 μg of supernatant from each sample was analyzed (37).

For stable transfections, tetracycline-off PC12 cells were cotransfected with 20 μg of pTRE-syt VII and 20 μg of pTK-hygromycin. Hygromycin was added after 48 h to a final concentration of 0.2 mg/ml. Medium was replaced with fresh hygromycin-containing medium every 4 days. Individual cell colonies were picked after 3 to 4 weeks of selection and placed in 96-well plates. Control cell lines were generated in parallel by cotransfection of 20 μg of pTRE lacking an insert and 20 μg of pTK-hygromycin. To compare expression levels of syt VII in wild-type PC12 cells, syt VII⁺ cell lines, and control cell lines, cells were harvested and lysed to obtain a crude total membrane fraction (37). Quantitative immunoblot assays were carried out as described previously (37) by using a fragment of syt VII (residues 1 to 260) fused to GST as a standard (20 μg of membrane preparation was assayed for each cell line). Signals were quantified by densitometry, and the abundance of syt VII in the membrane preparation was calculated from the standard curve made from recombinant protein. These data are plotted in Fig. 2 (lower panel); error bars represent the SEM from triplicate determinations. As controls, we also blotted for synaptobrevin II (monoclonal antibody 69.1), syt I, and syt IX. We assayed for tetracycline induction of syt VII expression but did not observe consistent increases in expression. Thus, all experiments were carried out without inducing the cells.

Immunocytochemistry.

syt VII⁺ PC12 cells were plated on 0.01% poly-l-lysine-coated coverslips at approximately the same density for each antibody labeling condition. Cells were washed once in phosphate-buffered saline (PBS) without Ca²⁺ or Mg²⁺ and fixed for 15 min in 4% paraformaldehyde with 4% sucrose in PBS. After one quick wash in PBS, residual paraformaldehyde was quenched by a 15-min incubation in 0.15 M glycine in PBS. Cells were simultaneously permeabilized and blocked overnight with 0.25 mg of saponin/ml in blocking buffer consisting of 10% goat serum, 6% bovine serum albumin (BSA), 0.1 M NH₄Cl, 1 mM MgCl₂, and 1 mM EGTA in 0.1 M morpholineethanesulfonic acid (pH 6.8). Primary and secondary antibodies were diluted in blocking buffer containing saponin. Cells were incubated for 1 h with a combination of the following primary antibodies: polyclonal rabbit anti-syt VII antibody (1:350) (37), monoclonal anti-syt I antibody (1:500; antibody 41.1), monoclonal anti-chromogranin B antibody (Cg; 1:500; provided by W. Huttner, Dresden, Germany), monoclonal anti-Lamp1 antibody (1:500; provided by N. Andrews, New Haven, CT), polyclonal rabbit anti-secretogranin II (Sg; 1:500; QED Biosciences, San Diego, CA), or monoclonal anti-syntaxin 1A antibody (1:500; provided by R. Jahn, Gottingen, Germany). The cells were then washed four times in PBS plus saponin and incubated for 1 h with Alexa488-conjugated anti-rabbit and Alexa594-conjugated anti-mouse secondary antibodies (Molecular Probes, Eugene, OR) diluted 1:800. After four 10-min washes in PBS with saponin, coverslips were washed once in distilled H₂O and mounted on slides with Prolong mounting reagent (Molecular Probes, Eugene, OR). Cells were imaged on a Nikon TE300 inverted microscope using a 100×, 1.4-numerical-aperture objective lens. Images were acquired by using a Princeton Instruments Micromax cooled charge-coupled device camera (Roper Scientific, Princeton, NJ) controlled by Metamorph software (Universal Imaging Corp, West Chester, PA). Z-series images were obtained at ×100 magnification with 200-nm sectioning. The resulting Z-stacks were deconvolved (100 to 200 iterations) by using Autodeblur/autovisualize software (AutoQuant Imaging Inc, Watervliet, NY). Colocalization was quantified by thresholding the green and red images and measuring the overlap between the two (red to green and green to red) for each optical section by using ImageJ (National Institutes of Health) software. Thresholds were set to include the top 33% of the total integrated fluorescence of the cell, which corresponds to ∼8% Lamp1, ∼15% syt VII, ∼11% chromogranin and secretogranin, and ∼14% syt I of the total cell volume. As such, if the distributions of, for example, syt VII and chromogranin, were random (uncorrelated), one would predict that they would overlap 1.2% of the time. In each case, we corrected for this factor. For each pair of antibodies six cells were analyzed, taken from at least two coverslips. For each cell, all voxels were analyzed for colocalization (∼50 optical sections per cell) to ensure z position within the cell did not influence the analysis. Colocalization was assessed by overlap of thresholds for the two labels. Pooled data are expressed as mean ± the SEM.

For the line scan data plotted in Fig. 3C, images of syntaxin 1A and syt VII were obtained and deconvolved as described above (using syt VII⁺ clone G10). Two 0.5-μm-thick lines were drawn across individual cells at roughly the angles shown in Fig. 3B. To better compare overlap, we normalized the syt VII and syntaxin fluorescence signals to the same arbitrary value. We then determined the degree of overlap between signals. This was carried out by using data from 10 cells, two line scans per cell. The overall degree of overlap was 12% (SEM = 1.95). A representative example of the normalized line scan data is shown in Fig. 3C.

FIG. 3.

syt VII is predominantly targeted to intracellular organelles in PC12 cells. (A) Low-magnification image of control and syt VII⁺ cells (clone G10) after staining with a syt VII polyclonal antibody. syt VII was stained with a rabbit anti-syt VII antibody and an Alexa-488-labeled secondary antibody. Camera settings were selected such that upregulated syt VII was readily detected; at these settings, endogenous syt VII was not detected. Hence, the syt VII antibody specifically recognizes upregulated syt VII. Note: these images were not deconvolved. (B) High-magnification deconvolved image of a syt VII⁺ PC12 cell (clone G10). To delineate the plasma membrane, syntaxin 1A was stained with HPC-1 as the primary antibody and Alexa-594 as the secondary antibody (red channel). syt VII was stained as in panel A (green channel). To estimate the relative levels of syt VII in the cell interior versus the cortical region of the cell (which includes the plasma membrane), line scans were taken across the longest dimensions of each cell (care was taken to avoid the nucleus). Two line scans were taken per cell, at roughly the angles shown in the figure. (C) Line scan data obtained from the representative cell shown in panel B; details are provided in Materials and Methods. Ten cells were analyzed in the same manner; only 12% of the syt VII signal overlapped with syntaxin 1A, and ≥88% of syt VII was intracellular.

Rotating disk electrode (RDE) voltammetry.

Cells were cracked (>90%) by a single pass through a 0.2510-in. bore ball homogenizer (H&Y Enterprise, Redwood City, CA) using a 0.2506-in. ball bearing. Metal concentration in the release buffers was calculated by using WEBMAXC 2.22 software. To buffer [Ba²⁺] and [Sr²⁺] we used 0.2 mM EGTA, and to buffer [Ca²⁺] we used 0.2 mM dipivaloyl-l-tartaric acid (DPTA) as the chelating agent. Note that for the Ca²⁺ dose responses, the actual final [Ca²⁺] was confirmed by using a Ca²⁺-sensitive electrode.

Release of catecholamine was monitored in real time with a rotating disk electrode voltammetry setup as described previously (12, 37). Release curves were fitted with single-exponential functions by using Axograph software, yielding time constants (τ). The rate (i.e., the reciprocal of the time constant, 1/τ), was plotted against [divalent cation]. The amplitude of response 80 s after addition of metal was also determined. These values were normalized to the maximum amplitude and plotted versus [divalent cation]. The data are shown as the mean ± the SEM (from triplicate determinations) in all plots and tables.

For the Sr²⁺ and Ba²⁺ titrations, data were fit to sigmoidal dose-response curves as described above for liposome binding assays. Since the Ca²⁺ dose-response data were not well fitted by sigmoidal dose-response equations, we connected the datum points and determined the effective concentration for 50% effect (EC₅₀) values by drawing a line through the 50% Y_max and reading the X value at the point where this line intersects the data curve. This analysis was carried out using data from three independent Ca²⁺ titrations for the experiments shown in Fig. 5C and D and Fig. 7B. The mean ± the SEM for these studies are provided in Tables 1 and 3.

FIG. 5.

Upregulation of syt VII in PC12 cells increases the Ca²⁺ sensitivity of catecholamine release. Ca²⁺-triggered catecholamine release from parental wild-type PC12 cells (control) or syt VII⁺ cells (the A1 clone) was monitored by RDE voltammetry. (A and B) Representative release profiles from the control (A) and syt VII⁺ cells (B). Arrows indicate the addition of Ca²⁺ to the cracked cell suspension; the final [Ca²⁺] is given next to each voltammetry trace. (C) Normalized amplitudes of response (amp) of the release profiles were plotted against [Ca²⁺]. (D) Reciprocal of the time constant (i.e., 1/τ) for release was plotted against [Ca²⁺]. [Ca²⁺]_1/2 values from the plots in panels C and D are listed in Table 1.

TABLE 1.

Ca²⁺ sensitivity of catecholamine release from control and syt VII⁺ (clone A1) cells

Control or syt VII+	Mean [Ca2+]1/2 (μM) ± SEM froma:
	amp	1/τ
Control	0.7 ± 0.1	0.8 ± 0.1
A1 syt VII+	0.4 ± 0.1*	0.6 ± 0.05*

^aThe asterisk indicates a significant difference between the syt VII⁺ clone A1 and the control. P < 0.05 using unpaired t test. amp, normalized amplitude.

TABLE 3.

Ca²⁺, Sr²⁺, and Ba²⁺ sensitivity of catecholamine release from control and syt VII⁻ cells

Divalent cation	Mean [divalent cation]1/2 concn (μM) ± SEMa from:
	amp		1/τ
	Control	syt VII−	Control	syt VII−
Ca2+	0.6 ± 0.1	1.0 ± 0.2*	0.7 ± 0.1	1.8 ± 0.2**
Sr2+	46 ± 8	83 ± 13*	20 ± 5	45 ± 9*
Ba2+	84 ± 10	114 ± 5**	86 ± 11	128 ± 7**

^aAsterisks indicate significant differences between syt VII⁻ clones and control. Double asterisks indicate P < 0.01, and single asterisk indicates P < 0.05 using the unpaired t test. amp, normalized amplitude.

We note that the [Ca²⁺]_1/2 values for release in the current study were 0.7 to 0.8 μM Ca²⁺. This value is significantly lower than our earlier estimate using this method (12). This difference might be due to the presence of intact cells in our previous study (12), which appear to release catecholamine in response to relatively high extracellular [Ca²⁺]. We have overcome this problem by using a ball homogenizer to ensure that >90% of the cells are made permeable.

RESULTS

syt I and syt VII exhibit distinct metal requirements for binding to PS-harboring liposomes.

In the first series of experiments we addressed the question of whether syt I and VII respond to similar or distinct concentrations of Ca²⁺. For these experiments, we immobilized the cytoplasmic domains (C2A and C2B) of each isoform onto beads and assayed their ability to bind ³H-labeled liposomes composed of 25% PS and 75% PC as a function of [Ca²⁺]. The intact cytoplasmic domains of each isoform were used because, at least in the case of syt I, the tandem C2 domains cooperate with each other for binding to both anionic lipids (3, 43) and t-SNAREs (2, 6, 37, 38). In addition to Ca²⁺, we also monitored C2A-C2B-membrane interactions as a function of [Ba²⁺] and [Sr²⁺]. Both of these metals have been shown to trigger release in a variety of cell types (18, 24, 34). Moreover, we make extensive use of Ba²⁺ and Sr²⁺ in the exocytosis assays, detailed below.

As shown in Fig. 1, the EC₅₀ values for Ca²⁺-, Sr²⁺-, and Ba²⁺-triggered binding of syt VII to PS/PC liposomes were markedly lower than for syt I (∼16-, 3-, and 9-fold lower, respectively). Thus, syt VII exhibits a higher apparent affinity than syt I for all divalent cations capable of triggering exocytosis (i.e., Ca²⁺, Sr²⁺, and Ba²⁺) (24) and is a bona fide high-affinity divalent cation sensor.

Upregulation and subcellular localization of syt VII in PC12 cells.

If syt's function as metal sensors in vivo, syt VII would be expected to trigger release at lower [Ca²⁺], [Ba²⁺], and [Sr²⁺] than syt I. We addressed this question using PC12 cells as a model system. In this cell line, the most abundant isoforms are syt I and IX (37), both of which are relatively low affinity divalent cation sensors (Fig. 1; see also reference 22, where it is reported that the [Ca²⁺]_1/2 for syt IX-liposome interactions is 14 μM). The only high-affinity isoform expressed in PC12 cells, syt VII, is at least 30 times less abundant than syt I (37).

In order to change the ratio between syt VII and syt I/IX (i.e., between high- and low-affinity sensors), we increased the expression level of syt VII so that it approached the abundance of syt I and IX. To this end, we transfected PC12 cells with an expression vector that encodes untagged syt VII. Three independent stable cell lines that overexpressed syt VII (designated as syt VII⁺) were characterized, as were three control cell lines transfected with the empty vector. Quantitative immunoblot analysis revealed that syt VII was upregulated to different extents in each of the VII⁺ clones (Fig. 2; syt VII⁺ clones A1, G10 and G3 exhibited ∼17-, 10-, and-6-fold increases in the level of syt VII expression, respectively). In clones A1, G10, and G3, syt VII had ∼50, 30, and 18% of the abundance of syt IX, and ∼30, 18, and 10% of the levels of syt I, respectively (note that the levels of syt I and IX were first determined in reference 37; similar values were obtained with the cells described in the present study [data not shown]). syt VII was not upregulated in any of the control lines.

We then explored the subcellular localization of untagged syt VII in the syt VII⁺ cell lines. It is currently the subject of debate as to whether syt VII-green fluorescent protein (GFP) fusion proteins are targeted to the plasma membrane (19, 35) or to intracellular organelles, including LDCVs (14, 16), in PC12 cells. To address this question, we first determined whether our syt VII antibody (used for the immunoblot analysis shown in Fig. 2) could be used for immunocytochemistry. As shown in Fig. 3A, the low levels of endogenous syt VII in control PC12 cells fell just below the limits of detection at the camera/microscope settings used. Under these conditions, strong syt VII staining was observed in syt VII⁺ clone G10, establishing that the antibody stains the upregulated protein. The same camera/microscope settings were used for the following syt VII localization experiments.

We then addressed the overall localization of syt VII (in clone G10) using higher-magnification images of cells in which the plasma membrane was demarcated by staining with an anti-syntaxin antibody (Fig. 3B). From this image, it is clear that the majority of syt VII is intracellular. To quantify these data, we carried out line scan (e.g., Fig. 3B) analysis of the syntaxin 1A and syt VII signals, and these data are plotted in Fig. 3C. Only 12% of the syt VII signal was colocalized with syntaxin 1A; hence, ≥88% of the syt VII signal is intracellular. We conclude that syt VII is not a plasma membrane protein (note that additional images, showing strong punctate staining of syt VII within cells are shown in Videos S1 and S2 in the supplemental material).

Next, we carried out quantitative immunolocalization experiments, again using syt VII⁺ clone G10, to determine which intracellular organelles harbor syt VII. As a negative control for our colocalization approach, we measured the overlap of the LDCV marker secretogranin II (Sg) (23), with the lysosomal marker Lamp1 (25). As expected, little overlap was observed (Fig. 4A, top row; Fig. 4C, ∼3 to 6% of the Lamp1 and Sg signals colocalized with each other). As a positive control, we costained for two well-documented LDCV markers, Sg and syt I (23, 29, 41), and extensive colocalization was observed (Fig. 4A, bottom row; Fig. 4C, ∼55 to 60% of the Sg and syt I signals colocalized with each other).

FIG. 4.

syt VII is localized to LDCVs and lysosomes in PC12 cells. (A) syt VII⁺ PC12 cells (clone G10) were stained for secretogranin II (abbreviated as Sg) and Lamp1 (first row) or for Sg and syt I (second row). Sg is an LDCV marker, and Lamp1 is a lysosomal marker; hence, the Sg-Lamp1 overlay provides a negative control (− control) for the colocalization analysis used in this study. Because Sg and syt I are well-documented LDCV markers, the Sg-syt I overlay provides a positive control for the colocalization analysis (+ control). In all of the immunofluorescence images shown, representative planes from three-dimensional reconstructions, and their overlays are provided. (B) syt VII⁺ PC12 cells (clone G10) were stained for syt VII and chromogranin B (abbreviated as Cg; first row). Cg and Sg are both LDCV markers (23). The monoclonal anti-Cg antibody was used for costaining with the polyclonal anti-syt VII antibody, while the polyclonal anti-Sg antibody was needed for costaining with the monoclonal anti-syt I antibody in panel A above. In the second and third row of images, syt VII was colabeled with syt I and Lamp1, respectively. (C) The overlays in panel A and B were quantified for the degree of colocalization as detailed in Materials and Methods. Two sets of data are given for each pair of proteins; labels indicate the fraction of the first protein that is colocalized with the second protein.

Having established the validity of our colocalization method, we examined the subcellular distribution of syt VII. We first double labeled syt VII⁺ cells (clone G10) with antibodies to syt VII and chromogranin B (Cg), which is another LDCV marker (23) (Fig. 4B, first row). Quantitative analysis revealed that ∼33% of the total syt VII signal was colocalized with Cg (Fig. 4C). In the reciprocal analysis, ∼49% of the Cg signal (i.e., the LDCV signal) was colocalized with syt VII (Fig. 4C). Thus, a significant fraction of LDCVs harbor upregulated syt VII. This finding was confirmed by colabeling syt VII and syt I (Fig. 4B, middle row); 45% of the syt VII colocalized with syt I, and 53% of syt I colocalized with syt VII (Fig. 4C). These data strongly indicate that LDCVs harbor both syt I and upregulated syt VII. These findings allow us to further address the impact of syt VII on the metal sensitivity of LDCV exocytosis, as detailed below.

To complete our syt VII localization experiments, we also determined whether syt VII is targeted to lysosomes by costaining with antibodies against the lysosomal marker Lamp1 (Fig. 4B, bottom row). Based on these experiments, 20% of the syt VII signal colocalized with Lamp1, and 35% of the Lamp1 signal colocalized with syt VII. Thus, a significant fraction of upregulated syt VII is targeted to lysosomes, a finding consistent with previous reports studying the localization of syt VII in nonneuronal cells (32).

Upregulation of syt VII increases the metal sensitivity of LDCV exocytosis.

Since syt VII is a high-affinity metal sensor (Fig. 1) that is targeted, at least in part, to LDCVs in PC12 cells (Fig. 4B and C), we sought to determine whether syt VII could increase the metal sensitivity of exocytosis. We addressed this question using RDE voltammetry (12). Cells were mechanically permeabilized so that the release machinery could be directly exposed to different concentrations of Ca²⁺. The amplitudes of the voltammetry signals (voltage) were determined at 80 s. Also, curves were fitted with single exponential functions to yield the time constants (τ) of release. Representative release profiles are shown for control (Fig. 5A) and syt VII⁺ clone A1 (Fig. 5B), where it is apparent that syt VII increased the Ca²⁺ sensitivity of release. To quantify this effect, normalized amplitudes (amp) and rate constants (1/τ) were plotted against [Ca²⁺]. As shown in Fig. 5C and D, both types of dose-response curves revealed that the [Ca²⁺]_1/2 for release from syt VII⁺ clone A1 was shifted to lower values compared to wild-type PC12 cells (summarized in Table 1; e.g., from the amplitude plot, the [Ca²⁺]_1/2 = 0.7 ± 0.1 μM for wild-type PC12 cells and [Ca²⁺]_1/2 = 0.4 ± 0.1 μM for clone A1 [P < 0.05, unpaired t test]). All three control PC12 cell lines (H1, B2, and E1 [data not shown]), as well as the parental cell line (Fig. 5C,D), yielded similar [Ca²⁺]_1/2 values that are consistent with data from a previous study (42). We also assayed for shifts in the [Ca²⁺]_1/2 in syt VII⁺ clones G10 and G3 and observed slight shifts to the left, but these changes were within the error of the assay (data not shown). These findings indicate that syt VII expression must reach 30% the levels of syt I to give rise to significant changes in the Ca²⁺ responsiveness of the secretory machinery.

We extended this analysis by carrying out RDE assays with Sr²⁺ and Ba²⁺ to trigger release. Although Sr²⁺ and Ba²⁺ can substitute to some extent for Ca²⁺, the abilities of these divalent cations to trigger release from intact cells differ (24). This may be due to a number of complicating factors, including differences in flux through voltage activated channels, differences in buffering, extrusion, and/or sequestration or differences in the relative abilities of Sr²⁺ and Ba²⁺ to activate Ca²⁺ sensors to trigger release (46, 47). These complicating factors were mitigated by the use of permeabilized cells, which bypass Ca²⁺ channels and make it possible to wash out mobile divalent cation buffers.

Again, three independent syt VII⁺ (A1, G10, and G3) and three independent control cell lines (H1, B2, and E1), as well as the wild-type parental cell line, were analyzed. Release profiles for control and syt VII⁺ PC12 cells are shown in Fig. 6A. Normalized amplitudes (amp) (Fig. 6B and C) and the rate constants (1/τ) (see Fig. S1A and B in the supplemental material) were plotted against [divalent cation]. The data were fitted with sigmoidal dose-response curves, and the EC₅₀ values and Hill coefficients were determined. Hill coefficients for all cell lines were similar (ca. 1.1 to 1.5). However, in both amplitude (Table 2) and 1/τ plots (see Table S1 in the supplemental material), syt VII⁺ cell lines exhibited significantly lower EC₅₀ values compared to wild-type PC12 cells. The tightly clustered [divalent cation]_1/2 values for the control PC12 cell lines and the parental PC12 cell line (Fig. 6B) indicate that there is little clonal variation, supporting the idea that the increase in syt VII expression mediates the shift in the Sr²⁺ and Ba²⁺ dose-response curves observed for all three syt VII⁺ clones.

FIG. 6.

Upregulation of syt VII in PC12 cells increases the Sr²⁺ and Ba²⁺ sensitivity of catecholamine release. Sr²⁺- and Ba²⁺-triggered catecholamine release from parental wild-type PC12 cells, control cell lines transfected with the empty vector, and syt VII⁺ cell lines was monitored by RDE voltammetry. (A) Representative release profiles for a control cell line that does not overexpress syt VII (left panel; clone B2), and a syt VII⁺ cell line (right panel; clone A1) at the indicated [Sr²⁺] are shown. The arrows indicate the addition of Sr²⁺ to the cracked cell suspension. (B) Normalized amplitudes of response (amp) of the release profiles for wild-type PC12 cells (PC12), as well as each control cell line (clones H1, B2, and E1), were plotted as a function of [Sr²⁺] (left panel) or [Ba²⁺] (right panel). The data were fitted with sigmoidal dose-response curves, yielding [Sr²⁺]_1/2 and [Ba²⁺]_1/2 values for release. (C) Release curves for syt VII⁺ cell lines (clones A1, G10, and G3) were analyzed as described in panel B. Plots obtained from wild-type PC12 cells are indicated by dotted curves. [Sr²⁺]_1/2 and [Ba²⁺]_1/2 values from the plots in panels B and C are listed in Table 2.

TABLE 2.

Sr²⁺ and Ba²⁺ sensitivity of catecholamine release from wild-type and stably transfected PC12 cell lines

Cell line	Mean concn (μM) ± SEMa
	[Sr2+]1/2	[Ba2+]1/2
PC12	46 ± 5	76 ± 15
H1 control	44 ± 8	84 ± 13
B2 control	43 ± 6	77 ± 12
E1 control	42 ± 5	85 ± 18
A1 syt VII+	14 ± 6**	33 ± 8*
G10 syt VII+	29 ± 4**	44 ± 12*
G3 syt VII+	28 ± 5**	45 ± 7*

^aAsterisks indicate significant differences between syt VII⁺ clones and the control. Double asterisks indicate P < 0.01 and single asterisks indicate P < 0.05 using the unpaired t test. [metal]_1/2 values were determined from amplitude-versus-[metal] plots.

Knockdown of syt VII reduces the metal sensitivity of LDCV exocytosis.

In the final series of experiments we decreased the level of syt VII expression in PC12 cells. If our model is correct, downregulation of syt VII should increase the fraction of exocytosis mediated by the low-affinity sensors, syt I and IX, thereby increasing the metal requirements for release.

To achieve this, we used an RNAi approach (11). We screened three short hairpin RNA (shRNA)-encoding constructs for their abilities to knock down the level of endogenous syt VII. The most effective shRNA is shown diagrammatically in Fig. 7A (left panel). After transient transfection of this construct, the level of syt VII expression was reduced by ∼60% to 70% (Fig. 7A, right panel). Knockdown was specific, since the shRNA had no effect on syt I or syt IX expression levels (Fig. 7A, right panel). Secretion from the knockdown cells (designated syt VII⁻) was analyzed as a function of [divalent cation] using RDE voltammetry. As a control, cells transfected with the empty pSHAG-1 vector were tested in parallel. As shown in Fig. 7B to D, EC₅₀ values were increased by 1.4- to 2.6-fold, depending on the divalent cation used, in syt VII⁻ cells but not in the control cells (in control cells, the EC₅₀ values were within error of these values obtained with wild-type untransfected cells [Tables 1 and 3]). These results provide further support for the idea that the metal sensitivity of secretion in PC12 cells is dictated by the relative abundance of syt isoforms expressed in these cells, in conjunction with the affinity of each isoform for Ca²⁺, Ba²⁺, or Sr²⁺.

DISCUSSION

There is limited information concerning the Ca²⁺ sensitivity of exocytosis from neurons, but current evidence indicates that this property can vary dramatically among different cell types. For example, in goldfish retinal bipolar neurons, reliable release does not occur at Ca²⁺ levels below 20 μM, and the rate of release was half-maximal at 194 μM Ca²⁺ (20). In contrast, in the rat calyx of Held, there was no clear threshold for release: reliable exocytosis occurred at 1 μM Ca²⁺ (4, 33). As a further example, astrocytes release glutamate at 84 to 140 nM Ca²⁺ (31). These widely ranging Ca²⁺ sensitivities imply the existence of Ca²⁺ sensors with distinct affinities for Ca²⁺. Members of the syt family of proteins are the leading candidates to function as Ca²⁺ sensors that regulate exocytosis. Thus, an emerging idea is that different isoforms of syt, with different affinities for Ca²⁺, underlie the distinct Ca²⁺-requirements exhibited by different cell types.

There are 16 isoforms of syt, but whether different isoforms have distinct apparent affinities for Ca²⁺ is unresolved (e.g., see references 26 and 27 versus reference 36). Using liposomes that contain PS as a binding partner, we demonstrate here that syt VII exhibits a higher apparent affinity for Ca²⁺, Ba²⁺, and Sr²⁺ than syt I (Fig. 1). We then upregulated syt VII in PC12 cells in order to test the idea that the ratio of high- and low-affinity isoforms of syt dictates the metal requirements for exocytosis. We chose PC12 cells because they express relatively high levels of the low-affinity isoforms syt I and IX (22) and express relatively low levels of the high-affinity sensor, syt VII (Fig. 2) (37).

The metal requirements of exocytosis from LDCVs were determined by using RDE voltammetry, which reports the release of catecholamines. syt VII can affect release only if it is targeted to membranes that participate in the LDCV fusion reaction. However, the localization of syt VII remained unclear, with different reports concluding that this isoform is targeted to lysosomes in nonneuronal cells (28, 32), to the plasma membrane (35) or to intracellular compartments—including the Golgi and LDCVs (14, 16)—in PC12 cells, or to secretory granules in pancreatic beta cells (17). Moreover, the idea that multiple syt isoforms are targeted to a common compartment such as a secretory vesicle (e.g., references 13, 15, and 30) has been questioned in a recent study focused on syt localization in Drosophila (1).

To address these issues, we carried out quantitative immunolocalization studies of untagged syt VII that was stably overexpressed in PC12 cells. One potential problem with previous localization experiments was the reliance on GFP fusion proteins that in many cases were transiently overexpressed in PC12 cells. The localization data reported here overcome the problems associated with tagged proteins, since we used untagged syt VII. In addition, our experiments made use of stable cell lines in which syt VII was unlikely to spill over into the wrong compartment because the upregulated syt VII levels (in the syt VII clone G10 used for immunolocalization) were only 18% the level of endogenous syt I. Our results revealed that a significant fraction of syt VII was colocalized with syt I on LDCVs (Fig. 4B and C), further establishing that a single type of secretory vesicle can harbor multiple isoforms of syt (13, 15, 30). Costaining with antibodies against syt VII and Lamp1 also suggested that a portion of the total pool of syt VII was targeted to lysosomes (Fig. 4B, bottom row; Fig. 4C) (28, 32). Some cortical syt VII immunoreactivity was observed, and this has been interpreted to reflect targeting to the plasma membrane (19, 35). However, only 12% of the total syt VII signal overlaps with syntaxin 1A, a bona fide plasma membrane protein (Fig. 3B and C). Whether this small fraction is indeed in the plasma membrane will require ultrastructural studies, but clearly this is a minor pool. We note that qualitative electron microscopy data indicate that endogenous syt VII is present in presynaptic nerve terminals (35). Immunogold labeling suggested that overexpressed syt VII fusion proteins were at least partially localized to synaptic vesicles (40). In fruit flies, a putative syt VII ortholog was not detected on synaptic vesicles (1).

Having established that syt VII is targeted, at least in part, to LDCVs, we determined the impact that this protein has on the metal requirements for release. The key finding in our study is that secretion from cells that overexpress syt VII is more sensitive to Ca²⁺, Ba²⁺, and Sr²⁺ than from control cells that express only low levels of syt VII (Fig. 5 and 6; see also Fig. S1 in the supplemental material). Thus, upregulation of the sole high-affinity isoform expressed in PC12 cells, syt VII (37, 48), decreases the metal requirements for LDCV exocytosis. We note that upregulation of syt VII resulted in a greater increase in the apparent sensitivity of the release machinery to Sr²⁺ and Ba²⁺ compared to Ca²⁺. Based on the liposome-binding studies in Fig. 1, one might have expected that upregulation of syt VII would have the strongest effect on Ca²⁺-stimulated release compared to Ba²⁺- and Sr²⁺-stimulated release. However, syt's are likely to regulate exocytosis via a series of effector interactions, in addition to binding PS (5), and these other interactions could be differentially regulated by Ba²⁺ and Sr²⁺ compared to Ca²⁺.

Our model predicts that downregulation of syt VII would increase the metal requirements for release, since less high-affinity sensor would be available to regulate exocytosis. Indeed, knockdown of syt VII expression shifted the Ca²⁺, Sr²⁺, and Ba²⁺ dose responses to the right (Fig. 7), presumably because a larger fraction of release was mediated by the low-affinity sensors, syt I and IX. These data further support the idea that the affinity of syt for metals, and the complement of syt's expressed in a given cell, determine the metal requirements for exocytosis. Consistent with this model, overexpression of syt IV, which is thought to bind Ca²⁺ weakly, increases the Ca²⁺ requirements for secretion from PC12 cells (42).

The C2A domain of syt VII is a potent inhibitor of exocytosis in PC12 cells (35, 37). This was interpreted as demonstrating that syt VII functions as a high-affinity Ca²⁺-sensor for exocytosis in this cell type (35). However, it is clear that the ability of a C2 domain to block secretion is unrelated to the expression profile of syt's in PC12 cells. For example, the C2A domain of syt III is a potent inhibitor of release but is not expressed in PC12 cells (37). C2 domains block release by binding to molecules that are essential for exocytosis but cannot provide information regarding syt expression patterns. In the case of syt C2 domains, inhibition appears to be mediated by binding to t-SNAREs and PIP₂ (37). The RNAi experiments shown in Fig. 7 provide the first evidence that endogenous syt VII functions as a high-affinity Ca²⁺ sensor in PC12 cells.

The data reported here begin to explain why there are so many isoforms of syt. Namely, the distinct Ca²⁺ requirements for release that have been observed in a number of different cell types might be mediated by different isoforms of syt. Within a given cell type, isoforms of syt might also be targeted to distinct compartments, endowing the fusion machinery in various compartments with different sensitivities to Ca²⁺. Finally, as shown here, multiple syt isoforms can be targeted to the same compartment, where the precise ratio of isoforms can fine-tune the metal requirements for exocytosis.