Creation of an Inositol 1,4,5-Trisphosphate-sensitive Ca2+ Store in Secretory Granules of Insulin-producing Cells*

A rapid rise in the cytosolic free Ca2+ concentration due to influx of extracellular Ca2+ or mobilization from intracellular stores is the primary trigger for exocytosis from secretory cells. Our understanding as to the pre- cise role of Ca2+ mobilization has been complicated by the presence of several types of Ca2+ stores in most cells. We now demonstrate that overexpression of the type 3 inositol 1,4,54risphosphate (IPS) receptor in insulin-se- creting PTC-3 cells results in the creation of a unique IPS-sensitive Ca2+ pool, restricted to the insulin secre- tory granules of these cells. The cellular actions of inositol 1,4,5-trisphosphate (IP,)' are mediated by specific receptors that function as ligand-acti-vated, Ca2+-selective channels (1). They have been identified in endoplasmic reticulum (21, nucleus (31, plasma membrane (4, 51, nerve terminals (61, and chromaffin granules (7). Molecular cloning studies have shown that the IP, receptors comprise a family of structurally related proteins. cDNAs encoding four different subtypes (IP3R-1 to -4) have been isolated, and each subtype has a distinct tissue distribution (1, 8-11).

A rapid rise in the cytosolic free Ca2+ concentration due to influx of extracellular Ca2+ or mobilization from intracellular stores is the primary trigger for exocytosis from secretory cells. Our understanding as to the precise role of Ca2+ mobilization has been complicated by the presence of several types of Ca2+ stores in most cells.

W e now demonstrate that overexpression of the type 3 inositol 1,4,54risphosphate (IPS) receptor in insulin-secreting PTC-3 cells results in the creation of a unique IPS-sensitive Ca2+ pool, restricted to the insulin secretory granules of these cells.
The cellular actions of inositol 1,4,5-trisphosphate (IP,)' are mediated by specific receptors that function as ligand-activated, Ca2+-selective channels (1). They have been identified in endoplasmic reticulum (21, nucleus (31, plasma membrane (4, 51, nerve terminals (61, and chromaffin granules (7). Molecular cloning studies have shown that the IP, receptors comprise a family of structurally related proteins. cDNAs encoding four different subtypes (IP3R-1 to -4) have been isolated, and each subtype has a distinct tissue distribution (1, [8][9][10][11]. We have previously shown that the IP, receptor subtype 3, IP3R-3, is expressed at high levels in rat insulinoma cells and is the predominant IP, receptor expressed in normal adult rat pancreatic islets (10). In this paper we describe studies on the contribution of IP3R-3 to the regulation of intracellular Ca2+ signaling. To do this, we transfected PTC-3 cells with a construct encoding IP3R-3 and generated clones expressing this protein.
* These studies were supported by the Howard Hughes Medical Institute, National Institutes of Health, Juvenile Diabetes Foundation International, and a long-term fellowship from the Human Frontiers Science Program (to S. J. G.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must U.S.C. Section 1734 solely to indicate this fact. therefore be hereby marked "aduertisement" in accordance with 18 8 To whom correspondence and reprint requests should be addressed:

EXPERIMENTAL PROCEDURES Dansfection of PTC-3 Cells
and Detection of IP3R-3 by Immunoblotting-The full-length cDNA sequence coding for rat IP3R-3 (nucleotides 1-8806) (10) was cloned into pCB6-Neo, which contains the human cytomegalovirus promotor and the neomycin resistance gene (a gift from D. Russell, M. Roth, and C. Brewer, University of Texas Southwestern Medical Center, Dallas, TX), to generate pCB6-IP3R3. PTC-3 cells were transfected by electroporation either with pCB6-IP3R3 (IP3R3-PTC cells) or with the pCB6-neo vector alone (C-PTC cells) as a control. Fifty pg of total membrane proteins were separated by 5% SDS-polyacrylamide gel electrophoresis and IP3R-3 detected by immunoblotting using an affinity-purified polyclonal rabbit antibody (IPR3AB3) (350 ng/ml) against the COOH-terminal 15 amino acids (residues 2656-2670, RQRLGFVDVQNCMSR) of the rat protein (10). The sequences of rat IP3R-1, -2, and -3 are unique in the region from which the peptide was selected for antibody production (10). Crude membranes were prepared from cultured cells, and Western blots were prepared and visualized by chemiluminescence techniques as described previously (10).
[Ca2+li was measured as described previously (12). Responses to bathapplied agonists were deemed acceptable for analysis according to the following criteria: 1) a stable [Ca2'l, for 30 s prior to drug application and 2) a rise in [Ca2+li of at least 15 n M following carbamyl choline (CCh) application, which returned to base line following washout. The peak Ca2+ is the maximal [Ca"], rise in the presence of CCh. The total Caz+ change was calculated by summing the differences between the [Ca"], and the mean base line at each time point during the response. Statistical significance was calculated using a two-tailed Student's t test.
Subcellular Localization of IP3R-3 by Immunogold Electron Microscopy-Cultured cells were fixed for 4 h at room temperature in Karnovsky's fixative containing 4% paraformaldehyde, 5% glutaraldehyde, 3.5 m M CaCl,, and 1% sorbitol, pH 7.4. Samples were then rinsed in 0.1 M cacodylate buffer, pH 7.4, for 12 h at 4 "C, dehydrated, and embedded in Lowicryl K4M (Chemische Werke Lowi, Waldkraiburg, Germany) at 65 "C according to the manufacturer's directions. Sections were placed on carbon-coated nickel grids and rinsed for 30 min in phosphate-buffered saline containing 3% bovine serum albumin. Grids were floated with the section side down for 2 h at room temperature on drops of affinity-purified IPR3AB3 antiserum (25 pg/ml) diluted in this buffer and then rinsed dropwise with the buffer. The grids were then treated in the same way with goat anti-rabbit IgG conjugated to 20-nm gold particles (Electron Microscopy Sciences, Fort Washington, PA) diluted 50-fold in the aforementioned buffer. After a water rinse, the grids were counterstained for 15 min with 3% uranyl acetate. The electron microscopy was done using a Siemens 101 electron microscope operated at 80 keV.
Materials-Cyclopiazonic acid and thapsigargin were obtained from RBI (Natick, MA). All other chemicals were of reagent grade.

RESULTS AND DISCUSSION
Immunoblotting of total membrane proteins from PTC-3 cells transfected with rat IP3R-3 (IP3R3-PTC) showed constitutive expression of the IP3R-3 protein (Fig. 1). The level of IP3R-3 expression in IP3R3-pTC cells at passage 2 post-transfection was comparable with that measured in RINm5F cells, which express high levels of this protein endogenously (10). The expression IP3R-3 in these cells was unstable and declined during repeated passage, despite maintaining the cells in neomycin, and IP3R-3 protein could not be detected at passage 5 ( Fig. 1

IP3R-3 in Secretory Granules
C-coupled muscarinic receptors (13) in IP3R3-PTC and control cells (C-PTC) at passages 2-3. There was no apparent difference in the base-line [Ca2+], between the two cell types, but the baseline did show considerable variation from cell to cell, for example some cells exhibited periodic, large oscillations in [Ca2'l, in control saline containing 2 mM Ca2+. Unless the baseline [Ca2+], was stable, the cells were not included in the analysis because it was not possible to separate the base-line oscillatory [Ca2'li responses from the effects of CCh. These oscillations were not more or less common in IP3R3-pTC cells.
Initial studies examined the changes in [Ca2+], in normal physiological saline and showed that IP3R3-pTC cells were more responsive to CCh than C-PTC cells (Fig. 2). The mean time to peak and the duration of the [Ca2'], signal following CCh stimulation were similar for IP3R-3-expressing and control cells. The differences between IP3R3-pTC and C-PTC cells were qualitatively similar when Ca2' was removed from the medium, consistent with the expressed IP3R-3 functioning as an intracellular IP,-activated Ca2+ channel. In both 0 and 2 mM extracellular Ca2+, a higher proportion of IP3R3-PTC cells responded to 10 p~ CCh, and the changes in [Ca2+li were larger in IP3R3-PTC than C-PTC cells (Fig. 2). As previously reported by Prentki et al. (14) for another insulin-secreting cell line (HIT), individual cells had distinct patterns of response to CCh. This diversity was observed in both IP3R3-PTC cells and C-PTC cells. Although the responses were larger and the responding cells more numerous in IP3R3-PTC clones there were no apparent differences in the pattern of [Ca2+l, changes between these cells and cells transfected with the vector alone. In the presence of thapsigargin, a selective inhibitor of the endoplasmic reticulum Ca2+-ATPase that depletes intracellular Ca2+ stores (151, the responses to CCh were almost completely abolished in C-PTC cells (Fig. 3,A andB). In contrast, when the same experiments were performed with IP3R3-PTC cells, nearly 40% of cells still exhibited a robust response to CCh in normal Ca2+-containing saline (Fig. 3, B and C). These responses were observed even in the absence of extracellular Ca2+, indicating that the response was due to release from intracellular stores. However, a second application of CCh in the presence of thapsigargin failed to produce further Ca2+ responses in IP3R3-PTC cells (Fig. 3 0 . Therefore thapsigargin was unable to "deplete" all CCh-mobilizable Ca2+ stores in IP3R3-PTC cells but did block refilling of these stores. It was possible that the absence of a response to CCh after thapsigargin treatment in C-PTC cells was due to the smaller effect of CCh in these cells. Fig. 3A shows that even when CCh elicits an abnormally large [Ca2+], change, thapsigargin prevents subsequent [Ca2+], responses to CCh. The effect of cyclopiazonic acid, a Ca2+-ATPase inhibitor that may be less selective than thapsigargin (161, was essentially the same (Fig. 30). We also tested whether the increased sensitivity to CCh in the IP3R3-PTC cells was entirely due to a Ca2+ store, which was not depleted by thapsigargin by calculating the total [Ca2+Ii change in response to thapsigargin. Curiously, IP3R3-PTC cells had significantly lower thapsigargin responses than C-PTC cells (IP3R3-PTC cells: peak [Ca2+], rise = 279 26 nM, total Ca2' change = 15,682 Ca2+ change = 33,292 2 156, n = 59). The above data show that the IPJR-3-regulated Ca2+ store is distinct from the major endogenous IP,-sensitive store and unusual in that it is intrinsically less leaky to Ca2+ because blockade of Ca2'-ATPases does not deplete the store but does prevent refilling (see Fig. 3C).
These results suggest that IP3R-3-sensitive Ca2+ stores may have a unique localization within IP3R3-PTC cells. In order to address this issue, we used Immunogold electron microscopy and an IPSR-3-specific antibody to examine the distribution of IP3R-3 in IP3R3-PTC cells and C-PTC cells (Fig. 4). There was Immunogold labeling of secretory granules in both IP3R3-PTC cells and C-PTC cells (Fig. 4) with the levels being significantly greater in the former. Over 80% of secretory granules were labeled in IP3R3-PTC cells (80.8% 2 3, n = 4 fields) whereas less than 50% of granules were labeled in the C-PTC cells (48. Response of cells to CCh after application of inhibitors of intracellular Ca2+-ATPases to deplete the intracellular Ca2+ stores.A, application of 500 nhl thapsigargin was sufficient to prevent a response to CCh in C-PTC cells. B, a response to CCh in the presence of thapsigargin was only observed in one C-PTC cell whereas 40% of IPBRB-PTC cells responded. Only cells that clearly responded to CCh immediately before thapsigargin application were included for analysis. Data are the means (+ S.E.) of the percent responding cells for each experiment; cell numbers are also given (*, indicatesp < 0.05 by unpaired t test). C, in IP3R3-pTC cells, responses were still observed in the presence of thapsigargin in Ca2+-free medium, but further applications of CCh had no effect (typical of 10 cells from three clones). D, responses to CCh were also observed in IP3R3-pTC cells following the application of cyclopiazonic acid, another Ca2+-ATPase inhibitor (typical of 17 cells from 3 clones). CCh did not elicit a response in C-PTC cells under the same conditions. Breaks in the trace indicate pauses in data collection when the cells were not illuminated. There was no labeling of the endoplasmic reticulum, mitochondria, or other structures in cells overexpressing IP3R-3. The Immunogold labeling was primarily localized to the periphery of the core or even the core itself. We believe that this may be a consequence of the fact that the granule membrane is poorly preserved in these preparations and collapses onto the core during fixation and embedding of the cells. These results are consistent with our previous observations in rat pancreatic islets showing that IP3R-3 is specifically localized to the secretory granules of insulin-secreting P-cells and somatostatin-secreting 3-cells (17).

IP3R-3 in
The selective localization of IP3R-3 receptors in secretory granules suggests a role for this Ca2+ store in the control of B "I Namh 5 7 insulin secretion. Our imaging data suggest that the secretory granule-associated IP3R-3 is functional and regulates the release of Ca2+ from this store. Insulin secretory granules are known to sequester Ca2' following glucose stimulation in isolated pancreatic islets (18), a mechanism that is considered a key event in stimulus-secretion coupling (19). The pumping of Ca2+ in secretory granules is believed by some authors t o occur through a Ca2+-ATPase (see Ref. 19). Our imaging data suggest that the secretory granule-associated IP3R-3 is functional and can regulate the release of Ca2+ from this store. Our results indicate that once CCh has been applied in the presence of thapsigargin and has led to the release of the remaining Ca2+ from granule stores, CCh is no longer effective. This observation is consistent with the existence of a thapsigargin-sensitive Ca2+-ATPase in the granule membrane whose activity is needed for a rapid refilling of the granule Ca2' store. The fact that secretory granules are able to retain their Ca2+ during continuous application of Ca2+-ATPase inhibitors confirms earlier reports showing that the passive permeability of the insulin secretory granule membrane to Ca2+ is much lower than the membrane of other Ca2+-storing organelles (20). Although IP3R-3 has been reported in rat pancreatic p-cells (17), it is unlikely that IP3R-3 is sine qua non for insulin release as regulated secretion does occur from normal PTC-3 cells (211, which have only low levels of this receptor. Nonetheless, it may be an important element in the modulation of insulin secretion by agents that act as agonists at phospholipase C-linked receptors (22, 23). Synthesis of IP, in response to such stimuli may produce local Ca2' release from granules that could potentiate either movement of granules through the cytoskeleton, fusion of granules with the plasma membrane, or both (23). Furthermore, changes in the levels of IP3R-3 may be an important element in the control of insulin release in both normal and pathophysiological circumstances.