INTRODUCTION

The pancreatic acinar cell is a polarized epithelial cell whose major function is to synthesize digestive enzymes, package them into secretory granules, and then release the granule contents by regulated exocytosis (1). A key event underlying secretagogue-stimulated digestive enzyme secretion is the inositol 1,4,5-trisphosphate (InsP₃)-mediated¹ release of Ca²⁺ from intracellular storage sites. Recently, utilizing digital imaging and confocal microscopy, it has become established that stimulation with the gut hormone cholecystokinin or the neurotransmitter acetylcholine results in a distinct spatial pattern of Ca²⁺ release, such that the release of Ca²⁺ is paradoxically initially observed at the luminal pole of the acinus and proceeds in a "wave" toward the basolateral (presumably receptor bearing) pole of the cell (2-4). At physiological concentrations of secretagogue, it has been shown that Ca²⁺ release can actually be confined to this luminal region without spreading throughout the whole cell (5). In addition to the overwhelming evidence for a role of InsP₃ to mediate release of Ca²⁺, studies have also indicated that release can be initiated through a process of Ca²⁺-induced Ca²⁺ release (3, 5, 8). The expression of ryanodine receptors has, however, not been physically demonstrated in the pancreatic acinar cell. It has been proposed that this pattern of Ca²⁺ release has significance for secretion since zymogen granules are confined to the luminal pole of the acinar cell and thus would provide a mechanism that increases [Ca²⁺]_i directly at the site of its projected action.

Although original cell fractionation studies reported the InsP₃-induced Ca²⁺ release site to be located in the endoplasmic reticulum (6, 7), a recent report has suggested that zymogen granules themselves fulfill this role (8). This latter conclusion was primarily based on the observation that InsP₃ was capable of releasing Ca²⁺ from a preparation enriched in granules (8). This proposal, however, remains controversial since the preparation was not extensively characterized, thus raising the possibility that it contained subcellular contaminants (8). In addition, immunolocalization of type III InsP₃ receptors in pancreatic acinar cells indicates that although receptors were indeed expressed in the luminal region of the acinar cell, these receptors do not appear to co-localize with a known zymogen granule protein (secretory carrier membrane protein, SCAMP) (9). In contrast to SCAMP, the type III receptor appeared to be distributed on structures very close to the luminal membrane. These data do not, however, exclude the possibility that an InsP₃ receptor other than the type III subtype is present on zymogen granules. Recently, the expression of InsP₃ receptors on granules from other secretory cells has also been questioned since it appears that an earlier report of type III InsP₃ receptors on secretory granules isolated from pancreatic  cells (10) can now be explained by nonspecific binding of the antibody to insulin (11).

In the present study, we have investigated the possibility that InsP₃ receptors are expressed on zymogen granules using two preparations, the first prepared by simple centrifugation and identical to that used to previously demonstrate InsP₃-induced release (8) and the second involving purification on a Percoll gradient by a protocol that has been used extensively to study integral zymogen granule proteins (12-18). In addition, using antisera specific to types I, II, and III receptors, the distribution of InsP₃ receptors in acinar cells has been defined.

MATERIALS AND METHODS

APCT-1, -2, and -3 were raised in rabbits against peptides corresponding to the C termini of rat types I, II, and III InsP₃ receptors, respectively, and affinity purified (19). This polypeptide region is highly divergent between the different InsP₃ receptors, and these antibodies have been previously shown to recognize and discriminate between types I, II, and III receptors (19) from a broad range of mammalian species. In some experiments, an additional type III receptor-specific monoclonal antibody directed against amino acids 22-230 of the human type III receptor was utilized (TL-3; a gift from Transduction Laboratories). A polyclonal antibody (amy-1), directed against the zymogen granule content protein, amylase, was purchased from Sigma.

Fura-2/free acid and lissamine rhodamine (sulforhodamine B) were purchased from Molecular Probes (Eugene, OR); bovine serum albumin (fraction V) from ICN Immunobiologicals (Lisle, IL); ECL detection system from Amersham Corp. (Arlington Heights, IL); inositol 1,4,5-trisphosphate, and all other materials were obtained from Sigma or Bio-Rad.

Pancreatic lobules were prepared and frozen in Tissue-Tek embedding medium (Miles, Elkhart, IN) with isopentane cooled in liquid nitrogen. Cryostat sections (5-10 µm thick) were placed on gelatin-coated slides, air dried, and then fixed in methanol at 20 °C for 10 min prior to immunofluorescence staining. Procedures for immunofluorescence localization of InsP₃ receptor isoforms followed methods previously described in detail (15, 20). Immunofluorescence staining of amylase was performed in an identical manner to that previously published (15). Specificity of immunofluorescence was determined by preincubating primary antisera at 4 °C overnight with 10 µg/ml respective peptide antigen before immunofluorescence staining (20). Immunolocalizations were analyzed by conventional epifluorescence and laser scanning confocal fluorescence microscopy using a Bio-Rad MRC 600 system. Digitized images were processed using Adobe^TM Photoshop 3.0 software.

Secretory granules were prepared from rat pancreas by one of two procedures, either a protocol essentially identical to Gerasimenko et al. (8) or by further purifying this preparation by centrifugation on a Percoll gradient (15). Briefly, pancreata were excised from male rats and rinsed in an ice-cold buffer containing 10 mM MOPS-Tris (pH 6.8), 0.1 mM MgSO₄, 3 mM ATP, 250 mM sucrose, 0.5 mg/ml soybean trypsin inhibitor, 1 mg/ml bovine serum albumin. The pancreas was finely minced and homogenized in the same buffer using a motorized Teflon/glass homogenizer. The homogenate was initially centrifuged at 1000 × g for 10 min. Following this spin, the supernatant was removed and further centrifuged at 2500 × g for 10 min. This spin resulted in a pellet consisting of a white core with a brown margin. Only the white portion of the pellet was utilized in experiments. For further purification, the white pellet was mixed with 50% Percoll containing 250 mM sucrose, 50 mM MES, 2 mM EGTA, 0.1 mM phenylmethylsulfonyl fluoride. After centrifugation at 25,000 × g for 90 min, a white band toward the bottom of the tube was recovered and washed three times in the same buffer without Percoll.

Ca²⁺ release was measured from each preparation by procedures similar to those used previously for the measurement of Ca²⁺ release from permeabilized cells (21, 22). The granule pellet was resuspended at a protein concentration of 0.25-1 mg/ml in release media (135 mM KCl, 10 mM MOPS-Tris, 1 mM MgSO₄, 3 mM ATP, 25 mM creatinine phosphate, 25 units/ml creatinine phosphokinase, 1 mg/ml bovine serum albumin, 0.5 mg/ml soybean trypsin inhibitor, and 100 µM Fura-2-free acid), which had been treated with Chelex-100 resin (Bio-Rad) to remove excess divalent cations. 25-µl aliquots of granules were placed in a small volume chamber mounted on the stage of an ATTOFLOR digital imaging system. Additions to the chamber were made either through a micropipette connected to a pressure ejector positioned in the chamber or directly into the chamber with a pipette. InsP₃ and all test agents were reconstituted into an identical buffer but also containing 25 µM sulforhodamine (excitation 560 nm, emission 630 nm) to confirm the pressure injection and to estimate by dye dilution the amount of agent applied to the bath. Ca²⁺ measurements were performed as described previously (21, 22) with increases in Ca²⁺ described by an increase in 340/380 nm emission ratio.

Whole cell lysates from pancreas, preparations of pancreatic acinar cells, together with the two preparations of granules were subjected to electrophoresis and subsequently transferred to nitrocellulose. The blots were incubated with either APCT-1, -2, -3, or TL-3 for the specific detection of types I, II, and III InsP₃ receptors, respectively. Immunoreactivity was visualized using peroxidase-conjugated secondary antibodies followed by detection using the ECL system. In preliminary experiments, the dilution of antibody was adjusted so that equal amounts of purified types I, II, or III receptor standards produced bands of approximately equal intensity as described previously (19).

Immediately after their preparation, partially purified and purified zymogen granules were fixed overnight at 4 °C in a mixture of 2.5% glutaraldehyde and 2% paraformaldehyde in 0.13 M sodium cacodylate buffer (pH 7.4). Subsequently, granule fractions were rinsed in cacodylate buffer, post-fixed for 1 h with 1% OsO₄ in 0.13 M cacodylate buffer (pH 7.4), dehydrated in ethanol, and embedded in Spurr's embedding resin (Electron Microscopy Sciences, Fort Washington, PA). Ultrathin sections (60-80 nm thick) were prepared with a Reichert-Jung Ultracut E ultramicrotome, stained with uranyl acetate and lead citrate, and viewed and photographed with a Philips CM100 electron microscope. Digitized images were processed using Adobe^TM Photoshop 3.0 software.

RESULTS AND DISCUSSION

The localization of type III receptors in the secretory "trigger zone" of acinar cells has been reported (9); however, the expression and possible localization of types I and II receptors in acinar cells are unknown. Thus the presence and subcellular distribution of each of the InsP₃ receptors were determined in cryostat sections of fresh frozen pancreatic lobules (23) fixed in methanol followed by detection using an fluorescein isothiocyanate-labeled secondary antibody and laser scanning confocal microscopy. Immunofluorescence staining was evident in acinar cells with the antibodies APCT-1, -2, and -3 or TL-3 designed to recognize types I, II, and III InsP₃ receptor subtypes, respectively. These data indicate that all three InsP₃ receptor subtypes are expressed in this tissue (Fig. 1, A-C, for polyclonal antibodies). Each of the receptors was localized most strikingly to regions close to or at the luminal membrane in the apical pole of the acinar cell (arrows in Fig. 1, A-C). This is a similar localization to that reported previously for the type III receptor (9). While type II receptors were exclusively distributed to the luminal region in the acinar cell, types I and III InsP₃ receptors appeared to be weakly expressed in nuclei (arrowheads in Fig. 1). For comparision, sections were stained with amy-1, a polyclonal antisera directed against the secretory protein amylase, known to specifically stain zymogen granules in pancreatic acinar cells. This antisera co-localizes with other zymogen granule proteins such as Rab3D and GP-2 (15). Figs. 1D and 2C (at high power) illustrate the clear difference between anti-InsP₃ receptor antibody staining associated with luminal membrane profiles and zymogen granule staining evident with the amylase antisera.

The fluorescence associated with the luminal distribution of type I InsP₃ receptor appeared subtly different from that associated with types II and III receptors. Whereas the types II and III receptor localizations in close proximity to the luminal membrane often appeared punctate (Fig. 1, B and C, and Fig. 2B for high magnification), type I receptor immunofluorescence appeared to be tightly and continuously localized to luminal membrane profiles (compare Figs. 1A, B, and C and see Fig. 2A for high magnification). For comparison, the granule localization of amylase is shown in Fig. 2C. The resolution obtainable with immunofluorescence does not allow a definitive identification of cellular structures associated with fluorescence staining. The staining pattern observed for types I and III receptors may simply reflect "clustering" of receptor expression at discrete sites at the luminal membrane. Alternatively, the punctate nature of the staining, consistently observed with types II and III antibodies is suggestive of a sub-luminal vesicular distribution. In contrast, the continuous staining of type I receptor is more consistent with localization of this receptor to the luminal plasma membrane (compare Fig. 2, A and B).

Staining with each polyclonal antisera was completely competed by overnight incubation of antibody at 4 °C with 10 µg/ml of the respective peptide used to raise the antibody (Fig. 3A, shown for APCT-2 incubated with APCT-2 peptide; Fig. 3B shows the corresponding Nomarski image for this field). In preliminary experiments, using high concentrations of APCT-3, diffuse staining was sometimes evident in the granular region (24). This staining is in all probability nonspecific since low level staining of the granular area was also observed after competition with APCT-3 peptide. Furthermore, monoclonal antibody TL-3, directed against the type III receptor also localized to the luminal membrane area and did not, even at high concentrations, localize to zymogen granules (Fig. 2B).

Although InsP₃ receptors were found not to be obviously associated with secretory granules in the acinar cells, the luminal distribution of InsP₃ receptor subtypes appears entirely consistent with the spatial pattern of Ca²⁺ signaling observed in these cells (2-4). That the Ca²⁺ signal is initially observed in the secretory pole of the cell and can be confined to this region even though the production of InsP₃ presumably occurs in a distal region may reflect the abundance of InsP₃ receptors in this area. The apparent presence of InsP₃ receptors in pancreatic acinar nuclei is also consistent with reports of Ca²⁺ signaling events in other non-excitable cell-types (25).

Since the above morphological data indicate that all InsP₃ receptors are expressed in a region of the acinar cell that is intimately involved in the secretory process, the possibility that zymogen granules themselves are the site of this expression was further investigated. Immunoblots were performed on granule preparations together with lysates generated from whole pancreas and pancreatic acinar cells. Consistent with the immunofluorescence data, each of the InsP₃ receptor antibodies recognized proteins of molecular weight of ~260 kDa in whole pancreatic lysate together with an acinar cell lysate (Fig. 4). In addition, types I, II, and III InsP₃ receptors could also be detected in granule fractions prepared by simple centrifugation in an identical manner to the study demonstrating release of Ca²⁺ from zymogen granules (8). The association of InsP₃ receptors with zymogen granules is not consistent with either immunofluorescence data reported in this study or data reported by Nathanson et al. (9), who demonstrated that type III receptor immunofluorescence did not co-localize with a known secretory granule protein.

To determine whether the InsP₃ receptors detected in this granule preparation reflected the presence of receptors on zymogen granules or was due to the presence of contaminants, zymogen granule fractions were further purified by centrifugation on a Percoll gradient. This procedure generates several distinct fractions, one of which has been shown to be an essentially pure zymogen granule preparation (12, 13). In contrast to granules prepared by simple centrifugation, InsP₃ receptor immunoreactivity could not be detected in the purified granule preparation (Fig. 4). Furthermore, when the pure granules were lysed by treatment with nigericin resulting in a preparation of granule membranes, a procedure which has been shown to enrich granule membrane proteins by up to 50-fold (15, 16), still no signal was observed with the InsP₃-specific antibodies (data not shown).

A possibility exists that InsP₃ receptor antigenicity was in some way altered by purification of granules on a Percoll gradient. This appears somewhat remote based on the widespread use of Percoll to prepare functionally intact cells or subcellular fractions (26-28). Indeed, this preparation of granules has been used in a variety of studies demonstrating the presence of integral zymogen granule membrane proteins by Western blotting techniques (14, 15), the existence of kinase activity on the granules (13), and for the demonstration of a granule ionic conductance (17, 18). Taken together, immunofluorescence and immunoblotting data indicate that InsP₃ receptors are unlikely to be present on zymogen granule membranes prepared from rat pancreatic acinar cells.

The above data do not exclude the possibility that zymogen granules express a novel InsP₃ receptor subtype not recognized by the battery of antisera used or that InsP₃ receptors are present on granule membranes in insufficient quantity to be resolved by immunofluorescence and immunoblotting. Experiments were therefore performed to determine if zymogen granules release Ca²⁺ in response to InsP₃. Granules prepared by simple centrifugation (8) consistently released Ca²⁺ in response to InsP₃ as shown by an increase of Ca²⁺ in the medium bathing the preparation (Fig. 5A) (12 experiments, 3 preparations). Interestingly, the release of Ca²⁺ in response to InsP₃ was always sustained, an observation which may be related to the scarcity of InsP₃ metabolizing enzymes or relatively poor Ca²⁺ reuptake back into the releasable pool. Addition of an equal volume of the solution used to dissolve the InsP₃ never increased Ca²⁺, indicating that the increase was not the result of contaminating Ca²⁺ in this solution. Addition of 0.1 µM thapsigargin to this preparation did not result in any release of Ca²⁺ (Fig. 5A; 5 experiments, 3 preparations). However, addition of 0.1 µM thapsigargin to a pancreatic homogenate or a suspension of the 1000 × g pellet discarded in the initial preparation of granules resulted in an increase in Ca²⁺ released in the media, indicating that thapsigargin is capable of releasing Ca²⁺ from these subcellular organelle preparations (1000 × g pellet is shown in Fig. 5C).

These data are consistent with the report by Gerasimenko et al. (8), who showed InsP₃-induced release of Ca²⁺ from both a suspension of organelles and from what was reported to be a single isolated granule. In contrast, when InsP₃ was added to preparations of granules that had been further purified by Percoll centrifugation, no release of Ca²⁺ was observed (Fig. 5B) (8 experiments, 3 preparations). Addition of ionomycin or nigericin resulted in an increase in medium Ca²⁺, indicating that the granules do indeed store Ca²⁺. Addition of thapsigargin did not release Ca²⁺ from this preparation (data not shown). To demonstrate that Percoll itself does not interfere with the binding of InsP₃ to its receptors and the subsequent release process, the granules prepared by simple centrifugation were incubated for 90 min in the release buffer containing 50% Percoll followed by the measurement of Ca²⁺ release after addition of InsP₃. Ca²⁺ release could always be elicited by InsP₃ in these preparations (Fig. 5D) (4 experiments), although the Ca²⁺ release in Percoll-containing samples when compared with paired non-Percoll-treated controls was slower to peak, and the magnitude of the release was reduced by 41 ± 5% (Fig. 5D; compare with 5A). These data show that Percoll, even when included in the Ca²⁺ mobilization buffer, does not interfere with the Ca²⁺ release process to an extent that InsP₃-induced release becomes refractory. In total, these data suggest that the source of InsP₃ releasable Ca²⁺ in granule fractions prepared by simple centrifugation is apparently lost when the granules are further purified by Percoll gradient centrifugation.

Little obvious differences between the two preparations could be seen viewing the preparations at the light-level with phase contrast microscopy. Both preparations appeared to consist of small vesicular structures less than 1 µm in diameter. Therefore, to gain insight into the differences between granules prepared by both methods, the morphology of the preparations was examined using transmission electron microscopy. As described previously, the preparation produced by purification on a Percoll gradient is dominated, almost to the exclusion of any contamination by zymogen granules (Fig. 6A, and see Refs. 11 and 12). The only non-granular organelle (estimated at less than 1% of the preparation) contamination consisted of mitochondria. In stark contrast, although the preparation produced by simple centrifugation (8) was enriched in zymogen granules, significant contamination was observed in each ultrathin section (Fig. 6B). In addition to zymogen granules, mitochondria, rough endoplasmic reticulum, together with nuclei were readily recognizable in each section of granules prepared in this fashion.

Contamination of the granule preparation prepared by simple centrifugation with InsP₃ receptor-expressing organelles apparently underlies the release of Ca²⁺ from this preparation. However, the nature of the contaminant is presently unclear. Since Ca²⁺ release from either preparation could not be initiated by treatment with thapsigargin, the zymogen granules, and the site of InsP₃ releasable Ca²⁺ isolated with the partially purified granules do not accumulate Ca²⁺ via a SERCA type Ca²⁺-ATPase. In addition, it follows that a Ca²⁺ pool, which is undoubtedly present in pancreatic acinar cells that is thapsigargin releasable (29), has been removed from this preparation.

In conclusion, immunocytochemical data regarding InsP₃ receptor distribution presented in this report is consistent with the view that the site of the physiologically important InsP₃-releasable Ca²⁺ pool is localized within the secretory "trigger zone" of the pancreatic acinar cell. This putative Ca²⁺ release/storage site appears not to be associated with zymogen granules but in all probability is situated in close association with the apical membrane.