(Received for publication, November 25, 1996, and in revised form, January 28, 1997)
From the Departments of Physiology and ¶ Anatomy
and Cell Biology, University of Michigan Medical School, Ann Arbor,
Michigan 48109 and the
Department of Pharmacology, State
University of New York Health Science Center at Syracuse, Syracuse, New
York 13210
A key event leading to exocytosis of pancreatic acinar cell zymogen granules is the inositol 1,4,5-trisphosphate (InsP3)-mediated release of Ca2+ from intracellular stores. Studies using digital imaging microscopy and laser-scanning confocal microscopy have indicated that the initial release of Ca2+ is localized to the apical region of the acinar cell, an area of the cell dominated by secretory granules. Moreover, a recent study has shown that InsP3 is capable of releasing Ca2+ from a preparation enriched in secretory granules (Gerasimenko, O., Gerasimenko, J., Belan, P., and Petersen, O. H., (1996) Cell 84, 473-480). In the present study, we have investigated the possibility that zymogen granules express InsP3 receptors and are thus Ca2+ release sites. Immunofluorescence staining, obtained with antisera specific to types I, II, or III InsP3 receptors and analyzed by confocal fluorescence microscopy revealed that all InsP3 receptor types were present in acinar cells. The type II receptor localized exclusively to an area close to or at the luminal plasma membrane. While types I and III InsP3 receptors displayed a similar luminal distribution, these receptors were also present at low levels in nuclei. The localization of InsP3 receptor was in marked contrast to the distribution of amylase, a zymogen granule content protein. In a zymogen granule fraction prepared in an identical manner to the aforementioned report demonstrating InsP3-induced Ca2+ release, immunoblotting demonstrated the presence of types I, II, and III InsP3 receptors. Ca2+ release from this preparation in response to InsP3, but not thapsigargin, could also be demonstrated. In contrast, when the zymogen granules were further purified on a Percoll gradient, InsP3 receptors were undetectable, and InsP3 failed to release Ca2+. Transmission electron microscopy performed on both preparations showed that the Percoll-purified granule preparation consisted of essentially pure zymogen granules, whereas the granules prepared without this step were enriched in granules but also contained significant contamination by mitochondria, endoplasmic reticulum, and nuclei. It is concluded that zymogen granules do not express InsP3 receptors and thus are not a site of Ca2+ release relevant to the secretory process in the pancreatic acinar cell.
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.
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.
ChemicalsFura-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.
ImmunocytochemistryPancreatic 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 AdobeTM
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.
Calcium Release MeasurementsCa2+ 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.
ImmunoblottingWhole 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).
Electron MicroscopyImmediately 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 AdobeTM Photoshop 3.0 software.
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).
Are InsP3 Receptors Expressed on Zymogen Granules?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.
Ca2+ Release Measurements in Granule PreparationsThe 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.
Granules Prepared by Simple Centrifugation Are Contaminated with Subcellular OrganellesLittle 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.
The authors thank John Williams for helpful discussions and Su Ge Luo, Noel Wys, and Christopher Baker for excellent technical assistance throughout the project.