(Received for publication, December 23 1996, and in revised form, April 18, 1997)
From the Departments of Physiology, and
§ Anatomy and Cell Biology, University of Michigan Medical
School, Ann Arbor, Michigan 48109
The heterotrimeric G-protein
Gq/11 was identified on pancreatic acinar zymogen
granules and its function in calcium-regulated exocytosis was examined.
Western blotting showed q/11, but not
s
or
o, to be localized to the zymogen granule membrane
along with G-protein
-subunit; all three
subunits were present
in a plasma membrane fraction and the
q/11 signal was
30-fold more enriched in the plasma membrane as compared with granule
membrane. Neither CCK receptors nor
subunits of the sodium pump,
both plasma membrane markers were present on granule membranes.
Immunohistochemistry of pancreatic lobules showed that
q/11 localized to the zymogen granule-rich apical region
of acinar cells together with a much stronger signal at the basolateral
plasma membrane. When the substance-P-related peptide GPAnt-2a, an
antagonist of Gq/11, was introduced into streptolysin-O
permeabilized acini to bypass the plasma membrane, the amylase release
induced by 10 µM free calcium was potentiated in a
concentration-dependent manner. By contrast, another
substance-P-related peptide, GPAnt-1, an antagonist of Go
and Gi, showed no effect on calcium-induced amylase release
from permeabilized acini. GPAnt-2a peptide also exerted an inhibitory
effect on the total GTPase activity of the purified zymogen granules
and a larger inhibitory effect on the GTPase activity of the
Gq/11 protein immunopurified from zymogen granules.
GPAnt-1, however, did not inhibit GTPase activity of either zymogen
granules or immunopurified Gq/11. These results suggest
that GPAnt-2a peptide augmented calcium-induced amylase release from
permeabilized acini by inhibiting GTPase activity of the
Gq/11 protein on zymogen granules. We conclude that
Gq/11 protein on zymogen granules plays a tonic inhibitory role in calcium-regulated amylase secretion from pancreatic acini.
Regulated exocytosis involves the highly controlled targeting,
docking, and fusion of secretory vesicles to the plasma membrane. Studies using cell permeabilization and non-hydrolyzable GTP analogues such as GTPS1 have indicated that
G-proteins play key roles in regulated exocytosis (1-3). Rab proteins,
members of the Ras-related small G-protein family, are well known to
participate in various steps of intracellular vesicle trafficking
including exocytosis (4-7). Isoforms of Rab3 have been localized to
secretory granules of neuronal and non-neuronal cells, although their
role in exocytosis is still not clear (7-9). Recently, we and others
have reported that Rab3D (10) is localized to secretory granules of
various tissues including exocrine pancreas and its localization
implies that it may be involved in regulated exocytosis (11-13). In
addition to small G-proteins, recent evidence suggest that
heterotrimeric G-proteins or their isolated
-subunits also play
important roles in intracellular vesicle trafficking and vesicle
formation in addition to their classical functions in receptor-coupled
signal transduction. For example, a heterotrimeric G-protein(s) was
shown to be required in endosome fusion in a cell-free system (14). In
LLC-PK1 epithelial cells, when Gi3 was overexpressed on
Golgi membranes, constitutive secretion was retarded (15). By using an
ADP-ribosylation method, Gs as well as Gi/o on
Golgi membranes were also demonstrated to participate in the regulation
of vesicle formation in PC12 cells (16). In addition, evidence suggests
heterotrimeric G-proteins may have a role in regulated exocytosis since
Go on secretory granules was recently shown to play an
inhibitory role in calcium-stimulated norepinephrine release from
chromaffin cells (17, 18). In insulin secreting B-cells, Gi
on secretory vesicles has been demonstrated to be involved in
mastoparan-induced insulin secretion (19). Although Gq/11
has recently been localized to Golgi membrane (20), its function in
vesicle trafficking is still unknown.
The exocrine pancreas is a model system widely used for studying
regulated exocytosis. Several small GTP-binding proteins have been
shown to be present on zymogen granules, and two have recently been
identified as Rab3D and Rab5 (12, 13, 21). However, little is known
about the presence or function of heterotrimeric G-proteins on zymogen
granules. Although the existence of a pertussis toxin-sensitive
G-protein on pancreatic zymogen granules was previously suggested (22)
its identity is still uncertain. In the current work, we have
identified a heterotrimeric G-protein(s), which localizes to zymogen
granules and appears to participate in regulated exocytosis. We found
that both q/11 and G-protein
-subunit exist on
zymogen granules. We also examined the effect of substance-P-related peptides on calcium-induced amylase release from streptolysin-O permeabilized acini. Substance-P-related peptide GPAnt-2a and GPAnt-1
were reported to antagonize Gq/11 and Go/i,
respectively (23) and thus these peptides are widely used for analyzing
the function of heterotrimeric G-proteins on both plasma membranes and
secretory vesicles (17, 18, 24). We found that GPAnt-2a, but not
GPAnt-1, enhanced calcium-stimulated amylase secretion from
streptolysin-O permeabilized pancreatic acini and that GPAnt-2a, but
not GPAnt-1, inhibited GTPase activity on zymogen granules. These
results suggest that Gq/11 on zymogen granules plays a
tonic inhibitory role in calcium-regulated amylase release from
pancreatic acini.
Streptolysin-O was purchased from Welcome
Diagnostic (Greenville, NC), chromatographically purified collagenase
from Worthington Biochemicals, bovine serum albumin (Fraction V) from
ICN Immunobiologicals (Lisle, IL), and protein A-agarose from Pierce.
The following compounds were purchased from Sigma: Mg-ATP, creatine
kinase, creatine phosphate, GTP, and activated charcoal.
[-32P]GTP (30 Ci/mmol) and
125I-Bolton-Hunter CCK-8 (2200 Ci/mmol) were purchased from
DuPont NEN (Boston, MA). GPAnt-1 and GPAnt-2A peptides were purchased from Sigma and Bio-Mol (Plymouth Meeting, PA), respectively.
Affinity-purified polyclonal anti-
q/11 antibody (C-19L),
anti-
o antibody (K-20), anti-
i (C-10L),
anti-G-protein-
-subunit antibody (T-20), and anti-phospholipase
C
1 antibody (G-12) were purchased from Santa Cruz Biotechnology
(Santa Cruz, CA). Anti-
s antiserum (RM-1) was purchased
from DuPont NEN. Monoclonal antibody against rat pancreatic GP-2 was a
gift from Dr. Anson Lowe (Stanford University). Polyclonal antibody
(31B) to the
subunit of the sodium pump has been described
previously (25). Secondary antibodies included peroxidase-coupled goat
anti-rabbit immunogloblin G (IgG, Amersham) and fluorescein
isothiocyanate-labeled goat anti-rabbit IgG (Sigma).
Purified zymogen granules were prepared by Percoll gradient separation by previously published procedures (13, 26-28). Their purity has been determined both by analysis of several subcellular enzymatic organelle markers and by electron microscopic analysis and shown to be enriched 100-400-fold compared with other organelles (26). Granule membranes were isolated by ultracentrifugation after lysis with nigericin as described previously (28). Plasma membrane-rich fractions were prepared by sucrose gradient centrifugation by previously published procedures (29). 125I-CCK binding to plasma membrane and zymogen granule membrane was carried out as described previously (29) except that membranes were collected on Whatman GFF filters and washed three times with ice-cold buffer to terminate binding. Protein content of membranes was determined with the Bio-Rad protein assay kit, using bovine serum albumin as standard.
Western Blotting of Zymogen Granule Membranes and Plasma MembranesElectrophoresis was performed as described (28). Variable amounts of protein from each sample were loaded per lane onto 7.5% or 10% SDS-polyacrylamide electrophoresis gels and run at 200 V. After gel electrophoresis, proteins were transferred to nitrocellulose membranes at 30 volts overnight. Western blotting was carried out as described previously, using the enhanced chemiluminescence reagent to visualize the secondary antibody.
Immunofluorescence MicroscopyFreshly prepared pancreatic
lobules were fixed for 2 h at 4 °C with a mixture of 2%
formaldehyde (prepared from paraformaldehyde) and 0.25% glutaraldehyde
in phosphate-buffered saline. Fixed tissue was rinsed in
phosphate-buffered saline, cryoprotected with sucrose, and frozen as
described previously (13, 30). Immunofluorescence localization of
q/11 in 5-µm thick cryostat sections followed the
procedures described previously in detail (13, 31). Polyclonal anti-
q/11 was used at 10 µg/ml. Specificity of
staining was assessed by preincubation of primary antibody with 10-fold
excess by weight of peptide used to generate the antibody. For
comparative purpose, some cryostat sections were exposed to monoclonal
antibody to the zymogen granule membrane marker, GP-2. Sections were
examined by epifluorescence microscopy and confocal fluorescence
microscopy (Bio-Rad MRC-600). Digitized images were processed using
Photoshop 3.0 software.
Acini were prepared by collagenase digestion as described previously (32). After isolation, acini were suspended in permeabilization buffer consisting of 20 mM PIPES (pH 7.0), 140 mM potassium glutamate, 0.91 mM MgCl2, 5 mM EGTA, 1 mg/ml bovine serum albumin, 0.1 mg soybean trypsin inhibitor, 1 mM ATP (Mg salt), and 0.5 IU/ml streptolysin-O. For permeabilizing acini and introducing G-protein antagonist peptides into acini, aliquots of the acinar suspension were incubated at 37 °C for 1.5 min in the presence of various amounts of the indicated peptide, then placed in ice-cold water bath for 5 min and aliquoted into 200-µl samples. Amylase release was initiated by adding 200 µl of permeabilization buffer supplemented with CaCl2 to give a final concentration of 1 mM free Mg2+ and the specified concentrations of free Ca2+ which were calculated using a computer program as described previously (32). After incubation for 5 min at 30 °C, samples were centrifuged for 10 s in an microcentrifuge. Amylase released into the supernatant during the incubation was quantified using the Phadebas Amylase Test (Pharmacia, Columbus, OH) and expressed as a percent of total amylase in the acini at the beginning of the incubation.
Immunoprecipitation of Gq/11 Protein from Purified Zymogen GranulesFor immunoprecipitation of q/11,
purified zymogen granules were sonicated for 15 s in lysis buffer
containing 25 mM Tris (pH 7.5), 1.5 mM
dithiothreitol, 3 mM benzamidine, 1 mM
phenylmethylsulfonyl fluoride, 2 mM EGTA, 10 µg/ml
leupeptin, and 0.5% Triton X-100. The insoluble fraction was removed
by centrifugation at 10,000 × g for 10 min at 4 °C.
The supernatant was incubated with 2 µg of anti-Gq/11
antibody at 4 °C for 2 h followed by incubation with protein
A-agarose beads for 1 h. The Gq/11-bound beads were washed with the same buffer and used for Western blotting or assay of
GTPase activity.
Purified zymogen granules
(200 µg of protein) or immunoprecipitated q/11 protein
from purified zymogen granules (2 mg) were suspended in 60 µl of
HEPES-buffered solution (50 mM HEPES, pH 7.2, 1 mM dithiothreitol). The reaction was initiated by adding 40 µl containing of 50 mM HEPES (pH 7.2), 1 mM
dithiothreitol, 2.5 mM EGTA, 250 mM NaCl, 0.5 mg/ml creatine kinase, 5.5 mM creatine phosphate, 2.5 mM ATP, 2.5 µM GTP, and 3 µCi of
[
-32P]GTP followed by incubation for 15 min at
30 °C. The reaction was terminated by the addition of 900 µl of
cold phosphoric acid (pH 2.3) supplemented with 25% activated charcoal
by weight. The samples were centrifuged at 10,000 × g
for 15 min at 4 °C and the [32P]Pi in the
supernatant (300 µl) was determined by liquid scintillation counting.
Nonenzymatic hydrolysis of [
-32P]GTP was subtracted
from all data.
To determine the heterotrimeric G-protein
-subunit(s) localized to zymogen granules, we performed Western
blotting of purified zymogen granule membranes and a plasma
membrane-rich fraction using antibodies specific to
q/11,
s,
o, and
i. Fig. 1 shows that
q/11,
s, and
o were all present in a plasma
membrane-rich fraction;
i was only faintly visualized in
plasma membranes and was therefore not studied further. Of the four G
proteins, only
q/11 was detected on zymogen granule
membranes in Western blotting. In addition, the Na+ pump
subunit was found on the plasma membrane but not granule membrane
(Fig. 1). To quantitate the relative abundance of
q/11 on plasma membrane and zymogen granule membrane we carried out a
dilution series of plasma membrane from 30 to 0.1 µg and compared the
immunoblot signal to 30 µg of granule membranes. In two separate preparations 30 µg of granule membranes were comparable to 1 µg of
plasma membranes. Based on the method of preparation and relative purity of granules (26) this is not likely due to contamination. To
confirm this quantitatively, we evaluated CCK receptor binding as a
well established pancreatic plasma membrane marker. Specific binding
was observed on plasma membranes (total binding 20,173 ± 211 cpm,
nonspecific 318 ± 29, mean ± S.E., n = 4).
By contrast, no specific binding was present on zymogen granule
membranes (total binding 306 ± 24 cpm, nonspecific 340 ± 42).
Recently, multiple heterotrimeric G-protein -subunits were shown to
be localized to the Golgi apparatus of pancreatic acinar cells using
immunohistochemistry, but no
-subunits were detected in the Golgi
apparatus (33). These investigators suggested that G-protein
-subunits are localized to the Golgi apparatus independent of
-subunits. To determine whether
q/11 exists alone or
as part of a complete G-protein complex, zymogen granule membranes were probed with a
-subunit antibody directed against a common region in
the COOH-terminal of all
-subunits. A 37-kDa band was found to be
present in both zymogen granule membranes and plasma membrane-rich fraction (Fig. 2A). These data suggest that
q/11 protein on zymogen granules maintains an
interaction with
-subunit as an intact heterotrimeric G-protein as
is the case for the plasma membrane. We also carried out Western
blotting using antiphospholipase C
1 antibody; phospholipase C
1
was detected in the plasma membrane-rich fraction, but not in purified
zymogen granule membranes (Fig. 2B). These observations in
total, confirm the purity of our zymogen granule and indicate that
q/11 in granule membrane preparations is not due to
contamination with plasma membranes.
Immunofluorescence Localization of Gq/11 in Pancreatic Acini
To further address the localization of q/11,
we analyzed the distribution of
q/11 protein in cryostat
sections of pancreatic lobules by immunofluorescence and confocal
fluorescence microscopy. As shown in Fig. 3A,
q/11 staining was pronounced along the basolateral membranes of acinar cells, and more weakly localized to the granule region in the apical cytoplasm. Although not quantitative, the weaker
immunofluorescence signal over the granule area compared with the
plasma membrane is consistent with the similar pattern resolved in
Western blots (Fig. 1). When anti-
q/11 antiserum was
preincubated with the peptide used to generate the antibody, the
specific staining associated with both the basolateral membrane and the
granules was abolished (Fig. 3B). Immunofluorescence
staining of the apical cytoplasm with anti-
q/11 appeared
at higher magnification to be punctate and to be associated
specifically with the zymogen granules (Fig. 3C). As shown
in this figure, no staining was apparent at the apical membrane of the
acinar cells.
Effect of G-protein Antagonist Peptides on Calcium-stimulated Amylase Secretion from Permeabilized Acini
To study the function
of Gq/11 on zymogen granule membranes, we utilized the
SP-related peptides GPAnt-1 and GPAnt-2a which are known to be
antagonists of Gi/o and Gq/11, respectively
(23). When these peptides are introduced into permeabilized acini and amylase secretion is stimulated by 10 µM free calcium, an
effect on heterotrimeric G-proteins at the plasma membrane to mobilize intracellular calcium can be excluded and the function of the G-proteins at steps distal to intracellular calcium mobilization can be
evaluated. GPAnt-2a peptide potentiated amylase secretion induced by 10 µM free calcium from streptolysin-O permeabilized acini
in a dose-dependent manner (Fig. 4). In
contrast, GPAnt-1 did not alter calcium-stimulated amylase secretion.
We also evaluated the peptide from the carboxyl-terminal of
q/11 used to generate the antibody and it had no effect
on secretion (not shown). None of the peptides affected basal amylase
secretion even at 100 µM. These data demonstrated that
GPAnt-2a peptide specifically potentiated calcium-stimulated amylase
release at a late step of exocytosis.
Effect of G-protein Antagonist Peptides on GTPase Activity on Zymogen Granules
To confirm that GPAnt-2a peptide augments
calcium-stimulated amylase secretion from permeabilized acini by
antagonizing G-proteins on zymogen granules, we evaluated the GTPase
activity of zymogen granules. As shown in Fig.
5A, 100 µM GPAnt-2a peptide
inhibited GTPase activity on zymogen granules approximately 25%.
GPAnt-1 peptide, however, had no effect on the GTPase activity on
zymogen granules. Since multiple small G-proteins are also known to be localized on zymogen granules (27, 34), total GTPase activity of
zymogen granules most likely represents the activity of multiple G-proteins. Therefore, the fact that GPAnt2a inhibited only 25% of the
GTPase activity on zymogen granules is not surprising.
To further determine that GPAnt-2a peptide specifically inhibits Gq/11 protein function on zymogen granules, we examined the inhibitory effect of G-protein antagonist peptides on GTPase activity of Gq/11 immunoprecipitated from purified zymogen granules. Fig. 5B shows immunoprecipitated Gq/11 protein from zymogen granules visualized by Western blotting. Using this immunoprecipitant, the effect of G-protein antagonist peptides on Gq/11 GTPase activity was assayed. As shown in Fig. 5C, GPAnt-2a peptide markedly inhibited immunoprecipitated Gq/11 GTPase activity by about 60%. GPAnt-1 peptide, however, had no effect on GTPase activity of Gq/11. Taken together, these data suggest that the inhibition of GTPase activity of zymogen granules by GPAnt-2A peptide results in the potentiation of calcium-stimulated amylase secretion from permeabilized acini.
Transduction and amplification of receptor-mediated signals to phospholipase C at the level of the plasma membrane is the only currently established function of Gq/11 (35, 36). Although some studies have predicted the participation of Gq/11 in intracellular vesicle trafficking (20, 37), such a role has not been conclusively demonstrated. In the present study, we have described a role for Gq/11 on secretory vesicles in exocytosis. We showed the localization of Gq/11 on zymogen granules by Western blotting, immunoprecipitation, and immunohistochemistry. We next demonstrated that Gq/11 antagonist peptide GPAnt-2a augmented calcium-stimulated amylase release from permeabilized acini. Furthermore, we observed that GPAnt-2a inhibited GTPase activity of both zymogen granules and immunopurified Gq/11 from zymogen granules. Accordingly, it is suggested that GPAnt-2a enhanced calcium-induced amylase secretion from permeabilized acini by inhibiting GTPase activity of Gq/11 on zymogen granules.
Of central importance to our evaluation of the role of
q/11 on zymogen granule membranes is the purity of the
granules and the certainty that the observed signal is not due to
contamination with Gq/11-rich plasma membrane. The
preparation of granules makes use of the high density of intact
granules which band in Percoll gradients separate from other organelles
and are enriched 250-500-fold relative to DNA (nuclei), glutamate
dehydrogenase (mitochondria), and NADH dehydrogenase (endoplasmic
reticulum) (26). In a recent study the purified granule preparation was
used to show the absence of inositol 1,4,5-trisphosphate receptors or
inositol 1,4,5-trisphosphate-induced Ca2+ release in
purified granules (38) although both are present in contaminating
membranes present in crude granules prepared by simple differential
centrifugation. Finally, in the present work
o,
s,
subunits of the sodium pump, phospholipase C
1, and CCKA receptors were all identified in the plasma membrane but not
zymogen granule membranes; if the granule membranes were contaminated
with as little as 1% of plasma membranes we would have detected
saturable CCK binding in the granule membranes. The results with CCK
binding are especially useful since there is a large and quantifiable
signal in plasma membranes. Finally, the immunohistochemistry reveals
specific labeling of the zymogen granules that could be completed by
preincubating antisera with the immunizing peptide. Thus, the data as a
whole makes a convincing case that
q/11 is present on
the zymogen granules.
G-proteins have been suggested to be involved in a late step of
exocytosis as a result of observations obtained using permeabilized cell systems. When a non-hydrolyzable GTP analogue, GTPS was introduced into neutrophils (1), mast cells (2), insulin secreting RIN
cells (3), and chromaffin cells (39), in the presence of free calcium,
exocytosis was markedly enhanced. These findings strongly supported the
involvement of G-proteins in the late steps of exocytosis. However,
since GTP
S can activate both heterotrimeric G-proteins and small
G-proteins, other reagents specific to trimeric G-proteins are required
to analyze their function in exocytosis. AlF4
and mastoparan, both of which are known to be specific activators of
heterotrimeric G-proteins, have been intensively used in permeabilized cell systems. When AlF4
was
introduced into permeabilized secretory cells, their calcium-stimulated exocytosis was enhanced (40, 41) or attenuated (17). These findings
strengthen the idea that heterotrimeric G-proteins are involved in the
late steps of exocytosis. Although
AlF4
is a powerful tool for
studying heterotrimeric G-proteins, it activates all isoforms. By
contrast, mastoparan, an amphiphilic tetradecapeptide purified from
wasp venom, is more specific to inhibitory heterotrimeric G-proteins
Gi and Go (42, 43). Application of mastoparan
to permeabilized cell systems suggest the direct linkage of
Gi or Go to the exocytotic machinery (44-46).
In addition to mastoparan, pertussis toxin is also a specific tool to
analyze Gi and Go. It catalyzes their
ADP-ribosylation, and results in their non-responsiveness to activating
signals. Thus pertussis toxin pretreatment affected the exocytosis
induced by exogenously introduced free calcium into permeabilized
chromaffin cells (47, 48). Therefore, the participation of mastoparan
and pertussis toxin sensitive G-proteins, Gi and
Go, in the late steps of exocytosis have been intensively
studied in these cells. Gi and Go were recently found on small synaptic vesicles of neuronal cells by
immunocytochemistry, and Go has also been found on large
dense core vesicles of adrenal chromaffin cells (49). In exocrine
tissues, both Gi and Go have been reported on
parotid secretory granule membranes by Western blotting (50). Since
various proteins on secretory vesicles play important roles at the step
of vesicle docking and fusion with plasma membrane (51, 52),
heterotrimeric G-proteins on secretory vesicles are now assumed to be
one of the G-proteins involved in the late steps of exocytosis.
Recently, Vitale et al. (17, 18) have shown that
Go on secretory granules has an inhibitory effect on the
final stages of exocytosis in chromaffin cells. In insulin-secreting
-cells, Konrad et al. (19) have described that
Gi on insulin containing vesicles plays a stimulatory role
in the mastoparan-induced insulin secretion. These data suggest that
different heterotrimeric G-proteins on secretory granules play distinct
roles in regulated exocytosis. Concerning other heterotrimeric
G-proteins, Gs is shown to be localized to secretory granule membranes of neuroendocrine tissues and parotid glands (20,
48). However, little is known about its function.
Gq/11 has not previously been reported on secretory granules although it was shown to exist in the Golgi apparatus in some neuroendocrine cells (20) and pancreatic acini (33). Its function in the Golgi is still unclear. In the current study, we have shown that Gq/11 is present on pancreatic zymogen granules and acts to inhibit calcium-triggered exocytosis. These findings are consistent with the previous evidence that G-proteins interact with intracellular calcium to bring about exocytosis (2, 53). Moreover, our findings that Gq/11 antagonist GPAnt2a potently augmented calcium-induced amylase secretion, but not basal secretion, reinforces the synergistic role of Gq/11 with intracellular free calcium in regulated exocytosis.
Currently, the process of exocytosis is thought to occur in two phases. The first step of docking and fusion of primed secretory vesicles to plasma membranes is completed in the first 5 min. The second step is the sequential release of vesicles from a reserve pool (54). According to this theory, our findings that GPAnt-2a potently enhanced the initial amylase release over a 5-min incubation suggests that the target protein of GPAnt-2a, Gq/11, acts on the docking and fusion steps of exocytosis in acinar cells. Taken together, it is reasonable to speculate that the Gq/11 target of GPAnt-2a is on the zymogen granules. To confirm our hypothesis, we examined the GPAnt-2a effect on both zymogen granules and Gq/11 isolated from zymogen granules. GPAnt-2a inhibited both GTPase activities. Thus we have concluded that Gq/11 on zymogen granules tonically inhibits calcium-induced amylase release from pancreatic acini.
Heterotrimeric G-proteins exist on plasma membranes as a complex of
,
, and
subunits at resting state, and separates into free
subunit and a
complex when activated;
,
subunits are assumed always to exist as a complex (55, 56). We have shown that
-subunit as well as
q/11 is localized to zymogen
granule membranes. Although we did not evaluate the presence of
subunits, it is reasonable to conclude that Gq/11 exists as
a complete complex of all three subunits on zymogen granules.
In the classical signal transduction pathway involving
Gq/11 on the plasma membrane, phospholipase C is the sole
effector protein known thus far. In our Western blotting studies, we
could detect phospholipase C1 in the plasma membrane-rich fraction but not in zymogen granule membranes (Fig. 2B). Although
this observation reinforces the purity of the zymogen granule membrane preparation, it concomitantly implies that the effector of the Gq/11 on zymogen granules still remains uncertain. F-actin
has been assumed to play inhibitory roles in exocytosis by blocking secretory vesicle movement toward the plasma membrane (57). Vitale
et al. speculated that the G-protein which is involved in
the negative control of exocytosis may be coupled with some component
controlling the subplasmalemmal cytoskeleton and the movement of
secretory granules to the exocytotic sites (17). Their speculation is
based on results concerning Go of chromaffin secretory
granules. However, the interaction between Go and F-actin is not yet elucidated. On the other hand, Ibarrondo et al.
(37) have recently reported the close association of Gq/11
with F-actin filaments in mammary tumor cell line. Moreover, it was
shown that assembly and disassembly of F-actin are essential to the
final steps of exocrine pancreatic exocytosis (54). Taken together, it
is more likely for Gq/11 that some protein(s) which
regulates the actin network organization might be the effector of the
G-proteins on secretory granules. Alternatively, the possibility still
remains that undefined isoforms of phospholipase C might exist on
zymogen granules as an effector of Gq/11. In any case,
further studies are needed to elucidate the effector of
Gq/11 on zymogen granules.
In conclusion, we have demonstrated a tonic inhibitory action of Gq/11 which localizes to zymogen granules in pancreatic acinar exocytosis. These observations provide new insights for understanding the synergistic regulation of exocytosis by intracellular calcium and G-proteins.
We thank R. W. Holz, R. R. Neubig, and S. M. Wade for their valuable suggestions and technical advice. The excellent technical assistance of Noel Wys is gratfully acknowledged.