Gastric gland stimulation triggers H+,K+-ATPase translocation from cytoplasmic tubulovesicles to apical plasma membrane in parietal cells, resulting in HCl secretion. We studied the mechanisms involved in tubulovesicle translocation with a permeabilized gland system. Streptolysin O (SLO)-treated glands were permeabilized such that exogenous fluorescently labeled actin incorporated into cytoskeleton in a pattern mimicking endogenous F-actin. As shown by accumulation of the weak base aminopyrine (AP), SLO-permeabilized glands are stimulated to secrete acid by addition of cAMP and ATP and inhibited by proton pump inhibitors. Direct visualization with the fluorescent pH probe Lysosensor showed acid accumulation in glandular lumen and parietal cell canaliculi. ME-3407, an antiulcer drug with inhibitory action implicated to involve ezrin, inhibited AP uptake in and effectively released ezrin from intact and SLO-permeabilized glands. In contrast, wortmannin, an effective secretion inhibitor in intact glands, had minimal effects on ezrin or AP accumulation in SLO-permeabilized glands. The finding that SNARE protein syntaxin 3 is associated with H+,K+-ATPase-containing tubulovesicles suggested that it is involved in membrane fusion. Addition of recombinant syntaxin 3, but not syntaxin 5 or heat-denatured syntaxin 3, dose-dependently inhibited acid secretion. Our studies are consistent with a membrane recycling hypothesis that activation of protein kinase cascades leads to SNARE-mediated fusion of H+,K+-ATPase-containing tubulovesicles to apical plasma membrane.
- SNARE hypothesis
- membrane fusion
the development of permeabilized cell preparations has provided a valuable experimental tool linking in vitro studies that use subcellular fractions with direct assays of cellular function. In parietal cell physiology, a number of important functions have been elucidated through the study of gastric glands permeabilized with detergents (reviewed in Ref. 13). One of the major drawbacks with many detergent-like molecules (digitonin or saponin) was that the treated glands were unable to transform from the resting to the stimulated state, although preparations stimulated before treatment could be maintained to secrete acid (10, 15). It was then discovered that pore-forming bacterial toxins permeabilize gastric glands while allowing the transition of resting to a stimulated state (22). These toxins have been valuable in generating permeable cell systems that facilitate better understanding of the parietal cell activation process (19, 21, 26). Depending on the size of the plasma membrane pore, one can introduce a variety of molecules into the cell to facilitate or interfere with normal functional activity. In previous work, we found α-toxin to be a superior permeabilization agent over digitonin and other pore-forming bacterial toxins. We determined (19) that stimulation was dependent on addition of cAMP, that addition of ATP led to synthesis of cAMP and therefore could trigger stimulation, and that addition of both of these nucleotides gave optimal results. As with other permeabilized preparations, α-toxin-treated glands required substrates of mitochondrial oxidative phosphorylation for regeneration of the ATP hydrolyzed for proton pumping (10). Because of the small pore size produced (∼3-nm diameter) α-toxin allows the entry of only small molecules, such as nucleotides and phalloidin (8). This toxin was useful in determining the metabolic properties (19) and phosphoproteins associated with cell activation (26) in isolated gastric glands. Recent work has shown that treatment of glands with the saponin ester β-escin yields viable glands that allow entry of small polypeptides (1). However, the development of a permeabilized system that allows the entry of larger (protein sized) molecules would be extremely useful for the dissection of the intracellular activation cascade.
We have revisited the use of another bacterial pore-forming toxin isolated from β-hemolytic group A streptococci, streptolysin O (SLO). Although pores formed by α-toxin only allow passage of molecules with a molecular mass <5 kDa (8), the pores formed by SLO are large enough to allow passage of proteins of up to 200–500 kDa (6). In the present study, we now show entry of large molecules [e.g., actin, ∼42 kDa and glutathioneS-transferase (GST)-syntaxin, ∼50 kDa] into SLO-permeabilized gastric glands. These SLO-treated glands retain the ability to transform from a resting to a stimulated state, and as in other permeabilized models of rabbit glands, stimulation occurs with the addition of substances that presumably increase levels of cAMP.
MATERIALS AND METHODS
Isolation of gastric glands.
Intact gastric glands were isolated from New Zealand White rabbits as described by Berglindh (3). Briefly, the stomach was perfused under high pressure with PBS (in mM: 2.25 K2HPO4, 6 Na2HPO4, 1.75 NaH2PO4, and 136 NaCl) containing 1 mM CaCl2 and 1 mM MgSO4. The gastric mucosa was scraped from the smooth muscle layer, finely minced, and then washed three times in PBS and twice with minimal essential medium buffered with 20 mM HEPES, pH 7.4 (HEPES-MEM). The minced mucosa was digested at 37°C for ∼30 min in a minimal amount (∼20 ml) of HEPES-MEM containing 15 mg of collagenase (Sigma Blend) and 20 mg each of BSA (fraction V, Sigma) and HEPES acid. The digestion was terminated by diluting the solution into 200 ml of HEPES-MEM followed by filtration through a nylon mesh to remove large pieces of undigested mucosa. Intact gastric glands were isolated by settling the digestion mixture for 10–15 min and then washed by settling three times in HEPES-MEM. In all subsequent gland experiments (intact and permeabilized), glands were resuspended at 5% cytocrit (vol/vol) in the appropriate buffer for final assay. All animal protocols were approved by the University of California, Berkeley Animal Care and Use Committee.
Fluorescent labeling of actin.
Muscle actin was isolated by first preparing a salt- and acetone-extracted muscle powder according to Pardee and Spudich (16). This powder was then extracted on ice in G buffer (in mM: 20 Tris-Cl pH 8, 0.2 CaCl2, 0.2 ATP, and 0.5 dithiothreitol). The solution was centrifuged for 30 min at 10,000g, and the resulting supernatant was filtered through a 0.45-μm filter. The G-actin solution was polymerized at 15°C in a two-step process optimized to reduce trapping of extraneous actin-binding proteins. In the first step, 10 mM KCl and 0.7 mM MgCl2 were added to the filtrate followed by a 1-h incubation. In the second step, KCl and MgCl2 were added to a final concentration of 50 mM and 2 mM, respectively, followed by another 1-h incubation. The filamentous actin was pelleted by a 2-h centrifugation at 100,000 g and then depolymerized by resuspending the pellet in G buffer. To obtain a high degree of purity, the resulting monomeric actin was centrifuged for 2 h at 100,000g to remove debris, followed by another round of polymerization and centrifugation. For labeling, F-actin was diluted to a concentration of 2–5 mg/ml and then dialyzed against a buffer containing (in mM) 30 PIPES (pH 6.8), 50 KCl, 0.2 CaCl2, and 0.2 ATP (11).N-hydroxysuccinimidyl-rhodamine (Pierce) was added at a concentration of 3.2 mg/ml, and the solution was incubated in the dark for 1 h with gentle shaking. The labeling reaction was quenched by addition of 10 mM glutamine followed by overnight dialysis against G buffer at 4°C. After 30 min of centrifugation at 10,000 gto remove debris, the actin underwent another round of polymerization and depolymerization. The final product (dialyzed against G buffer) was stored in aliquots at −80°C.
SLO permeabilization of gastric glands.
SLO was obtained from S. Bhakdi (University of Mainz, Mainz, Germany) and from H.-P. Moore (University of California, Berkeley, CA). Stock solutions of SLO were prepared at 1 mg/ml (750,000 lytic units/mg) in 0.1% BSA-10 mM HEPES (pH 7.2) and stored at −80°C. Reducing agents were not required for activity because the single thiol group of wild-type SLO had been removed (17). Reconstituted SLO solutions were kept on ice when in use. Intact glands in HEPES-MEM were washed two times by settling at 4°C in ice-cold K buffer (in mM: 10 Tris base, 20 HEPES acid, 100 KCl, 20 NaCl, 1.2 MgSO4, 1 NaH2PO4, and 40 mannitol, pH 7.4). SLO was added to a final concentration of 1 μg/ml, and the glands (at 5% cytocrit) were mixed by inversion and then incubated on ice until the glands had completely settled (10–15 min). The K buffer supernatant containing excess SLO was removed, and the glands were washed two more times by settling at 4°C in ice-cold K buffer. Finally, glands were resuspended in K buffer solution containing 1 mM pyruvate and 10 mM succinate and then incubated at 37°C for 3 min to allow the SLO pores to form. SLO-treated glands were used immediately or maintained on ice until use.
14C-labeled aminopyrine uptake assays.
Stimulation of intact or SLO-permeabilized gastric glands was quantified using the aminopyrine (AP) uptake assay. The performance of, and calculations for, the AP uptake assay in glands were essentially as described by Berglindh (3) with modifications described by Wallmark et al. (25). For both intact and permeabilized preparations, glands were kept in the resting state with cimetidine (100 μM). Intact preparations were stimulated with IBMX (30 μM) and either histamine (100 μM) or dibutyryladenosine 3′,5′-cyclic monophosphate (DBcAMP; 1 mM). Permeabilized preparations were stimulated with cAMP (100 μM) plus ATP (1 mM). Other drugs were also added where indicated; the H+,K+-ATPase-specific proton pump inhibitor SCH-28080 (10 μM; Ref. 23) and the kinase inhibitors ME-3407 (300 μM; Ref. 22) and wortmannin (10 μM). Recombinant proteins (see Purification of recombinant syntaxin proteins) were also added. Gland preparations were incubated for 20 min at 37°C with shaking (∼160 oscillation/min) with or without various reagents as specified and then centrifuged briefly to pellet the glands. The pellets were dried and weighed, and aliquots of both the supernatant and pellet were counted in a Beckman liquid scintillation counter. These data were used to calculate the AP accumulation ratio (ratio of intracellular to extracellular AP concentration) according to the method of Berglindh (3). To normalize AP uptake values among the various preparations, the data are expressed as a fraction of the stimulated control in some experiments.
Purification of recombinant syntaxin proteins.
Recombinant syntaxin 3 and syntaxin 5 were expressed in bacteria as GST fusion proteins. Briefly, rat syntaxin 3 [amino acids (aa) 1–263, full length], H3 domain (aa 188–263), and syntaxin 5 cDNAs were cloned into pGEX-2T vector (2). One liter of lysis buffer (LB; 10 g NaCl, 10 g tryptone, and 5 g yeast extract, pH 7.4) was inoculated with bacteria transformed with either GST-syntaxin 3 (GST-syn3) or GST-syntaxin 5 (GST-syn5). When the optical density (A600) of the bacteria culture reached between 0.7 and 1.0, expression of protein was induced by addition of 0.5 mM isopropyl β-d-thiogalactopyranoside (IPTG). Bacteria were harvested by centrifugation 3 h after induction, resuspended in PBS containing 5 μg/ml proteinase inhibitors (leupeptin, pepstatin, and chymostatin; 5 μg/ml), and sonicated for four bursts of 10 s each using a probe-tip sonicator. The lysis solution was cleared of insoluble material by centrifugation at 10,000 g for 20 min. The soluble fraction was applied to a column packed with glutathione-agarose beads (Sigma), followed by extensive washes with PBS (20× bead volume). The fusion proteins were then eluted in PBS containing 10 mM glutathione and then applied to a gel filtration column (Econ-Pac 10 DG; Bio-Rad) to remove glutathione and exchange PBS buffer for K buffer. The fusion protein was estimated to be 90% pure by SDS-PAGE; major contaminants were degraded fragments of syntaxins. Protein concentrations were determined by Bradford assay (5).
In some experiments, histidine-tagged syntaxin 3 (HIS-syn3) was used as an alternative source of recombinant syntaxin 3. A rat syntaxin 3 cDNA was cloned into pRSET (Invitrogen). Induction and lysis of bacteria were performed as described above for GST-syntaxin, except that bacteria were lysed in (in mM) 300 NaCl, 50 NaH2PO4, and 20 imidazole, pH 8. The soluble fraction containing soluble HIS-syn3 protein was applied to a column packed with nickel-Sepharose beads (Qiagen), followed by extensive washes with lysis buffer containing 50 mM imidazole. Purified HIS-syn3 was eluted in lysis buffer containing 250 mM imidazole and then applied to a gel filtration column. The fusion protein (in K buffer) was estimated to be 90–95% pure by SDS-PAGE; the major contaminants were degraded fragments of syntaxin 3. Protein was determined by Bradford assay (5). In some assays, aliquots of HIS-syn3 were heated at 85°C for 2 min to inactivate the protein by denaturing (DNT-syn3).
Immunofluorescence and microscopy.
Intact and permeabilized glands were prepared identically to glands used in AP uptake experiments except that no AP was added. After incubation for 20 min at 37°C, glands were fixed by addition of formaldehyde (final concentration of 3.7%). After 10 min, the fixed glands were washed by settling in PBS. Glands were then mounted on poly-l-lysine-coated glass coverslips, permeabilized in 0.5% Triton X-100 in PBS for 15 min, and blocked in 2% BSA in PBS for 15 min. All subsequent antibody dilutions were made in PBS containing 2% BSA. Ezrin was detected by 60-min incubation with mouse monoclonal antibody 4A5 (9). The primary antibody was detected by FITC-labeled goat anti-mouse IgG (Jackson Labs) in a subsequent 60-min incubation. Actin was detected by coincident incubation with 80 nM rhodamine-labeled phalloidin (Sigma). Coverslips were supported on slides by grease pencil markings and mounted in Gel/Mount (Biomeda). Confocal microscopy was performed with a Zeiss Axioplan microscope and a Bio-Rad MRC 1024 laser system, using COMOS software to collect images.
Live imaging of acid-secreting glands.
Accumulation of acid in live gastric glands was assayed using a cell-permeable fluorescent dye, Lysosensor Yellow/Blue DND-160 (Molecular Probes). The excitation and emission spectra of Lysosensor change in response to the pH of its surroundings. The product literature indicates that Lysosensor has little fluorescence when excited at ≥400 nm above pH 5 but has increasing fluorescence as the pH drops below 5. Because of these changes in spectra, Lysososensor in the gland experiments was excited at 440 nm, and images were taken at 530 ± 15 nm. After permeabilization, SLO-treated glands were maintained on ice in K buffer containing 1 mM pyruvate, 10 mM succinate, and 5 μM Lysosensor. Glands were then plated onto acid-washed glass coverslips and placed on a heated microscope stage (Warner Instruments) in the same buffer. Data were collected on an inverted Nikon microscope with a video camera system (Sutter Instruments) controlled by Axon 4.0 software. Images were taken at time points before and after stimulation with 1 mM ATP and 0.1 mM cAMP. The ionophore nigericin (0.1 mM final concentration) was added at the conclusion of the experiments to show that the increase in fluorescence was due to the proton gradient formed.
In previous work, we compared several methods for permeabilizing gastric glands, such as α-toxin, digitonin (22), and even SLO (unpublished results); however, the SLO-permeabilized glands stimulated poorly. In this original protocol, both the α-toxin- and SLO-permeabilized glands were maintained at 37°C for 1 h, including both the permeabilization and subsequent acid accumulation steps. Although this was an optimal treatment for α-toxin-induced poration, we hypothesize that the long exposure of glands to SLO could have resulted in the permeabilization of intracellular organelle membranes or in disruption of glandular structure. In support of the first argument, the small size of α-toxin pores excludes the entry of its own monomer (34,000 Da). However, SLO-induced pores are large enough to allow passage of SLO monomer (57,000 Da). Subsequently, we learned that SLO could bind to cholesterol-containing membranes on ice, whereas the oligomerization process that leads to pore formation only occurred at higher temperatures (4). We therefore modified our protocol to include a step of incubation of glands with SLO at 4°C that allows the toxin to partition into the plasma membrane but not to generate pores. After washout of unbound SLO, the permeabilization was initiated by warming glands at 37°C. SLO-induced permeabilization occurs rapidly (within 3 min) when the glands are warmed to 37°C.
In our current experiments, we determined that a concentration of SLO in the range of 1–2 μg/ml was optimal for permeabilization of gastric glands (not shown). Glands treated with much lower SLO concentrations (e.g., 0.04 and 0.2 μg/ml) gave reduced values in our assay of AP uptake. These SLO-treated glands were poorly permeabilized as judged by the entry into parietal cells of small-molecular-size chemicals such as fluorescent-labeled phalloidin, a filamentous actin-binding compound. When glands were treated with much higher concentrations of SLO (e.g., 5 μg/ml), the permeability of glandular cells was improved as judged by more intense labeling of the actin cytoskeleton by rhodamine-phalloidin. The ability of these glands to accumulate AP, however, was greatly reduced compared with that of glands treated with 1 μg/ml SLO. Therefore, 1 μg/ml of SLO was chosen as a standard condition for permeabilizing the glands.
The AP uptake data in Fig. 1 Ashow that SLO-permeabilized glands, like α-toxin-permeabilized glands, were able to secrete acid on the addition of ATP and cAMP. Addition of 1 mM ATP and 0.1 mM cAMP to SLO-treated glands resulted in an 11.8-fold increase in AP ratio over unstimulated glands (102.9 ± 10.1 vs. 8.7 ± 0.6, respectively). Although the absolute AP ratios were lower in SLO-treated glands, the fold stimulation over resting was comparable to values seen in intact glands. For intact preparations AP accumulation ratios increased 7.5-fold in glands stimulated with histamine and IBMX compared with resting glands (166.3 ± 18.4 vs. 22.2 ± 5.4, respectively). Direct stimulation of the protein kinase A (PKA) pathway in intact glands with DBcAMP and IBMX resulted in an 11.7-fold increase in AP ratio over resting glands (233.7 ± 24.7 and 19.9 ± 5.1, respectively). As with other permeabilized gland models (1, 10, 15, 19,22), acid accumulation in SLO-treated glands requires substrates for oxidative phosphorylation to regenerate ATP required by the gastric proton pump. SLO-permeabilized glands incubated with ATP and cAMP but lacking pyruvate and succinate showed no significant increase in AP ratio (7.4 ± 0.4) compared with those without the nucleotides (8.7 ± 0.6).
Figure 1 B compares the time course for cAMP/ATP-dependent stimulation of acid secretion in SLO-permeabilized glands during the standard 20-min assay at 37°C. Overall, the AP uptake kinetics for SLO-permeabilized glands appeared similar to those seen in intact glands. At time 0, the AP ratios for the SLO (6.5 ± 1.1) and intact (7.4 ± 2.5) preparations were nearly identical. Furthermore, both preparations showed the greatest increase in AP accumulation within the first 5 min, with intact and SLO glands reaching 73% and 81% of their peak values, respectively. AP ratios continued to increase at the 10- and 15-min time points, but at a slower rate. At 20 min, a significant difference between intact and SLO-permeabilized gland preparations is apparent (P < 0.04 by t-test assuming unequal variance). The SLO preparations reached their highest AP values at 15 min (102.9 ± 3.8), declining back to 83.7 ± 9.9 at 20 min (82% of the peak AP ratio). In contrast, the AP ratio for intact glands continued to increase between 15 and 20 min, from 109 ± 18.7 to 126.9 ± 13.0 (89% to 100%, respectively).
Figure 2 shows the dependence of AP uptake in SLO permeabilized glands on ATP concentration with and without the inclusion of cAMP. In the absence of cAMP there was a progressive increase in AP uptake as ATP concentration was increased up to 1 mM. Addition of 0.1 mM cAMP significantly potentiated the stimulation by ATP at lower concentrations of ATP and even produced a 5.1-fold increase in AP uptake ratio in the absence of added ATP. The cAMP potentiation was diminished or abolished at higher levels of ATP. In addition, 1 mM of either AMP or ADP could substitute for 1 mM ATP in stimulation of AP accumulation by the SLO-permeabilized glands (data not shown). These data are very much like those obtained with α-toxin-permeabilized glands, in which it was shown that any adenine nucleotide source could provide the necessary energy intermediate (19). Even cAMP serves both as a second messenger to trigger PKA activation and a suitable source of ATP, through conversion to AMP by phosphodiesterase and subsequent resynthesis to ATP. This is illustrated by the finding that cAMP-triggered AP accumulation, which reaches 50–60% of the maximum (Fig. 2), was completely abolished when cellular phosphodiesterase was inhibited by IBMX in the SLO-permeabilized glands (data not shown).
Because addition of nucleotide appeared to bypass the H2receptor pathway, we investigated whether this nucleotide-dependent uptake of AP reflected accumulation of acid in an apical/canalicular location. A fluorescence microscope was used to image live SLO-treated glands in the presence of the Lysosensor probe, a weak base that accumulates in acid spaces and fluoresces yellow at low pH. In general, resting glands had only weak fluorescence, confined mainly to the interior of the parietal cells. Occasionally, brilliantly stained cells were apparent even in the resting state, but these were atypical. Within 3 min of addition of cAMP/ATP, the overall Lysosensor staining was much greater, and fluorescence was detectable in the lumen of the gland for the first time (Fig.3 b). This fluorescence increased over time; between 9 and 13 min, the glandular lumens were the predominant fluorescent structures (Fig. 3, c andd). In general, addition of cAMP/ATP caused fluorescence within parietal cells that qualitatively resembled apical/canalicular staining of gastric glands (Fig. 3,b–e). This staining was not seen in preparations without added nucleotide (Fig.3 a) or preparations in which protonophores (e.g., nigericin or thiocyanate) were included along with cAMP/ATP (Fig. 3 f). Within 5 min after addition of 0.1 mM nigericin, fluorescence was completely abated.
Rhodamine-labeled actin (Rh-actin), with a molecular mass of ∼42 kDa, was used to determine whether large polypeptides could enter SLO-permeabilized glands. Within 5 min of addition of Rh-actin, much of the gland lumen and the canalicular surface of parietal cells became fluorescently labeled. At shorter times, the intensity of the labeling was somewhat uneven and tended to show preference toward the lumen. For example, Fig. 4 B shows that after 3 min of exposure to Rh-actin, the lumen of SLO-permeabilized glands was intensely labeled, comparable to the staining obtained with FITC-phalloidin after the glands were fixed. In contrast, the intensity of parietal cell canalicular staining was relatively lower than the phalloidin staining pattern. At somewhat longer times (e.g., 15 min), Rh-actin fluorescence (Fig. 4 C) was qualitatively the same as for the actin cytoskeleton labeled with FITC-phalloidin (Fig.4 D) in both canaliculi and gland lumen. It is possible that the relative staining differences in the early exposure times might reflect some inherent difference in the actin turnover of the respective structures, but it may also be due to the fact that the lumen is formed from both parietal and the smaller nonparietal cells (blackened regions between parietal cells in the phalloidin staining pattern) and thus reflects the sum of many individual fluorescent components. The important point here is that the rather large molecular probe, actin, can easily pass through the SLO pores and reach the apical plasma membrane of glandular cells.
The effects of several inhibitors of secretory function were assayed to compare the characteristics of SLO-permeabilized glands (nucleotide-dependent stimulation) to those of intact glands (histamine/IBMX stimulation). SCH-028080 inhibits AP accumulation by directly inhibiting the proton pump H+,K+-ATPase (23). Wortmannin and the antiulcer drug ME-3407 are both inhibitors of acid secretion that do not act directly on the proton pump but rather through some intermediate signaling step such as phosphorylation or the action of ezrin (24). As shown in Fig.5, addition of 10 μM SCH-28080 to either intact or SLO-permeabilized glands completely abolished AP accumulation and actually reduced AP uptake to levels below that seen in nonstimulated glands. Addition of 0.3 mM ME-3407 also reduced AP uptake for both intact and SLO-permeabilized glands, although the degree of inhibition tended to be somewhat more effective for the intact glands. Surprisingly, although 10 μM wortmannin effectively inhibited intact glands (15 ± 6% of stimulated controls), AP uptake by the SLO-permeabilized glands was only slightly reduced by wortmannin (86 ± 11% of stimulated controls; P> 0.05).
To further explore this difference in action between wortmannin and ME-3407, we looked at the effects of both drugs on the cytoskeletal membrane-linking protein ezrin. ME-3407 was shown previously to cause a delocalization of ezrin away from the plasma membrane and into the cytoplasm (24). Here, we compared intact and SLO-permeabilized glands that had been prepared for immunofluorescence using an anti-ezrin monoclonal antibody. TRITC-conjugated phalloidin was used as a second probe to localize the F-actin in the glandular cells. In general, the ezrin staining was weaker in SLO-permeabilized glands compared with intact preparations (Fig.6), most likely because of some ezrin exiting the cell through SLO pores. However, the remaining ezrin signal was clearly localized to the canalicular membranes of parietal cells in both intact and SLO-treated glands (Fig. 6) and in both the resting (not shown) and stimulated states. ME-3407 caused a remarkable loss of ezrin staining in the intact and SLO-treated glands. Furthermore, the remaining ezrin signal was diffusely distributed throughout the parietal cell cytoplasm and entirely lacking in association with membrane structures. Wortmannin treatment may also have reduced ezrin staining in intact glands but not to the same extent as ME-3407 (Fig.6). After wortmannin treatment, canalicular structures were still visible in many, but not all, parietal cells, and the diffuse ezrin signal in the cytoplasm was increased. In contrast, wortmannin had little effect on the ezrin staining pattern of SLO-permeabilized glands, i.e., canalicular structures remained distinct and clearly defined (Fig. 6).
The above experiments clearly demonstrate that SLO treatment permits entry of relatively large molecular components (e.g., 42-kDa actin) and that treated glands have a robust acid secretion response when incubated with cAMP/ATP. The transition between the resting and stimulated states very likely involves tubulovesicle/apical plasma membrane fusion events, and previous studies implicated a role for soluble N-ethylmaleimide-sensitive fusion attachment protein (NSF) receptor (SNARE) proteins in parietal cell activation. We wanted to test whether protein components that could theoretically interact with the SNARE machinery of the parietal cell might alter the secretory response. Although previous work (7, 17) identified the presence of several syntaxin isoforms (e.g., types 1, 2, 3, and 4), 25-kDa synaptosome-associated protein (SNAP-25), and vesicle-associated membrane protein (VAMP)-2 in the gastric parietal cell, only syntaxin 3 and VAMP-2 were highly represented on the H+,K+-ATPase-rich tubulovesicle membranes. Recombinant proteins, expressed in bacteria from plasmids containing two syntaxin isoforms (syntaxin 3 and syntaxin 5) were purified using a GST. GST-syn3 and GST-syn5 contain either rat syntaxin 3 or rat syntaxin 5, respectively, fused to the COOH terminus of GST. Recombinant proteins were added to SLO-permeabilized glands in the presence and absence of cAMP/ATP. The GST-syn5 fusion protein served as a control for nonspecific effects caused by the addition of recombinant protein. The syntaxin 5 isoform has been implicated in endoplasmic reticulum to Golgi transport (20) and should have minimal effect on the trafficking of H+,K+-ATPase-containing tubulovesicles. Addition of GST-syn5 caused relatively small changes in AP uptake (at most ∼20% decrease), and there was no dose-dependent inhibitory effect. In contrast, GST-syn3 caused a dose-dependent inhibition of acid secretion in SLO-permeabilized glands, as measured by AP uptake. No significant inhibition was noted at 2.5 μg protein/ml, but 5 μg protein/ml caused a 28% reduction in acid secretion and maximal inhibition (89–91%) occurred at 10 and 20 μg/ml (Fig. 7).
It is possible that the proteins purified from bacteria using the GST affinity tag could contain a small amount of contaminating bacterial protein or other toxins. Although not necessarily detrimental to reconstitution or other cell-free assays, these bacterial contaminants can be cytotoxic. We therefore used an alternative construct of syntaxin 3 with a different affinity tag, which allowed purification by a completely different procedure. The HIS-syn3 construct contains a 6-histidine tag fused to the NH2 termini of rat syntaxin 3. The HIS tag purification system allows for washing at higher stringency (near denaturing) conditions. The data of Fig.8 show that HIS-syn3 has inhibitory characteristics similar to those of GST-syn3. No significant inhibition was seen at 2.5 μg protein/ml, but higher concentrations of HIS-syn3 caused a dose-dependent inhibition of AP accumulation. As a further control, in parallel experiments HIS-syn3 was boiled to denature the protein (DNT-syn3) before addition to the SLO-permeabilized glands. No significant inhibition of acid secretion was observed with any amount of DNT-syn3 (2.5–20 μg protein/ml).
The soluble coiled coil domain of syntaxin 1a (the H3 domain) is known to directly interact with SNAP-25, forming a binary complex (21,28). It was shown previously that addition of a syntaxin 1a H3 peptide construct inhibits fusion of norepinephrine-containing granules with the plasma membrane of PC12 cells (28), presumably by competing with the endogenous protein. To delineate the molecular function of the syntaxin 3 isoform in parietal cell secretion, we sought to test whether the H3 domain of syntaxin 3 is involved in forming a complex between syntaxin 3 and SNAP-25. The H3 domain (aa 188–263) of syntaxin 3 was purified to near 90% purity with glutathione-agarose beads, and the same protocol was used to isolate the full-length GST-syn3 (Fig.9 A). When added to SLO-treated glands the H3 domain of syntaxin 3 exerted very little effect on acid secretion at doses up to 20 μg/ml, whereas the full-length GST-syn3 caused the same dose-dependent inhibition on acid secretion seen in studies above (Fig. 9 B). Thus it appears that apical trafficking of tubulovesicle is efficiently inhibited by full-length syntaxin 3 but not the H3 domain alone.
Permeabilization of gastric glands by SLO allows access to the intracellular environment of parietal cells without appreciable loss of cell function, especially the activation of acid secretion. The preparation has many similarities to other permeabilized gastric gland models and has some additional advantages. First, nucleotides such as cAMP trigger acid secretion. This appears to validate the consensus theory that activation of the cAMP-dependent pathway is sufficient to activate the secretory machinery. Second, permeabilized glands have a secondary requirement for either ATP or an ATP-regenerating system to supply energy for the proton pump. These two conclusions are supported by the synergistic effect of ATP and cAMP on acid secretion. Furthermore, although cAMP alone triggers stimulation, at least some of this cyclic nucleotide is being converted into ATP by the cell to fuel H+,K+-ATPase. This is illustrated by the inability of cAMP to induce acid accumulation either in the presence of saturating amounts of IBMX or in the absence of mitochondrial substrates. Third, the initial rate of acid secretion in SLO-permeabilized glands stimulated with ATP/cAMP mirrors the rate measured in intact glands stimulated with histamine/IBMX. The drop-off in acid secretion by permeabilized glands seen at later times (∼20 min) could result from a weakening of the gland structure, leading to breakage of cell to cell contacts and subsequent loss of accumulated acid from the gland lumen. In support of this, SLO-treated glands appeared to be much shorter after a 20-min incubation at 37°C. Fourth, permeabilization is fairly uniform, because rhodamine-labeled actin quickly incorporates into the cytoskeleton of both parietal and nonparietal cells, resulting in staining that is qualitatively the same as phalloidin staining of F-actin. The uptake of rhodamine-labeled actin appeared to occur first at the glandular lumen, a structure contributed by parietal and nonparietal cells, and then at the canalicular structures throughout all parietal cells a few minutes later. Images of SLO-treated glands treated with Lysosensor dye also suggest uniform permeabilizations, because cAMP/ATP triggers acidification in a large number of parietal cells within the gland. Therefore, it is unlikely that acid secretion is due to only a few stimulated parietal cells. Fifth, this model allows the incorporation of large-molecular-mass proteins, such as actin (42 kDa) and GST-syntaxin (52 kDa).
Although the cAMP-mediated stimulation of SLO-permeabilized glands is in many ways similar to histamine/IBMX stimulation of intact glands, two known kinase inhibitors have different effects on these two systems. Previous experiments showed that both ME-3407 and wortmannin inhibit acid secretion in intact glands, and it was suggested that they had a common target in the enzyme myosin light chain kinase (24). ME-3407 was also shown to have the further effect of delocalizing ezrin from its usual membrane-bound location in parietal cells. In SLO-permeabilized glands, we found that ME-3407 both inhibited AP uptake and caused the delocalization of ezrin. Wortmannin, on the other hand, had rather minimal effect on ezrin staining in either intact or SLO-treated glands. Furthermore, it had radically different effects on acid secretion in the intact and SLO-permeabilized preparations. Consistent with earlier work (24) wortmannin inhibited AP uptake by intact glands but, unpredictably, had no significant effect on SLO-permeabilized glands. This suggests that ME-3407 and wortmannin have different targets unlike previously proposed. The wortmannin-sensitive factor may be released into the extracellular medium by permeabilization, whereas the ME-3407-sensitive factor appears to be retained by the permeabilized gland.
On the basis of these findings, it is not unreasonable to suggest that the inhibitory effect of ME-3407 on acid secretion may be due to some ezrin-related effect. With this in mind, the staining pattern of ezrin in SLO glands is noteworthy. Although total ezrin levels were reduced after SLO permeabilization, the remnant ezrin staining pattern is reminiscent of staining seen in intact glands. Higher laser power was required to get the same level of intensity, but there remained distinct canalicular staining in parietal cells. The finding of reduced ezrin levels is at odds with the hypothesis that ezrin is an integral component of the stimulation process (see Ref. 23). However, it may be that the remnant ezrin is sufficient for the process of transforming the parietal cell from the resting to the stimulated state. Or, it is possible that membrane-associated ezrin is not required for the immediate events leading to acid secretion, although ezrin may still be required as a membrane organizing protein. Future work in quantitating the cellular levels of ezrin will be useful in distinguishing among these possibilities.
The inhibitory effect of exogenously added syntaxin 3 protein provides a point of evidence toward the large pore size in the SLO model, as well as lending additional support to the involvement of soluble NSF receptors (SNAREs) in the recruitment of H+,K+-ATPase-containing tubulovesicles to the apical plasma membrane. The SNARE hypothesis holds that membrane fusion involves the pairing of specific proteins in both the vesicle and target membranes. The proteic complex consists of four helical bundles, contributed by components in the vesicle (v-SNAREs) and target membrane (t-SNAREs). This taxonomy can lead to some confusion, because certain SNARE proteins make promiscuous pairings, acting as either a v- or a t-SNARE. As a case in point, the SNARE members VAMP-2 and syntaxin 3 have both been shown to be present on the same tubulovesicle membrane fraction that contains H+,K+-ATPase (7, 17), suggesting that they act as v-SNAREs in parietal cells. However, syntaxin 3 is associated with the plasma membrane in neuronal cells, where it has been proposed to act as a t-SNARE (2). The inhibition we see with exogenously added syntaxin 3 is specific because neither denatured syntaxin 3 nor native syntaxin 5 showed any dose-dependent inhibition. Syntaxin 5 was not expected to have any effect, because it is associated with the assembly of pre-Golgi intermediates (20), and denatured syntaxin 3 would not be expected to interact correctly with cellular proteins. These results suggest that syntaxin 3 acts as a part of the protein scaffold required for fusion of vesicles to the plasma membrane, resulting in the partitioning of H+,K+-ATPase to the apical plasma membrane. The inhibitory effect presumably occurs through competitive binding with target sites in the plasma membrane. The added syntaxin 3, although full length, is not inserted into the tubulovesicle membrane. However, pairing of recombinant syntaxin 3 with its cognate receptors prevents endogenous syntaxin 3 from forming a functional protein complex that is essential for the fusion of tubulovesicles to the apical plasma membrane. A similar argument has been proposed by Lehnardt et al. (12), who showed that exogenously added α-SNAP inhibits tubulovesicle membrane fusion and subsequent acid secretion.
Using Madin-Darby canine kidney cells, Low et al. (13) showed that overexpression of syntaxin 3, but not syntaxin 2 or 4, inhibits trafficking from the trans-Golgi network and endosome to the apical plasma membrane, demonstrating a functional difference between syntaxin 3 and other syntaxin isoforms. These data are similar to the present observations, where syntaxin 3, but not syntaxin 5, is essential for tubulovesicle-plasma membrane fusion in gastric parietal cells. Because the H3 domain of syntaxin 1a forms a binary complex with SNAP-25 in vitro (28) and, in separate experiments, the H3 domain was found to inhibit a fusion-based assay in permeabilized PC12 cells (21), we tested the homologous H3 domain of syntaxin 3 in our system. We expected that the H3 domain would be more effective at inhibiting secretion because it lacks the NH2-terminal domain that inhibits complex formation (21). We were surprised to find that the H3 domain of syntaxin 3 did not inhibit parietal cell secretion to the same extent that full-length syntaxin 3 did; however, several possible explanations presented themselves. First, Scales et al. (21) found that the “equivalent” H3 domains from other syntaxin isoforms exert different inhibitory effects on fusion events in PC12 cells. Although syntaxin isoforms 1a, 3, and 4 are all present at the plasma membrane, the H3 domain from syntaxin 3 was much less effective at inhibiting norepinephrine release than that of syntaxin 4, despite its higher identity to syntaxin 1a (21). Together, these results suggest that, although the H3 domain of syntaxin 1a is sufficient for high-affinity binding to its SNARE partners, other syntaxin isoforms may have sites of SNARE interaction located outside of their equivalent H3 domains. These results also support the idea that the SNARE proteins themselves contribute to the specificity of membrane fusion, adding a level of regulation on top of the fact that specific SNAREs are spatially localized to different subcellular compartments in a pattern characteristic to a given cell type (13, 17). Second, the abundant expression of syntaxin 3 on the cytoplasmic compartment of tubulovesicular membranes (17) suggests that this syntaxin isoform may function in the gastric parietal cell as a v-SNARE instead of a t-SNARE, as is more commonly found. Therefore, it is possible that syntaxin 3 may form a SNARE complex with protein(s) other than SNAP-25 on the apical plasma membrane in the parietal cells. Third, an efficient complex between syntaxin 3 and SNAP-25 may require other binding interface(s) in addition to the H3 domain given its low efficiency in inhibiting norepinephrine release in PC12 cells (21). Distinction among these possibilities will require studies of finer cytolocalization and the deployment of additional peptide domains in permeabilized systems such as developed here. In particular, further studies will attempt to discover the identities of the cognate SNAREs that form a complex with syntaxin 3 at the apical plasma membrane.
In summary, we have established an SLO-permeabilized gastric gland system in which functional transformation of parietal cells can be triggered by addition of cAMP/ATP. The ability to introduce macromolecules easily into this permeable model will facilitate the dissection of molecular interactions underlying parietal cell acid secretion. With this permeabilized cell model, we offered evidence that syntaxin 3 is required for trafficking of tubulovesicles to the apical plasma membrane.
We thank Mark Bennett for the syntaxin constructs and Yu-Tsun Chen, Joseph Eisenberg, and Anastasia Tolles for assistance in experimental procedures.
This work was supported in part by National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-10141 and DK-38972 (to J. G. Forte), National Research Service Award DK-09984-01 (to D. A. Ammar), and Grant DK-56292 (to X. Yao).
Address for reprint requests and other correspondence: X. Yao, Dept. of Molecular and Cell Biology, MC #3200, Univ. of California, Berkeley, CA 94720 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2002 the American Physiological Society