Gastric parietal cells possess an amplified apical membrane recycling system dedicated to regulated apical recycling of H-K-ATPase. While amplified in parietal cells, apical recycling is critical to polarized secretory processes in most epithelial cells. To clarify putative regulators of apical recycling, we prepared immunoisolated parietal cell H-K-ATPase-containing recycling membranes from human stomachs and analyzed protein contents by tryptic digestion and mass spectrometry. We identified and validated by Western blots many of the proteins previously identified on immunoisolated rabbit tubulovesicles, including Rab11, Rab25, syntaxin 3, secretory carrier membrane proteins (SCAMPs), and vesicle-associated membrane protein (VAMP)2. In addition, we detected several previously unrecognized proteins, including Rab10, VAMP8, syntaxin 7, and syntaxin 12/13. We also identified the K+ channel component KCNQ1. Immunostaining of human gastric mucosal sections confirmed the presence of each of these proteins in parietal cells and their colocalization with H-K-ATPase on tubulovesicles. To investigate the role of the identified soluble N-ethylmaleimide-sensitive factor attachment receptor (SNARE) proteins in apical recycling, we transfected them as DsRed2 fusions into an enhanced green fluorescent protein (EGFP)-Rab11a-expressing Madin-Darby canine kidney (MDCK) cell line. Syntaxin 12/13 and VAMP8 caused a collapse of the EGFP-Rab11a compartment, whereas a less dramatic effect was observed in cells transfected with syntaxin 3, syntaxin 7, or VAMP2. The five DsRed2-SNARE chimeras were also transfected into a MDCK cell line overexpressing Rab11-FIP2(129-512). All five of the chimeras were drawn into the collapsed apical recycling system. This study, which represents the first proteomic analysis of an immunoisolated vesicle population from native human tissue, demonstrates the diversity of putative regulators of the apical recycling system.
- vesicle-associated membrane protein
the functional interactions of polarized epithelial cells with their distinct apical and basolateral environments are dictated by the repertoire of pumps, channels, and receptors present on respective plasma membrane surfaces. Investigators over the past decade have increasingly recognized the importance of membrane protein internalization and recycling as a major determinant in epithelial physiology. Regulation of the rapidity of recycling or the distribution of internalized cargoes to recycling or degradative pathways can translate into different levels of dynamic adaptability and signal “throughput.” The gastric parietal cell represents the most prominent example of a regulated apical recycling system. Parietal cells manifest a vast tubulovesicular membrane network, which sequesters H-K-ATPase required for acid secretion in an intracellular compartment within the quiescent parietal cell. Upon stimulation, fusion of this H-K-ATPase-containing tubulovesicular pool with the apical membrane causes a fivefold expansion of the secretory canalicular membrane (5, 6, 63). Several groups have identified some of the components of these tubulovesicles, including Rab11a (11, 26), Rab25 (11, 24), syntaxin 3 (35, 53), and vesicle-associated membrane protein (VAMP)2 (11).
We (11) previously identified proteins on rabbit tubulovesicle membranes immunoisolated using a monoclonal antibody to the α-subunit of gastric H-K-ATPase. We have now expanded those studies by immunoisolating H-K-ATPase-containing tubulovesicle membranes from resting human stomachs obtained at the time of organ donation. We examined the protein constituents of immunosolated human tubulovesicle membranes by analysis of total tryptic digests with two-dimensional liquid chromatography (LC) and tandem mass spectrometry (MS). These experiments revealed numerous groups of regulatory proteins associated with parietal cell apical recycling vesicles, including multiple Rab proteins and soluble N-ethylmaleimide-sensitive factor attachment receptor (SNARE) proteins. These results demonstrate the complex regulation of tubulovesicular membrane trafficking through epithelial apical recycling systems.
MATERIALS AND METHODS
The mouse anti-gastric H-K-ATPase α-subunit antibody (HK 12.18) was as previously described (62, 63). The mouse anti-pan-secretory carrier membrane protein (SCAMP) antibody (7C12) was a gift from Dr. David Castle (University of Virginia, Charlottesville, VA). The chicken anti-pantophysin antibody was a gift of Dr. Rudolf E. Leube (Johannes Gutenberg-University Mainz). The rabbit anti-syntaxin 13 antibody was a gift from Dr. Rytis Prekeris (University of Colorado Health Sciences Center, Denver, CO) (56). The mouse anti-syntaxin 1 antibody was from BD Biosciences; rabbit anti-syntaxin 2, anti-syntaxin 3, and anti-VAMP8 antibodies were from Abcam. Goat anti-syntaxin 7, anti-KCNE2 (MiRP1, C-20), and anti-KCNQ1 (C-20) antibodies were from Santa Cruz Biotechnology. The rabbit anti-VAMP2 antibody was from Stressgen, the rabbit anti-syntaxin 7 antibody was from Synaptic Systems, and the mouse anti-golgin 97 antibody was from Molecular Probes. Rabbit polyclonal antibodies against Rab11a (VU57), Rab11b (VU76), and Rab25 were raised against specific peptide sequences for the COOH-terminal variable domains and showed specificity for each of the Rab11 family members. The following peptides were used to immunize rabbits (Covance): for Rab11a antisera, a peptide spanning amino acids 181–211 (MSDRRENDMSPSNNVVPIHVPPTTENKPKVQC; the C was added at the COOH terminus for conjugation purposes) was used; for Rab11b antisera, a peptide spanning amino acids 188–211 (DISVPPTTDGQRPNKLQKC) was used; and for Rab25 antisera, a peptide spanning amino acids 182-207 (QNSTRTSAITLGNAQAGQDPGLGEKR) was used. Rab11a (VU57) and Rab11b (VU76) antisera demonstrated no additional bands when tested on Western blots (data not shown), so no further purification was performed. Rab25 antisera were purified by affinity chromatography using the same peptide as for immunization. All three antisera were only blocked by their corresponding peptide and only bound to their corresponding green fluorescent protein (GFP) chimera on Western blots (data not shown), thus exhibiting specificity.
Tubulovesicule membrane preparation from the human stomach.
Human stomachs were obtained from four organ donors at the time of organ harvest for cadaveric organ transplantation. The following protocol was approved by the Vanderbilt Medical Center Institutional Review Board, with family informed consent obtained by the Tennessee Donor Services. Patients were maintained on intravenous histamine H2 receptor blockade until the time of organ harvest. All patients were Heliobacter pylori negative, as demonstrated by a negative CLO test (Kimberly-Clark, Roswell, GA) on a sample of antral mucosa. Our four donors had an age range of 20–50 yr old and consisted of three men and one woman. The stomach was opened along the greater curvature and rinsed with 0.9% NaCl. Full-thickness tissue blocks for sectioning and staining were removed (see below), and the fundic mucosa was then dissected away from the submucosa using blunt and sharp dissection. Tubulovesicular membranes were isolated by a modification of procedures developed for the preparation of rabbit tubulovesicles (17). The tissue was weighed, cut into pieces, and rinsed with PBS. Tissue pieces were resuspended at a ratio of 1 g mucosa to 3 volumes of homogenization buffer (113 mM mannitol, 37 mM sucrose, 0.4 mM EDTA, and 5 mM MES; pH 6.7) containing Sigma mammalian protease inhibitor cocktail (1:200) and Sigma phosphatase inhibitor cocktails 1 (1:200) and 2 (1:200) and then dispersed by pulsing in a blender. The dispersed tissue was then homogenized using eight strokes of a Teflon-on-glass Potter homogenizer at 800 rpm with the torque limiter turned off (Master Servodyne, Cole Palmer). The resulting homogenate was centrifuged at 4,000 g for 10 min, and the supernate was transferred to new tubes and centrifuged at 17,000 g for 20 min. The supernate was transferred to new tubes and centrifuged at 100,000 g for 60 min. All centrifugations were performed at 4°C. The final 100,000-g pellet was resuspended to a final volume of 25 ml with resuspension buffer (300 mM sucrose, 0.2 mM EDTA, and 5 mM HEPES; pH 7.4) containing protease and phosphatase inhibitors, transferred to a Teflon-on-glass homogenizer, and resuspended with three strokes at 400 rpm. An aliquot was removed, and the remaining material was layered onto six discontinuous sucrose gradients (20%, 27%, and 33% sucrose in 0.2 mM EDTA and 5 mM HEPES; pH 7.4) and centrifuged at 135,000 g for 2.5 h in a Sorvall Discovery 90SE Ultracentrifuge in a SureSpin 12 ml swinging bucket rotor. After centrifugation, the visible membrane layers at the sucrose interfaces (P20, P27, and P33) were removed, diluted 1:5 with resuspension buffer, and centrifuged at 100,000 g for 1 h. The gradient pellet was washed with resuspension buffer. All pellets were resuspended in resuspension buffer, assayed for protein concentration (Pierce BCA protein assay), aliquoted, and stored at −80°C.
Immunoisolation of H-K-ATPase vesicles.
Human tubulovesicles were immunoisolated using a modification of the methods originally developed for rabbit tubulovesicles (11). Sheep anti-mouse IgG Dynabeads (125 μl, Dynal) were washed three times for 5 min with PBS and then resuspended in 500 μl of PBS, and 1.5 μl of mouse anti-H-K-ATPase (HK 12.18) or mouse IgG2a (Sigma) were added and incubated with rotation overnight at 4°C. Beads were washed three times for 5 min with PBS and then two times for 5 min with 0.2 M triethanolamine (pH 8.2), resuspended in fresh coupling solution [20 mM dimethyl pimelimidate (Pierce) and 0.2 M triethanolamine; pH 8.2] and incubated at room temperature with rotating for 30 min. Beads were quenched by first incubating with rotating for 15 min at room temperature with 50 mM Tris (pH 7.5) and then by incubating for 5 min at room temperature with rotating in 20 mM ethanolamine. Beads were washed a final three times for 5 min with PBS. For the immunoisolation, 25 μg of protein from the P27 membrane fraction were diluted into a final volume of 500 μl PBS with protease inhibitor cocktail (Sigma) added to the antibody-conjugated beads and incubated 2 h at room temperature with rotating. For proteomic analysis, four separate immunoisolations were performed and then combined at the first of the following washes. Beads were washed three times for 20 min in PBS with rotation, and the material was eluted from the beads with 40 μl of 1% CHAPS and 50 mM Tris (pH 7.5) for 15 min with rotating at room temperature. SDS-PAGE sample buffer [200 mM Tris (pH 7.5), 0.9% SDS, 12 mM EDTA, 4% sucrose, 10 mM DTT, and pyronin Y] was added, and samples were heated for 5 min at 70°C and then applied to a 10% SDS-PAGE gel with no stacker. Electrophoresis continued until the dye front had migrated into the gel ∼1–2 cm. The gel was then stained with Gel Code Blue (Pierce), and the protein band was excised from the gel, minced, and washed. Proteins were subjected to in-gel trypsin digestion (31), and the resulting peptides were eluted from the gel and submitted for LC-MS-MS analysis. For direct analysis by Western blotting, antibodies were not covalently attached to the beads, and the beads were either eluted in CHAPS or directly resuspended in the SDS-PAGE sample buffer.
LC-MS-MS analysis and protein identification.
LC-MS-MS analysis of the resulting peptides was performed using a Thermo Finnigan LTQ ion trap mass spectrometer equipped with a Thermo MicroAS autosampler and Thermo Surveyor HPLC pump, Nanospray source, and Xcalibur 1.4 instrument control. Peptides were separated on a packed capillary tip (100 μm × 11 cm) with C18 resin (Monitor C18, 5 μm, 100 Å, Column Engineering, ON, Canada) using an in-line solid-phase extraction column that was 100 μm × 6 cm packed with the same C18 resin [using a frit generated with liquid silicate Kasil 1 (16)] similar to that previously described (40) except the flow from the HPLC pump was split prior to the injection valve. The flow rate during the solid phase extraction phase of the gradient was 1 ml/min and during the separation phase was 700 nl/min. Mobile phase A was 0.1% formic acid, and mobile phase B was acetronitrile with 0.1% formic acid. A 95-min gradient was performed with a 15-min washing period (100% A for the first 10 min followed by a gradient to 98% A at 15 min) to allow for solid phase extraction and removal of any residual salts. After the initial washing period had passed, a 60-min gradient was performed where the first 35 min was a slow, linear gradient from 98% to 75% A, followed by a faster gradient to 10% A at 65 min and an isocratic phase at 10% A to 75 min. MS-MS spectra of the peptides were performed using data-dependent scanning in which one full MS spectrum, using a full mass range of 400–200 amu, was followed by 3 MS-MS spectra. In addition to this one-dimensional LC-MS analysis, peptides were also subjected to two-dimensional LC-LC-MS analysis in which the peptides were first fractionated using strong cation exchange chromatography (100 μm × 7cm Luna SCX column, Phenomonex, Torrance, CA) using a 0–500 mM (pH 3.0–8.0) ammonium formate gradient in 25% acetonitrile as described previously (1) except a final salt bump of 0.5 M ammonium formate during the last 10 min of the 65 min was substituted for the KCl bump. After collection of the flowthrough fraction, each of the nine separate fractions was collected, the last three fractions were combined, and each fraction (including the flow through fraction) was subjected to a reverse-phase separation directly inline with the LTQ as described above. Proteins were identified using the cluster version of the SEQUEST algorithm (72) using the human subset of the Uniref 100 database (www.uniprot.org). The database was concatenated with the reverse sequences of all the proteins in the database to allow for the determination of false positive rates. Protein matches were preliminarily filtered using the following criteria: if the charge state of the peptide is 1, the cross-correlation score (xcorr) is ≥1, the ranking based on preliminary score (RSp) is ≤5, and the preliminary score (Sp) is ≥350. If the charge state is 2, the xcorr is ≥1.8, the RSp is ≤5, and the Sp is ≥350. If the charge state is 3, the xcorr is ≥2.5, the RSp is ≤5, and the Sp is ≥350. Once filtered based on these scores, all protein matches that had less than two peptide matches were eliminated. These filtering criteria achieved a false positive rate of <1% in all datasets.
Immunostaining of the paraffin-embedded human stomach.
Full-thickness tissue samples were dissected from the fundus for paraffin embedding. The tissue was fixed in 4% paraformaldehyde, 30% sucrose, and PBS overnight at 4°C. The tissue was then washed three times for 10 min in PBS, placed in a cassette, and stored in 70% ethanol until paraffin embedding and sectioning were carried out in the Vanderbilt Human Tissue Acquisition and Pathology Shared Resource Core. The 5-μm paraffin-embedded sections were warmed for 30 min at 58°C and then allowed to cool for another 30 min. Sections were dewaxed three times for 10 min in Histo-Clear (National Diagnostics) and then rehydrated three times for 5 min in 100% ethanol, two times for 5 min in 95% ethanol, one time for 5 min in 75% ethanol, one time for 5 min in water, and three times for 5 min in PBS. Sections were blocked for 30 min in 1% normal donkey serum and PBS. Primary antibodies were diluted in 0.1% normal donkey serum and 0.1% Tween 20-PBS and incubated overnight at 4°C in a humidified chamber. Sections were washed three times for 10 min in 0.1% Tween 20 and PBS, and secondary antibodies (Alexa 488 anti-mouse and Alexa 568 anti-rabbit, anti-goat, or anti-chicken; Molecular Probes) were diluted 1:400 in 0.1% normal donkey serum, 0.1% Tween 20, and PBS and incubated for 1 h at room temperature. Sections were washed as described for the primary antibodies with an additional PBS wash, rinsed in water, and then mounted with ProLong Gold anti-fade with 4′,6-diamidino-2-phenylindole (DAPI; Molecular Probes). Sections were imaged on an Olympus FV1000 confocal scanning microscope.
Isolation and staining of human gastric glands.
Human gastric glands were isolated using a modification of a method initially described for rabbits (4). A piece of the fundic mucosa that had been grossly dissected away from the submucosa using blunt and sharp dissection was further dissected away from the submucosa with sharp dissection. The tissue was minced in PBS and rinsed three times with PBS. The finely minced tissue was transferred to a round bottom flask with a stirring star containing 50 ml NRMc [114.4 mM NaCl, 5.4 mM KCl, 5.0 mM Na2HPO4, 1.0 mM NaH2PO4, 1.2 mM MgSO4, 15 mM HEPES, 1 mM CaCl2, and 10 mg/l phenol red (pH 7.4), 0.2% BSA, 0.2% glucose, and 1 mM pyruvate sodium salt] with 0.04 g collagenase (Sigma blend) and 0.025 g pronase (Calbiochem). The flask was incubated at 37°C for 30 min with constant stirring and flushed with 100% O2. The pH was maintained during the incubation by the addition of 1 N NaOH. The incubation was terminated by the addition of 50 ml NRMc with 1 mM DTT. The solution was transferred to 50-ml tubes and spun for 1 min at 1,500 g at room temperature. The glands were washed twice with NRMc, brought up to 10 ml with NRMc, and split into two 14-ml round bottom tubes with a snap cap and a micro-stir bar. One tube received 100 μM ranitidine (Sigma; resting), and the other tube received 100 μM histamine (Sigma; stimulated). Tubes were incubated at 37°C for 30 min with constant stirring. The glands were washed two times with PBS, fixed for 15 min at room temperature with 4% paraformaldehyde, and washed three times with PBS. The glands were stored at 4°C in PBS with 0.002% sodium azide until needed. For immunostaining, 50 μl of the gland preparation were spotted onto Colormark plus microscope slides (Erie Scientific) and dried on a slide warmer at 37°C. The dried spots were rehydrated with water and then rinsed in PBS. The glands were then stained as for tissue culture cells below.
The human sequences of the following proteins were amplified using Pfu DNA polymerase (Promega) and subcloned into pDsRed2-C1 (Clontech). The corresponding Image clones used (American Type Culture Collection) were as follows: syntaxin 1a (2820551), syntaxin 3 (3939676), syntaxin 7 (5469250), syntaxin 12 (3851266), VAMP2 (5781966), and VAMP8 (3538672). All constructs were sequenced at the Vanderbilt Sequencing Core to assure no mutations had been introduced.
Cell culture and transfections.
Parent Madin-Darby canine kidney (MDCK)-II cells and MDCK cells stably expressing enhanced GFP (EGFP)-Rab11a (38) were grown in DMEM supplemented with 10% heat-inactivated FBS, 2 mM l-glutamine, and penicillin-streptomycin solution. The EGFP-Rab11a cell line was also supplemented with G-418 (0.5 mg/ml). Cells were split 1:20 onto 12-mm (0.4-μm pore) Transwell-clear filters (Costar, clear polyester) and allowed to polarize for 4 days before being treated and stained. The EGFP-Rab11-FIP2(129-512) (30) cell line was grown as for the EGFP-Rab11a cell line with the addition of 20 ng/ml doxycycline to inhibit transgene expression. For experiments, cells were plated and kept in media without doxycycline and with 10% FBS certified without tetracycline (Hyclone). The DsRed2-chimeric constructs were transiently transfected into cells at the time of plating using Effectine (Qiagen).
Fixation and immunocytochemistry.
Cells were washed with PBS and then fixed with 4% paraformaldehyde for 15 min at 4°C. Cells were permealized, blocked with 1% normal donkey sera (Jackson Immunological) and 0.3% Triton X-100 in PBS for 20 min, incubated with primary antibodies diluted in 0.1% normal donkey serum, 0.1% Tween 20, and PBS, and then incubated for 2 h at room temperature or overnight at 4°C. Cells were washed three times for 10 min with 0.1% Tween 20 and PBS and then incubated with the appropriate secondary antibodies (Cy5; Jackson ImmunoResearch) diluted 1:200 in 0.1% normal donkey serum, 0.1% Tween 20, and PBS for 1 h at room temperature. Cells were washed three times for 10 min with 0.1% Tween 20 and PBS, one time for 10 min with PBS, rinsed in water, and mounted in Prolong Gold anti-fade with DAPI (Molecular Probes). For the visualization of actin, Alexa 647-phalloidin was added (1:200) with the secondary antibodies. MDCK cells and EGFP-chimeric cell lines were imaged with either a Zeiss 510 Meta or an Olympus FV1000 scanning confocal fluorescence microscope.
Western blot analysis.
Protein samples were resolved on 12% SDS-PAGE gels following a standard Laemmli protocol (36). All incubations were performed at room temperature. Proteins were transferred onto Immobilon-P membranes (Millipore) and blocked for 30 min with 5% dry milk powder (DMP) and Tris-buffered saline (TBS) with 0.05% Tween 20 (TBST). Blots were incubated with primary antibody diluted in 2.5% DMP and TBST for 1 h, washed four times for 10 min in 2.5% TBST, and then incubated for 1 h with horseradish peroxidase-conjugated donkey anti-rabbit, anti-mouse, anti-chicken, or anti-goat IgG (Jackson Immunological) diluted 1:10,000 in 1% DMP and TBST. Blots were then washed three times with TBST and then one time with TBS, and specific labeling was detected by enhanced chemiluminescence (Supersignal, Pierce) with autoradiography using Kodak BioMax ML film.
Analysis of immunoisolated tubulovesicle by protein MS.
We obtained gastric fundic mucosa from four separate organ donors. Since the subjects were maintained on intravenous histamine H2 receptor blockade prior to organ harvest, resting morphology was observed in gastric parietal cells (Fig. 1A). The human stomach submucosa was fractionated in a manner similar to previously published methods developed in the rabbit (11, 68). The majority of H-K-ATPase and Rab11a was found in the lighter membranes partitioning at the 20% and 27% sucrose interfaces (Fig. 1B). While some of the Rab11b was also seen in these fractions, the majority was present in a heavier fraction, as previously seen in the rabbit (38). Since the majority of the Golgi membrane immunoreactivity for golgin 97 was found in the P20 sucrose fraction (Fig. 1B), the P27 sucrose interface membrane fraction was used as the starting material for the immunoisolations.
We immunoisolated H-K-ATPase-containing tubulovesicles from four separate donor stomachs using a monoclonal antibody against the α-subunit of H-K-ATPase, and vesicle proteins were eluted from the magnetic beads by solublization in 1% CHAPS. This procedure left the majority of H-K-ATPase still associated with the magnetic beads with the antibody (data not shown). Eluted proteins were electrophoresed into SDS-PAGE gels, and in-gel tryptic digestion was then performed. Tryptic peptides were analyzed by LC-MS-MS. As an internal criterion for analytical validity, only preparations that demonstrated the presence of intrinsic factor within the preparation were studied. Intrinsic factor is a relatively low-abundance protein that is contained within the lumen of human parietal cell tubulovesicles (61). Table 1 shows all potential vesicle proteins, and Table 2 shows all potential cargo and membrane proteins, identified in at least three of the four donors. Both tables show the confidence level of the protein identification as predicted by the ProteinProphet program (50) (only proteins with a confidence level of ≥0.9 were included), numbers of unique peptides identified, and numbers of donors where peptides corresponding to that protein were observed. Due to the size of the complete list containing all categories [membrane associated, trafficking, endoplasmic reticulum (ER)/Golgi, mitochondrial, cytoskeletal, signaling, and unknowns], the full table of data is presented in Supplemental Table 1. An alphabetical listing of the identified proteins with the corresponding Uniref accession number is contained in Supplemental Table 2. This table also shows all the peptide identification information, including the peptide sequence, location within the protein, the xcorr, and the charge state for each peptide. Each peptide identified was run back through the UniRef 100 database, and any alternative UniRef protein identification for that peptide is shown in Supplemental Table 2.
We identified many proteins previously associated with rabbit tubulovesicles. One of these is clathrin, a known vesicle protein, which had previously been localized to rabbit and hog gastric tubulovesicles (48, 52). For the highly homologous members of the Rab11 family, we identified 1 peptide shared by all 3 of the family members, 10 identified peptides corresponding to either Rab11a or Rab11b, 3 identified peptides specific for Rab11a, 2 identified peptides specific for Rab11b, and 6 identified peptides specific for Rab25 (Table 1). Besides members of the Rab11 family, we identified three other Rab small GTPases: Rab1a, Rab5c, and Rab10. VAMP2, previously observed on rabbit tubulovesicles (11, 53), was also observed on human tubulovesicles along with VAMP3 and VAMP8 (Table 1). As seen with the Rab11 family, we identified both shared and specific peptides for the homologous VAMP family members. While we identified one peptide that could belong to VAMP1, VAMP2, or VAMP3, we never saw a specific peptide for VAMP1. We identified two peptides shared by VAMP2 and VAMP3, one specific peptide for VAMP2, and another specific peptide for VAMP3. We concluded that both VAMP2 and VAMP3 potentially are contained on H-K-ATPase-containing vesicles, but probably not VAMP1. Peptides for the Rab11 and VAMP families were seen in all donors, and Table 1 shows the number of donors the specific or shared peptide was observed in.
Previously, SCAMPs have been identified on rabbit tubulovesicles (11, 53) using a pan-SCAMP antibody. In human immunoisolated tubulovesicles, we identified peptides from SCAMP1, SCAMP2, and SCAMP3 (Table 1). While previous studies had identified syntaxin 3 on rabbit gastric tubulovesicles (11, 35, 53), in this work on human gastric tubulovesicles, we observed not only syntaxin 3 but also syntaxin 7 and syntaxin 12/13. We also identified other SNAREs and vesicle trafficking proteins not previously recognized as on H-K-ATPase-containing membranes, including α-soluble N-ethylmaleimide-sensitive fusion protein (NSF) attachment protein, synaptogyrin-2, pantophysin, EPS15 homology domain-containing 2, and vacuolar ATPase.
Table 2 lists potential cargo and membrane proteins that were identified, including both the α- and β-chains of H-K-ATPase. Previous investigations have implicated the presence of K+ channel KCNQ1/KCNE2 in H-K-ATPase-containing tubulovesicles in the rat (33, 37). In this survey of human H-K-ATPase-containing tubulovesicles, we identified KCNQ1, and, in two of the four donors, we observed peptides for KCNE2 (since this did not meet our criteria, KCNE2 was not included in our final tables). Some potential contamination by gastric chief cell proteins was observed, including identification of anterior gradient homolog 2, carboxypeptidase D, gastric lipase, pepsinogen, and pepsin A. These results indicate that further validation of the identified proteins is needed to assign a specific protein to H-K-ATPase-containing tubulovesicles with high confidence.
Analysis of immunoisolated H-K-ATPase-containing membranes by Western blots.
Since proteomics is sensitive but not quantitative, we performed independent immunoisolations and assayed the results by Western blot analysis to validate and assess the amount of protein immunoprecipitated with vesicles. As shown in Fig. 2, the majority of H-K-ATPase, Rab11a, Rab25, syntaxin 3, syntaxin 7, syntaxin 12/13, VAMP8, pantophysin, and KCNE2 were almost entirely immunoisolated with the H-K-ATPase antibody. We observed partial immunoisolation of VAMP2 and SCAMPs. In contrast, we observed little immunoisolation of Rab11b and syntaxin 1. Whereas calreticulin and other ER proteins were detected in immunoisolated vesicles by MS (Supplemental Table 1), Western blot analysis demonstrated that little calreticulin was isolated with H-K-ATPase-containing vesicles. These results are consistent with the presence of some H-K-ATPase with the ER in the P27 membrane fraction used for the immunoisolation.
Localization of tubulovesicle proteins by immunocytochemistry in sections of the human stomach.
Where possible, we performed immunocytochemical localization of tubulovesicle proteins in sections of the human stomach. We first stained human stomach sections with antibodies to syntaxin 3, syntaxin 7, syntaxin 12/13, and VAMP8 as well as Rab11a (Fig. 3A). As previously reported for Rab11a in rabbit gastric glands (25, 26) and for syntaxin 3 in rabbit parietal cells (12, 35, 53), both colocalized with H-K-ATPase in the tubulovesicular compartment. Syntaxin 3 also localized to small triangular cells that were not positive for H-K-ATPase in human stomach sections (Fig. 3A, arrow). These cells were morphologically similar to enterochromaffin-like (ECL) cells. Syntaxin 7 appeared to be parietal cell specific and also localized with H-K-ATPase (Fig. 3A). Syntaxin 12/13 was also parietal cell specific in the human stomach, but, unlike the other syntaxins, it only partially colocalized to the tubulovesicular compartment with H-K-ATPase (Fig. 3A). VAMP8 exhibited less intense staining in human stomach sections that was parietal cell specific and colocalized with H-K-ATPase.
Previous investigations have implicated the presence of the K+ channel KCNQ1/KCNE2 in H-K-ATPase-containing tubulovesicles in the rabbit (33, 37). In our proteomic survey of human H-K-ATPase-containing tubulovesicles, we identified KCNQ1 and, with less confidence, KCNE2. To see if these K+ channel components were associated with H-K-ATPase-containing tubulovesicles in human parietal cells, we stained human stomach sections for KCNQ1 and KCNE2 (Fig. 3B). Both K+ channel subunits colocalized with H-K-ATPase in the tubulovesicular compartment.
To explore further the observation of peptides for the ER protein calreticulin, we also stained human stomach sections for calreticulin (Fig. 3B). As with the immunoprecipitation (Fig. 2), calreticulin did not colocalize with H-K-ATPase, again indicating that proteins from a proteomic list must be further validated prior to actual assignment of a protein to a particular cellular compartment.
Localization of syntaxins in human glands following stimulation with histamine.
To ascertain the dynamic localization of the identified proteins upon stimulation, we prepared isolated fundic glands from human stomachs and treated them with either ranitidine or histamine for 30 min prior to fixation. As seen in Fig. 4, both syntaxin 3 and syntaxin 7 localized with H-K-ATPase to tubulovesicles in the resting glands and both SNAREs relocated with the proton pump to the apical canaliculus in stimulated glands. As in stomach sections, syntaxin 3 stained other cells, most probably ECL cells, in the gastric glands, whereas syntaxin 7 appeared to be parietal cell specific. Syntaxin 12/13 again only partially colocalized with H-K-ATPase in the resting glands and appeared to have greater colocalization with H-K-ATPase in the apical canaliculus after stimulation.
Association of parietal cell apical recycling system proteins with the apical recycling system in MDCK cells.
Since primary parietal cells are difficult to culture and transfect and to examine the generalizability of our observations in parietal cells to another model of apical recycling, we used MDCK cells to examine targeting of DsRed2-chimeric constructs. We transiently overexpressed DsRed2 chimeras of syntaxin 3, syntaxin 7, syntaxin 12/13, VAMP2, and VAMP8 and assessed whether they could alter the localization of EGFP-Rab11a (Fig. 5). The DsRed2-syntaxin 7 construct drew the EGFP-Rab11a into tight disklike membranous cisternae, whereas the DsRed2-syntaxin 3 construct colocalized with EGFP-Rab11a in looser aggregates. The DsRed2-syntaxin 12/13 construct drew some of the EGFP-Rab11a into small tight structures, indicating potentially different functions along the Rab11a-guided apical recycling pathway. Both DsRed2-VAMP2 and DsRed2-VAMP8 constructs also caused the accumulation of EGFP-Rab11a into loose aggregates similar to the structures observed with the DsRed2-syntaxin 3 construct (Fig. 5).
To determine further whether these newly identified protein components of H-K-ATPase-containing tubulovesicles could function in apical recycling in other polar cells, we collapsed the Rab11a apical recycling endosome with a truncated EGFP-Rab11-FIP2(129-512) chimeric protein (29) and assessed the effects of this dominant negative recycling inhibitor on the expressed DsRed2 chimeras of syntaxin 1, syntaxin 3, syntaxin 7, syntaxin 12/13, VAMP2, and VAMP8. We (29, 30) have previously shown that this truncated Rab11-FIP2 construct can alter the apical recycling compartment and pull known members (i.e., Rab11a and pIgA) into this compartment; if the identified SNAREs are present along the apical recycling pathway in MDCK cells, we would expect to observe them similarly pulled into this collapsed compartment. Syntaxin 1, which did not coisolate with H-K-ATPase (Fig. 2) (12) did not localize with the truncated EGFP-Rab11-FIP2 chimera (Fig. 6). DsRed2-syntaxin 7 and DsRed2-VAMP8 constructs colocalized with the EGFP-FIP2(129-512) chimera in tight collapsed cisternae. While the DsRed2 chimeric constructs of syntaxin 3, syntaxin 12/13, and especially VAMP2 colocalized with the EGFP-FIP2(129-512) chimera, they also expanded the cisternae into a looser aggregation of vesicles.
The gastric parietal cell functions as a massively amplified apical recycling system specialized to provide second messenger-regulated delivery of H-K-ATPase to the apical secretory canaliculus of the parietal cell. The intracellular morphology of the parietal cell is unique, with intracellular tubulovesicular membranes containing H-K-ATPase underlying a widely ramifying intracellular canalicular membrane invagination from the apical surface. Stimulation of parietal cells with secretory agonists such as histamine leads to expansion of the apically oriented secretory canaliculus and delivery of functional H-K-ATPase to the canalicular membrane (5, 6, 22, 68). While others have previously suggested that the expansion of the secretory canaliculus with acid secretion was caused by an osmotic swelling of the collapsed membrane (54, 55), a more recent study (71) has supported the viewpoint that fusion of tubulovesicular elements into the secretory canaliculus is the central event in presentation of H-K-ATPase at the parietal cell apical surface. The so-called “membrane recycling hypothesis” was first proposed by John Forte in 1977 (22). While the original hypothesis was conceived based on morphological evidence, studies over the past decade, predominantly in rabbit parietal cells, have increasingly identified crucial regulators of membrane trafficking and recycling enriched in isolated tubulovesicular membranes (11, 12, 35, 53). Most prominently, the enrichment of Rab11a in gastric tubulovesicles membranes (26) established these membranes as a highly amplified recycling membrane population. Indeed, subsequent studies have established Rab11a as a preeminent marker of plasma membrane recycling vesicles in a vast number of systems (9, 10, 12, 13, 20, 28, 69). In addition, the identification of the SNARE proteins VAMP2 and syntaxin 3 on immunoisolated rabbit tubulovesicle membranes strongly implicated regulated docking and fusion events in the process of H-K-ATPase exocytosis (11, 35, 53). The present investigation of immunoisolated human parietal tubulovesicles has now established an even greater diversity of vesicle trafficking proteins, with at least two vesicle-vesicle SNAREs (v-SNAREs) and three vesicle-target SNAREs (t-SNAREs) present on isolated membranes. The proteomic analysis also established the presence of a number of further putative regulators of vesicle trafficking and fusion, including Rab10, pantophysin, synaptogyrin 2, Mal2, and SCAMPs. All of these investigations support the presence of a complex compendium of vesicle trafficking regulators on the highly amplified parietal cell apical recycling system.
Previous investigations of parietal cell tubulovesicle proteins were generally guided by a candidate approach. The original identificationof Rab11a and Rab25 in rabbit parietal cells was made from rapid amplification of cDNA ends studies (24, 66) of transcripts coding for Rab proteins in mRNA prepared from >95% pure rabbit parietal cell preparations. Similarly, the identification of VAMP2, syntaxin 3, and SCAMPs was predicated on the probing of tubulovesicle preparations with available antibodies (11, 35, 53). The proteomic analysis of immunoisolated tubulovesicles has yielded a broader group of putative regulatory proteins, many of which would not have been predicted a priori. We validated a number of these proteins using available commercial and collaborative antibody preparations.
However, these studies using shotgun proteomics underline many of the benefits and drawbacks of proteomic analyses of isolated organelles. First, while the immunoisolation protocol used a specific monoclonal antibody against H-K-ATPase and a membrane preparation highly enriched for H-K-ATPase recycling vesicles, ER and mitochondrial markers were clearly present. While the vast majority of the ER marker calreticulin did not immunoisolate with the H-K-ATPase antibody, a small amount was detected in the immunisolated fraction (Fig. 2), perhaps due to the presence of newly synthesized proton pump molecules in these membranes. The proteomic list thus represents a starting point for future studies, and Western blot analysis with independent immunoisolations and immunocytochemical colocalization with H-K-ATPase must be performed to validate all proteins with secondary confirmation. Second, the topological orientation of the HK 12.18 monoclonal antibody epitope used for membrane immunoisolation (62) ensured that only cytoplasmic side-out tubulovesicles were recovered in our protocol. Tubulovesicles that undergo eversion and luminal side-out refusion during homogenization were excluded from immunoprecipitation, and so their vesicular cargo was not detected in our analysis. Third, variation in the membrane preparations, owing to their derivation from intact human tissues, may alter the identification of some proteins. Since all of the organ donor subjects were maintained on histamine H2 receptor blockade prior to organ harvest, the majority of the parietal cells were maintained in the resting state. This was critical for the yield of immunoisolated tubulovesicle membranes. We anticipated that the greatest variations in yield would accrue in peripherally associated proteins, and, indeed, the majority of identified proteins were integral membrane proteins. Fourth, because H-K-ATPase represents such a high percentage of the total protein in the preparation, we chose to elute proteins from immunoisolation beads with 1% CHAPS, thus leaving the majority of the total proton pump still attached to the antibodies on the beads. This process decreased the homogenous signal from the proton pump and increased the yield of minor peptides. Nevertheless, this procedure likely leaves behind proteins complexed directly with the proton pump. These potential components of the recycling machinery are the focus of ongoing investigations. Clearly, some tubulovesicle proteins, including Rab11-FIPs and myosin Vb, are not represented here. Peptides for Rab11-FIP1, Rab11-FIP2, and myosin Vb were occasionally detected, but not in a consistent enough manner to warrant inclusion on the list. We also did not identify other proteins previously associated with H-K-ATPase tubulovesicles including the Cl− channel ClC-2 (34, 60) or the tetraspanin CD63 (18) (although we did identify another unknown, tetraspanin 8; Supplemental Table 1. No proteome is truly complete, and in all proteomic studies, the identification of proteins is dependent on a number of factors including abundance, digestion patterns, and the solubility, charge, and size of the tryptic peptides. We continue to develop further modifications of these procedures to expand the coverage of identified proteins.
The interactions of v-SNAREs and t-SNAREs are associated with vesicle-vesicle and vesicle-target membrane docking and fusion events (19, 21, 32, 45). The present investigations indicate the apical recycling system of parietal cells contains a broader repertoire of SNARE proteins than previously recognized. Importantly, we also provided evidence that these three t-SNAREs and two v-SNAREs are also associated with the apical recycling system in polarized MDCK cells. Previous studies have noted the presence of syntaxin 3 in tubulovesicles (11, 35, 53) as well as in the recycling membranes in hepatocytes (23) and polarized MDCK cells (43). In MDCK cells, syntaxin 3 has been implicated in trafficking from the trans-Golgi network to the apical membrane (67) and is involved in the establishment and maintenance of polarity (44, 59). Our biochemical confirmation of syntaxin 3 on human gastric parietal cell tubulovesicles is evidence of the validity of the proteomic analysis.
The association of syntaxin 7 and syntaxin 12/13 with polarized recycling systems is less well established. Previous investigations have noted that syntaxin 7 regulates trafficking from early endosomes to late endosomes (49) and complexes with VAMP8 (70) in nonpolar cells. In a gastric epithelial cell line, syntaxin 7 localized to the vacuoles induced by the VacA cytotoxin of H. pylori, and mutant syntaxin 7 blocked VacA-induced formation of endosome/lysosome hybrids (64). In this work, we have shown that syntaxin 7 is involved in apical recycling in polar cells. In nonpolar cells, syntaxin 12/13 localizes to early and recycling endosomes and complexes with VAMP2 (56, 65). In PC12 cells, syntaxin 13 localizes with the transferrin receptor, and the addition of an anti-syntaxin 13 antibody to permeabilized PC12 cells inhibits transferrin receptor recycling (56). Syntaxin 13 appears to play a role in early endosome fusion (8) and is recruited to the docking and fusion complex on the membrane through binding to early endosome antigen 1 (47). Both syntaxin 7 and syntaxin 13 are involved in the maturation of the phagosome, exerting their influences at different times: syntaxin 13 is acquired early and rapidly recycles off the phagosome, whereas syntaxin 7 is recruited later and continues to accumulate throughout maturation (15).
Previous investigations (11, 53) have demonstrated the association of VAMP2 with parietal cell tubulovesicle membranes. The present study demonstrated that VAMP8 is also strongly enriched on parietal cell tubulovesicles. Since syntaxin 12/13 can associate with VAMP2 (56) and syntaxin 7 can associate with VAMP8 (70), the presence of these SNARE pairs appears appropriate. Nevertheless, the exact make up of SNARE docking complexes remains unclear, as do their roles in regulating specific aspects of trafficking through plasma membrane recycling systems.
As noted above, studies have established the critical role of Rab11a in regulating apical recycling. We did identify all three Rab11 family members as well as three other Rab proteins: Rab1a, Rab5c, and Rab10. Rab1a may be involved with newly synthesized H-K-ATPase in the ER, similar to the presence of other ER proteins, as seen in Supplemental Table 1. The presence of Rab5c could indicate the presence of some H-K-ATPase-containing vesicles recycling through the early endosome, since treatment with the histamine H2 blockers would induce recycling in previously stimulated parietal cells (Fig. 1A). While the exact roles of Rab10 are less well established, recent work from two groups has placed Rab10 within the endocytic recycling system in Caenorhabditis elegans intestinal cells (14) and in MDCK cells (2). How specific small GTPases may interact with particular SNARE assemblies remains to be determined.
The full compendium of protein cargoes trafficking through the apical recycling system is unclear. A previous study (22) has established that H-K-ATPase constitutes 85% of the total plasma membrane protein. In addition, other studies (58, 61) have documented intrinsic factor as a luminal secretory protein within the tubulovesicles. While we did observe these two established cargoes, we also identified other putative cargoes. Recent investigations have suggested that a K+ channel composed of a heterodimer of KCNQ1 and KCNE2 is present in rat and mouse tubulovesicle membranes and may represent the pathway for K+ conductance through the parietal cell apical membrane (27, 33, 37). While these investigations demonstrated the presence of KCNQ1 in enriched parietal cell tubulovesicle membranes and immunocytochemical colocalization with H-K-ATPase, they did not examine immunoisolated populations of vesicles. The present study using immunoisolated human tubulovesicle membranes indicates that KCNQ1 and KCNE2 are indeed present on tubulovesicles. We also observed immunocytochemical colocalization of both KCNQ1 and KCNE2 with H-K-ATPase in sections of the human stomach. These results suggest that the K+ channel KCNQ1/KCNE2 is also a cargo in tubulovesicles and likely represents a major pathway for apical K+ extrusion during parietal cell stimulation. We did not identify another K+ channel previously identified on rabbit gastric tubulovesicles, Kir2.1(46). The identities of other cargoes remain to be determined. Proteins such as the scavenger receptor may also represent cargoes in tubulovesicles, but confirmation of these proteins awaits the development of appropriate antibody reagents. Finally, it is notable that a number of proteins traditionally associated with chief cell granules were present on the proteomic list, including pepsinogen and gastric lipase. So far, we have not confirmed the presence of these proteins in immunoisolated vesicles by Western blot analysis due to inadequacies in the antibody preparations presently available. However, it is notable that chief cell granule membrane markers such as Rab3d (51) were not found in our proteomic study. It thus seems possible that some proteins released into the lumen by chief cells located at the base of glands could be internalized into tubulovesicles during endocytosis. We identified other proteins that have been localized to other membrane compartments, such as the basolateral proteins Na-K-ATPase and transferrin receptor and the Golgi protein mannose-6-phosphate receptor. While these proteins appear to be abundant by the number of proteins detected (Supplemental Table 1), their presence could be due to small contamination from the ER/Golgi compartment, similar to what we observed with calreticulin (Figs. 2 and 3). Another possibility is the presence of a sorting recycling endosomal compartment where H-K-ATPase mixes with other cargoes prior to being sorted along the specific pathway.
As noted above, the massively amplified apical recycling system in parietal cells is highly enriched for apical recycling proteins, including Rab11a (19, 26). Indeed, the enrichment of apical recycling systems has served as an important source for the identification of regulators of apical recycling. The critical actin motor myosin Vb and the first of the Rab11 family interacting proteins, Rab11-FIP1A, were identified thorough two-hybrid screening of a rabbit parietal cell cDNA library (29, 39). These proteins, which were first demonstrated in association with parietal cell tubulovesicles, were then recognized as generalized regulators of apical recycling in polarized epithelial cells and plasma membrane recycling in nonpolarized cells (29, 39, 41, 42, 57). Similarly, in the present study, we observed that VAMPs and syntaxins associated with tubulovesicles were also associated with the apical recycling system in MDCK cells. While the complete list of putative tubulovesicle-associated proteins may contain a number of common regulators of recycling, we also anticipate that some of the regulators will be specific to parietal cell function. Indeed, it is notable that, while much of epithelial cell recycling is constitutive or dependent on receptor/cargo-induced internalization, H-K-ATPase delivery to the canalicular membrane is tightly regulated by second messenger levels, most prominently intracellular cAMP and Ca2+. Thus, entries in the parietal cell tubulovesicle proteome list may provide clues to mediators of regulated recycling. We anticipate that comparisons of the human parietal cell tubulovesicle proteome with those for other regulated recycling membranes, such as those trafficking aquaphorin-2 (3) or glucose transporter 4 (7), may yield the identities of mediators of regulated recycling pathways. Further investigations based on the wealth of information available from proteomic studies will be necessary to define how proteins identified in these studies cooperate to regulate apical recycling.
This work was supported by National Institutes of Health (NIH) Grants DK-070856, DK-48370, and DK-43405 (to J. R. Goldenring) and DK-34092, DK-43138, and DK64371 (to A. J. Smolka). Confocal fluorescence imaging was performed in part through the use of the Vanderbilt University Medical Center Cell Imaging Shared Resource, which was supported by NIH Grants CA-68485, DK-20593, DK-58404, and HD-15052. The authors also thank Vanderbilt University and the Vanderbilt-Ingram Cancer Center for institutional support of the Proteomics Laboratory in the Mass Spectrometry Research Center through the Academic Venture Capital Fund.
This study is dedicated to the generosity of families consenting for donation of transplantable organs and the tireless efforts of the members of the Tennessee Donor Services. We thank Dr. Ravi Chari for the assistance in establishment and implementation of the stomach donation protocol. We thank Drs. Rytis Prekeris, Rudolf Leube, and David Castle for the gifts of antibody reagents. We thank both Jade Johnston and Kristin Cheek for the technical assistance in the Proteomics Laboratory in the Mass Spectrometry Research Center and Dr. Jeff Franklin for insightful discussions.
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 © 2007 the American Physiological Society