Cholinergic agonists are major stimuli for fluid secretion in parotid acinar cells. Saliva bicarbonate is essential for maintaining oral health. Electrogenic and electroneutral Na+-HCO3− cotransporters (NBCe1 and NBCn1) are abundant in parotid glands. We previously reported that angiotensin regulates NBCe1 by endocytosis in Xenopus oocytes. Here, we studied cholinergic regulation of NBCe1 and NBCn1 membrane trafficking by confocal fluorescent microscopy and surface biotinylation in parotid epithelial cells. NBCe1 and NBCn1 colocalized with E-cadherin monoclonal antibody at the basolateral membrane (BLM) in polarized ParC5 cells. Inhibition of constitutive recycling with the carboxylic ionophore monensin or the calmodulin antagonist W-13 caused NBCe1 to accumulate in early endosomes with a parallel loss from the BLM, suggesting that NBCe1 is constitutively endocytosed. Carbachol and PMA likewise caused redistribution of NBCe1 from BLM to early endosomes. The PKC inhibitor, GF-109203X, blocked this redistribution, indicating a role for PKC. In contrast, BLM NBCn1 was not downregulated in parotid acinar cells treated with constitutive recycling inhibitors, cholinergic stimulators, or PMA. We likewise demonstrate striking differences in regulation of membrane trafficking of NBCe1 vs. NBCn1 in resting and stimulated cells. We speculate that endocytosis of NBCe1, which coincides with the transition to a steady-state phase of stimulated fluid secretion, could be a part of acinar cell adjustment to a continuous secretory response. Stable association of NBCn1 at the membrane may facilitate constitutive uptake of HCO3− across the BLM, thus supporting HCO3− luminal secretion and/or maintaining acid-base homeostasis in stimulated cells.
- protein kinase C
- muscarinic type-3 receptor
saliva is essential for maintaining oral health and function (57). Impairment of isotonic fluid secretion in parotid acinar cells causes dry mouth syndrome in millions of patients (41). The parasympathetic nervous system is responsible for the production of salivary gland fluid and electrolyte secretion following stimulation by the neurotransmitter ACh (6). The ACh signaling via the muscarinic type-3 receptors (M3R) is primarily responsible for the initiation of fluid and ion secretion in parotid acinar cells (11, 62).
Normal bicarbonate secretion makes it possible to neutralize harmful acid produced by oral bacteria, thus preventing oral infections. The Na+-HCO3− cotransporters (NBC) encoded by SLC4 gene family are recognized now as important mechanisms for HCO3− flux in many secretory epithelia (1, 4, 9, 12, 21, 24–26, 29, 33, 42, 46, 47, 52, 53, 55).
Moreover, mutations and deletions of NBC genes are major culprits of poor dentition in humans and mice (16, 19, 23). NBC was functionally detected in bovine and sheep parotid acinar cells before cloning of the SLC4 gene family (45, 50, 59). These reports were confirmed by a more recent report that electrogenic NBCe1 (SLC4A4) is present and functional in bovine parotid acinar cells (66). Additionally, several groups reported abundant expression of NBCe1 and NBCn1 in parotid acinar cells of different species (53). However, nothing is known about regulation of NBC by cholinergic secretory stimuli in parotid acinar cells. In fact, so little attention has been given to potential roles that NBC could play in the fluid secretion process that current models of fluid secretion proposed for parotid acinar cell do not include NBC (20, 39, 41). It is critical to clarify the roles of NBC in parotid acinar cells for a thorough understanding of the molecular physiology of the fluid secretion process.
We (43, 44) reported earlier that signaling via the angiotensin receptor (AT1R) stimulates endocytosis and inhibits functional activity of NBCe1 in Xenopus oocytes. Here, we expanded our early observations to investigate regulation of the surface expression of NBCe1 and NBCn1 in untreated and cholinergically stimulated parotid acinar cells. For these studies, we have used the ParC5 cell line, developed by Dr. David O. Quissell, which is functionally similar to native parotid acinar cells (48, 49, 63). Using confocal fluorescent microscopy in fixed ParC5 cells, NBCe1 and NBCn1 antibodies and markers for polarized epithelial cells and endocytosis, and surface biotinylation, we found the following. 1) The basolateral membrane (BLM) of rat ParC5 cells expresses NBCe1 and NBCn1, which would support HCO3− ions influx into the cell via BLM. In addition, a net negative charge enters the cell via electrogenic BLM NBCe1. 2) NBCe1, but not NBCn1, undergoes constitutive endocytosis. 3) Cholinergic stimulation induces PKC-dependent endocytosis of NBCe1. Removal of electrogenic NBCe1 from the BLM decreases HCO3−, and negative charge entry may support recovery of the acinar cell from massive loss of salt and water (63). 4) In contrast, cholinergic stimulation did not affect NBCn1, which may continuously transport HCO3− inside to possibly neutralize intracellular acidification in secreting acini. Our data suggest that NBCe1 and NBCn1 are important molecular mechanisms for the fluid secretion process in parotid acinar cells.
GF-109203X (GF), W-13, carbachol (CCh), and monensin were purchased from Sigma (St. Louis, MO). Texas red-conjugated transferrin (Tfn-TR), zona occludens 1 monoclonal antibody (ZO-1), and E-cadherin monoclonal antibody (E-Cad) were from Invitrogen (Carlsbad, CA). The EEA1 monoclonal antibody was from BD Transduction Laboratories (San Jose, CA). PMA was purchased from Calbiochem (La Jolla, CA). KIA, a rabbit polyclonal antibody against electrogenic NBCe1 (SLC4A4), was a kind gift from Dr. Walter F. Boron, Yale University School of Medicine, New Haven, CT. The COOH terminus of NBCe1 variants are identical, thus the antibody does not distinguish among renal NBCe1-A and pancreas NBCe1-B (7). NBCn1-106 and ntNBCn12977, two rabbit polyclonal antibodies raised against the COOH terminus of the rat sequence of electroneutral NBCn1, were kind gifts from Dr. Søren Nielson, University of Aarhus, Denmark (64). Another NBCn1 affinity purified rabbit polyclonal antibody raised against the COOH terminus of the rat sequence of electroneutral NBCn1 was a kind gift from Dr. Inyeong Choi, Emory University (49a). NBCn1 antibodies do not distinguish among NBCn1B, NBCn1C, NBCn1D, and NBCn1E because the COOH terminus of all four rat NBCn1 variants are identical (21, 30, 64). Secondary antibodies Alexa 488 and Alexa Fluor-Texas red were from Invitrogen. Vectashield mounting media was purchased from Vector Laboratories (Burlingame, CA).
ParC5 cell culture.
We used a well-established rat parotid acinar cell line, ParC5 (passages 10–40; Refs. 48, 49, 63). Cells were cultured at 37°C in humidified atmosphere of 5% CO2 and 35% O2. Cells were split by dry trypsinization with 0.05% trypsin (Gibco, Carlsbad, CA) and grown to confluency before experiments were performed. ParC5 cells were grown in DMEM/F-12 media with 15 mM HEPES and l-glutamine (Invitrogen) with the following added: 5 μg/ml insulin (Invitrogen), 10−7 M retinoic acid (Sigma), 5 mM glutamine (Invitrogen), 0.4 μg/ml hydrocortisone (Sigma), 5 μg/ml transferrin (Invitrogen), 10% standard FBS (HyClone), 1% nonessential amino acids (Invitrogen), and 50 μg/ml gentamicin (Invitrogen). Experiments were performed by first exchanging media with serum-free media before drug treatments. Serum-free media consisted of DMEM/F-12 media with 15 mM HEPES and l-glutamine (Invitrogen).
Monensin was made as a 10,000× stock solution in MeOH. The final working concentration of MeOH was 0.01%. CCh, atropine, and W-13 were made as 1,000× stock solution in distilled water (dH2O). PMA was made as a 1,000× stock solutions in DMSO. The final working concentration of DMSO was 0.1%. All drugs were diluted in ParC5 culture media to final concentrations before use. Where indicated, drugs were diluted in serum-free media. As controls, we used 0.1% DMSO and 0.01% MeOH.
Immunofluorescence and confocal microscopy.
Cells were cultured on 12-mm glass coverslips in 24-well plates or, where indicated, plated on Transwell filters (Corning, Corning, NY). After 3 days (filters) or 4 days (coverslips) in culture, cells were incubated with monensin, MeOH, W-13, PMA, CCh, ACh, and/or, in some experiments, with GF or atropine, as indicated. Cells were serum-starved for 15 min before treatment. Following treatment, cells were washed with PBS and fixed with 2% paraformaldehyde for 15 min at room temperature. For immunofluorescence staining, cells were permeabilized with 0.1% Triton X-100 in PBS for 10 min and incubated in blocking solution (0.05% Triton X-100, 1% bovine serum albumin, and 10% FBS in PBS). After blocking was completed, cells were incubated with anti-NBCe1 (1:500), anti-NBCn1 (1:500), anti-E-Cad (1:1,000), anti-ZO-1 (1:250), or anti-EEA1 (1:500) primary antibodies, followed by a series of washes (1% bovine serum albumin and 10% FBS in PBS), and then the respective Alexa Fluor-labeled secondary antibodies for 1 h. Subsequently, cells and coverslips were washed in PBS and then dH2O and mounted with Vectashield on slides for qualitative analysis through confocal microscopy.
Fluorescent images were taken using a Spinning Disk confocal microscope (Olympus) controlled by SlideBook 188.8.131.52 (Light Microscopy Facility at University of Colorado Denver; http://www.uchsc.edu/lightmicroscopy/). Confocal images were captured in z-stack intervals of 0.5 μm using a ×60 oil immersion objective (1.45 numerical aperture). Under mercury illumination, the filter sets were: FITC, excitation 460–480 nm, emission 495–540 nm; tetramethylrhodamine isothiocyanate (TRITC), excitation 535–555 nm, emission 570–625 nm. Qualitative and quantitative analysis was done to ascertain whether colocalization occurred between Alexa 488 and Alexa Fluor-Texas red or Cy3.
ParC5 cells were grown in 35-mm dishes and used at 100% confluency. Cells were washed twice with cold PBS containing 0.1 mM CaCl2 and 1 mM MgCl2 and incubated for 20 min on ice with 1 mg/ml sulfosuccinimidyl 2-(biotinamido)-ethyl-1,3-dithiopropionate (EZ-Link Sulfo-NHS-SS-Biotin; Pierce, Rockford, IL) in PBS, followed by a second incubation with fresh NHS-SS-Biotin. After biotinylation, the cells were washed twice with cold PBS, incubated for 30 min on ice with 0.1 M glycine in PBS, and washed with PBS. The cells were then solubilized by scraping with a rubber policeman in lysis buffer [50 mM NaCl, 2 mM MgCl2, 10 mM Tris·HCl (pH 6.8), 10% glycerol, 1 mM EGTA, 1 mM EDTA, 10 mM sodium fluoride, 1 mM phenylmethylsulfonyl fluoride, 10 μg/ml leupeptin, 544 μM iodoacetamide, 10 μg/ml aprotinin, 1% Triton X-100, 1% sodium deoxycholate] and by further incubating for 30 min at 4°C. Cell lysates were precleared by centrifugation at 16,000 g; the biotinylated proteins were precipitated with NeutrAvidin beads (Pierce), washed five times with lysis buffer, and denatured by heating the beads in sample buffer at 95°C for 5 min. NeutrAvidin beads were resolved on 7.5% SDS-PAGE, and the proteins were transferred to the membrane. Western blotting was performed with polyclonal antibodies to NBCe1 or NBCn1 followed by species-specific secondary antibodies conjugated with horseradish peroxidase. The enhanced chemiluminescence kit was purchased from Pierce. NBCe1 or NBCn1 proteins were detected using Kodak Image Station 440CF, and their intensities were measured using ImageJ software (http://rsb.info.nih.gov/ij/). For negative controls, we incubated cells in the absence of EZ-Link Sulfo-NHS-SS-Biotin and precipitated cell lysates in the absence of NeutrAvidin beads.
Two-electrode oocyte voltage clamp.
All studies were performed in accordance with guidelines from and submitted to and approved by the University of Colorado Denver Animal Care and Use Committee. Female Xenopus laevis frogs (NASCO) were anesthetized with 1.5 mg/ml tricaine. The ovarian lobes were surgically removed, dissected, and treated with 2 mg/ml collagenase type IA, and oocytes were incubated as described previously (43, 44). The cDNAs encoding human NBCe1 and M3R receptors were each subcloned into the pGH19 expression vector. DNAs were transcribed in vitro using the mMessage Machine kit (Ambion, Austin, TX) to generate synthesized capped mRNAs. Oocytes were injected with 50 nl of 0.5 ng/nl NBCe1 mRNA; 25 nl of 1 ng/nl NBCe1 mRNA plus 25 nl of 1 ng/nl M3R mRNA; or 50 nl of dH2O. Three days later, oocytes were subjected to electrophysiological experiments as described before (43, 44). Briefly, oocytes were voltage-clamped at room temperature using a two-electrode oocyte clamp (Warner Instruments, New Haven, CT) and microelectrodes made by pulling borosilicate glass capillary tubing (Warner Instruments) on a microelectrode puller. The oocytes were impaled with microelectrodes filled with 3 M KCl (resistance = 0.3–1.0 MΩ). We clamped oocytes to −50 mV holding potential (Vh) and measured NBC current (INBC) in response to a 60-s depolarization from −50 to 0 mV. The currents were filtered at 20 Hz (4-pole Bessel filter) and digitized. An oocyte was placed in a chamber with a 4 ml/min constant superfusion using bath solutions described previously (43, 44). Bath solutions were delivered with syringe pumps (Harvard Apparatus, South Natick, MA), and solutions were switched with pneumatically operated valves (Clippard Instrument Laboratory, Cincinnati, OH).
Collapsed z-stacks were done by creating a projection image in ParC5 cells using SlideBook 184.108.40.206. At least three independent experiments with consistent results were done for each experimental condition. Each experiment contained three separate coverslips of cells from which we took at least five images from different areas of each coverslip. Representative images of these experiments are shown in the figures. The Pearson R correlation coefficients were calculated to quantify endocytosis of NBCe1 measured as increased colocalization of internalized NBCe1 with early endosomes marked with EEA1. Using SlideBook 220.127.116.11 Cross Channel software (Olympus Spindisk fluorescent microscope), we acquired the Pearson R correlation coefficients that match the intensity of the green fluorescence FITC filter channel (NBCe1 or NBCn1) with the red TRITC filter channel (EEA1). Pearson R coefficient 0.0 signifies no correlation trend, 1.0 means complete co-correlation, and −1.0 implies anticorrelation. We normalized intensity of biotinylated protein bands in treated cells to that in untreated control cells. In Xenopus oocyte experiments, we normalized amplitude of the “test” to “control” voltage-clamped NBCe1 current. All averages are reported as means ± SD. For ratios, averages are presented as log-normal means. The statistical significance data were determined using an unpaired Student's t-test (28). Differences were considered significant at a level of P < 0.05.
Localization of endogenous electrogenic NBCe1 and electroneutral NBCn1 to the BLM of polarized ParC5 cells.
Here, we investigated specifically where NBCe1 and NBCn1 cotransporters are localized in polarized ParC5 cells. Cells were grown on Transwell filters for 3–4 days to achieve polarization (see experimental procedures) and costained with antibodies against NBCe1 and the BLM marker E-Cad or the apical marker ZO-1. Note that only truly polarized cells form tight junctions between the individual cells and display clear ZO-1 staining (15). In Fig. 1 A-1, the fluorescent staining for endogenous electrogenic NBCe1 (green) colocalizes almost entirely with E-Cad (red) staining both in apical view (AV) images and lateral view (LV) images. On the other hand, staining for ZO-1 (red) does not colocalize with, and is visibly separate from, the staining for NBCe1 (Fig. 1A-2). Images are collapsed z-stacks of the entire cell through the z-axis (AV) and collapsed z-stacks of the region of interest (white line) through the y-axis (LV). Therefore, within both of these images, one can see that there is clearly no apical presence of NBCe1 staining in ParC5 cells. This is most notable in the LV images showing strong colocalization of NBCe1 and E-Cad and total separation of NBCe1 and ZO-1.
In parallel experiments, we also stained cells with antibodies against endogenous electroneutral NBCn1 along with antibodies against the BLM marker E-Cad and the apical membrane marker ZO-1 (Fig. 1B; here, we used NBCn1-106 antibody). Similar to the results found with NBCe1, we found that NBCn1 is also localized at the BLM of ParC5 cells. This is evidenced by colocalization of NBCn1 staining with E-Cad staining and a clear separation of NBCn1 and ZO-1 staining. We also found the same staining pattern by using two other NBCn1 antibodies (see experimental procedures; data not shown). In our immunolabeling control experiments, in which we preabsorbed each antibody with its immunizing peptide, no NBC staining was detected (data not shown). Again, a lack of apical staining for NBCn1 suggests there is no NBCn1 present at the apical surface of ParC5 cells. These data illustrate that rat parotid acinar ParC5 cells express electrogenic NBCe1 and electroneutral NBCn1, both of which are localized at the BLM of these cells.
Recycling inhibitors significantly increased accumulation of NBCe1 but not NBCn1 in early endosomes.
We (44) previously reported that recombinant NBCe1 undergoes constitutive endocytosis in Xenopus oocytes. Here, we noticed that although most endogenous NBCe1 and NBCn1 are present at the BLM, there can also be found NBCe1 and, to a lesser extent, NBCn1 staining within the cytoplasm of the untreated ParC5 cells (Figs. 1 and 2). This suggested that NBCe1 and possibly NBCn1 may go through some degree of constitutive endocytosis in ParC5 cells. To test this idea, we used two recycling inhibitors: the carboxylic ionophore monensin and the calcium-binding protein calmodulin (CaM) inhibitor W-13, which we (44) previously used to study recombinant NBCe1. It is known that monovalent carboxylic ionophore monensin raises pH within endosomes and inhibits recycling of internalized proteins to the plasma membrane (5, 58, 65). It is also known that CaM controls the recycling step of the endocytic pathway of several membrane proteins, and its antagonist, W-13, blocks the exit of internalized cargo from early endosomes resulting in the formation of large endosomes (3, 13, 22, 32, 61). Others have shown that CaM is involved in the regulation of functional activity of endogenous NBCe1, whereas we reported that CaM is involved in trafficking of recombinant NBCe1 in Xenopus oocytes (4, 17, 44, 54).
To study the effect of monensin on NBCe1 distribution, we treated ParC5 cells with 50 μM monensin added to serum-free culture media for 60 min and kept in a humidified atmosphere of 5% CO2 and 35% oxygen at 37°C (Fig. 2A-1, 2nd row from the top). In a parallel control experiment, we applied the vehicle MeOH at 0.01% for 60 min at 37°C (Fig. 2A-1, 3rd row from the top). Subsequently, we fixed cells and stained with antibodies against NBCe1 and the early endosomal marker EEA1. We compared confocal microscopy images of cells untreated and monensin-treated for differences both in NBCe1 staining and in colocalization of NBCe1 with EEA1. As seen in Fig. 1A, in untreated ParC5 cells, the staining of NBCe1 is almost entirely colocalized with E-Cad and therefore at the BLM. In Fig. 2A-1 (top row), in addition to strong BLM staining of NBCe1, we also detected NBCe1 staining that colocalized with EEA1-stained early endosome in an untreated cell (Fig. 2A-1, arrow). In the monensin-treated cells, we found a marked decrease of NBCe1 at the BLM, significant accumulation of NBCe1 in intracellular endosomal-like formations, and increased colocalization with EEA1-stained early endosomes (Fig. 2A-1, 2nd row from the top). This was in remarkable contrast to untreated (Fig. 2A-1, top row) or vehicle-treated (Fig. 2A-1, 3rd row from the top) cells. Figure 2A-1 (see bottom row) shows a typical experiment in which we applied 100 μM W-13 for 60 min to serum-starved ParC5 cells at 37°C. We found that W-13 caused a significant loss of NBCe1 from the BLM, a parallel accumulation in endosomal-like formations, and increased colocalization of cotransporters with EEA1-stained early endosomes (Fig. 2A-1, insets). To quantify our results, we used SlideBook 18.104.22.168 Cross Channel software (Olympus fluorescent Spindisk microscope) to acquire the Pearson R correlation analysis (see experimental procedures). As shown in a bar graph in Fig. 2A-2, the Pearson R correlation coefficients that match colocalization of NBCe1 with EEA1 were low in untreated (0.16 ± 0.06, n = 124) and vehicle-treated cells (0.15 ± 0.05, n = 80). The colocalization of NBCe1 with EEA1 were ∼3-fold increased in cells treated with monensin (0.32 ± 0.06, n = 60) and W-13 (0.47 ± 0.06, n = 84). These results suggest that endogenous electrogenic NBCe1 undergoes constitutive endocytosis in ParC5 cells. With application of these recycling inhibitors, internalized NBCe1 is, possibly, unable to return to the surface and is essentially trapped within endosomes. As the recycling inhibitor W-13 is also a CaM antagonist, these findings may prove useful in future experiments to study the role of Ca2+ and CaM in NBCe1 trafficking and regulation.
The effect of the recycling inhibitor monensin on the electroneutral NBCn1 is shown in Fig. 2B. This figure shows that in contrast to NBCe1, NBCn1 does not undergo constitutive endocytosis. We found that after application of 50 μM monensin for 1 h, there was no detectable loss of NBCn1 from the BLM (Fig. 2B-1, bottom row). The bar graph in Fig. 2B-2 shows that the Pearson R correlation coefficient of the colocalization of NBCn1 with EEA1 is low both in untreated cells (0.13 ± 0.02, n = 61) and monensin-treated cells (0.104 ± 0.044, n = 68). This suggests that NBCn1 does not undergo constitutive endocytosis or does so at a very slow rate. To investigate this latter possibility, we applied monensin overnight and found little to no change in staining from untreated cells (data not shown). Therefore, we conclude that either the constitutive endocytosis of NBCn1 is extremely slow or simply does not occur.
Surface biotinylation quantitative analysis showed that recycling inhibitors cause removal of NBCe1 but not NBCn1 from the BLM.
To extend support of our confocal microscopy data, we asked whether recycling inhibitors cause removal of the electrogenic NBCe1 but not the electroneutral NBCn1 from the BLM. For this, we performed surface biotinylation studies using a membrane-impermeable derivative of biotin (see experimental procedures). First, we studied surface expression of the electrogenic NBCe1. Western blot analysis with anti-NBCe1 polyclonal antibodies (KIA; 1:500) detected biotinylated NBCe1 protein as an ∼130-kDa band on SDS-PAGE (Fig. 5A-1). We found that recycling inhibitors caused a significant decrease of the intensity of the surface-biotinylated NBCe1 protein. Thus the intensity of the NBCe1 band from cells treated with 50 μM monensin as well as with 100 μM W-13 was 23.8% ± 19.6% (n = 4) and 33.3% ± 19.5% (n = 4), respectively, of that of untreated cells (Fig. 5A-2). The intensity of the NBCe1 band in vehicle-treated (0.01% MeOH) cells was 88.9% ± 11.1% (n = 4) of that of untreated cells. This suggests that the possible inhibition of the constitutive recycling of NBCe1 induces removal of the cotransporter from the BLM of ParC5 cells. Next, we examined the electroneutral NBCn1. Western blot analysis with anti-NBCn1 affinity purified polyclonal antibodies (NBCn1-Ct; 1:500) detected surface-biotinylated NBCn1 protein as an ∼127-kDa band on SDS-PAGE (Fig. 5B-1). In striking contrast to NBCe1, we found that recycling inhibitor monensin did not alter the level of the surface expression of NBCn1 in ParC5 cells. The surface-biotinylated NBCn1 band from cells treated for 60 min with 50 μM monensin was 99.1% ± 26.9% (n = 6) of that of untreated cells, indicating that this recycling inhibitor does not induce removal of NBCn1 from the BLM of ParC5 cells. Negative controls showed a complete absence of biotinylated NBCe1 or NBCn1 (data not shown). Taken together, our confocal fluorescent microscopy findings and surface biotinylation data suggest that the electrogenic NBCe1 undergoes constitutive endocytosis, whereas the electroneutral NBCn1 is stably present at the BLM of ParC5 cells.
PMA stimulates endocytosis of NBCe1 but not NBCn1.
Given the differences in constitutive endocytosis of NBCe1 and NBCn1, we hypothesized that there may also be differences in stimulated endocytosis of these cotransporters. For this, we utilized PMA, which is commonly used to stimulate endocytosis of membrane proteins in mammalian cells (34, 40, 56), in an attempt to induce endocytosis of NBCe1 and NBCn1 in ParC5 cells. Results for endogenous electrogenic NBCe1 are summarized in Fig. 3 A-1 (middle row) and A-2 (2nd bar). In this experiment, ParC5 cells were treated with 1 μM PMA added to serum-free culture media for 15 min and kept in a humidified atmosphere of 5% CO2 and 35% oxygen at 37°C. Subsequently, we fixed cells and stained with antibodies against NBCe1 and EEA1. We compared confocal microscopy images of untreated cells and PMA-treated cells (Fig. 3A-1, top and middle rows) for differences in staining of NBCe1 and in its colocalization with EEA1. In PMA-treated cells, we found a marked loss of the BLM staining of NBCe1 and its redistribution into endosome-like formations, including EEA1-stained early endosomes within the cytoplasm of the cell (Fig. 3A-1, middle row). To quantify our results, we performed a Pearson R correlation analysis (see experimental procedures). As shown in the bar graph in Fig. 3A-2, the Pearson R correlation coefficient of colocalization of electrogenic NBCe1 with EEA1 was low (0.15 ± 0.05, n = 499 cells) in control untreated cells (see “UNTR” bar). PMA increased by ∼2.5-fold the colocalization of NBCe1 with EEA1 with a Pearson R correlation coefficient of 0.38 ± 0.07 (n = 170 cells) (see “PMA” bar in Fig. 3A-2). These results indicate that PKC activator PMA induces significant endocytosis of NBCe1.
Next, we investigated the effect of PMA on endogenous electroneutral NBCn1 (see Fig. 3B-1). ParC5 cells were treated with PMA and stained with antibodies against NBCn1 and EEA1. In PMA-treated cells, we found that NBCn1 (Fig. 3B-1, middle row) in striking contrast to NBCe1 (Fig. 3A-1, middle row) remains at the BLM, does not redistribute within the cytoplasm, and does not colocalize with EEA1-stained early endosomes. As summarized in the bar graph in Fig. 3B-2, the Pearson R correlation coefficient of colocalization of NBCn1 with EEA1 in the PMA-treated cells (0.15 ± 0.04, n = 100) was similar to that in the control untreated cells (0.17 ± 0.03, n = 270). These results suggest that PMA does not stimulate endocytosis of NBCn1.
CCh stimulates endocytosis of NBCe1 but not NBCn1.
Cholinergic stimulation is a major stimulus for isotonic salt and fluid secretion in rat parotid salivary cells (18). Here, we used muscarinic receptor (MR) agonists ACh and CCh to examine whether cholinergic stimulation regulates trafficking of endogenous NBCe1 and NBCn1. Results for endogenous electrogenic NBCe1 are shown in Fig. 3A-1 (bottom row). This figure shows a typical experiment in which we applied 50 μM CCh for 15 min to live cells and then fixed and stained cells with antibodies against NBCe1 and EEA1. Similar to PMA, we found that in CCh-treated cells, the staining at the BLM was reduced, and colocalization of NBCe1 and EEA1 increased. Again, in response to CCh, NBCe1 redistributed to endosomal-like formations within the cell. Similar robust endocytosis of NBCe1 was detected in ParC5 cells treated with neurotransmitter ACh (data not shown). As shown in the bar graph in Fig. 3A-2, cholinergic stimulation increases by ∼2.6-fold the Pearson R correlation coefficient of colocalization of electrogenic NBCe1 with EEA1 from 0.15 ± 0.05 (n = 499 cells) in control untreated cells to 0.40 ± 0.08 (n = 167) in CCh-treated cells. These data suggest that in 15 min following application of cholinergic agonists, NBCe1 undergoes a significant stimulated endocytosis in ParC5 cells.
In striking contrast to NBCe1, we found that CCh does not stimulate endocytosis of endogenous electroneutral NBCn1 in ParC5 cells (see Fig. 3B-1). Briefly, we applied 50 μM CCh for 15 min to live cells and then fixed and stained cells with antibodies against NBCn1 and EEA1. We found that NBCn1 staining remains strong at the BLM, its accumulation in endosomal formations was absent, and there is no increase in the colocalization of NBCn1 and EEA1 in CCh-treated cells (Fig. 3B-1, bottom row) compared with untreated cells (Fig. 3B-1, top row). The Pearson R correlation coefficient of colocalization of NBCn1 with EEA1 in the CCh-treated cells (0.18 ± 0.04, n = 170) was similar to that in the control untreated cells (0.17 ± 0.03, n = 270) (see bar graph in Fig. 3B-2).
Surface biotinylation quantitative analysis showed that PMA and CCh remove NBCe1 but not NBCn1 from the BLM.
By surface biotinylation, we demonstrated that PMA and CCh removed the electrogenic NBCe1 but not the electroneutral NBCn1 from the BLM of ParC5 cells. We found a significant decrease in the intensity of the surface-biotinylated NBCe1 band from cells treated with PMA or CCh (Fig. 5A-3). As summarized in the bar graph in Fig. 5A-4, intensity of the NBCe1 band from cells treated with 1 μM PMA or 50 μM CCh was 36.2% ± 21.1% (n = 4) and 50.2% ± 13.1% (n = 4), respectively, of that of untreated cells. In striking contrast to NBCe1, we found that neither PMA nor CCh altered the level of the surface expression of NBCn1 in ParC5 cells. Intensities of the surface-biotinylated NBCn1 band from cells treated with 1 μM PMA or 50 μM CCh were 110.2% ± 8.9% (n = 6) and 101.2% ± 17.4% (n = 6) of that of untreated cells (see bar graph in Fig. 5B-2). Negative controls showed a complete absence of biotinylated NBCe1 or NBCn1 (data not shown). Taken together, our confocal fluorescent microscopy findings and surface biotinylation data suggest that PMA and CCh induce endocytosis of electrogenic NBCe1 but not electroneutral NBCn1 in ParC5 cells.
CCh-stimulated endocytosis of NBCe1 occurs through the MR.
It was reported that cholinergic muscarinic receptors are expressed in the BLM of ParC5 cells (8, 63). To ensure that the observed above effects of CCh are mediated via muscarinic receptors, we used the cholinergic receptor inhibitor atropine. Figure 4B (bottom row) shows staining of NBCe1 and EEA1 in ParC5 cells treated with a mixture of 1 μM atropine and 50 μM CCh. Atropine completely prevents loss of NBCe1 from the BLM (Fig. 4B, bottom row) in contrast to that found in cells treated with CCh alone (bottom rows in Figs. 3A-1 and 4A). We also observed a noticeable decrease of colocalization of NBCe1 and EEA1 staining. Thus the Pearson R correlation coefficient dropped from its value of 0.40 ± 0.08 (n = 167) in the CCh-treated cells to its value of 0.14 ± 0.06 (n = 143 cells) in the cells treated with a mixture of atropine with CCh. Note that its value was 0.15 ± 0.05 (n = 499 cells) in the control untreated cells (see bar graph in Fig. 3A-2). Additionally, our surface biotinylation quantitative analysis showed that atropine prevents CCh-induced removal of NBCe1 from the BLM. As summarized in the bar graph in Fig. 5 A-4, a normalized intensity of the biotinylated NBCe1 band in the cells treated with a mixture of atropine with CCh was 90.4% ± 13.9% (n = 4), which is a ∼2-fold increase from its value of 50.2% ± 13.1% (n = 4) in the CCh-treated cells. Thus our confocal microscopy findings and surface biotinylation data indicate that the CCh-induced endocytosis of NBCe1 is indeed due to stimulation of muscarinic receptors in ParC5 cells.
PMA- and CCh-stimulated endocytosis of NBCe1 is PKC dependent.
Previously, we (43) reported that PMA-induced inhibition of activity of NBCe1 in Xenopus oocytes is PKC dependent. Here, we investigate the PKC dependency of NBCe1 endocytosis caused by PMA, which activates the conventional and novel isoforms of PKC. In the experiment in Fig. 4B (top row), we applied a mixture of 1 μM PMA and 500 nM GF (a specific PKC inhibitor) to ParC5 cells for 15 min and then fixed cells and stained with antibodies against NBCe1 and EEA1. We found that BLM staining remained strong, and there was no noticeable increase in endosomal formation staining with NBCe1. We also found no increase in colocalization of NBCe1 and EEA1 staining (see Fig. 4, A and B). Thus, in cells treated with a mixture of PMA and GF, the Pearson R correlation coefficient was 0.10 ± 0.05 (n = 80), which is similar to its value of 0.15 ± 0.05 (n = 499) in untreated cells (see bar graph in Fig. 3A-2). Additionally, our surface biotinylation data show that the PKC inhibitor GF prevents PMA-induced removal of NBCe1 from the BLM (see Fig. 5A-3). As summarized in the bar graph in Fig. 5A-4, the intensity of the biotinylated NBCe1 band in the cells treated with a mixture of PMA and GF was 85.3% ± 21.3% (n = 4) of that in untreated cells. This was in striking contrast to the intensity of the biotinylated NBCe1 band in cells treated with PMA alone, which was 36.2% ± 21.1% (n = 4) of that in untreated cells (see Fig. 5A-4). Thus our confocal microscopy findings and surface biotinylation data support the view that PMA-induced endocytosis of NBCe1 in ParC5 cells is PKC dependent.
We (44) also reported that PKC mediates angiotensin II (ANG II)/AT1-stimulated endocytosis of NBCe1 in Xenopus oocytes. It is known that, similar to AT1R activation, binding of agonists to MR activates PKC in ParC5 cells (51, 67). Here, we aimed to investigate the PKC dependence of the CCh-stimulated endocytosis of NBCe1. First, we examined the PKC dependence of the cholinergic regulation of the activity of NBCe1 coexpressed with the M3R in Xenopus oocytes. For this, we performed study using the MR agonist ACh. Results of these experiments are summarized in the bar graph in Fig. 4C. Briefly, oocytes were superfused with 5% CO2-33 mM HCO3− solution, voltage clamped to −50 mV for 10 min, and depolarized from −50 to 0 mV to record either control currents before treatment or test currents following treatment. At 0 min, control currents were measured, and oocytes were treated with 500 nM ACh or with a mixture of 500 nM ACh and 500 nM GF. NBCe1 currents were recorded every 5 min. Figure 4C shows that the NBCe1 currents decrease over time with ACh application. With addition of the PKC inhibitor GF to the ACh treatment, the NBCe1 current stays statistically the same over at least a 15-min period. Thus this study shows that CCh-induced inhibition of activity of NBCe1 in Xenopus oocytes was also PKC dependent.
Second, to investigate the PKC dependence of the CCh-stimulated endocytosis of NBCe1, we applied a mixture of 50 μM CCh and 500 nM GF to ParC5 cells followed by staining with antibodies against NBCe1 and EEA1 (Fig. 4B, middle row). We found that BLM staining remained strong, and there was no noticeable increase in endosomal formation staining with NBCe1. We also found that PKC inhibitor significantly decreases CCh-induced colocalization of NBCe1 and EEA1. In the cells treated with a mixture of CCh and GF, the Pearson R correlation coefficient dropped to a value of 0.11 ± 0.04 (n = 65) from its value of 0.40 ± 0.08 (n = 167) in the CCh-treated cells. Note that its value was 0.15 ± 0.05 (n = 499 cells) in the control untreated cells (see bar graph in Fig. 3A-2). Additionally, our surface biotinylation studies showed that the PKC inhibitor GF prevents CCh-induced removal of NBCe1 from the BLM. As summarized in the bar graph in Fig. 5A-4, a normalized intensity of the biotinylated NBCe1 band in the cells treated with a mixture of GF with CCh was 86.1% ± 10.1% (n = 4) compared with its value of 50.2% ± 13.1% (n = 4) in the CCh-treated cells. Therefore, our confocal microscopy findings and surface biotinylation data demonstrated that PKC inhibitor GF completely prevents PMA- and CCh-induced endocytosis of NBCe1, suggesting PKC dependence of NBCe1 membrane trafficking in ParC5 cells.
PMA and CCh increase colocalization of NBCe1 with transferrin-stained endosomes.
Since transferrin is a well-studied ligand of the transferrin receptor that endocytoses through the clathrin-dependent mechanism (35), we used Tfn-TR to investigate possible mechanisms of internalization of NBCe1 in ParC5 cells (Fig. 4A). We found that internalized NBCe1 colocalizes with Tfn-TR within the same endosomes in PMA- and CCh-treated cells (see insets in Fig. 4A, middle and bottom rows). These data suggest possible involvement of clathrin in stimulated endocytosis of NBCe1 in ParC5 cells. Further studies manipulating clathrin or its mechanisms must be done, however, to determine whether it is, in fact, involved.
BLM staining of both electrogenic NBCe1 and electroneutral NBCn1 in polarized ParC5 cells.
Cholinergic MR agonists are major secretory stimuli with HCO3− luminal secretion being an important component of isotonic fluid secretion in rat parotid acinar cells (6, 20, 39). A dual function of acinar cells is to carry out HCO3− flux and to maintain neutral intracellular pH (pHi) during stimulated secretion. HCO3− transport via NBC could support HCO3− secretion and/or normalize pHi of HCO3−-secreting cells. Indeed, electrogenic NBCe1 and electroneutral NBCn1 are found in the HCO3−-secreting epithelia of pancreas, guts, distal renal tubule, and salivary glands (1, 4, 9, 12, 21, 24–26, 29, 33, 42, 46, 47, 52, 53, 55). Here, we investigated the subcellular localization and regulation of membrane trafficking of NBCe1 and NBCn1 in the ParC5 cells that are functionally similar to native parotid acinar cells (48, 49, 63). The ParC5 cell is also an established model for MR-mediated secretion (8). We observed strong expression of NBCe1 and NBCn1 in the BLM of polarized ParC5 cells (Fig. 1). Our results showing BLM staining of electrogenic cotransporter is in agreement with reports by others that NBCe1 is present at the BLM in rat, human, mouse, and guinea pig parotid acinar cells (27, 31, 42, 53).
In contrast to our detection of electroneutral cotransporter at the BLM in ParC5 cells, Gresz et al. (21) did not find any NBCn1 in rat parotid acini. Gresz et al. (21) used an indirect immunoperoxidase labeling of dewaxed and rehydrated paraffin tissue sections, whereas here we used confocal immunofluorescence microscopy in fixed cultured cells. We also support our finding with Western blot analysis of the biotinylated NBCn1 protein extracted from ParC5 cells (Fig. 5B). Therefore, this discrepancy may be explained by differences in the sample preparation and/or in the sensitivity of these different approaches. In agreement of our finding of the BLM staining of NBCn1 in HCO3−-secreting ParC5 cells is the reported detection of NBCn1 at the BLM in HCO3−-secreting renal distal medullary B intercalated cells (46, 47). Therefore, HCO3− entry via BLM NBC could be one of the important features of the salivary fluid secretory response.
Constitutive endocytosis of electrogenic NBCe1 but not electroneutral NBCn1 in ParC5 cells.
Functional activity of bicarbonate transporters depends on the amount of their proteins at the cell surface, which could be regulated by endocytosis, as we (43) previously showed in Xenopus oocytes. Here, we detected a striking difference in membrane trafficking of electrogenic NBCe1 vs. electroneutral NBCn1 in ParC5 cells. In particular, our evidence strongly suggests that NBCe1, but not NBCn1, undergoes constitutive endocytosis. In support of this, a 60-min blockage of constitutive recycling by monensin or W-13 leads to loss of surface NBCe1 and redistribution of the cotransporter from the BLM into EEA1-marked cytosolic compartments (Figs. 2A and 5A). In contrast, electroneutral NBCn1 was constantly present at the cell surface of ParC5 cells after 60-min (Figs. 2B and 5B) or overnight treatment with monensin (data not shown).
PKC-dependent endocytosis of electrogenic NBCe1 but not electroneutral NBCn1 in stimulated ParC5 cells.
Next, we investigated the effect of cholinergic agonists on the membrane trafficking of NBC in ParC5 cells. We found here that the MR agonist CCh stimulates massive endocytosis of electrogenic NBCe1. A 15-min CCh application leads to a loss of NBCe1 from the cell surface and its redistribution into endosomes, some of which are marked with EEA1 early endosomal marker (Fig. 3A-1, bottom row). ACh, another MR agonist, also induced similar endocytosis of NBCe1 in ParC5 cells (data not shown). Working with Xenopus oocytes, we found that ACh inhibits NBCe1 activity, which could reflect similar activity inhibition due to endocytosis of NBCe1 in ParC5 cells (Fig. 4C). This speculation is supported by our reports (43, 44) of AT1R-induced endocytosis of NBCe1 and parallel inhibition of its activity in Xenopus oocytes. Thus endocytosis of NBCe1 could significantly reduce translocation of HCO3− ions and a net negative charge across the BLM in ParC5 cells. We found that the MR blocker, atropine, completely inhibits CCh-induced (Figs. 3A-2, 4B, and 5A) and ACh-induced (data not shown) endocytosis of NBCe1. Therefore, signaling via G protein-coupled MR receptor mediates endocytosis of NBCe1 in stimulated ParC5 cells. This new finding is in agreement with our previous report (43) that signaling via G protein-coupled AT1R stimulates endocytosis of NBCe1 in Xenopus oocytes.
Stimulation of MR and AT1R G protein-coupled receptors activates PKC, which can be delivered to the plasma membrane to phosphorylate its multiple target proteins (14). Here, we showed that the PKC-specific inhibitor GF prevents CCh-stimulated endocytosis of NBCe1 in ParC5 cells (Figs. 3A, 4B, and 5A). To additionally support the view that PKC regulates endocytosis of NBCe1, we used the PKC activator PMA known to stimulate endocytosis of many membrane proteins in mammalian cells (56). We found that PMA stimulates PKC-dependent endocytosis of NBCe1, which can be blocked by GF (see Fig. 3A, 2nd row, Fig. 4A, 2nd row, and Fig. 4B, 1st row). These new data are in agreement with our previous reports (43, 44) that AT1R-activated PKC binds to recombinant NBCe1, which undergoes PKC-dependent endocytosis in Xenopus oocytes. Here, we used Tfn-TR, an agonist of the transferrin receptor known to endocytose via the clathrin-dependent pathway (56). We found that internalized NBCe1 was colocalized with Tfn-TR-positive endosomes in ParC5 cells treated with CCh or PMA (Fig. 4A). Further study is needed to clarify whether PKC-dependent CCh- and/or PMA-stimulated endocytosis of NBCe1 is also clathrin dependent.
In striking contrast, the electroneutral NBCn1 was constantly present at the BLM after 15-min of CCh or PMA application (Figs. 3B and 5B). Longer 30-min treatment with CCh or PMA produced no changes in BLM expression of NBCn1 (data not shown). Interestingly, Kwon et al. (29) also reported that the abundance of the BLM NBCn1 was unchanged in response to ANG II treatment in the medullary thick ascending limb (mTAL) cells of rat kidney.
Why would NBCn1 not go through constitutive endocytosis, and why would NBCn1 not endocytose in response to cholinergic stimulation of ParC5 cells? We can speculate that NBCn1, unlike NBCe1, is not regulated by PKC due to low probability of putative PKC phosphorylation sites in its amino acid sequence (analysis by NetPhos 2.0 at http://www.cbs.dtu.dk; data not shown). Therefore, lack of PKC regulation of NBCn1 could possibly prevent its endocytosis in response to cholinergic stimulation of ParC5 cells. Further support for this speculation is found in our earlier report (44) that in Xenopus oocytes, AT1R-activated PKC failed to stimulate endocytosis of excitatory amino acid transporter (EAAT3), which is not regulated by PKC. These speculations may be able to explain the absence of activated endocytosis but would not suffice to explain constitutive endocytosis. However, the amino acid sequence of NBCn1 has a putative PDZ-binding site, the last 4 amino acids (ETSL) of its COOH terminus. We can therefore speculate that NBCn1 could be anchored to the BLM cytoskeleton by PDZ adapter proteins. This could stabilize NBCn1 at the BLM in ParC5 cells, similar to a model of potassium channel stabilization proposed by Tamkun's group (60).
Possible physiological roles of NBCe1 and NBCn1 in ParC5 cells.
Differences in the regulation of membrane trafficking of NBCe1 vs. NBCn1 indicates that electrogenic and electroneutral processes of Na+ and HCO3− translocation across the BLM would play different roles in HCO3− and/or total salt and fluid luminal secretion and/or in pHi regulation in ParC5 cells. Both NBCe1 and NBCn1 translocate HCO3− across the BLM to possibly support a net HCO3− secretion in acinar cells. In addition to Na+ and HCO3− influx, the electrogenic NBCe1 translocates a net negative charge across the BLM as reported in bovine parotid acinar cells (66). Therefore, transport via NBCe1 would contribute to electronegative membrane potential, which drives luminal secretion in parotid acinar cells (20). As we observed, NBCe1 undergoes removal from the BLM of ParC5 cells in 5–15 min following cholinergic stimulation. Interestingly, we (63) reported earlier that within the first 30 s of agonist stimulation, ParC5 cells undergo transient increase in anion, i.e., Cl− and HCO3−, luminal secretion, which, after approximately 2–5 min, was gradually decreased to a low steady-state level of secretion lasting at least for 10 min. This transient decrease of anion secretion coincides with reports by others of a transition between the initial and the sustained phases of stimulated fluid secretion in parotid acinar cells (36–38). Therefore, we speculate that the timing of endocytosis and inhibition of NBCe1 could be a part of acinar cell adjustment to a new steady state characteristic of a continuous secretory response.
In striking contrast, the electroneutral NBCn1 is stably present at the BLM throughout cholinergic stimulation, possibly continuously supporting its portion of the HCO3− flux via the BLM. For NBCn1, a low cellular Na+ concentration is usually sufficient to drive inwardly directed ionic transport (2, 10). We speculate that the role of the BLM NBCn1 would be to allow for an accompanying HCO3− entry to maintain neutral pHi in secreted parotid acinar cells. This is similar to the BLM K+ channels, which allow for an accompanying K+ efflux to maintain intracellular electroneutrality in acinar cells (20).
Parotid acinar cells secrete salt and fluid in response to parasympathetic muscarinic stimulation. It is the resulting anion loss, including massive HCO3− loss, across the apical membrane that drives transepithelial fluid secretion in acinar cells. Therefore, the BLM Na+, HCO3−, and a net negative charge entry via two types of NBC could be important features of the salivary fluid secretory response. Undoubtedly, future studies are needed to measure activity of NBCe1 in stimulated parotid acinar cells, especially during the first 30 s of increased anion, i.e., Cl− and HCO3−, secretion. Future studies are also needed to measure the activity of NBCn1 in resting and stimulated acinar cells. Since current models of fluid secretion proposed for parotid acinar cells do not include NBC (20, 39, 41), determining the roles that NBC plays within these models would ultimately lead to a more thorough understanding of the molecular physiology of the fluid secretion process in parotid acinar cells.
This work is supported by The Pilot Project Grant in Women's Health Research, University of Colorado Denver (I. I. Grichtchenko) and National Institute of Dental and Craniofacial Research Grant DE-015648 (M. E. Reyland).
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 © 2008 the American Physiological Society