Acute inhibition of the betaine transporter by ATP and adenosine in renal MDCK cells

Abstract

Extracellular ATP interacts with purinergic P2 receptors to regulate a range of physiological responses, including downregulation of transport activity in the nephron. ATP is released from cells by mechanical stimuli such as cell volume changes, and autocrine signaling by extracellular ATP could occur in renal medullary cells during diuresis. This was tested in Madin-Darby canine kidney (MDCK) cells, a model used frequently to study P1 and P2 receptor activity. ATP was released within 1 min after transfer from 500 to 300 mosmol/kgH₂O medium. A 30-min incubation with ATP produced dose-dependent inhibition (0.01–0.10 mM) of the renal betaine/GABA transporter (BGT1) with little effect on other osmolyte transporters. Inhibition was reproduced by specific agonists for P2X (α,β-methylene-ATP) and P2Y (UTP) receptors. Adenosine, the final product of ATP hydrolysis, also inhibited BGT1 but not taurine transport. Inhibition by ATP and adenosine was blocked by pertussis toxin and A73122, suggesting involvement of inhibitory G protein and PLC in postreceptor signaling. Both ATP and adenosine (0.1 mM) produced rapid increases in intracellular Ca²⁺, due to the mobilization of intracellular Ca²⁺ stores and Ca²⁺ influx. Blocking these Ca²⁺ increases with BAPTA-AM also blocked the action of ATP and adenosine on BGT1 transport. Finally, immunohistochemical studies indicated that inhibition of BGT1 transport may be due to endocytic accumulation of BGT1 proteins from the plasma membrane. We conclude that ATP and adenosine, through stimulation of PLC and intracellular Ca²⁺, may be rapidly acting regulators of BGT1 transport especially in response to a fall in extracellular osmolarity.

plasma betaine is derived mainly from the diet but can be augmented by synthesis in the liver and kidney (10). It is freely filtered and reabsorbed by Na⁺ and H⁺ cotransport systems in the luminal membrane of the proximal tubule (62). Betaine, together with myo-inositol, sorbitol, and glycerophosphorylcholine, serves as an organic osmolyte that is accumulated by the cells of the inner medulla of animal and human kidneys to maintain osmotic balance (2, 49). Betaine has a similar protective function in the brain (4) and may also be important as a chaperone to stabilize protein structure during stress (46). Acute water diuresis was accompanied by a decrease in medullary betaine content in rabbits (60) and a prompt (1 h) increase in urinary betaine excretion in humans (49). Conversely, dehydration produced an increase in medullary betaine content (20). This suggests that transport of betaine, and other osmolytes, is a physiologically important and dynamic process in the kidney.

The renal betaine/GABA transporter (BGT1) protein is located primarily in cells in the medullary thick ascending limb and collecting duct and also the macula densa (7, 37). At least in polarized cells in culture, it is upregulated and inserted primarily in the basolateral plasma membrane during hypertonic stress, a process that requires 12–24 h (64). Transcriptional activation of the BGT1 gene during hypertonic stress has been studied in detail (61), but regulation of the transport activity of the BGT1 protein in situ has remained unknown. Two recent studies (28, 33), including one from this laboratory, showed that BGT1 transport activity in the plasma membrane of cultured kidney cells can be downregulated acutely (within 30–40 min) by changes in extra-/intracellular Ca²⁺ and by PKC activation. Acute regulation is consistent with the observed dynamic changes in medullary betaine content at the organ level and would allow adjustments to the always high but changing osmolarity in the medulla.

Extracellular ATP is known to increase intracellular Ca²⁺ in renal cells, including MDCK (1, 13, 18), and purinergic P2 receptors are widespread in the nephron (19, 31). The aim of this study was to determine if extracellular ATP serves as an autocrine regulator of BGT1 transport via P1 and P2 receptors and intracellular Ca²⁺.

MATERIALS AND METHODS

Materials.

α,β-methylene-ATP (Me-ATP), 5′-[γ-thio]triphosphate (ATP-γS), IBMX, α-methylaminoisobutyric acid (MeAIB), N^ω-nitro-l-arginine methyl ester (l-NAME), pertussis toxin, suramin, U73122, U73343, and UTP were from Sigma-Aldrich. BAPTA-AM was from Calbiochem/EMD. Fura 2-AM was from Invitrogen/Molecular Probes. 2-Chloro-N⁶-cyclopentyladenosine (CCPA) and 1,3-dipropyl-8-cyclopentylxanthine (DPCPX) were from Tocris Bioscience.

Cell culture.

Canine MDCK cells (CCL-34), originally from the American Type Culture Collection (Rockville, MD), were provided by the laboratory of Dr. J. A. McAteer (Indiana University School of Medicine), The cells were cultured in a 1:1 mixture of DMEM-Ham's F-12 containing 10% bovine calf serum, 15 mM HEPES, 25 mM NaHCO₃ (pH 7.4), 100 IU/ml penicillin, and 100 μg/ml streptomycin, as in previous studies (28, 30). When grown in collagen gels, these cells readily form cysts that progressively increase in size (32, 52), consistent with high-resistance type I MDCK cells that resemble collecting duct segments (43). Cells of the mouse inner medullary collecting duct cell line (mIMCD3), a gift from Dr. Christopher Rivard (University of Colorado Health Sciences Center, Denver, CO), were grown in similar medium containing 10% FBS (42, 44). Cells were grown in monolayer culture on glass coverslips for immunohistochemistry and fura 2 fluorescence measurements and in 24-well plastic plates for transport measurements. Hypertonic stress was induced by incubation in normal growth medium containing NaCl added to a final osmolality of 500 mosmol/kgH₂O. This medium was changed to hypertonic serum-free growth medium containing 0.1% BSA for 2 h before addition of ATP and other test compounds. All tested agonists and antagonists were applied for 30 min and were used from stock solutions in either DMSO or ethanol that were diluted 1:1,000 after addition to experimental solutions. An equivalent volume of solvent was added to the controls.

Transport measurements.

The osmolality of all solutions used for transport was matched to the osmolality of the growth medium. Hypertonic solutions (500 mosmol/kgH₂O) were achieved by addition of sucrose. As described previously in detail (28, 30), transport activity of BGT1 in MDCK monolayers was assessed as NaCl-dependent uptake of [³H]GABA, and amino acid transport system A was determined as Na⁺-dependent uptake of α-[¹⁴C]MeAIB. Transport activity of the taurine transporter (TAUT) was determined as NaCl-dependent uptake of 0.1 mM [³H]taurine (55). Test compounds were not present during transport assays. Osmolyte transport systems were induced by hypertonic stress for either 5 h (system A) or 24 h (BGT1 and TAUT) before transport assays (26).

Intracellular Ca²⁺ measurements.

Changes in intracellular Ca²⁺ concentration were measured with the Ca²⁺-sensitive probe fura 2, which was loaded into cells as the acetoxymethyl ester (28, 50). MDCK cells on coverglasses were loaded with 2.5 μM fura 2-AM in solution containing the following (in mM): 118 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 25 NaHCO₃, 5 glucose, and 10 HEPES (pH 7.4) at 37°C. After 30 min, the cells were rinsed in fresh solution and placed in the superfusion chamber on the stage of an epifluorescence microscope. Cells were studied at room temperature using the InCa calcium imaging system (Intracellular Imaging, Cincinnati, OH), as used and described previously (22, 28). Fura 2 was excited by alternating 340 and 380 nm light, and fluorescence emitted at 510 nm was collected by a CCD camera attached to a computer for data acquisition by the InCa ratiometric fluorescence program. In some experiments, Ca²⁺ influx pathways were monitored using extracellular Mn²⁺ as a surrogate ion for Ca²⁺ (3, 41, 59). The Mn²⁺ quench technique evaluates divalent cation (Ca²⁺/Mn²⁺) influx with greater sensitivity because 1) fura 2 has a 40-fold greater affinity for Mn²⁺ than Ca²⁺, 2) Mn²⁺ is not transported by other Ca transporters, and 3) since there are no Mn²⁺ stores, no intracellular release is measured. The extracellular medium contained 2 mM MnCl₂ instead of CaCl₂. Entry of Mn²⁺ entry was measured as quenching of fura 2 at an excitation wavelength of 360 nm, the Ca²⁺-insensitive isosbestic wavelength. For each experiment, the maximum fluorescence signal of at least 9–10 cells was averaged for analysis, after correction for background fluorescence.

ATP assay.

Cell suspensions were used so that the cells could be rapidly separated from released ATP to minimize any degradation (1, 8). Confluent MDCK cells in two T-75 flasks were placed in hypertonic growth medium (500 mosmol/kgH₂O) for 24 h. The cells were collected by trypsinization in hypertonic solution and low speed centrifugation and washed twice in hypertonic PBS. The final pellet was suspended in 0.5 ml of Ca²⁺-free solution containing the following (in mM): 125 NaCl, 5 KCl, 3 MgCl₂, 25 NaHCO₃, 5 glucose, 10 HEPES (pH 7.4), 0.1% BSA, and additional NaCl to give a final osmolality of 500 mosmol/kgH₂O. The suspension was allowed to equilibrate at 37°C for 30 min using gentle rotation. Aliquots (125 μl) of the cell suspension were added to Eppendorf tubes containing an equal volume of either the same medium (controls) or water containing 3 mM MgCl₂ (to preserve divalent cations; Ref. 5) and mixed by inversion. After 30 s, the tubes were centrifuged at 13,000 g for 1 min to pellet the cells, 150 μl were removed immediately from the upper region of the supernatant, frozen on dry ice, and stored at −80°C. ATP in the supernatant was determined by a luminescence assay in a 96-well microplate using firefly luciferase and other reagents in the ATPlite kit from Perkin Elmer. Sample blanks and reagent blanks were included to correct for background luminescence.

Immunostaining and Western blots.

MDCK cells on glass coverslips were fixed in either cold methanol or 4% paraformaldehyde (PFA) in PBS and processed for antibody staining and confocal microscopy as described previously (27, 28). Cells fixed in PFA were permeabilized by 5-min incubation in PBS containing 0.2% Triton X-100 and 100 mM glycine and then rinsed in PBS. The affinity-purified BGT1 antibody kindly provided by Dr. H. Moo Kwon (University of Maryland) was used (1:200) on methanol fixed cells after blocking with 1% gelatin/PBS, as in our previous studies (6, 28). A second BGT1 antibody (Proteintech Group) was used (1:200) on PFA-fixed cells after blocking with 2.5% BSA in PBS. The peptide antigen for both antibodies was residues 595-613 of dog BGT1. The secondary antibody was goat anti-rabbit IgG conjugated to FITC (Jackson ImmunoResearch) and diluted 1:100. Cells were washed and counterstained for 5 min in propidium iodide (2 μg/ml) to visualize nucleic acids and nuclei. Triton extracts of MDCK cells were prepared and processed for Western blotting as described previously (27, 28). Rabbit antibodies to the adenosine A₁ receptor (Abcam) were used at 1:1,000 dilution, and the secondary antibody was goat anti-rabbit IgG conjugated to HRP at 1:5,000 dilution.

Data are means ± SD of at least three separate experiments. In each transport experiment, the mean value was derived from triplicate determinations. Where appropriate, different groups were compared by Student's t-test or by ANOVA and Tukey's test for multiple comparisons, using Instat v.3.06 software (GraphPad). A probability of P < 0.05 was considered statistically significant.

RESULTS

There was rapid release of ATP from suspensions of MDCK cells in hypertonic medium when the extracellular osmolarity was decreased by addition of an aqueous 3 mM MgCl₂ solution. Release of ATP within 30 s was increased 16-fold compared with control suspensions that were mixed with an equal volume of hypertonic solution (Fig. 1). The 330 pmol of ATP/mg cell protein that were released during 30-s hypotonic stress were recovered in 0.245 ml supernatant and equate to a final extracellular concentration of 1.3 μM.

Fig. 1.

Madin-Darby canine kidney (MDCK) cells release ATP within 30 s when extracellular osmolarity is decreased. After a 24-h adaptation to hypertonic growth medium, MDCK cells were resuspended in hypertonic solution and mixed with an equal volume of either the same solution (controls) or water containing 3 mM MgCl₂ for 30 s before rapid centrifugation. ATP was measured in the supernatants by the luciferase assay. *P < 0.001, compared with controls. Data are means ± SD from 3 separate experiments.

After hypertonic adaptation of MDCK monolayers for 24 h, the addition of ATP at a range of concentrations (0.01–1.00 mM) for 30 min produced a dose-dependent inhibition of BGT1 transport activity (Fig. 2, top left). Maximum inhibition (45%) occurred at an extracellular ATP concentration of 0.1 mM, which was used routinely in further experiments. The lowest concentration of ATP that produced significant inhibition was 0.03 mM, which is an order of magnitude greater than that calculated to occur in the previous experiment (Fig. 1). However, the extracellular ATP levels may be much greater in the interstitium of the kidney medulla in vivo. At the same concentration range, ATP also inhibited TAUT but the inhibition was not dose dependent and never exceeded 20% (Fig. 2, top right). A 30-min exposure to 0.1 mM ATP produced no significant change in amino acid transport system A (Fig. 2, bottom).

Fig. 2.

Exogenous ATP causes dose-dependent inhibition of betaine/GABA transporter (BGT1) transport activity but not taurine transport (TAUT) or system A transport. Incubations with ATP were at 37°C and transport assays were at 25°C. Cells were adapted to hypertonic medium for either 5 h (system A) or 24 h (BGT1; TAUT) before use. Data are based on 5-min uptakes and are means ± SD from 3–4 separate experiments. *P < 0.05, compared with controls (ATP absent).

The inhibitory action of 0.1 mM ATP on BGT1 transport was reproduced by a 30-min incubation with the nonhydrolyzable analog ATP-γS, indicating that intact ATP was inhibitory (Fig. 3, right). ATP action was likely mediated by purinergic P2 receptors since a similar inhibition of BGT1 transport occurred after a 30-min treatment with either Me-ATP (P2X receptor agonist) at 0.03 mM or UTP (P2Y receptor agonist) at 0.1 mM (Fig. 3, left; Ref. 45).

Fig. 3.

Inhibitory action of ATP on BGT1 transport activity is reproduced by the ATP-γS analog (right) and by agonists for P2X (Me-ATP) and P2Y (UTP) receptors (left). ATP-γS, adenosine 5′-[γ-thio]triphosphate; Me-ATP, α,β-methylene-ATP. Data are based on 5-min uptakes and are means ± SD from 3–4 separate experiments. *Significantly different (P < 0.05) from untreated control group.

Additional experiments showed that ADP or AMP at 0.1 mM reproduced the action of ATP on BGT1 transport during a 30-min incubation with MDCK cells. Furthermore, when extracellular ATP levels were monitored during incubation of MDCK monolayers with 0.05 mM ATP, only 1% of the initial amount of ATP was detected after 60 min and the half-life of ATP destruction was 3–4 min (Fig. 4, left). Adenosine is the final product of ATP degradation by phosphatases and ectonucleotidases (19), and a 30-min incubation of MDCK cells with adenosine at a concentration range of 0.01–1.00 mM produced dose-dependent inhibition of BGT1 transport (Fig. 4, right). The inhibition at 0.1 mM was 38%. Like ATP, adenosine did not significantly inhibit the transport activity of TAUT. Inclusion of adenosine deaminase, which converts adenosine to inosine, during the incubations decreased the inhibition of BGT1 at all adenosine concentrations (Fig. 4, right). This suggests that the inhibitory action of ATP during a 30-min treatment may be due to the combined action of adenosine nucleotides and adenosine itself, possibly activating both P1 and P2 purinergic receptors (16, 19). This may explain, at least in part, why the inhibitory action of 0.1 mM ATP was significantly greater after a 30-min incubation (43 ± 5%) compared with incubation for only 10 min (25 ± 5%; means ± SD; n = 4; P < 0.05).

Fig. 4.

Left: destruction of ATP during incubation of MDCK monolayers with 0.05 mM ATP at 37°C. At each time point, 10-μl aliquots of the incubation solution were removed, centrifuged for 1 min at 13,000 g, and stored at 4°C before ATP assay with luciferase. Right: dose-dependent inhibition of BGT1 transport activity during 30-min incubation of MDCK monolayers with adenosine (0.01–1.00 mM) at 37°C. Inclusion of adenosine deaminase (AD; 5 U/ml) reduced the level of inhibition. Data are means values from 2 independent experiments, each performed in triplicate.

Some actions of ATP on renal medullary cells may be mediated by cAMP or nitric oxide (25, 48). We reported previously that stimulation of cAMP levels by IBMX and forskolin did not affect BGT1 transport in MDCK cells (28). New data showed that the inhibitory action of 0.1 mM ATP (30-min treatment) on BGT1 transport in MDCK cells was reproduced in the presence of 0.3 mM IBMX, suggesting that this action of ATP was independent of an increase in cAMP. Similarly, ATP action on BGT1 transport was not blocked in the presence of 0.5 mM l-NAME, a nonspecific inhibitor of nitric oxide synthases (9), indicating that the inhibition was not mediated by nitric oxide (data not shown).

The inhibition of BGT1 transport by ATP and adenosine was observed also in mIMCD3 cells. Both P2X and P2Y receptors and adenosine responsiveness were previously demonstrated in this cell line derived from medullary collecting duct (63). The inhibition by both nucleotides was blocked completely when the cells were pretreated with pertussis toxin (1 μg/ml) for 5 h (Fig. 5), suggesting involvement of the inhibitory G protein (G_i) in intracellular signaling (3, 24). Treatment with toxin alone had no effect on BGT1 transport.

Fig. 5.

ATP and adenosine (ADO) at 0.1 mM also inhibited BGT1 transport in mIMCD3 cell monolayers, and the inhibitory effect was blocked when the cells were pretreated with pertussis toxin (PTX; 1 μg/ml) for 5 h. Transport in cells not exposed to nucleotides (group C) was 198 ± 43 in controls and 186 ± 26 pmol·mg⁻¹·10 min⁻¹ after PTX (means ± SD; n = 4 experiments) and was not significantly different. *Significantly different (P < 0.05) from untreated control group C.

ATP is known to increase intracellular Ca²⁺ in renal cells, including MDCK (1, 13, 18), and we (28) previously reported that increased intracellular Ca²⁺ is associated with inhibition of BGT1 in MDCK cells. We confirmed that ATP produced a rapid increase in intracellular Ca²⁺ in MDCK cells, based on fura 2 fluorescence measurements. There was a prompt increase in the fluorescence ratio (340/380 nm) in response to 0.1 mM ATP (Fig. 6, top left), and the response was significantly greater in the presence of extracellular Ca²⁺ (2 mM) compared with its absence (extracellular Ca²⁺ replaced by 0.05 mM EGTA). The change in ratio (calculated from baseline to peak) was 0.62 ± 0.11 with Ca²⁺ compared with 0.20 ± 0.04 without Ca²⁺ (P < 0.005; n = 17). This is consistent with ATP stimulating both Ca²⁺ release from intracellular stores via P2Y receptors and Ca²⁺ influx from the extracellular solution. This was confirmed by the finding that, in Ca²⁺-free superfusate containing MnCl₂, addition of ATP produced rapid quenching of fura 2 fluorescence at 360 nm, consistent with opening of Ca²⁺ influx pathways and Mn²⁺ entry (Fig. 6, top right). The dual action of ATP was observed in the same group of cells when ATP addition was begun in Ca²⁺-free superfusate to release Ca²⁺ from intracellular stores. The increased fluorescence decayed due to sequestration and/or efflux of Ca²⁺. At this point, the initial perfusate was replaced with normal medium containing both ATP and 2 mM Ca²⁺. This produced a second increase in the fura 2 fluorescence ratio that decayed more slowly than the first, possibly because Ca²⁺ influx swamped the pathways for sequestration and/or efflux. (Fig. 6, bottom). One possible influx pathway is via P2X receptor channels (19). The prompt response to ATP indicates that at least part of the increase in Ca²⁺ was due to intact ATP.

Fig. 6.

ATP (0.1 mM) produced a prompt increase in the fura 2 fluorescence ratio in MDCK monolayers perfused with solution containing 2 mM Ca²⁺ (top, left). This was due in part to increased Ca²⁺ influx because ATP stimulated rapid quenching of fura 2 by extracellular Mn²⁺ (2 mM) in the absence of extracellular Ca²⁺ (top, right). ATP treatment in Ca²⁺-free solution also increased the fluorescence ratio, suggesting mobilization of intracellular Ca²⁺ stores (bottom). This was followed by a second fluorescence peak when extracellular calcium was restored, consistent with Ca²⁺ influx. A single curve is shown that represents the mean ratio derived from the number of cells indicated on the graphs.

Similar experiments revealed that adenosine also increased intracellular Ca²⁺ by mobilizing intracellular stores and increasing Ca²⁺ influx. However, adenosine was less effective than ATP when compared directly. While normal medium containing 2 mM Ca²⁺ was perfused, addition of 0.1 mM adenosine increased the fura 2 fluorescence ratio by 0.36 ± 0.03 from baseline to peak (n = 28 cells), compared with an increase of 0.92 ± 0.05 (n = 30 cells; means ± SD; P < 0.001) in response to 0.1 mM ATP.

Activation of the A₁ adenosine receptor in kidney cells may be linked via PLC to mobilization of intracellular Ca²⁺ (11–13). We next tested the action of the A₁ receptor agonist CCPA (47, 53) on BGT1 transport in MDCK cells. A 30-min incubation of MDCK monolayers with 20 μM CCPA significantly inhibited BGT1 transport by 30 ± 7% (n = 3) but did not inhibit TAUT. CCPA (1 μM) also increased intracellular Ca²⁺. The increase in the fura 2 fluorescence ratio (baseline to peak) was 1.0 ± 0.2 (n = 8) with Ca²⁺ in the extracellular solution. The increase was due, in part, to influx of Ca²⁺ as determined by Mn²⁺ quenching of fura 2. Thus, the specific A₁ agonist reproduced the actions of adenosine. Western blotting of triton extracts from MDCK cells confirmed the presence of A₁ adenosine receptor protein (results not shown).

Antagonists were used in an attempt to confirm the roles of P2 and A₁ receptors in inhibition of BGT1 transport. Surprisingly, the inhibitory action of both Me-ATP and UTP agonists (30-min treatment) on BGT1 transport was enhanced in the presence 0.1 mM suramin, an antagonist of both P2X and P2Y receptors (8, 54; Fig. 7, top left). However, suramin has been reported to have a broad range of actions, which includes increasing intracellular Ca²⁺ in kidney cells at a similar concentration range (3, 63). The mechanism may involve mobilization of intracellular stores (23, 38). Suramin (0.1 mM) also produced a prompt increase in intracellular Ca²⁺ in MDCK cells (Fig. 7, top right). Similarly, the inhibitory action of adenosine (50 μM, 30 min) on BGT1 transport was not blocked by DPCPX (50 μM), an antagonist of the A₁ adenosine receptor (47, 53; Fig. 7, bottom left). This may be due in part to the finding that DPCPX (20 μM) also increased intracellular Ca²⁺ (Fig. 7, bottom right). The mechanism is not known but may result in influx of extracellular Ca²⁺, based on the relatively slow rise in the fura 2 fluorescence ratio. Overall this pharmacological approach served to emphasize the importance of intracellular Ca²⁺ in the acute regulation of BGT1 transport activity, without confirming the proposed role for P2 and A₁ receptors.

Fig. 7.

Top: nonspecific P2 receptor antagonist suramin (SUR; 0.1 mM) potentiated the inhibitory action of the P2X and P2Y agonists MeATP (0.05 mM) and UTP (0.10 mM) on BGT1 transport. *P < 0.01, compared with transport in the presence of MeATP or UTP alone (means ± SD; n = 4). SUR (0.1 mM) also produced a prompt increase in the fura 2 fluorescence ratio (mean value from 42 cells) with extracellular Ca²⁺ present. Bottom: A₁ adenosine receptor antagonist DPCPX (50 μM) did not block the inhibition of BGT1 transport by adenosine (50 μM). *P < 0.01 compared with controls (C; means ± SD; n = 3). DPCPX (20 μM) also increased the fura 2 fluorescence ratio (mean value from 9 cells) with extracellular Ca²⁺ present.

An alternative approach was to use U73122, the membrane-permeable inhibitor of PLC. It was previously shown in MDCK cells that ATP stimulated the production of inositol trisphosphate, and this stimulation was blocked by pretreatment with U73122 (15, 63). In the present study, inhibition of BGT1 transport by ATP or adenosine was blocked completely in MDCK cells pretreated with 5 μM U73122 for 30 min. For example, BGT1 transport after a 30-min exposure to 0.1 mM ATP was 70 ± 6% (means ± SD; n = 4; P < 0.001) of the value in untreated controls. In contrast in the presence of U73122, BGT1 transport in cells exposed to ATP was 102 ± 9% (n = 3) of the value in controls exposed only to U73122. In the absence of ATP, U73122 had no significant effect on BGT1 transport, which was 1,661 ± 187 pmol·mg⁻¹·10 min⁻¹ (n = 4) in untreated controls. The normal inhibitory action of ATP was reproduced in the presence of 5 μM U73343, the inactive analog of U73122 (15). These findings, together with the pertussis toxin studies in mIMCD3 cells (Fig. 5), provide indirect support for a role of P2Y and A₁ receptors (coupled to G_i proteins) in mediating ATP and adenosine action on BGT1 transport.

When MDCK cells were preloaded with BAPTA-AM to chelate intracellular calcium (1, 40), there was no significant increase in the fura 2 fluorescence ratio in response to addition of 0.1 mM ATP, as expected, (Fig. 8, top) and the inhibitory actions of both ATP and adenosine on BGT1 transport were blocked completely (Fig. 8, bottom).

Fig. 8.

Top: time course of changes in the fura 2 fluorescence ratio in response to addition of 0.1 mM ATP in MDCK control cells (left) and in cells preloaded for 30 min with 25 μM BAPTA-AM (right). Extracellular solution contained 2 mM Ca²⁺. Bottom: ATP produced no significant change in the fluorescence ratio in BAPTA-loaded cells (means ± SE of 30 cells). *Significantly different (P < 0.01) compared with the baseline fluorescence ratio (left). Neither ATP nor ADP (0.1 mM; 30 min) inhibited the BGT1 transport activity in BAPTA-loaded cells (means ± SD of 4 separate experiments). *Significantly different (P < 0.03) from the %inhibition in control cells (right). ADO, adenosine.

Distribution of BGT1 protein in MDCK cell monolayers was determined by confocal microscopy. After paraformaldehyde fixation, BGT1 was primarily detected in the plasma membrane after hypertonic stress (500 mosmol/kgH₂O) for 24 h compared with the intracellular distribution in cells maintained in isotonic medium (Fig. 9, A and B), as we have reported previously (6, 28). Treatment of hypertonic cells for 30 min with 0.1 mM ATP or 0.1 mM adenosine produced punctuate green fluorescence in the cytoplasm of many cells, consistent with endocytic internalization of BGT1 protein (Fig. 9, C–D). This effect of ATP was confirmed in separate experiments using a different fixative (methanol) and a different BGT1 antibody (Fig. 9, E and F). Preloading the cells with BAPTA-AM did not change the plasma membrane localization of BGT1 but blocked almost completely the internalization induced by ATP (Fig. 9F) and ADP (not shown).

Fig. 9.

Distribution of BGT1 protein (green) in MDCK cell monolayers is primarily in the plasma membrane after hypertonic stress (500 mosmol/kgH₂O) for 24 h (B) in marked contrast to cells maintained in isotonic medium (A). Treatment of hypertonic cells for 30 min with 0.1 mM ATP (C) or 0.1 mM adenosine (D) produced punctuate green fluorescence (arrows) in the cytoplasm of many cells, consistent with endocytic internalization of BGT1 protein. In contrast, almost no intracellular punctuate green fluorescence was observed when BAPTA-loaded cells were treated with 0.1 mM ATP (F), compared with the action of ATP on control cells (E). Cells in A–D were fixed in paraformaldehyde, permeabilized, and BGT1 was detected with BGT1 antibody from Proteintech. Cells in E–F were methanol fixed and BGT1 was detected with antibody provided by Dr. H. M. Kwon. Nucleic acids and nuclei were counterstained with propidium iodide (red). Bar = 20 μm.

DISCUSSION

ATP release is induced by mechanical stimuli, such as hypotonic swelling, shear stress, or mechanical strain, and has been observed in almost all eukaryotic cells (5). Potential release mechanisms include utilization of specific transporters, ion channels, and exocytosis (5, 45, 56, 57). Extracellular ATP interacts with purinergic receptors, which are widespread in renal blood vessels and nephron segments (19, 31), to regulate a broad range of physiological responses. For example, ATP acts as an autocrine factor that couples an increase in cell volume to opening of ion channels in MDCK cells (1, 58). It has been estimated that release of only 0.1% of the intracellular ATP pool will activate local purinergic receptors (45). Here we confirm rapid release of ATP from MDCK cells in response to a decrease in extracellular osmolarity (Fig. 1).

The inhibitory action of extracellular ATP on BGT1 transport was observed in MDCK and mIMCD3 cell lines, both known to possess purinergic receptors (18, 60). ATP action was relatively specific since the system A amino acid transporter was not altered and the TAUT transporter was inhibited much less over the same concentration range (Fig. 2). In this in vitro system, extracellular ATP was degraded (Fig. 4), indicating that other adenosine nucleotides (ADP and AMP) and adenosine may contribute to the inhibition of BGT1 transport. In general, P2 receptors have a greater affinity for ATP and ADP than for AMP or adenosine, while the reverse is true for P1 receptors (19). It was previously reported that the actions of ATP and adenosine are additive (16), and, at least in MDCK cells, it appears that extracellular adenosine may promote additional release of ATP (35, 36). If this also occurs in the kidney medulla, it would provide a feedback system for increased production of extracellular adenosine and the various nucleotides and may achieve the concentration necessary to activate purinergic receptors in vivo. Rapid destruction of extracellular ATP observed in MDCK cells (Fig. 4) could be important in vivo to allow potential restimulation of P2 receptors by new ATP (17).

Purinergic P2 receptors are either ligand-gated cation channels (P2X) or are coupled to heterotrimeric G proteins (P2Y). Multiple subtypes of P2X receptor channels and P2Y receptors are present throughout the nephron (19), including both basolateral and apical membranes of the medullary collecting duct (45). A consistent finding in the renal collecting duct is that major transport activity is downregulated when luminal and/or basolateral P2Y receptors are activated (31). P2Y activation either mobilizes intracellular Ca²⁺ or alters cAMP levels to modulate activity of Ca²⁺- and cAMP-sensitive ion channels (54). P2X receptor channels directly admit extracellular Ca²⁺. Both P2X and P2Y receptor subtypes have been implicated in inhibition of Na⁺ channels in collecting duct cells (34). This study with MDCK cells has verified both mobilization of intracellular Ca²⁺ and influx of extracellular Ca²⁺ in response to extracellular ATP or adenosine. P2X and P2Y receptors, as well as adenosine A₁ receptors, are present in MDCK cells, although the P2Y receptor subtypes may be predominant (54). Blockade of the inhibitory actions of ATP and adenosine on BGT1 transport by U73122 suggests a role for PLC and, by implication, inositol trisphosphate in the increase in intracellular Ca²⁺. This indicates that ATP action on BGT1 may be mediated primarily by activation of P2Y receptors. Blockade by pertussis toxin indicates the receptors may be coupled to G_i proteins, which presumably activate PLC through βγ-dimers released from Gα_i (15, 21). Adenosine, by binding to A₁ receptors coupled to G_i proteins, could activate PLC through a similar mechanism (19, 21, 35). Using MDCK cells, we previously found rapid downregulation of renal BGT1 in response to increased intracellular Ca²⁺, but no response to increased cAMP (28), which rules out a cAMP dependent mechanism.

The possible significance of these findings is that in vivo an increase in intracellular Ca²⁺ could occur through autocrine signaling by ATP, which may be released from medullary cells when extracellular hypertonicity falls, e.g., during diuresis. The resultant inhibition of BGT1 transport would complement the rapid loss of betaine through swelling activated channels (14, 39, 51). The action of ATP is less efficacious than phorbol esters (28) but is just as rapid. It occurs even in hypertonic medium (Fig. 8), a condition that normally inhibits endocytosis (29). This suggests that extracellular ATP may be an important physiological regulator of BGT1 transport in the hypertonic renal medulla.

The increase in intracellular Ca²⁺ is the response common to all the purinergic agonists and antagonists that were tested. It appears to be essential for internalization of BGT1 protein from the plasma membrane since blocking the increase in Ca²⁺ with BAPTA also prevented internalization of BGT1 protein and inhibition of transport (Fig. 8). The molecular mechanism may involve protein kinase C-mediated phosphorylation of BGT1, which disrupts the tethering of BGT1 protein to the plasma membrane, as previously discussed in detail (33). Whether the internalized BGT1 is targeted for degradation or recycling remains to be determined.

In summary, we have presented evidence to support the concept that renal BGT1 transport activity can be dynamically regulated. Both ATP and adenosine may be important physiological regulators of BGT1, especially for facilitating downregulation in response to a decrease in the extracellular osmolarity in the renal medulla.

GRANTS

This work was supported by an IUPUI Research Support Funds Grant, a grant-in-aid from the American Heart Association Midwest Affiliate (S. A. Kempson) and National Heart, Lung, and Blood Institute Grant HL-062552 (M. Sturek). J. M. Edwards was supported by a Translational Research Fellowship from Indiana University School of Medicine. A. Osborn was a participant in Indianapolis Project SEED, sponsored by the American Chemical Society.