ATP Activates cAMP Production via Multiple Purinergic Receptors in MDCK-D1 Epithelial Cells

Extracellular nucleotides regulate function in many cell types via activation of multiple P2-purinergic receptor subtypes. However, it has been difficult to define which individual subtypes mediate responses to the physiological agonist ATP. We report a novel means to determine this by exploiting the differential activation of an autocrine/paracrine signaling pathway. We used Madin-Darby canine kidney epithelial cells (MDCK-D1) and assessed the regulation of cAMP formation by nucleotides. We found that ATP, 2-methylthio-ATP (MT-ATP) and UTP increase cAMP production. The cyclooxygenase inhibitor indomethacin completely inhibited UTP-stimulated, did not inhibit MT-ATP-stimulated, and only partially blocked ATP-stimulated cAMP formation. In parallel studies, ATP and UTP but not MT-ATP stimulated prostaglandin production. By pretreating cells with indomethacin to eliminate the P2Y2/prostaglandin component of cAMP formation, we could assess the indomethacin-insensitive P2 receptor component. Under these conditions, ATP displayed a ten-fold lower potency for stimulation of cAMP formation compared with untreated cells. These data indicate that ATP preferentially activates P2Y2 relative to other P2 receptors in MDCK-D1 cells (P2Y1 and P2Y11, as shown by reverse transcriptase polymerase chain reaction) and that P2Y2 receptor activation is the principal means by which ATP increases cAMP formation in these cells. Blockade of autocrine/paracrine signaling can aid in dissecting the contribution of multiple receptor subtypes activated by an agonist.

The identification of multiple receptor subtypes for physiologic agonists can lead to a difficult problem. How can one define the preferred receptor interaction for a physiologic agonist? As an example, a wide variety of cells have been shown to possess receptors for extracellular nucleotides. ATP, which is an important extracellular-signaling molecule, is stored and released from sympathetic neurotransmitter vesicles and from stressed/damaged cells. Receptors that interact specifically with ATP (classified as P 2 -purinergic receptors) are present in many tissues and cell types (1)(2)(3)(4). The P 2Y1 and P 2Y2 subtypes are two widely expressed G-protein-coupled purinergic receptors that differ in their specificity for nucleotides. P 2Y1 receptors bind ATP, and the synthetic agonist 2-methylthio-ATP (MT-ATP), 1 whereas P 2Y2 receptors respond to both ATP and UTP (2,5). Additionally, ATP can be metabolized by ecto-ATPases to adenosine, the agonist for P 1 -purinergic receptors.
We have studied P 2 receptor-mediated signaling events in a well differentiated kidney epithelial cell line, Madin-Darby canine kidney (MDCK-D1), derived from distal tubule/collecting duct. P 2 -purinergic receptors expressed on these cells respond to extracellular nucleotides by increasing the activity of various phospholipases (phospholipases C, D, and A 2 ), the activity of protein kinase C, the concentration of intracellular calcium, the production of cAMP, and transepithelial ion transport (6 -12).
In several other cell types, stimulation of P 2 receptors decreases intracellular cAMP production via a pertussis toxinsensitive mechanism (13)(14)(15)(16)(17)(18). In contrast, ATP was shown to activate adenylyl cyclase in HL60 cells (19,20). Using MDCK-D1 cells, we found that P 2 receptor agonists increase cAMP production (10) and tested hypotheses regarding several potential mediators for the increase in cAMP production, protein kinase C, G s , adenosine, and arachidonic acid metabolites. Our results indicated that ATP and UTP, acting at P 2 -purinergic receptors, increase cAMP via an indomethacin-sensitive pathway, implying that cyclooxygenase-derived products mediate this response.
In the current study, we sought to distinguish between the actions of ATP and other nucleotides that act at different P 2 receptor subtypes. Our data indicate that P 2Y2 receptor-mediated cAMP formation occurs via an indomethacin-sensitive pathway. In contrast, MT-ATP (acting at other classes of receptors, perhaps P 2Y1 and/or P 2Y11 receptors) preferentially couples to indomethacin-insensitive pathways to increase cAMP production. The physiological agonist ATP increases cAMP via preferential interaction with P 2Y2 receptors.

MATERIALS AND METHODS
Cell Culture-MDCK-D 1 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% serum as described previously (8). Cells were used in assays at 60 -80% confluency. Basal cAMP levels were increased at cell densities greater than this.
Measurement of cAMP Accumulation-Before treatment of cells, growth medium was removed and cells were equilibrated for 30 min at * This work was supported by National Institutes of Health Grants (GM 40781, GM 31987, and HL 53773), a grant from the California affiliate of the American Lung Association (to S. R. P.), and a grant (Forschungsstipendium) from the Deutsche Forschungsgemeinschaft, Bonn, Germany (to L. C. R.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ‡ Present address: Innere Medizin IV, Universitaetsklinik Freiburg, Hugstetter Str. 55, D-79106 Freiburg, Germany.
To whom correspondence should be addressed. Tel.: 619-534-2295; Fax: 619-822-1007; E-mail: pinsel@ucsd.edu. 37°C in serum-free Dulbecco's modified Eagle's medium containing 20 mM HEPES buffer (DMEH, pH 7.4). Subsequently, cells were incubated in fresh DMEH and various agents as described in the figure legends. Unless otherwise indicated, incubations with agonist were conducted for 5 min at 37°C in the presence of 200 M isobutylmethylxanthine, a phosphodiesterase inhibitor, and terminated by aspiration of medium and addition of 7.5% trichloroacetic acid. Trichloroacetic acid extracts were frozen (Ϫ20°C) until assay. Intracellular cAMP levels were determined by radioimmunoassay (Calbiochem) of trichloroacetic acid extracts following acetylation according to the manufacturer's protocol. Production of cAMP was normalized to the amount of acid-insoluble protein assayed by the Bio-Rad protein assay.
Measurement of PGE 2 Release-Cells were treated with indicated concentrations of nucleotide for 5 min in a total volume of 0.5 ml. Incubation medium was removed and assayed for PGE 2 content using an enzyme immunoassay kit (Cayman Chemical, Ann Arbor, MI). The amount of PGE 2 present in the medium in the absence of nucleotide was subtracted from each point.
Reverse Transcriptase (RT) Reaction-RNA (10 g) was reverse transcribed in 25 l of RT buffer (Life Technologies, Inc.) together with 0.002 OD units of random hexamers, 10 mM dithiothreitol, 800 M dNTP, and 200 units Moloney murine leukemia virus RT. After a 1-h incubation at 37°C, reactions were stopped by boiling for 4 min and dilution to 100 l in RNase-free water. To control for possible DNA contamination of the samples, reactions were also performed in the absence of RT enzyme.
PCR and Sequencing-PCR conditions were identical for all primers (primers: P 2y1 forward 5Ј-ACCCTGTACGGCAGCATCCTSTTCCTCAC-3Ј, reverse 5Ј-AGGWAGSAGASGGCGAAGAC-3Ј expected product size 422 base pairs; P 2y2 : forward 5Ј-AGTCCCCCGTGCTCTACTTT-3Ј, reverse 5Ј-GTCAGTCCTGTCCCACCTGT-3Ј expected product size 539 base pairs; P 2y11 : forward 5Ј-CTGGTGGTTGAGTTCCTGGT-3Ј, reverse 5Ј-GTTGCAGGTGAAGAGGAAGC-3Ј expected product size 234 base pairs): 10 l of reverse transcribed RNA or 1 g of genomic DNA was added to a solution of 20 M each of forward and reverse primer, 2.5 mM MgCl 2 buffer (Perkin-Elmer), PCR buffer (50 mM KCl, 10 mM Tris-HCl, pH 8.3), 0.2 mM dNTPs (Amersham Pharmacia Biotech), 5 units of Amplitaq Gold Polymerase (Perkin-Elmer) and dH 2 O in a total volume of 50 l. Temperature cycling proceeded as follows: 1 cycle at 95°C for 10 min to activate the enzyme, 95°C for 30 s, 60°C for 90 s and 72°C for 90 s, for 40 cycles, followed by 72°C for 10 min. PCR products were then subjected to gel electrophoresis on a 1% Seakem-agarose gel (FMC Bioproducts). The bands were extracted using a Qiaquick gel extraction kit (Qiagen). The DNA was resuspended in Tris-EDTA buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0), and 1 volume of the gel-extracted PCR product was purified using a Centricon concentrator (Amicon). Purified fragments were sequenced (ABI automated DNA sequencer, model 377) using the same forward primers that were used to generate the PCR fragments.

RESULTS AND DISCUSSION
In previous studies, we showed that extracellular nucleotides that act at P 2 -purinergic receptors increase cAMP formation in MDCK-D1 kidney epithelial cells (10). The ability of ATP and UTP to stimulate cAMP was inhibited by the cyclooxygenase inhibitor indomethacin, indicating that this response was secondary to the release of arachidonic acid and most likely by its conversion to one or more prostaglandins. Because ATP is generally considered the endogenous ligand for P 2 -purinergic receptors, our initial assumption was that P 2 receptors coupled to adenylyl cyclase activation secondary to the production of prostaglandins. Using indomethacin at a concentration that completely inhibits the ability of exogenous arachidonic acid to stimulate cAMP formation (Fig. 1A), we found that the increase in cAMP in response to UTP was also abolished (Fig. 1B). In contrast, the cyclooxygenase inhibitor did not significantly inhibit cAMP formation in response to the P 2Y1 receptor agonist MT-ATP. In the presence of 1 M indomethacin, the ability of ATP to increase cAMP was reduced, but only by Ϸ80%. These results suggested the involvement of two signaling pathways in coupling P 2 receptors to adenylyl cyclase activation: P 2Y2 receptors, whose ability to increase cAMP formation appears to be entirely dependent upon the formation of prostaglandin; and another P 2Y receptor, whose activation of adenylyl cyclase is apparently independent of prostaglandin production.
We sought to determine the identity of the prostaglandins that coupled P 2 receptors with adenylyl cyclase activation. MDCK cells produce and release PGE 2 , prostaglandin I 2 (prostacyclin), and prostaglandin F 2␣ (21,22). Of these, only PGE 2 and prostacyclin, tested using a stable prostacyclin analog (6␣-PGI 1 ), were positively coupled to adenylyl cyclase activation ( Fig. 2A) (10,21). To correlate prostaglandin production with P 2 receptor activation, we measured the PGE 2 released following the addition of purinergic receptor agonists. At concentrations that increase cAMP formation, both ATP and UTP substantially increased PGE 2 release from MDCK cells (EC 50 Ϸ 30 M, Fig. 2B). In contrast, receptor activation with MT-ATP did not increase PGE 2 release even at concentrations that maximally stimulated cAMP formation (Figs. 2B and 4A). The production of PGE 2 was completely blocked by indomethacin (1 M) pretreatment (data not shown). These data correlate well with the sensitivity of cAMP production to indomethacin for P 2Y2 receptor agonists (Fig. 1). Thus, by measuring cAMP formation in MDCK cells pretreated with indomethacin, we could specifically assess activation of indomethacin-insensitive P 2Y receptors independent of the contribution of P 2Y2 receptor activation. Having defined conditions that allowed us to distinguish between indomethacin-insensitive and indomethacin-sensitive (P 2Y2 ) receptor activation, we assessed the relative contributions of these two receptor pathways in cAMP accumulation stimulated by the physiological agonist ATP. We compared the ability of ATP to increase cAMP formation via both pathways (i.e. absence of indomethacin) with that of the indomethacininsensitive pathway alone (i.e. following pretreatment with indomethacin). As shown in Fig. 3A, pretreatment of cells with indomethacin had no significant effect on the ability of MT-ATP to stimulate cAMP formation (EC 50 Ϸ 25 M in each case). In contrast, ATP displayed a substantial difference in its ability to activate adenylyl cyclase following pretreatment of MDCK cells with indomethacin (Fig. 3B). In untreated cells, ATP, acting via both pathways, displayed an apparent EC 50 Ϸ 10 M. However, following indomethacin treatment, ATP, acting selectively via indomethacin-insensitive receptors, demonstrated a greatly reduced potency to increase cAMP (EC 50 Ͼ 100 M). This result indicates that in untreated cells, the cAMP formed in response to ATP arises by the preferential activation of P 2Y2 receptors.
The substantial inhibition of ATP-stimulated cAMP formation by indomethacin suggested that ATP was activating multiple P 2 receptor pathways, but with greater apparent potency at P 2Y2 receptors. Given the previous evidence for expression of P 2Y1 and P 2Y2 receptors in MDCK-D1 cells (8,23) (and the efficacy of MT-ATP in stimulating cAMP formation), we initially hypothesized that the indomethacin-insensitive response to MT-ATP resulted from activation of P 2Y1 receptors. To test this assumption, we used suramin as a P 2Y1 receptor antago-nist. In cells preincubated with indomethacin, suramin competitively inhibited MT-ATP-stimulated cAMP formation (Fig.  4A) and displayed a pA 2 Ϸ 5.2 (i.e. an antagonist concentration required to increase EC 50 for agonist 2-fold), which closely matches that previously reported for competition at P 2Y1 receptors (1). Suramin also completely inhibited the indomethacininsensitive component of ATP-stimulated cAMP formation (Fig. 4B). These data are consistent with the notion that ATP acts at P 2Y1 receptors to increase cAMP by a prostaglandinindependent pathway.
Recently, a new P 2Y receptor subtype, P 2Y11 , which responds to ATP and MT-ATP, was identified and shown to couple to adenylyl cyclase activation (24). However, ATP activated the cloned P 2Y11 receptor with much greater potency (EC 50 ϭ 30 M) than we observed for the indomethacin-insensitive cAMP response in MDCK cells. Nevertheless, we assessed whether MDCK-D1 cells express this receptor subtype. Indeed, RT-PCR analysis indicated that P 2Y1 , P 2Y2 , and P 2Y11 receptors are all expressed in MDCK-D1 cells (Fig. 5). PCR of DNA and RT-PCR of RNA yielded bands of the anticipated size for each of the three receptors. Omission of reverse transcriptase from the RT-PCR reaction failed to yield any product indicating the absence of contaminating DNA. The identity of the products (in the case of the P 2Y1 , the larger of the two products) was confirmed by direct sequencing using the forward primer. Thus, although the pharmacology of the indomethacin-insensitive cAMP response is consistent with that of a P 2Y1 receptor, we cannot exclude the possible contribution of P 2Y11 receptors in mediating the indomethacin-insensitive response to ATP.
Previous studies have demonstrated the ability of P 2 receptor agonists to regulate transepithelial ion transport in MDCK cells (11,12). PGE 2 and other agents that increase cAMP are known to couple receptors to ion transport in epithelial cells. In MDCK cells, the ability of P 2Y2 , but not P 2Y1 , receptor agonists to elicit ion transport was indomethacin-sensitive (12). Similarly, the ability of exogenous arachidonic acid to alter ion transport was inhibited by indomethacin. The results that demonstrate the indomethacin-sensitive cAMP production elic-ited by P 2Y2 agonists provide the most likely molecular basis for the ability of these receptors to couple to ion transport. Moreover, our finding that MDCK cells also express P 2Y11 receptors raises the possibility that this receptor subtype may regulate ion transport via activation of adenylyl cyclase.
In MDCK-D1 cells, cAMP is a messenger that is considerably "downstream" from the initial occupancy by agonist of the P 2Y2 receptors. Nucleotide-mediated stimulation of cAMP formation requires receptor occupancy, activation of cyclic phospholipase A 2 (via apparent involvement of Ca 2ϩ and multiple protein kinase C isoforms (see Ref. 9), cyclooxygenase-mediated formation of PGE 2 , PGE 2 release from cells and autocrine/paracrine activation of PGE 2 receptors, G s , and adenylyl cyclase). Thus, indomethacin, an inhibitor of an intermediate step in this scheme, can be used to define the relative ability of ATP to act at P 2Y2 receptors in MCDK-D 1 cells. Inhibitors that act on other components of the signaling pathways, which are not shared by the different P 2 -purinergic receptors, might also be used in this manner. We predict that selective inhibitors of the receptors that couple prostaglandin binding to activation of adenylyl cyclase (presumably EP2 receptors) (25) should also distinguish the indomethacin-sensitive (P 2Y2 ) and indomethacin-insensitive components of ATP action.
The existence of multiple subtypes appears to be very common for G-protein-coupled receptors. Pharmacological approaches that involve use of receptor-selective agonists and antagonists have not kept pace with the discovery of receptor subtypes by molecular cloning strategies. In the case of P 2purinergic receptors, the absence of high affinity antagonists and the limited specificity of agonists have made it difficult to define precisely the cellular function of different receptor subtypes (4). In many cases receptor subtypes that recognize the same physiologic agonist preferentially activate different signaling pathways. In addition to the P 2 -purinergic receptors and the physiologic agonist ATP, other examples include receptors for adenosine, norepinephrine/epinephrine, dopamine, acetylcholine (muscarinic receptors), histamine, prostaglandins, and serotonin as well as receptor for certain peptides and peptide hormones (e.g. angiotensin and vasopressin) (26). Most work to date has emphasized linkage to different classes of G-proteins as the explanation for differences in signaling. The current studies show that one can use blockade of downstream signals that result from differences in signaling cascades to define contribution of different receptor subtypes that recognize the same physiologic agonist. Such downstream differences may prove useful for the analysis of other receptor systems, in particular those for which response has been attributed, at least in part, to generation of cyclooxygenase-derived products (e.g. see Refs. [27][28][29][30]. RT-PCR of RNA (ϩ) and PCR of DNA (DNA) run alongside a molecular weight marker, yielded bands of the anticipated size for the P 2y1 , P 2y2 , and P 2y11 receptors, 422-, 539-, and 234-base pairs, respectively. Omission of the reverse transcriptase from the RT-PCR reaction failed to yield any products (Ϫ).