P2Y11Receptors Activate Adenylyl Cyclase and Contribute to Nucleotide-promoted cAMP Formation in MDCK-D1Cells

Extracellular nucleotides activate P2Y receptors, thereby increasing cAMP formation in Madin-Darby canine kidney (MDCK-D1) cells, which express P2Y1, P2Y2, and P2Y11 receptors (Post, S. R., Rump, L. C., Zambon, A., Hughes, R. J., Buda, M. D., Jacobson, J. P., Kao, C. C., and Insel, P. A. (1998)J. Biol. Chem. 273, 23093–23097). The cyclooxygenase inhibitor indomethacin (indo) eliminates UTP-promoted cAMP formation (i.e. via P2Y2 receptors) but only partially blocks ATP-promoted cAMP formation. The latter response is completely blocked by the nonselective P2Y receptor antagonist suramin. We have sought to identify the mechanism for this P2Y receptor-mediated, indo-resistant cAMP formation. The agonist rank order potencies for cAMP formation were: ADPβS ≥ MT-ADP > 2-MT-ATP > ADP, ATP, ATPγS > UTP, AMP, adenosine. We found a similar rank order in MDCK-D1 cells overexpressing cloned green fluorescent protein-tagged P2Y11 receptors, but the potency of the agonists was enhanced, consistent with a P2Y11 receptor-mediated effect. cAMP generation by the P2Y1 and P2Y11 receptor agonist ADPβS was not inhibited by several P2Y1-selective antagonists (PPADS, A2P5P, and MRS 2179). Forskolin synergistically enhanced cAMP generation in response to ADPβS or PGE2, implying that, like PGE2, ADPβS activates adenylyl cyclase via Gs, a conclusion supported by results showing ADPβS and MT-ADP promoted activation of adenylyl cyclase activity in MDCK-D1 membranes. We conclude that nucleotide-promoted, indo-resistant cAMP formation in MDCK-D1 cells occurs via Gs-linked P2Y11 receptors. These data describing adenylyl cyclase activity via endogenous P2Y11 receptors define a mechanism by which released nucleotides can increase cAMP in MDCK-D1 and other P2Y11-containing cells.

Cells commonly co-express multiple receptor subtypes that recognize the same physiological agonist, but it is difficult to define which among such receptor subtypes mediates a particular response. This can be a particularly vexing problem if subtype-selective agonists and antagonists are not available. One such example is P2Y receptors, which respond to ATP and other nucleotides, are expressed in a variety of tissues and cell types, and for which few subtype-selective antagonists exist (2)(3)(4). Our laboratory has undertaken a series of studies re-lated to signal transduction by P2Y receptors in the renal epithelial cell line, MDCK 1 -D 1 (reviewed in Refs. 5 and 6). P2Y receptors in MDCK-D 1 cells modulate membrane potential and short circuit current, and the receptors regulate phospholipases, intracellular [Ca 2ϩ ], prostaglandin E 2 (PGE 2 ) formation, and cAMP accumulation (1,(7)(8)(9)(10)(11)(12)(13)(14).
Cyclooxygenase (COX) inhibitors, such as indomethacin (indo), have proven useful for studying P2Y receptors in MDCK-D 1 cells (1,12). Agonist-stimulated cAMP accumulation in MDCK-D 1 cells occurs via both indo-sensitive and -insensitive pathways. Response to the P2Y 2 agonist UTP is entirely indo-sensitive, whereas response to ATP is partially sensitive and to 2-methylthio-ATP (MT-ATP) is insensitive (1). These findings suggest that UTP, and ATP in part, stimulate P2Y 2 receptors to cause COX-mediated (perhaps by both COX1 and COX2, see Ref. 15) formation of arachidonic acid metabolites (e.g. PGE 2 ), which activate EP receptors to stimulate cAMP formation, while ATP and MT-ATP can also enhance cAMP formation via an indo-insensitive P2Y receptor pathway (1).
The present studies were designed to characterize more fully the nature of the latter pathway. Our working hypothesis, based in part on initial results obtained with MDCK-D 1 cells (12), was that the indo-resistant response might represent a P2Y 1 receptor effect. The current data show results not consistent with this hypothesis but instead suggest a key role for another receptor, the P2Y 11 receptor, in indo-resistant cAMP formation. The findings directly document a role for P2Y 11 receptors in stimulation of adenylyl cyclase activity and in potentially contributing to autocrine-paracrine regulation by nucleotides.
Measurement of cAMP Accumulation in Normal and GFP-tagged cP2Y 11 Receptor-overexpressing MDCK-D 1 Intact Cells-Prior to the treatment of the cells, the growth medium was removed, and cells were equilibrated for 30 min at 37°C in serum-free 20 mM HEPES-buffered Dulbecco's modified Eagle's medium (DMEH, pH 7.4). Subsequently, cells were incubated in fresh DMEH and various agents as shown in the figures. Incubations with the agonists ADP␤S, MT-ADP, ATP␥S, ATP, * 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.
MT-ATP, ADP, AMP, adenosine, UTP, PGE 2 , and forskolin and the antagonists PPADS, A2P5P, and MRS 2179 were conducted for 7 min at 37°C in the presence of 200 M IBMX, a phosphodiesterase inhibitor, and terminated by placing on ice and replacing the medium with 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 (Lowry and Bradford methods).
Preparation of Membranes-Membranes were prepared from MDCK-D 1 cells as follows: cells were grown to confluency on 15-cm plates and washed twice with phosphate-buffered saline, the second wash containing 0 .01% EDTA. Cells were then incubated for 15 min at 37°C in phosphate-buffered saline containing 0.01% EDTA. Detached cells were centrifuged at 1,000 rpm, and the resulting pellet was suspended in 10 ml of 4°C lysis buffer (20 mM Tris-HCl, pH 7.4, 10 mM MgCl 2 , 2 mM EDTA, 2 mM EGTA, 1 mM dithiothreitol) containing protease inhibitors (10 M leupeptin, 500 nM pepstatin, 200 M benzamidine, 200 M 4-(2-aminoethyl) benzenesulfonylfluoride-HCl (AEBSF). Suspended cells were then placed under 450 p.s.i for 10 min to lyse the cells. The resulting lysed cell mixture was centrifuged at 2,000 rpm for 15 min at 4°C to remove cell nuclei. The supernatant was separated from the nuclear pellet and centrifuged for 90 min at 15,000 rpm at 4°C. The resulting pellet was suspended in buffer A (20 mM Tris-HCl, pH 7.4, 5 mM MgCl 2 , 1 mM EDTA, 1 mM dithiothreitol), which included protease inhibitors as indicated previously. A protein assay (Bio-Rad) was run to determine membrane protein concentration. Membranes were stored at Ϫ80°C.
Membrane Adenylyl Cyclase Assays-75 g of membrane were placed in 1.5-ml Eppendorf tubes on ice. The volume was brought up to 250 l using substrate (30 M AMP-PNP), agonist, and buffer B (50 mM HEPES, pH 7.4, 1 mM EDTA, 5 mM MgCl 2 , 1 mM IBMX, 1 mM dithiothreitol, 10 M GTP, and protease inhibitors). Reaction tubes were then incubated in a 37°C water bath for 30 min. The reaction was stopped by boiling for 1 min, and samples were then centrifuged for 10 min at 4°C at 13,000 rpm. Supernatant cAMP levels were then determined using the radioimmunoassay protocol described above.
Statistics-Data shown generally represent mean Ϯ S.E. of multiple determinations. Curves shown in Figs. 1 and 3 were obtained using the sigmoidal dose response function of Graph Pad Prizm (Graphpad Software Inc., San Diego, CA); analysis using this program yielded EC 50 values shown in text and tables.

RESULTS
As a first step toward investigating the basis for indo-resistant cAMP formation in MDCK-D 1 cells, we determined the ability of different P2 agonists to elicit this response. Stimulation of cells with UTP was maximally able to increase cAMP levels ϳ3-fold over basal levels, but consistent with previous results (1,12), pretreatment with 1 M indo abolished this response (Fig. 1A). Treatment of cells with indo also partially decreased ATP-stimulated cAMP accumulation but was less effective than the inhibition of the response to UTP, particularly at higher concentrations of ATP (Fig. 1B). The methylthio-derivative of ATP, MT-ATP, produced a largely indoresistant increase in cAMP levels (Fig. 1C). ATP can be sequentially dephosphorylated to ADP, AMP, and adenosine by ecto-ATPases and nucleotidases (16,17). To assess whether ATP was responsible for the observed increase in cAMP levels or whether an ATP metabolite might be responsible, we treated cells with an ATPase-resistant form of ATP, ATP␥S (16), and obtained results comparable with those observed with ATP ( Fig. 1D), suggesting that ATP need not be metabolized to raise cAMP levels.
Like ATP, ADP elevated indo-resistant cAMP levels in a concentration-dependent manner (Fig. 1E). ADP␤S, a nonhydrolyzable analog of ADP, and MT-ADP both increased indo-resistant cAMP levels as much or more than did ADP, ATP, or ATP␥S (Fig. 1, F and G). The ADP␤S-stimulated cAMP increase was entirely indo-resistant. Stimulation of MDCK-D 1 cells with AMP and adenosine, both metabolic breakdown products of ADP, did not elevate cAMP (Ref. 12 and data not shown).
The observed pattern of response for the different agonists, in particular with reference to ATP, ADP, ADP␤S, and MT-ATP, suggested that indo-resistant cAMP accumulation might be the result of an action at the P2Y 1 receptor (2-4). We thus tested several P2Y 1 antagonists to determine whether this receptor might mediate indo-resistant cAMP accumulation. The P2Y-nonselective antagonist suramin blocked this response to 10 M ADP␤S in MDCK-D 1 cells (Fig. 2), but several other putative P2Y 1 -selective antagonists proved ineffective, including PPADS and A2P5P (2,18,19). Additionally, the highly selective P2Y 1 antagonist MRS 2179 failed to inhibit indo-resistant cAMP accumulation in MDCK-D 1 cells even though in parallel control studies MRS 2179 inhibited canine P2Y 1 -mediated inositol trisphosphate formation in P2Y 1 -transfected 1231N1 cells (Ref. 20 and data not shown). These findings suggested that the receptor responsible for the indo-insensitive effect is not the P2Y 1 receptor but instead is another receptor, perhaps the P2Y 11 receptor, which in MDCK-D 1 cells also responds to adenine bisphosphates (14,21).
Investigation into whether the P2Y 11 receptor can cause indo-resistant cAMP elevation in MDCK-D 1 cells is hampered by the lack of antagonists to this receptor. Therefore, as an alternative strategy, we used P2Y 11 receptors that we had cloned from MDCK-D 1 cells expressed as P2Y 11 -GFP receptors in MDCK-D 1 cells (14) and assessed cAMP accumulation. In the cells overexpressing P2Y 11 -GFP receptors, we found a 5-to 10-fold increase in sensitivity of indo-resistant cAMP formation in response to several nucleotides, compared with responses observed with native MDCK-D 1 cells (Table I and Fig. 3). The relative potency of the different nucleotides in P2Y 11 -GFPoverexpressing MDCK-D 1 cells was similar to that for native MDCK-D 1 cells, although compared with other agonists, ADP had a somewhat greater enhancement in apparent potency in the P2Y 11 -overexpressing cells (Table I). UTP treatment gave no response; thus overexpression of P2Y 11 receptors did not alter P2Y 2 response (data not shown). Taken together with data in the native cells ( Figs. 1 and 2), this enhancement of the indo-resistant cAMP formation in P2Y 11 -overexpressing MDCK-D 1 cells is consistent with the notion that P2Y 11 , and not P2Y 1 , receptors mediate this response.
To define the molecular mechanism underlying the indoresistant cAMP elevation in MDCK-D 1 cells, we assessed whether this response results from the coupling of a P2Y receptor, presumably the P2Y 11 receptor, to an increase in adenylyl cyclase activity. In previous studies it has been difficult to ascertain whether P2Y 11 receptors increase cAMP via a receptor/G s -mediated activation of adenylyl cyclase activity or indirectly via alterations in calcium or other regulators of cAMP formation (14,(22)(23)(24). We used two approaches to assess this possibility. The diterpene forskolin enhances coupling of G s -linked receptors to adenylyl cyclase, thereby enhancing the ability of G s receptors to raise cellular cAMP levels (12,(25)(26)(27). Co-incubation of indo-pretreated MDCK-D 1 cells with 10 M ADP␤S and 0.1 M forskolin led to a cAMP elevation that was greater than the sum of the individual responses for forskolin and ADP␤S. Co-incubation of forskolin and PGE 2 , a known G s activator in MDCK-D 1 cells (12), showed an even greater enhancement of cAMP levels than was caused by co-incubation with ADP␤S and forskolin (Fig. 4). Nevertheless, the synergistic effect obtained with forskolin and agonists, when expressed as -fold increase over the sum of separate forskolin-and agonist-stimulated cAMP accumulations, was similar or even greater for ADP␤S-stimulated cAMP generation than for PGE 2 response (Fig. 4, inset). These results suggest that ADP␤S, like PGE 2 , activates G s to promote cAMP formation in MDCK-D 1 cells.
As a more direct test for the ability of a P2Y receptor to couple to an increase in adenylyl cyclase activity, we sought to assay enzyme activity, but we reasoned that such an assay would require use of a nucleotide as a substrate, which unlike ATP was not active at the P2Y receptors that increase cAMP levels in MDCK-D 1 cells. We chose as an alternate substrate AMP-PNP, which has previously been used as a substrate for adenylyl cyclase assays (28) and which we found in preliminary studies did not promote indo-resistant cAMP accumulation (following 30 min of treatment of cells with 30 M AMP-PNP, data not shown). We also found that AMP-PNP yielded linear time-and protein-dependent cAMP generation in MDCK-D 1 membranes, whereas MT-ADP and ADP␤S, nucleotides that promoted cAMP accumulation in intact cells, did not (data not shown).
MT-ADP, ADP␤S, and PGE 2 all elevated cAMP levels in the presence of AMP-PNP, whereas UTP had no effect (Fig. 5). Thus, while MT-ADP and ADP␤S lack the ability to serve as substrates for adenylyl cyclase, in combination with AMP-PNP they, like PGE 2 , are agonists that can increase adenylyl cyclase activity, consistent with activation of P2Y 11 receptors linked to G s . In contrast, UTP, which stimulates G q -coupled P2Y 2 receptors (2,3,13), did not lead to increased adenylyl cyclase activity. Taken together with other data shown here, these findings provide direct evidence for P2Y 11  wide variety of cell types (2)(3)(4). Of the seven unique P2Y receptor subtypes (P2Y 1 , P2Y 2 , P2Y 4 , P2Y 6 , P2Y 11 , P2Y 12 , and P2Y 13 ) all, with the exception of P2Y 12 and P2Y 13 subtypes, couple via G q/11 -dependent mechanisms to the activation of phospholipase C and increases in cellular [Ca 2ϩ ], but the various receptor subtypes have quite different abilities to regulate formation of cAMP. Certain of the receptors (e.g. P2Y 12 , P2Y 13 ) are known to utilize G i -dependent mechanisms to inhibit cAMP formation (29,30), whereas other P2Y receptors may regulate adenylyl cyclase activity via effects on regulators of adenylyl cyclase, such as [Ca 2ϩ ] or protein kinase C.
MDCK-D 1 cells, in common with several other cell types, express several different P2Y receptor subtypes (1,6,12,31,32). Early work on MDCK-D 1 cells, prior to the molecular cloning and precise identification of the different P2Y receptor subtypes, emphasized the ability of added nucleotides to alter membrane potential, ion flux, or short circuit current (7,8). Recent studies have indicated that MDCK-D 1 cells release ATP and that this released nucleotide helps contribute to basal activity of signal transduction pathways (15,33). Of the three different P2Y receptors that we have thus far detected on MDCK-D 1 cells (P2Y 1 , P2Y 2 , and P2Y 11 , see Ref. 1), our previous work documented that P2Y 2 receptors mediate the indosensitive (i.e. COX-dependent) cAMP accumulation via the ability of the receptors to increase formation of PGE 2 and the subsequent activation of EP receptors linked to G s and adenylyl cyclase activity (1, 12, 13).
Several lines of evidence from the current studies, designed to identify alternative mechanisms by which ATP might increase cellular cAMP levels, are consistent with the conclusion that MDCK-D 1 P2Y 11 receptors are responsible for the COXindependent (indo-resistant) increases in cAMP formation promoted by ATP and other nucleotides: 1) the rank order of potency of agonists in promoting this response; 2) the insensitivity to blockade by several P2Y 1 -selective antagonists; 3) the increased potency of several agonists in P2Y 11 -overexpressing MDCK-D 1 cells; 4) the synergistic enhancement in response by co-incubation of cells with nucleotides plus forskolin; and 5) the ability of two P2Y 11 agonists, MT-ADP and ADP␤S, to enhance adenylyl cyclase activity in MDCK-D 1 cell membranes. Taken together, we believe these results provide strong evidence for P2Y 11 receptor-promoted stimulation of adenylyl cyclase activity, presumably secondary to enhanced enzyme activity by receptor-mediated activation of G s . This is the first data of which we are aware to document that P2Y 11 receptors directly activate adenylyl cyclase activity. At the time of the initial cloning of P2Y 11 receptors, Communi et al. (22,23) utilized two different cell types to heterologously express the cloned P2Y11 receptors and to conclude that these receptors couple to both G s and G q . However, this conclusion was open to alternative interpretations, given the use of differ-   ent cell types to assess receptor coupling to the different pathways. Data by those workers and others who have reported on the ability of P2Y 11 receptors to stimulate cAMP formation (e.g. 24,34) have involved studies with intact cells, for which the possibility of system-dependent influences has been noted (35,36). A number of factors can influence cAMP formation, degradation, or export from cells. Some of those factors are likely to be altered by activation of the G q -linked pathways, such as [Ca 2ϩ ], activity of protein kinase C and other downstream kinases (e.g. MAP kinase), and phosphodiesterase activity, etc. Our results that show activation of adenylyl cyclase activity in MDCK-D 1 cell membranes provide definitive evidence for the notion that P2Y 11 receptors stimulate adenylyl cyclase activity, presumably by direct activation of G s . The method that we used to draw this conclusion, involving use of AMP-PNP as the substrate for adenylyl cyclase assays, may prove useful to others interested in assessing effects of P2Y 11 receptors on adenylyl cyclase activity.
Overall, our results provide evidence that ATP and ADP, two physiologic nucleotides that can exist in the extracellular space, are able to raise cAMP levels in native cells via activation of P2Y 11 receptors. As such, these results provide a mechanism, in addition to activation of P2Y 2 or adenosine receptors, by which exogenous or endogenously released nucleotides can increase cellular levels of this important cyclic nucleotide. Given the evidence that MDCK-D 1 and a number of types of cells both release ATP and possess P2Y 11 receptors, nucleotide-mediated activation of P2Y 11 receptors provides a means for autocrine-paracrine regulation of epithelial and other cell types.