Post on 15-Mar-2020
Page 1
Identification and characterization of a cell-surface receptor, P2Y15, for AMP
and adenosine
Hisayo Inbe, Shinichi Watanabe, Miwa Miyawaki, Eri Tanabe, and Jeffrey A. Encinas*
Bayer Yakuhin, Ltd., Research Center Kyoto, 6-5-1-3 Kunimidai, Kizu-cho, Soraku-gun, Kyoto
619-0216, Japan
* To whom correspondence should be addressed:
Jeffrey A. Encinas
Bayer Yakuhin, Ltd., Research Center Kyoto
6-5-1-3 Kunimidai, Kizu-cho, Soraku-gun, Kyoto 619-0216, Japan
Tel.: +81-774-75-2468; Fax: +81-774-75-2506
E-mail: jencinas@post.harvard.edu
Running title: Identification of a Receptor for AMP and Adenosine
Keywords: AMP, adenosine, P2Y15, G protein-coupled receptor
JBC Papers in Press. Published on March 4, 2004 as Manuscript M400360200
Copyright 2004 by The American Society for Biochemistry and Molecular Biology, Inc.
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 2
Summary
AMP and adenosine are found in all cell types and can be released by cells or created extracellularly
from the breakdown of ATP and ADP. We have identified an orphan G protein-coupled receptor
with homology to the P2Y family of nucleotide receptors that can respond to both AMP and
adenosine. Based on its ability to functionally bind the nucleotide AMP, we have named it P2Y15.
Upon stimulation, P2Y15 induces both Ca2+ mobilization and cyclic AMP generation, suggesting
coupling to at least two different G proteins. It is highly expressed in mast cells and is found
predominantly in the tissues of the respiratory tract and kidneys, which are known to be affected by
AMP, adenosine, and adenosine antagonists. Until now, AMP’s effects have been thought to depend
on its dephosphorylation to adenosine, but we demonstrate here that P2Y15 is a bona fide AMP
receptor by showing that it binds [32P]-AMP. Since AMP and adenosine have bronchoconstrictive
effects that can be inhibited by theophylline, we tested whether theophylline and other adenosine
receptor antagonists can block P2Y15. We found inhibition at a theophylline concentration well
within the therapeutic dose range, indicating that P2Y15 may be a clinically important target of this
drug.
Abbreviations used in this paper: AMP, adenosine 5’-monophosphate; ATP, adenosine 5’-
triphosphate; ADP, adenosine 5’-diphosphate; GPCR, G protein-coupled receptor; RT-PCR, reverse
transcription-polymerase chain reaction; ADA, adenosine deaminase; IBMX, 3-isobutyl-1-
methylxanthine; 8-SPT, 8-(p-sulfophenyl)theophylline; 8-PT, 8-phenyltheophylline; DPCPX, 8-
cyclopentyl-1,3-dipropylxanthine; CSC, 8-(3-chlorostyryl)caffeine; NECA, 5’-(N-
ethylcaboxamido)adenosine; Chloro-IB-MECA, 2-chloro-N6-(3-iodobenzyl)-adenosine-5’-N-
methyluronamide.
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 3
Introduction
Adenosine can mediate diverse physiological effects including bronchoconstriction, inhibition of
platelet aggregation, inhibition of lipolysis, induction of sedation, vasodilation, suppression of
cardiac rate and contractility, and stimulation of gluconeogenesis. To date, five cell-surface
receptors have been identified for adenosine, namely, adenosine receptors A1, A2A, A2B, and A3
(collectively referred to as P1 receptors), and the growth hormone secretagogue receptor GHSR
(1,2). AMP has a similarly diverse repertoire of effects (3-6), including bronchoconstriction,
stimulation of DNA synthesis, mitogenesis, and stimulation of chloride secretion, but no receptor
for AMP has previously been reported. Numerous cell types can release adenosine, including mast
cells (7), kidney brush border cells (8) and cardiac cells (9), while AMP can be released by such cell
types as activated platelets (10), neutrophils (4), and eosinophils (11). AMP can also be generated
extracellularly from the hydrolysis of ATP and ADP by ecto-ATPases (12) and ecto-ATP
diphosphohydrolases (13), and can be further dephosphorylated by ecto 5’-nucleotidases to produce
adenosine (14). A number of widely-used drugs have been developed, such as theophylline (15) and
cromolyn (16), that can modulate what are thought to be the effects of adenosine in diseases such as
asthma, but their mechanisms of action are not yet fully understood.
Other extracellular nucleotides similarly induce a wide variety of responses in many cell types,
including muscle contraction and relaxation, vasodilation, neurotransmission, platelet aggregation,
ion transport regulation, and cell growth. The effects are exerted mainly through two types of cell
surface molecules: P2Y type G protein-coupled receptors (GPCRs), and P2X type ligand-gated ion
channels. Nine nucleotide-stimulated P2Y type GPCRs have been characterized to date in humans:
P2Y1, P2Y11, P2Y12, and P2Y13 which are activated by the adenine nucleotides ATP or ADP;
P2Y4, P2Y6, P2Y14,and CYSLT1 which are activated by the uridine nucleotides UTP or UDP (or
in the case of P2Y14, UDP-glucose); and P2Y2 which is activated by both adenine and uridine
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 4
nucleotides (17-20). None of these receptors has been shown to be able to bind adenosine or AMP.
Here we report on the characterization of a GPCR with close homology to the P2Y receptors that
can bind and respond to both AMP and adenosine.
Experimental Procedures
P2Y15 cDNA Cloning. Protein sequences of known P2Y receptors were used to search for
homologs in the Genbank database of the National Center for Biotechnology Information
(http://www.ncbi.nlm.nih.gov) using the program tblastn. The search identified an intronless
genomic sequence subsequently also found by others and designated in Genbank as the orphan
receptor GPR80 (21). To clone the gene, human genomic DNA was used as template, and PCR was
performed using primers 5’-GCCAAACTGAACTCTCTTGTTTTCTTGC-3’ and 5’-
GCCCTGGCTTTGGCACATGATTAC-3’ and a blend of HotStarTaq (Qiagen, Hilden, Germany)
and Pfu Turbo (Stratagene, La Jolla, CA) polymerases. PCR products were cloned into pCRII-
TOPO (Invitrogen, Carlsbad, CA), cycle-sequenced with an ABI Prism Dye Terminator Cycle
Sequencing Reaction Kit (Applied Biosystems, Foster City, CA), and analyzed on an ABI Prism
377 sequencing system (Applied Biosystems). For functional studies, the cDNA was subcloned into
a pDisplay vector (Invitrogen) to append an N-terminal HA epitope and Igκ signal sequence.
Expression profiling. 25 µg of total RNA from the following were used as template in reactions to
synthesize first-strand cDNA for expression profiling: Human Total RNA Panel I-V (Clontech
Laboratories, Palo Alto, CA), normal human lung primary cell lines (BioWhittaker Clonetics,
Walkersville, MD), several common cell lines (ATCC, Washington, DC), and various cells purified
from peripheral blood. First-strand cDNA was synthesized using oligo (dT) (Nippon Gene Research
Laboratories, Sendai, Japan) and the Superscript™ First-Strand Synthesis System for RT-PCR (Life
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 5
Technologies, Rockville, MD) according to the manufacturer’s protocol. For these samples, 1/1250th
of the synthesized first-strand cDNA was subsequently used as template for quantitative PCR.
Additional samples were purchased as presynthesized cDNAs (Human Immune System MTC Panel
and Human Blood Fractions MTC Panel, Clontech Laboratories), and for these, 10 ng of cDNA was
used as template for quantitative PCR.
Quantitative PCR was performed in a LightCycler (Roche Molecular Biochemicals, Indianapolis,
IN) with oligonucleotide primers 5’-TTCGGATCGAATCTCGCCTGCT-3’ and 5’-
TGCTTGCTCAAGGTTCCCGCTTA-3’ in the presence of the DNA-binding fluorescent dye
SYBR Green I. Results were then converted into copy numbers of the gene transcript per ng of
template cDNA by fitting to a standard curve. The standard curve was derived by simultaneously
performing the quantitative PCR reaction on PCR products of known concentrations amplified
beforehand from the target gene.
To correct for differences in mRNA transcription levels per cell in the various tissue types, a
normalization procedure was performed using similarly calculated expression levels of five
different housekeeping genes: GAPDH, hypoxanthine guanine phophoribosyl transferase, beta-actin,
porphobilinogen deaminase, and β2M. Expression levels of the five housekeeping genes in all tissue
samples were measured in three independent reactions per gene using the LightCycler and a
constant amount (25 µg) of starting RNA.
Expression levels were also measured using CodeLinkTM microarrays (Amersham Biosciences,
Buckinghamshire, England). Total RNA was prepared from human umbilical cord blood-derived
mast cells (kindly provided by Professor H. Nagai, Department of Pharmacology, Gifu
Pharmaceutical University, Gifu, Japan; prepared as described in (22)) and purified leukocyte
fractions, and then used to synthesize biotin-labeled cRNA using the Amersham cRNA synthesis kit.
cRNA yield was quantified by measuring absorbance at 260 nm, and then the cRNA was
fragmented in 40 mM Tris–acetate (TrisOAc) pH 7.9, 100 mM KOAc and 31.5 mM MgOAc, at
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 6
94°C for 20 min. Ten micrograms of fragmented cRNA from each sample was used for
hybridization to a CodeLinkTM UniSet 20K Human Expression Bioarray chip (Amersham
Biosciences). cRNAs bound to the microarrays were stained with streptavidin-Cy5 and the
processed slides were scanned with an Axon GenePix 4000B Scanner. Images for each slide were
analyzed using the CodeLinkTM Expression Analysis Software (Amersham Biosciences).
Determination of ligand specificity. HA-tagged P2Y15 was transfected into HEK293 cells (ATCC)
using Lipofectamine (Invitrogen). Expression on the cell surface was verified by staining cells with
phycoerythrin-labeled anti-HA antibody (Santa Cruz Biotechnology, Santa Cruz, CA) and
measuring fluorescence on a FACSort (Becton-Dickinson, Franklin Lakes, NJ). Stably transfected
clones were generated by selection in G418 (500 µg/ml) and reconfirmed for cell-surface
expression of the P2Y15 protein. Ligand screening was performed in a Ca2+ mobilization assay as
follows: Stably transfected P2Y15 GPCR-expressing cells were seeded into 96-well plates and
incubated overnight at 37 °C. The culture medium was aspirated and replaced with 100 µl of
loading buffer consisting of 0.1% BSA, 20 mM HEPES, 1 mM probenecid, 0.01% pluronic F127,
and 1 µM Fluo-3-AM (Molecular Probes, Eugene, OR) in HBSS, and incubated for 1 hour at room
temperature. The cells were then washed gently 3 times with wash buffer consisting of 0.1% BSA,
20 mM HEPES, and 1 mM probenecid in HBSS. The washed cells were placed in an FDSS6000
functional drug screening system (Hamamatsu Photonics, Hamamatsu, Japan) and changes in
cellular fluorescence were measured after adding serial dilutions of potential ligands. A panel of
about 130 potential ligands for testing was assembled by selecting known ligands of the GPCRs
most closely related to P2Y15 GPCR as well as several naturally occurring chemical relatives of the
ligands. The panel included various bioactive lipids, eicosanoids, peptides, cannabinoids,
chemokines, nucleosides, nucleotides and chemically related substances which were generally
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 7
purchased from either Sigma (St. Louis, MO) or R&D Systems (Minneapolis, MN).
Ligand confirmation analyses were performed by repeating the Ca2+ mobilization experiments
using the identified ligands to stimulate P2Y15 transfectants, identically constructed P2Y8
transfectants, and non-transfected HEK293 cells, all of which had been cultured for two hours with
or without 1 µM pertussis toxin. Antagonist assays for inhibition of calcium responses to AMP and
adenosine were performed essentially as above except that serial dilutions of antagonist compounds
were added five minutes prior to the addition of ligand. Agonist assays for stimulation of calcium
responses were performed in the same manner as the ligand screen.
AMP conversion assays. Transfected and nontransfected cells in DMEM medium were seeded at
105 cells per well into 96-well plates and incubated for 3 hours. The medium was then exchanged
with fresh medium and 1.85 µM [3H]-AMP (Amersham) was added to the wells. After incubation
for 5 or 60 min, 6 µl of the medium, together with AMP and adenosine standards, were applied to
thin layer chromatography sheets, separated with isobutyl alcohol/isoamyl alcohol/2-
ethoxyethanol/ammonia/H2O (9:6:18:9:15) as solvent, and visualized under UV light as described
by Yegutkin et al. (23). The spots corresponding to AMP and adenosine were cut from the sheets
and the amounts of each were quantified by scintillation counting.
Receptor binding assays. 105 cells per well in 96-well plates were washed twice for 1 h with
DMEM medium. Wheatgerm agglutinin SPA beads (Amersham) were then added at 1 mg/well,
followed 1 h later by the addition of increasing concentrations of [3H]-Adenosine or [3H]-AMP
(Amersham), in a constant volume of HBS (10 mM Hepes, 130 mM NaCl, 5 mM KCl, 1 mM
MgCl2, 1 mM CaCl2 and 1 g/l glucose (6 µl/well)). After incubating at 4 °C for 16 h, the plates were
centrifuged for 10 min at 1500 rpm and then scintillation measured on a TopCount automated
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 8
scintillation counter. Non-specific binding measurements and competitive binding experiments were
carried out under the same conditions but with either an excess of cold ligand (2.5 mM) or
increasing concentrations of cold ligand, respectively. Adenosine binding in competitive binding
experiments was measured after incubation for 1 hr at 4 °C. Binding of [32P]-AMP was carried out
under the same conditions except that instead of using SPA beads, at the end of the incubation, cells
were washed three times by vacuum filtration and 100 µl scintillation fluid was added to the wells.
Kd values were determined by non-linear regression using the program Prism (Graph Pad Software,
San Diego, CA).
Cyclic AMP production assay. Cyclic AMP production after stimulation of cells was measured with
the Tropix cAMP-screen (Applied Biosystems) according to the manufacturer’s protocol. Briefly,
stable transfectants and control cells (1 x 105 cells/well) were cultured for two hours with or without
1 mM pertussis toxin, then treated for 30 min with 10 mM forskolin and serial dilutions of AMP or
adenosine. The cells were then lysed and the cAMP produced was measured by a cAMP-specific
ELISA. Concentrations of cAMP produced were calculated by comparing against cAMP standards
measured simultaneously. The effect of adenosine deaminase (ADA) on ligand-stimulated cyclic
AMP production was measured essentially as above except that pertussis toxin was excluded and
cells were treated for 30 min with 10 mM forskolin prior to the addition of 25 mM adenosine or
100 mM AMP. Serial dilutions of ADA (Roche Diagnostics, Tokyo, Japan) were added to the
cultures 10 minutes prior to the addition of ligand.
Results
Cloning and Sequencing. In an effort to find new receptors for extracellular nucleosides and
nucleotides, we searched for homologs of known P2Y nucleotide receptors in the Genbank database
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 9
of the National Center for Biotechnology Information using the program tblastn. P2Y receptors are
receptors for diphosphate and triphosphate adenine and uridine nucleotides, though none respond to
AMP or adenosine. Our search identified an intronless genomic sequence subsequently also found
by others and designated variously as the orphan receptor GPR80 (21) or GPR99 (24). We are now
renaming it P2Y15 as a new member of the P2Y family. A putative mouse ortholog of the gene has
recently appeared in Genbank under accession number XP_139267. We subsequently found the rat
ortholog by using the mouse protein sequence in a tblastn query against rat genome sequences. An
alignment of the human, mouse, and rat protein sequences is shown in figure 1.
The conceptually translated protein product of the human P2Y15 gene shows 36% identity over
its full length with the nucleotide receptor P2Y1, increasing to an overall sequence similarity of
58% when amino acids with related physicochemical properties are included. Homology of the
protein sequence with other P2Y nucleotide receptors P2Y2, P2Y4, P2Y6, and P2Y11 likewise
shows an overall identity ranging from 25 to 35% and a similarity ranging from 43 to 57%. A
similar, though slightly lower, level of homology is seen among the mouse and rat orthologs
(excluding P2Y11, which has not yet been found in rodents). The human gene transcript encodes a
polypeptide of 337 amino acids with a calculated molecular mass of 38.3 kD. A phylogenetic
analysis comparing the protein sequence with other GPCRs places the molecule among a cluster of
other P2Y receptors, distant from the known receptors for adenosine (figure 2). The gene sequence
is found on the genomic contig NT_009952 which has been localized to human chromosome
13q32.
Tissue Distribution of P2Y15. As a first step to investigating the the function of P2Y15, we
examined the distribution of P2Y15 messenger RNA expression in several different human tissues,
cell types, and commonly used cell lines. We designed oligonucleotide primers near the 3’ end of
the coding region that could specifically amplify P2Y15 cDNA and used these in a quantitative
reverse transcription-polymerase chain reaction (RT-PCR) analysis to measure relative transcript
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 10
levels. Among the tissues tested, trachea, salivary glands, kidney, fetal brain and lung showed the
highest expression levels (figure 3a). The high expression in the respiratory tract prompted us to
look in depth at the cell types within the trachea and lung that might be responsible for the
predominant expression there. Little or no expression, however, could be detected in any of the
tested primary cell populations or transformed cell lines derived from lung tissues or from immune
cell subtypes (figure 3b).
We therefore investigated the possibility that the gene is expressed in a minor population of cells in
the respiratory tract such as mast cells or eosinophils. Because mast cells and eosinophils typically yield
only small amounts of mRNA, we analyzed the expression of P2Y15 in these cells by microarray
analysis which allows the analysis of more genes than quantitative RT-PCR and provides multiple
controls to verify the quality of the sample preparations. Analysis of gene expression in CD4 and CD8 T
cells, tonsil B cells, neutrophils, eosinophils, and mast cells showed that the P2Y15 gene is expressed
specifically and at very high levels in mast cells (figure 3b). Indeed, the high level of P2Y15 expression
in mast cells places it among the top 1% of the approximately 20,000 different gene probes included in
the microarray. Similar specificity and high level expression in these cell types was not seen with any of
the other P1 adenosine receptors (figure 3c) or P2 nucleotide receptors (data not shown).
Identification of AMP and Adenosine as Functional Ligands for P2Y15. To identify the ligand of
P2Y15, we generated stable transfectants with HEK293 cells and then tested for calcium mobilization in
response to a panel of ligands. Among the potential ligands tested, only AMP and adenosine were able to
induce a response in the transfectants while not inducing a similar response in either nontransfected
HEK293 cells or HEK293 cells stably transfected with the control orphan GPCR P2Y8 in an identical
vector construct. We detected a calcium response with an EC50 of 920 nM for AMP and 670 nM for
adenosine (figure 4a). The calcium response to either ligand was not significantly affected by pertussis
toxin (data not shown). Both stable transfectants and nontransfected cells mobilized calcium in response
to ATP, ADP, and UTP, consistent with previous reports of HEK293 endogenously expressing P2Y1 and
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 11
P2Y2 receptors (25). Further analysis by RT-PCR showed that the nucleotide receptors P2Y4, P2Y12,
and P2Y13 and the adenosine receptors A2A and A2B (previously reported in (26)) are also expressed in
HEK293 cells (data not shown). Despite the endogenous expression of the adenosine receptors, however,
calcium mobilization responses to adenosine in nontransfected cells could only be detected at very high
adenosine concentrations, and showed only a very weak response (figure 4a).
To determine the effect of P2Y15 stimulation on adenylate cyclase activity, we measured cyclic
AMP accumulation in our P2Y15-HEK293 stable transfectants in response to AMP and adenosine.
Stimulation with either ligand alone gave only minimal responses barely above the detection limit.
In the presence of 10 µM forskolin, however, both AMP and adenosine induced the generation of
cyclic AMP in a dose dependent manner, with an EC50 of 214 nM for AMP and 327 nM for
adenosine (figure 4b). Nontransfected HEK293 cells similarly generated cyclic AMP in response to
adenosine, likely due to the stimulation of endogenously expressed adenosine receptors, but did not
respond strongly to AMP. The production of cyclic AMP in response to either ligand was not
affected by pretreatment of the cells for two hours with 1 µM pertussis toxin (data not shown),
indicating that P2Y15 does not couple with an adenylate cyclase-inhibiting G protein. The
responsiveness of the P2Y15 transfectants to AMP in the cyclic AMP assay did not appear to be due
to an increased conversion rate of AMP to adenosine, since transfected and non-transfected cells
showed similarly low rates of endogenous nucleotidase activity (figure 4c).
Saturation binding analysis of the ligands to the P2Y15 receptor in stable transfectants gave Kd
values of 12.0 µM for [2-3H]-adenosine (figure 5a) and 18.6 µM for [2-3H]-AMP (data not shown).
Since AMP can be dephosphorylated to adenosine by ectonucleotidases, we repeated the binding
analysis with adenosine 5’-[32P] monophosphate to confirm that the binding being measured was
AMP and not adenosine. This resulted in a similar binding curve with a Kd of 18.8 µM (figure 5b),
indicating that AMP itself, and not a breakdown product, was binding to the receptor. We then
performed competitive binding assays to determine whether one ligand could antagonize the
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 12
binding of the other ligand to the P2Y15 transfectants. While unlabeled AMP was able to block a
large proportion of the binding of 10 µM [2-3H]-adenosine to transfectants (Ki = 32.0 µM), it did
not block the binding as completely as unlabled adenosine and had little effect in blocking 3H-
adenosine binding to nontransfected HEK293 cells (figure 5c). On the other hand, unlabled
adenosine was able to block the binding of 10 µM [2-3H]-AMP to transfectants (Ki = 39.8 µM) with
a potency similar to that of unlabled AMP (figure 5d).While these results provide evidence that both
AMP and adenosine bind to P2Y15, the results also demonstrate a lack of specific AMP binding
sites on the nontransfected cells since neither AMP nor adenosine could antagonize the binding of
[2-3H]-AMP to nontransfected HEK293 cells beyond the background level.
To further clarify whether AMP itself, without its conversion to adenosine, is able to induce a
response in P2Y15 transfectants, we measured AMP- and adenosine-induced cyclic AMP
production in the presence of adenosine deaminase (ADA), an enzyme that breaks down adenosine
to inosine. As expected, ADA inhibited the adenosine-induced response in a dose-dependent manner,
but showed little if any effect against the AMP-induced response, indicating that adenosine is not a
mediator of the AMP response (figure 6).
Characterization of Antagonists and Agonists of the P2Y15 Receptor. To test whether any of the
known antagonists or agonists of adenosine receptors could antagonize P2Y15, we performed
calcium mobilization assays in the presence of varying concentrations of such compounds.
Antagonists tested were the non-selective adenosine receptor antagonists theophylline, caffeine, 3-
isobutyl-1 methylxanthine (IBMX), and 8-(p-sulfophenyl)theophylline (8-SPT); selective A1
receptor antagonists 8-phenyltheophylline (8-PT) and 8-cyclopentyl-1,3-dipropylxanthine
(DPCPX); the selective A2A receptor antagonist 8-(3-chlorostyryl)caffeine (CSC); and the selective
A2B receptor antagonists enprofylline and alloxazine. With the exception of alloxazine and DPCPX
which were agonistic in both transfected and non-transfected cells, all of these compounds were
able to block the calcium mobilization induced by AMP and adenosine, with Ki values for blocking
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 13
AMP ranging from 200 nM for 8-SPT to 3,309 nM for caffeine (figure 7a, table 1) and Ki values for
blocking adenosine ranging from 962 nM for 8-theophylline to 64,060 nM for enprofylline (figure
7b, table 1). Since many of these compounds can also act as phosphodiesterase inhibitors and cause
an increase in cyclic AMP levels that can potentially inhibit Ca2+ mobilization, the compounds were
also tested for their ability to inhibit ADP-induced Ca2+ mobilization in the same cells. None of the
compounds, however, had significant inhibitory effects at concentrations less than 1 mM (data not
shown). Adenosine receptor agonists tested were the non-selective agonist 5’-(N-
ethylcaboxamido)adenosine (NECA); the selective A1 receptor agonist N6-cyclopentyladenosine;
the selective A2A receptor agonist CGS-21680 hydrochloride; and the selective A3 receptor agonist
2-chloro-N6-(3-iodobenzyl)-adenosine-5’-N-methyluronamide (Chloro-IB-MECA). Among these
componds, only NECA and N6-cyclopentyladenosine showed measurable agonist effects (table 2).
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 14
Discussion
We describe here the identification of a cell-surface receptor that can respond to both AMP and
adenosine. Although the responses to both ligands are pharmacologically similar, we consider the
receptor a new member of the P2Y family since the primary structure of the receptor protein more
closely resembles the P2Y nucleotide receptors than the P1 adenosine receptors. Moreover, the
sequence contains the conserved basic residue arginine at position 268 that has been reported to be
essential for binding the phosphate moiety of nucleotide ligands by P2Y receptors (27). The ability
of the receptor to bind the nucleoside adenosine, however, is unique among the P2Ys and may
require a broadening of the criteria for the P2Y classification.
Functionally, the receptor is able to respond to stimulation by inducing the mobilization of
calcium ions and causing the generation of cyclic AMP. This dual functionality is similar to that of
the adenosine receptor A2B (28,29) and the nucleotide receptor P2Y11 (19) both of which show
evidence of being dually coupled to activators of phospholipase C (calcium flux) and adenylate
cyclase (cyclic AMP production). For both A2B and P2Y11, the activation of phospholipase C and
adenylate cyclase is thought to be achieved through Gq class and Gs G proteins, respectively.
Because P2Y15-induced calcium flux and cyclic AMP generation are both insensitive to pertussis
toxin, which inactivates Go and Gi classes of G proteins, P2Y15 is likely to be similarly coupled to
the pertussis toxin-insensitive Gq class and Gs G proteins. The nature of the P2Y15’s coupling to
adenylate cyclase-stimulatory Gs proteins in our system is not completely straightforward, however,
since detectable levels of cyclic AMP production could only be seen in the presence of the adenylate
cyclase activator forskolin. This apparent requirement for forskolin, which synergistically
potentiates the cyclic AMP production, may be specific to the HEK293 host cells used, but in any
case is reminiscent of the forskolin potentiation that has similarly been reported for β-adrenergic
receptors (30).
The similarity to other receptors, particularly A2B which can respond to adenosine as a ligand
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 15
and which is expressed on the HEK293 host cell used in the experiments presented here, was a
potentially confounding factor in our analysis and therefore had to be taken into account when
interpreting our results. In calcium mobilization experiments, despite previous reports of HEK293
calcium responses generated upon stimulation with adenosine (28,29), the response of endogenous
receptor to AMP and adenosine in our HEK293 cells was so slight as to be negligible. The
transfection procedure itself also appeared to have no influence on the calcium response since
transfectants generated with P2Y8 instead of P2Y15 also showed negligible responses to AMP and
adenosine. On the other hand, in cyclic AMP production assays, nontransfected cells gave a
response to adenosine that was indistinguishable from that seen in the transfectants. Although it was
clear that AMP could cause a distinct increase in cyclic AMP production in the P2Y15 transfectants,
we cannot say with certainty that adenosine stimulation of P2Y15 had the same effect. Confirmation
of the ability of adenosine to stimulate P2Y15 to activate adenylate cyclase must therefore await the
successful functional expression of P2Y15 in a cell type that does not express A2 adenosine
receptors.
In receptor binding assays, while AMP clearly showed saturatable binding kinetics only in P2Y15
transfectants, adenosine was able to bind specifically to receptors on both nontransfectants and
P2Y15 transfectants, albeit at a higher level in the transfectants. Competitive binding experiments
using cold AMP to compete against labeled adenosine binding, however, showed that adenosine
binds to sites on P2Y15 transfectants that can be competed with AMP, presumably P2Y15 receptors,
but on the nontransfectants binds only to sites that cannot be competed by AMP, such as other
adenosine receptors, giving compelling evidence that adenosine binds to P2Y15. Although the
affinity of the receptor for AMP and adenosine is relatively low, the ligand binding results we
obtained are in line with the range of Kd values that have been reported so far for some of the other
adenosine receptors and P2Y receptors. For example, agonist binding to the A2B adenosine receptor
is typically in the double digit micromolar range (1) and [35S]ATP[γS] binding to P2Y receptors on
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 16
tracheal gland cells has been reported as 2.5 µM and 20 µM (31). Furthermore, the lack of
antagonism of the P2Y15 receptor by DPCPX and alloxazine, specific antagonists of A1 and A2B
receptors respectively, and the lack of significant agonism of the P2Y15 receptor by N6-
cyclopentyladenosine, CGS-21680 hydrochloride, and Chloro-IB-MECA, specific agonists of A1,
A2A, and A3 respectively, strongly suggest that the signaling responses measured were generated
through activation of P2Y15 and not another adenosine receptor.
Nevertheless, a concern when conducting experiments with AMP is the potential for enzymatic
breakdown of the molecule to adenosine. In most reports that have demonstrated the activity of
AMP, for example as an inducer of bronchoconstriction in the lungs of patients with asthma (3,32)
or as a paracrine activator of intestinal chloride ion secretion produced by neutrophils and
eosinophils (11,33), due to the absence of a receptor for AMP to which this activity could be
attributed, it has been assumed that AMP is converted to adenosine before the resultant effects are
produced. Our analysis of AMP to adenosine conversion rates did not show any evidence of
enhanced AMP breakdown in P2Y15-expressing cells. To determine, therefore, whether AMP itself
can bind P2Y15 or must first be converted to adenosine, we performed receptor binding assays
using 32P-labeled AMP and found specific, saturatable binding to the transfectants. Since the
radiolabel on the molecule would be lost upon dephosphorylation and conversion to adenosine, our
binding assay shows that AMP can bind to P2Y15 without conversion.
The expression of P2Y15 in mast cells, the respiratory tract, and kidney is consistent with effects
that have been reported for AMP, adenosine, and adenosine antagonists in these tissues. In the
respiratory tract, an immediate bronchoconstriction is typically experienced by patients with asthma
upon the inhalation of AMP or adenosine (3), and adenosine has been found to be increased in the
bronchoalveolar lavage fluid from airways of patients with chronic inflammatory conditions of the
lung, such as asthma and chronic obstructive pulmonary disease (34). Recently, adenosine, or more
commonly AMP which is more easily solubilized, has been utilized in bronchoprovocation tests for
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 17
the diagnosis and monitoring of asthma (3,32) due to the tests’ better disease specificity and
correlation with inflammation state than other bronchoprovocators. Conversely, adenosine receptor
antagonists, such as theophylline, have been used for over 50 years as effective bronchodilators (15).
Evidence on the mechanism of adenosine and AMP mediated bronchoconstriction has indicated an
extracellular site of action and the stimulation or potentiation of mast cell mediator release (35).
Recent studies to determine which adenosine receptor is responsible for the bronchoconstriction
response, however, have failed to conclusively implicate a particular receptor, and on the contrary,
experiments in rats have ruled out the role of any of the known P1 adenosine receptors, leading to
the conclusion that an unknown receptor must be involved (36). While it is still possible that the
effects of AMP in the human respiratory tract are dependent upon its breakdown to adenosine and
the subsequent stimulation of P1 adenosine receptors, with the evidence provided here to show the
existence of an AMP receptor, the alternative possibility of a direct effect by AMP must now also be
considered. Regarding the expression in the kidneys, adenosine antagonists such as theophylline
and caffeine are well known to possess diuretic effects. These effects are thought to be mediated
primarily through the A1 and A2A adenosine receptors, but the pharmacology of these two
receptors cannot satisfactorily account for the effects seen (37). The existence of a third receptor in
the kidneys that can be affected by adenosine antagonists may help to explain the effects.
Adenosine receptor antagonists, such as theophylline, enprofylline, and caffeine, are among the
world’s most widely used drugs. Their molecular mechanism, however, remains undefined, as do
their sites of action, which include adenosine receptors, phosphodiesterases, histone deacetylases,
and other sites that have yet to be found (38,39). To determine whether P2Y15 is a target of such
antagonists, several adenosine receptor antagonists were tested and all but two showed antagonist
activity against P2Y15. The concentration of theophylline at which calcium responses could be
inhibited (Ki of 0.7 µM against AMP; 5.6 µM against adenosine) is well below that considered to be
the therapeutically optimal plasma concentration (usually 55-110 µM) (15) for this drug. Similarly,
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 18
plasma concentrations of caffeine as low as 5 µM have been shown to relieve histamine-induced
bronchoconstriction (40), and at this concentration caffeine can effectively inhibit AMP-induced
calcium responses (Ki of 3.3 µM against AMP). Unexpectedly, however, the antagonists often
showed a large difference in the concentrations required to block AMP-mediated signaling
compared with those required to block adenosine-mediated signaling, even though AMP and
adenosine show similar EC50 values for signaling and similar Kd values for binding. The underlying
reason for this is unclear, but the discrepancy suggests that the mechanism of action of the
antagonists is more complex than can be explained by simple affinity based competition.
The identification of a receptor that binds and responds to both AMP and adenosine and is
susceptible to blocking by adenosine receptor antagonists can help us to better understand the
complex physiological effects of AMP and adenosine. Since the safety and effectiveness of
adenosine antagonists in treating respiratory diseases and other ailments has been limited by
toxicity in the central nervous system and heart (41) where most adenosine receptors are found, it
will be of considerable interest to determine whether P2Y15 can be specifically targeted in order to
develop better treatments.
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 19
Acknowledgements
The authors of this article would like to thank Eiko Okazaki for her technical support.
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 20
References
1. Fredholm, B. B., AP, I. J., Jacobson, K. A., Klotz, K. N., and Linden, J. (2001) Pharmacol Rev 53,
527-552
2. Tullin, S., Hansen, B. S., Ankersen, M., Moller, J., Von Cappelen, K. A., and Thim, L. (2000)
Endocrinology 141, 3397-3402
3. Polosa, R., and Holgate, S. T. (1997) Thorax 52, 919-923
4. Madara, J. L., Patapoff, T. W., Gillece-Castro, B., Colgan, S. P., Parkos, C. A., Delp, C., and Mrsny,
R. J. (1993) J Clin Invest 91, 2320-2325
5. Wu, S. N., Lin, Y. T., and Chen, S. S. (1992) Jpn J Physiol 42, 35-47
6. Lewis, M. D., Webster, J., Ham, J., Davies, J. S., and Scanlon, M. F. (1996) J Clin Endocrinol Metab
81, 1296-1298
7. Marquardt, D. L., Gruber, H. E., and Wasserman, S. I. (1984) Proc Natl Acad Sci U S A 81, 6192-
6196
8. Williams, T. C., and Jarvis, S. M. (1991) Biochem J 274 ( Pt 1), 27-33
9. Decking, U. K., Schlieper, G., Kroll, K., and Schrader, J. (1997) Circ Res 81, 154-164
10. Jarvis, G. E., Evans, R. J., and Heath, M. F. (1996) Eur J Pharmacol 315, 203-212
11. Resnick, M. B., Colgan, S. P., Patapoff, T. W., Mrsny, R. J., Awtrey, C. S., Delp-Archer, C., Weller, P.
F., and Madara, J. L. (1993) J Immunol 151, 5716-5723
12. Yegutkin, G. G., and Burnstock, G. (2000) Biochim Biophys Acta 1466, 234-244
13. Kaczmarek, E., Koziak, K., Sevigny, J., Siegel, J. B., Anrather, J., Beaudoin, A. R., Bach, F. H., and
Robson, S. C. (1996) J Biol Chem 271, 33116-33122
14. Misumi, Y., Ogata, S., Hirose, S., and Ikehara, Y. (1990) J Biol Chem 265, 2178-2183
15. Weinberger, M., and Hendeles, L. (1996) N Engl J Med 334, 1380-1388
16. Joad, J. P. (1990) Am J Respir Cell Mol Biol 2, 193-198
17. Di Virgilio, F., Chiozzi, P., Ferrari, D., Falzoni, S., Sanz, J. M., Morelli, A., Torboli, M., Bolognesi,
G., and Baricordi, O. R. (2001) Blood 97, 587-600
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 21
18. Hollopeter, G., Jantzen, H. M., Vincent, D., Li, G., England, L., Ramakrishnan, V., Yang, R. B.,
Nurden, P., Nurden, A., Julius, D., and Conley, P. B. (2001) Nature 409, 202-207
19. Communi, D., Gonzalez, N. S., Detheux, M., Brezillon, S., Lannoy, V., Parmentier, M., and
Boeynaems, J. M. (2001) J Biol Chem 276, 41479-41485
20. Abbracchio, M. P., Boeynaems, J. M., Barnard, E. A., Boyer, J. L., Kennedy, C., Miras-Portugal, M.
T., King, B. F., Gachet, C., Jacobson, K. A., Weisman, G. A., and Burnstock, G. (2003) Trends
Pharmacol Sci 24, 52-55
21. Lee, D. K., Nguyen, T., Lynch, K. R., Cheng, R., Vanti, W. B., Arkhitko, O., Lewis, T., Evans, J. F.,
George, S. R., and O'Dowd, B. F. (2001) Gene 275, 83-91
22. Saito, H., Ebisawa, M., Tachimoto, H., Shichijo, M., Fukagawa, K., Matsumoto, K., Iikura, Y., Awaji,
T., Tsujimoto, G., Yanagida, M., Uzumaki, H., Takahashi, G., Tsuji, K., and Nakahata, T. (1996) J
Immunol 157, 343-350
23. Yegutkin, G. G., and Burnstock, G. (1999) J Recept Signal Transduct Res 19, 437-448
24. Wittenberger, T., Hellebrand, S., Munck, A., Kreienkamp, H. J., Chica Schaller, H., and Hampe, W.
(2002) BMC Genomics 3, 17
25. Schachter, J. B., Sromek, S. M., Nicholas, R. A., and Harden, T. K. (1997) Neuropharmacology 36,
1181-1187
26. Sunahara, R. K., Dessauer, C. W., and Gilman, A. G. (1996) Annu Rev Pharmacol Toxicol 36, 461-
480
27. Moro, S., Hoffmann, C., and Jacobson, K. A. (1999) Biochemistry 38, 3498-3507
28. Gao, Z., Chen, T., Weber, M. J., and Linden, J. (1999) J Biol Chem 274, 5972-5980
29. Linden, J., Thai, T., Figler, H., Jin, X., and Robeva, A. S. (1999) Mol Pharmacol 56, 705-713
30. Morin, D., Sapena, R., Tillement, J. P., and Urien, S. (2000) Pharmacol Res 41, 435-443
31. Merten, M. D., Saleh, A., Kammouni, W., Marchand, S., and Figarella, C. (1998) Eur J Biochem 251,
19-24
32. van den Berge, M., Kerstjens, H. A., and Postma, D. S. (2002) Clin Exp Allergy 32, 824-830
33. Lennon, P. F., Taylor, C. T., Stahl, G. L., and Colgan, S. P. (1998) J Exp Med 188, 1433-1443
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 22
34. Driver, A. G., Kukoly, C. A., Ali, S., and Mustafa, S. J. (1993) Am Rev Respir Dis 148, 91-97
35. Marquardt, D. L. (1997) in Asthma (Barnes, P. J., Grunstein, M. M., Leff, A. R., and Woolcock, A. J.,
eds), pp. 585-591, Lippincott-Raven Publishers, Philadelphia
36. Hannon, J. P., Tigani, B., Wolber, C., Williams, I., Mazzoni, L., Howes, C., and Fozard, J. R. (2002)
Br J Pharmacol 135, 685-696
37. Weaver, D. R., and Reppert, S. M. (1992) Am J Physiol 263, F991-995
38. Howell, R. E. (1990) J Pharmacol Exp Ther 255, 1008-1014
39. Ito, K., Lim, S., Caramori, G., Cosio, B., Chung, K. F., Adcock, I. M., and Barnes, P. J. (2002) Proc
Natl Acad Sci U S A 99, 8921-8926
40. Henderson, J. C., O'Connell, F., and Fuller, R. W. (1993) Thorax 48, 824-826
41. Melamed, J., and Beaucher, W. N. (1995) Ann Allergy Asthma Immunol 75, 516-520
42. Hofmann, K., and Stoffel, W. (1993) TMpred - Prediction of Transmembrane Regions and
Orientation.
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 23
Figure legends
Figure 1. Alignment of human, mouse, and rat P2Y15 amino acid sequences. Identical residues
are shown with white backgrounds, similar residues with gray backgrounds, and dissimilar residues
with black backgrounds. Identity between either of the rodent sequences and the human sequence is
86%, and between the mouse and rat is 96%. Seven transmembrane regions as predicted by the
computer program TMpred (42) are indicated with heavy overlines and numbered TM1-7. The
Genbank accession number of the rat P2Y15 sequence is AY191367.
Figure 2. Unrooted phylogenetic analysis comparing the P2Y15 protein sequence with
adenosine receptors, other P2Y receptors, and closely related GPCRs. Phylogenetic analysis
was performed with the Neighbor Joining algorithm and the dendrogram drawn with the computer
program Vector NTI. Known ligands are indicated.
Figure 3. Tissue and cellular distribution of P2Y15. Expression in various human tissues (a) and
cells (b) was assessed by quantitative RT-PCR. The x axis represents the approximate number of
copies of messenger RNA transcript per 10 ng of total RNA after normalization to a set of five
housekeeping genes. (c) Expression of P2Y15 and the P1 adenosine receptors (gene names
ADORA1, -2A, -2B, and -3) in various leukocyte subsets was analyzed by hybridization of cRNAs
generated from cellular total RNAs to microarrays containing approximately 20,000 gene-specific
probes. The x axis represents the relative fluorescence intensity of cRNAs bound to each gene-
specific probe.
Figure 4. P2Y15 is a functional receptor for AMP and adenosine. AMP (filled symbols) and
adenosine (open symbols) stimulated calcium mobilization (a) and cyclic AMP generation (b) in a
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 24
dose-dependent manner in transfected HEK293 cells expressing P2Y15 (circles). Nontransfected
HEK293 cells (squares), by contrast, showed little calcium mobilization in response to either ligand,
and showed cyclic AMP generation in response to adenosine but not to AMP. Calcium mobilization
was measured in 10-well replicates in an FDSS6000 functional drug screening system and plotted
as the integral of the ratio of signal to background over a 60 second time period. Cyclic AMP
generation in the presence of 10 µM forskolin was measured in duplicate with the Tropix cAMP
screen. The data for cyclic AMP generation are representative of three separate experiments. (c) No
differences could be seen between conversion rates of AMP to adenosine by nontransfected
HEK293 cells (white bars) and P2Y15-expressing cells (black bars). 1.85 µM [2-3H]-AMP was
added to triplicate cultures of nontransfectants and P2Y15 transfectants then recovered from the
medium after incubation times of 5 and 60 minutes. Thin layer chromatography analysis showed
only two bands, which migrated together with AMP and adenosine standards, indicating minimal
conversion to other products such as IMP, ADP, or ATP.
Figure 5. P2Y15 can bind both AMP and adenosine. Specific binding of increasing
concentrations of [2-3H]-adenosine (a) and [32P]-AMP (b) to transfected HEK293 cells expressing
P2Y15 (circles) and nontransfected HEK293 cells (squares) is shown. The binding of [2-3H]-
adenosine (c) or [2-3H]-AMP (d) to P2Y15-transfected cells and nontransfected cells could be
competed in both cases by unlabled AMP (filled symbols) or adenosine (open symbols). All binding
assays were performed in quadruplicate.
Figure 6. ADA can inhibit adenosine-induced but not AMP-induced P2Y15 signaling. Cyclic
AMP production induced by 25 µM adenosine in transfectants expressing P2Y15 was effectively
inhibited in a dose-dependent manner by ADA. Cyclic AMP production in response to 100 µM
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 25
AMP, however, showed little change even at ADA concetrations up to 4 U/ml. Cyclic AMP
production was measured in 4-well replicates in the presence of 10 µM forskolin as described in
figure 4.
Figure 7. Antagonists of adenosine block AMP- and adenosine-induced P2Y15 signaling.
Calcium mobilization induced by 10 µM AMP (a) or 10 µM adenosine (b) was effectively blocked
in a dose-dependent manner by the indicated adenosine receptor antagonists. Curves for alloxazine
and DPCPX, which showed agonistic effects in both transfectants and non-transfectants, are
excluded. Calcium mobilization was measured in 6-well replicates as described in figure 4. The data
are representative of three separate experiments.
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 26
Table 1. Ki values for antagonists of AMP- and adenosine-stimulated Ca2+ mobilization derived by
non-linear regression of the curves shown in figure 6a and 6b.
Adenosine receptor AMP antagonism Adenosine antagonism
Antagonist selectivity Ki (nM) Ki (nM)
Theophylline Non-selective 771 5,620
Caffeine Non-selective 3,309 23,100
IBMX Non-selective 540 23,800
8-SPT Non-selective 200 962
8-PT A1 384 6,330
DPCPX A1 - -
CSC A2A 277 18,100
Enprofylline A2B 3,033 64,060
Alloxazine A2B - -
Abbreviations: IBMX, 3-isobutyl-1-methylxanthine; 8-SPT, 8-(p-sulfophenyl)theophylline; 8-PT, 8-
phenyltheophylline; DPCPX, 8-cyclopentyl-1,3-dipropylxanthine; CSC, 8-(3-chlorostyryl)caffeine;
-, no detectable antagonistic effect.
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Page 27
Table 2. EC50 values for P2Y15 agonist-stimulated Ca2+ mobilization.
Agonist
Adenosine receptor
selectivity EC50 (nM)
AMP 920
Adenosine 670
NECA Non-selective 6,959
N6-cyclopentyladenosine A1 72,000
CGS-21680 hydrochloride A2A -
Chloro-IB-MECA A3 -
Abbreviations: NECA, 5’-(N-ethylcaboxamido)adenosine; Chloro-IB-MECA, 2-chloro-N6-(3-
iodobenzyl)-adenosine-5’-N-methyluronamide; -, no detectable agonistic effect.
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from
Hisayo Inbe, Shinichi Watanabe, Miwa Miyawaki, Eri Tanabe and Jeffrey A. Encinasadenosine
Identification and characterization of a cell-surface receptor, P2Y15, for AMP and
published online March 4, 2004J. Biol. Chem.
10.1074/jbc.M400360200Access the most updated version of this article at doi:
Alerts:
When a correction for this article is posted•
When this article is cited•
to choose from all of JBC's e-mail alertsClick here
by guest on March 19, 2020
http://ww
w.jbc.org/
Dow
nloaded from