| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
IUPHAR Nomenclature Report |
Department of Pharmacological Sciences, University of Milan, Milan, Italy (M.P.A., M.F.); Autonomic Neuroscience Centre, Royal Free & University College Medical School, London, United Kingdom (G.B., G.E.K.); Institut de Recherche Interdisciplinaire en Biologic Humanie et Moléculaire and Department of Medicinal Chemistry, Erasme Hospital, University of Libre, Bruxelles, Belgium (J.-M.B.); Department of Pharmacology, University of Cambridge, Cambridge, United Kingdom (E.A.B.); Molecular Pharmacology, Inspire Pharmaceuticals, Inc., Durham, North Carolina (J.L.B.); Department of Physiology and Pharmacology, University of Strathclyde, Strathclyde, United Kingdom (C.K.); Institut National de la Santé et de la Recherche Médicale U.311, Etablissement Francais du Sang-Alsace, Strasbourg, France (C.G.); Laboratory of Bioorganic Chemistry, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland (K.A.J.); and Department of Biochemistry, University of Missouri, Columbia, Missouri (G.A.W.)
Abstract
Abstract I. Brief Historical Background of Nucleotides and Their Receptors II. Molecular Structure of P2Y Receptors A. Nomenclature and Molecular History of P2Y Receptors B. Structural Aspects C. Orphan Receptors Related to P2Y Receptors III. Second Messenger Systems and Ion Channels A. Coupling to G Proteins and Intracellular Signaling Pathways B. P2Y Receptor Coupling to Ion Channels 1. Significance. 2. Approaches to Analysis of the Channel Interactions of Molecularly Identified P2Y Receptors. 3. Voltage-Activated Channels Regulated by P2Y Receptors. a. Ca2+ channels. b. The M-current K+ channel. 4. Activation or Inactivation of G Protein-Gated K+ Channels by P2Y Receptors. 5. Conclusions on the Interactions with Identified Ion Channels. C. Other Potential Interactions with Ion Channels IV. Principles of P2Y Receptor Classification V. Agonists and Antagonists A. Chemical Structure of Agonist and Antagonist Ligands 1. ADP-Preferring P2Y Receptors: 2. ATP-Preferring P2Y Receptor: 3. UTP-Recognizing P2Y Receptors: 4. UDP-Preferring P2Y Receptor: 5. UDP-Sugar-Preferring P2Y Receptor: B. Molecular Modeling Studies VI. P2Y Receptor Subtypes A. P2Y1 B. P2Y2 C. P2Y4 D. P2Y6 E. P2Y11 F. P2Y12 G. P2Y13 H. P2Y14 VII. Receptor Distribution and Function A. Excitable Cells, Nerves, Glial Cells, and Muscle B. Immune Cells C. Endocrine, Adipose, and Exocrine Cells D. Gut, Liver, and Biliary System E. Kidney and Bladder F. Lung G. Bone and Cartilage H. Skin I. Endothelial Cells J. Special Senses K. Platelets 1. The P2Y1 Receptor Initiates Platelet Activation and Aggregation. 2. The P2Y12 Receptor Completes and Amplifies Platelet Activation and Aggregation. VIII. Source of Naturally Occurring Ligands and Mechanisms of Transport and Breakdown A. Basal Unstimulated Nucleotide Release 1. Constitutive Release of ATP. 2. Constitutive Release of UDP-Glucose. B. ATP Release by Excitable and Secretory Tissues C. ATP Release by Nonexcitatory Cells 1. Stress/Hypoxia/Mechanical Stimulation. a. ATP binding cassette proteins. b. Stretch and voltage-activated Cl- channels. 2. Nucleotide Release via Vesicular Trafficking. 3. Agonist-Promoted ATP Release. D. ATP Release by Tissue Damage E. Extracellular Nucleotide Metabolism 1. Ecto-Nucleoside Triphosphate Diphosphohydrolases. 2. Ecto-Nucleotide Phosphophosphates/Phosphodiesterases. 3. Hydrolysis of UDP-Glucose. 4. Hydrolysis of Diadenosine Polyphosphates. 5. 5'-Nucleotidase. 6. Nucleoside Diphosphokinase 7. Alkaline Phosphatase. 8. Adenylate Kinase. IX. Interactions between P2Y and Other Receptors A. Modes of Interaction between G Protein-Coupled Receptors B. Receptor Dimerization C. Receptor Cross-Talk 1. G Protein-Coupled Receptors. 2. Receptor Tyrosine Kinases. 3. Ligand-Gated Cation Channels. X. Gene Activation Regulated by P2Y Receptors A. Scope of the Gene Activations B. Synaptically Released ATP Can Act in the Control of Gene Transcription XI. Potential Therapeutic Applications
There have been many advances in our knowledge about different aspects of P2Y receptor signaling since the last review published by our International Union of Pharmacology subcommittee. More receptor subtypes have been cloned and characterized and most orphan receptors deorphanized, so that it is now possible to provide a basis for a future subdivision of P2Y receptor subtypes. More is known about the functional elements of the P2Y receptor molecules and the signaling pathways involved, including interactions with ion channels. There have been substantial developments in the design of selective agonists and antagonists to some of the P2Y receptor subtypes. There are new findings about the mechanisms underlying nucleotide release and ectoenzymatic nucleotide breakdown. Interactions between P2Y receptors and receptors to other signaling molecules have been explored as well as P2Y-mediated control of gene transcription. The distribution and roles of P2Y receptor subtypes in many different cell types are better understood and P2Y receptor-related compounds are being explored for therapeutic purposes. These and other advances are discussed in the present review.
I. Brief Historical Background of Nucleotides and Their Receptors
The first description of the extracellular signaling by purines was by Drury and Szent-Györgyi (1929
), and purinergic receptors were defined in 1976 (Burnstock, 1976). After an early hint (Spedding and Weetman, 1976), receptors for purines were subdivided into P1 (adenosine) and P2 (ATP and ADP) receptors (Burnstock, 1978), and later subdivision of P2 receptors into P2X and P2Y subtypes was made on the basis of pharmacology (Burnstock and Kennedy, 1985). It was recognized that some P2Y receptors responded to pyrimidines as well as purines (Seifert and Schultz, 1989). After cloning of P2 receptors and studies of transduction mechanisms in the early 1990s, the basis for subdivision into P2X and P2Y receptor families was confirmed and extended (Abbracchio and Burnstock, 1994) and seven subtypes of P2X receptors and eight subtypes of P2Y receptors are currently recognized (Ralevic and Burnstock, 1998; North, 2002; Burnstock, 2004).
II. Molecular Structure of P2Y Receptors
A. Nomenclature and Molecular History of P2Y Receptors
Regarding the currently used nomenclature, P2Y is used for functional mammalian receptor proteins and functional nonmammalian species. The lower case, p2y, is used for mammalian orphan receptors or functional nonmammalian receptors without a mammalian ortholog. The subscript number (1-n) following P2Y or p2y sequentially list proteins in their chronological order of cDNA cloning. The first P2 receptors were cloned in 1993 (Lustig et al., 1993; Webb et al., 1993). They corresponded to receptors previously characterized by pharmacological criteria: P2Y1 (formerly P2Y) and P2Y2 (formerly P2U). Since then several other subtypes were isolated by homology cloning and assigned a subscript on the basis of cloning chronology (P2Y4, P2Y6, and P2Y11). The long-awaited Gi-coupled ADP receptor (P2Y12) of platelets was finally isolated by expression cloning (Hollopeter et al., 2001), and P2Y13 and P2Y14 receptors were characterized during a systematic study of orphan receptors (Chambers et al., 2000; Communi et al., 2001a). As of today, there are eight accepted human P2Y receptors: P2Y1, P2Y2, P2Y4, P2Y6, P2Y11, P2Y12, P2Y13, and P2Y14 (Abbracchio et al., 2003) (Table 1; Fig. 1) (see also section IV.). The missing numbers represent either nonmammalian orthologs or receptors having some sequence homology to P2Y receptors but for which there is no functional evidence of responsiveness to nucleotides. In particular p2y3 (Webb et al., 1996a) may be a chicken ortholog of P2Y6 (Li et al., 1998), whereas p2y8 (Bogdanov et al., 1997) and tp2y (Boyer et al., 2000) could be the Xenopus and turkey orthologs of P2Y4, respectively. p2y7 (Akbar et al., 1996) is a leukotriene B4 receptor (Herold et al., 1997; Yokomizo et al., 1997); however, recently cross-reaction between agonists for some leukotriene receptors and some P2Y receptors has been found (see section IX.C.1), requiring further investigation. p2y5 (Webb et al., 1996b; Li et al., 1997) and p2y10 (Rao et al., 1999) must be considered as orphan receptors, although it has been reported (King and Townsend-Nicholson, 2000) that human p2y5 expressed in oocytes gives functional responses to ATP. p2y9 was reported to be a novel receptor for lysophosphatidic acid, distant from the Edg family (Noguchi et al., 2003). P2Y15 was recently introduced to designate the orphan receptor GPR80/GPR99 on the basis that it would be a receptor for adenosine 5'-monophospahte (AMP2) (Inbe et al., 2004), but it is now firmly established that it is actually a receptor for 
-ketoglutarate (He et al., 2004; Qi et al., 2004; Suarez Gonzalez et al., 2004), as was also recently underlined in a report by the Subcommittee (Abbracchio et al., 2005).
|
|
In contrast with P2X receptors, P2Y receptor genes do not contain introns in the coding sequence, except for the P2Y11 receptor. Site-directed mutagenesis of the P2Y1 and P2Y2 receptors has shown that some positively charged residues in transmembrane domains (TM) 3, 6, and 7 are crucial for receptor activation by nucleotides (Erb et al., 1995; Jiang et al., 1997b) (Fig. 2) (section V.). They probably interact with the negative charges of the phosphate groups of nucleotides, since it is known that the receptor ligands are nucleotide species uncomplexed to magnesium or calcium. Actually, the eight P2Y receptors identified so far have a H-X-X-R/K motif in TM6. The P2Y1, P2Y2, P2Y4, P2Y6, and P2Y11 receptors share a Y-Q/K-X-X-R motif in TM7, whereas another motif, K-E-X-X-L is found in P2Y12, P2Y13, and P2Y14 receptors (Abbracchio et al., 2003) (see also Fig. 2 and sections IV. and V.). More recently, for P2Y12, P2Y13, and P2Y14 receptors, one additional K residue in extracellular loop 2 has been suggested to be particularly important for nucleotide binding (Costanzi et al., 2004) (see also below; section V.).
|
C. Orphan Receptors Related to P2Y Receptors
The bioinformatic analysis of the human genome has revealed the existence of approximately 800 G proteincoupled receptors (GPCRs) (Fredriksson et al., 2003), of which 367 would be receptors for endogenous ligands, the remaining ones being olfactory and other chemosensory receptors (Vassilatis et al., 2003). Among the 367 "endoG-PCRs," more than 150 remain orphans. Several of the currently available orphan GPCR sequences are structurally related to P2Y receptors (Lee et al., 2001; Wittenberger et al., 2001; Joost and Methner, 2002; Vassilatis et al., 2003): GPR87, H963, and GPR34. In particular, the sequence of those receptors contains the structural motifs in TM6 and TM7 described earlier (Fig. 2).
Human GPR87 has been shown to be highly expressed in placenta and thymus. The mouse ortholog has also been cloned and has been found to be expressed in brain and liver (Wittenberger et al., 2001).
Human H963 has been shown to be expressed in fibroblasts, human peripheral blood mononuclear cells, and T cells (S. Ferrario, D. Lecca, C. Volonté, and M. P. Abbracchio, unpublished data). Neither the mouse nor the rat orthologs have been cloned; however, a recent study has suggested that GPR34 is not a nucleotide receptor (Sugo et al., 2006).
Human GPR34 is expressed in brain, heart, placenta, small intestine, pancreas, spleen, thymus, kidney, and skeletal muscle (Schöneberg et al., 1999). The mouse ortholog has also been described, and prominent expression has been found in liver and testis (Schöneberg et al., 1999).
Despite active research in several laboratories, the ligands of these receptors have not been identified. For example, despite the demonstration that GPR34 and ADP-like receptors (e.g., P2Y12 and P2Y13) have a common evolutionary origin (Schulz and Schöneberg, 2003), in the inositol phosphate assay, Cos-7 cells coexpressing GPR34 and Gqi4 did not show any response to ADP application (Schöneberg et al., 1999).
III. Second Messenger Systems and Ion Channels
A. Coupling to G Proteins and Intracellular Signaling Pathways
Coupling of the various P2Y receptors to specific G proteins was initially inferred from indirect evidence [measurement of intracellular levels of inositol phosphates, calcium, or cAMP and determination of pertussis toxin (PTX) sensitivity]. Direct evidence was recently obtained by measuring the effect of ADP on GTP hydrolysis in vesicles reconstituted with P2Y1 and either Gq
1
2 or G1112 (Waldo and Harden, 2004). Similar experiments demonstrated that P2Y12 couples to Gi2 more effectively than to Gi1 and Gi3 and not at all to Go or Gq (Bodor et al., 2003). One given P2Y receptor can couple to functionally distinct G proteins. For instance, in HEL cells, activation of phospholipase C (PLC) by the P2Y2 receptor is inhibited completely by a G16 antisense oligonucleotide but also partially by PTX (Baltensperger and Porzig, 1997). Similarly, in gastric smooth muscle cells, it appears that P2Y2 couples to PLC-1 via Gq/11 and to PLC-3 via Gi312-derived subunits (Murthy and Makhlouf, 1998). The P2Y2 receptor also has been shown to interact with v integrin to promote Go-mediated chemotaxis in astrocytoma cells (Bagchi et al., 2005). The P2Y12 receptor activates phosphatidylinositol 3-kinase (PI3-K) via Gi, but also RhoA and Rho kinase (Soulet et al., 2004). This action, which is insensitive to PTX, might be mediated by G12/13, recently shown to play a critical role in platelet activation (Moers et al., 2003). Coupling of the same P2Y receptor to distinct G proteins and signaling pathways provides the possibility of agonist-specific signaling involving distinct active conformations of the receptor. For instance, activation of the P2Y11 receptor by ATP leads to a rise in cAMP and in inositol trisphosphate (IP3) and cytosolic calcium, whereas activation by uridine 5'-triphosphate (UTP) has been reported to produce calcium mobilization without IP3 or cAMP increase (White et al., 2003). The P2Y13 receptor can simultaneously couple to G16, Gi, and, at high concentrations of ADP, to Gs (like other Gi-coupled receptors such as the 2-adrenergic receptor): these three signaling pathways are characterized by different ratios of ADP to 2-methylthio-ADP (2-MeSADP) potency, suggesting the existence of ligand-specific conformations (Marteau et al., 2003). The activation of several P2Y receptors is commonly associated with the stimulation of several mitogen-activated protein (MAP) kinases, in particular extracellular signal-regulated protein kinase (ERK) 1/2. According to the cell context and the particular subtype, other classes of the MAP kinases, protein kinase (PK) C, calcium, and PI3-K are found to be involved to a variable extent (Soltoff et al., 1998; Huwiler et al., 2000; Communi et al., 2001a; Santiago-Pérez et al., 2001; Sellers et al., 2001).
B. P2Y Receptor Coupling to Ion Channels
1. Significance.
In recent years, GPCRs in neurons and other excitable cells have been found to modulate the activity of voltage-gated ion channels in the cell membrane through certain actions of activated G proteins. Such actions are now well established in closing (or in certain cases in opening or potentiating) various classes of K+ channels (Hille, 1994) and voltage-gated Ca2+ channels (Dolphin, 2003). Various voltage-gated Na+ channels also have been observed to be modulated in certain cases via GPCR actions (Cantrell and Catterall, 2001; Maurice et al., 2001; Mantegazza et al., 2005), but this has not been reported for P2Y receptors. Although GPCR downstream signaling can lead to indirect effects on ion channels via activation of protein kinases, some GPCR regulation of several types of ion channel is more specific and more direct. This action has been investigated so far for only a very small fraction of the GPCR class. For P2Y receptors, specific couplings to certain K+ and Ca2+ channels have been observed and analyzed. This coupling will be an important component of P2Y receptor signal transduction, but one that will usually be invisible in studies of second messenger or downstream pathways, since those channel interactions can occur in short timescales (down to 
100 ms) by a direct or quasi-direct pathway in the cell membrane.
Our consideration here of P2Y signaling through cell membrane channels is necessarily focused on cases in which the P2Y subtypes concerned have been identified. Thus, in a number of recent studies, the disturbing factors of enzymatic breakdown or interconversion of the nucleotides applied as agonists (Lazarowski et al., 2000) or the activation of adenosine receptors via such breakdown (Masino et al., 2002) have been experimentally minimized, allowing demonstration of ion channel responses upon activation of native P2Y receptors in brain neurons, with clear evidence for their identity (Wirkner et al., 2002; Khakh et al., 2003; Koizumi et al., 2003; Luthardt et al., 2003; Zhang et al., 2003b; Bowser and Khakh, 2004; Kawamura et al., 2004). That evidence shows that ATP (or UTP or their products ADP or UDP) present at synapses, plus ATP diffusing from astrocytes, activates P2Y receptors on distinct subsets of brain neurons, regulating their activities by the coupling of those receptors to specific ion channels, as detailed below.
Although ion channel couplings of P2Y receptors are primarily of importance in neurons, they have in a few cases been detected also in various other tissues, e.g., in cardiac muscle cells (Vassort, 2001). As yet that category has been little explored.
2. Approaches to Analysis of the Channel Interactions of Molecularly Identified P2Y Receptors.
Some studies of channel coupling by P2Y receptors have been made by heterologous expression in commonly transfected host cell lines such as CHO or HEK293, or in the Xenopus oocyte. However, usually both the P2Y receptor and the identified ion channel under study must be introduced into them, and the final relationship and protein environment of those components may be far from that in any native neuron, in which individual GPCR types can be located specifically with their effectors in microdomains (Delmas et al., 2004).
The problems there can be minimized if a suitable neuronal host cell can be found. A number of requirements for this exist (e.g., endogenous P2 receptors to be insignificant therein), and all of those conditions have been found to be met in the superior cervical ganglion (SCG) cell from the sympathetic nervous system of the young rat or mouse (Brown et al., 2000a). This cell type is well equipped with endogenous ion channels of the types found in neurons generally (Ikeda, 1996; Filippov et al., 1997). Its size readily allows nuclear injection of a receptor cDNA, a route that favors normal processing and trafficking of the protein. Transfection difficulties with neurons are avoided, and recordings of the channel couplings can be made in each receptor-expressing cell, as reviewed below. Because single cells are constantly perfused with medium and subsequently with the (purified) agonist and the assay period is extremely short, the method avoids potential problems well known in other P2 receptor activity studies, i.e., accumulation of nucleotides released from cell populations or their acute release by cell disturbance, as well as losses of the added agonist by metabolism.
3. Voltage-Activated Channels Regulated by P2Y Receptors.
Among the channels with which the SCG cell membrane is well endowed are two types of voltage-gated channel that are important in receptor-based regulation of neuronal activity, the Ca2+ channel of the N-type and the M-current K+ channel. The M-current K+ channels are heteromers of subunits of the Kv7 family and are critical for setting the responsiveness of neurons to synaptic inputs (Selyanko et al., 2001, and references therein). Closing of the M-current K+ and/or N-type Ca2+ channels by action of certain P2Y receptors has been shown to occur.
a. Ca2+ channels.
For Gi/o protein-coupled receptors in general, inhibition of the N-type Ca2+ current has been shown to occur through G subunits, acting directly on the channel (reviewed by Dolphin, 2003). This action holds for the Gi-linked P2Y12 receptor in the SCG system, as shown by the demonstration that closure of the N-type Ca2+ channel via the P2Y12 receptor is fully sensitive to PTX and is totally abolished when G-transducin, a G-scavenging protein, is coexpressed (Simon et al., 2002). For the Gi/o-linked P2Y13 receptor, inhibition of voltage-gated Ca2+ channels would again be expected. Indeed, evidence has been obtained in HEK293 cells (into which the N-type Ca2+ channel had been introduced by transfection) to suggest that a native P2Y13 receptor is there and acts thus (Wirkner et al., 2004).
The action at N-type Ca2+ channels of activated P2Y1 and P2Y2 receptors (Table 1) is very similar to that of the endogenous M1 receptor (Gq/11-linked) in the same cells, with all three receptors showing a PTX-sensitive and a PTX-insensitive component. q is presumably required for the latter component, as was established using knockout mice for the M1 inhibition of this channel (Haley et al., 2000), although those P2Y receptors could conceivably also use 11 in some other cells. The PTX-sensitive component requires (at least for M1; Haley et al., 2000) the Go protein. With P2Y1, in contrast, Gi/o does not act here (Table 2). However, with P2Y1,2,6 receptors, both the PTX-insensitive and the PTX-sensitive N-type Ca2+ channel responses are abolished when subunits are sequestered by G-transducin (Simon et al., 2002; Filippov et al., 2004). Hence, the and the components of selected trimeric G protein(s) must operate together in this type of P2Y receptor action, as summarized in Table 2.
|
Higher potency of the agonists for a given subtype than in second messenger assay systems is generally observed in the channel transduction, which follows a direct route within the cell membrane; e.g., for 2-MeSADP at the P2Y1 receptor the EC50 value for Ca2+ channel closure is 0.57 ± 0.05 nM (Filippov et al., 2000). (All experimental values quoted in section III. are at 20-22°C). Adenosine triphosphates (e.g., 2-MeSATP and ATP) when in pure form also are agonists in this P2Y1 reaction, in conditions in which their agonistic diphosphates are excluded throughout; this has been controversial for P2Y1 receptors in some other expression systems, but the potency and efficacy of the triphosphates vary strongly with the density of this receptor subtype (see discussion and references in Filippov et al., 2000). The N-type Ca2+ channel closure via the P2Y12 receptor behaves likewise, with 2-MeSATP an agonist at P2Y12 receptors, too, with exceptional potency (Simon et al., 2002) (see also section VI.F. for further discussion). This behavior, compared with the variation of behavior seen with these agonists at P2Y12 receptors in different native situations (Barnard and Simon, 2001), again suggests a particularly strong dependence of the observed efficacy of these subtypes on receptor density in the membrane and on the directness of the transduction pathway involved.
Interestingly, the agonist selectivity of a P2Y receptor can also be changed in the channel interactions from that observed for it in transductions downstream. Thus, whereas the transfected P2Y6 receptor was reported to be UDP-selective and to have negligible action by UTP in its IP3 formation (Nicholas et al., 1996), both those nucleotides are strong agonists in the closure of the N-type Ca2+ channel and likewise for the M-current K+ channel response (Filippov et al., 1999). With impurity and metabolism of the UTP being excluded, the KA of rat P2Y6 for UTP in its channel activity is 20.1 ± 1.4 nM. A precedent for this phenomenon lies in the transduction dependence of relative agonist potencies described above for the P2Y13 receptor (section III.A.); it is interesting for the P2Y6 receptor because it indicates alternative binding site conformations for two native agonists, UTP and UDP.
b. The M-current K+ channel.
The M-current can be inhibited through the activation of a number of GPCRs of the Gq/11-linked class (Brown and Yu, 2000) but not of other classes. The G protein subunit involved in GPCR-mediated closing of this channel by M1 muscarinic receptors in rat or mouse SCG neurons is Gq, as shown both by anti-Gq antibody depletion and by Gq-gene knockout (Haley et al., 2000). For P2Y receptors, this parallel, together with the complete inability of the Gi/o-linked P2Y12 receptor to affect the M-current plus the PTX-insensitive closure of that channel by all four of the Gq/11-linked subtypes examined (Table 2), indicates that this action can be ascribed to a Gq pathway also. Whether in other neurons G11 may also act thus is not yet excluded.
The pathway from Gq-GTP to the M-type channel Kv7 proteins has been deduced to be, in general, via PLC-mediated depletion of phosphatidylinositol-4,5-bisphosphate (PIP2), removing its stabilizing interaction with those proteins and hence allowing channel closure (Ford et al., 2003; Zhang et al., 2003a; Suh et al., 2004; Winks et al., 2005). PIP2 thus acts as a second messenger, but one that is diffusible only within the membrane and acting negatively. Restoration of the membrane PIP2 occurs by a PI4-kinase and PI5-kinase pathway (Winks et al., 2005). The same PIP2/PI4-kinase mechanism has been supported for the closure of the N-type Ca2+ channel by similar observations on Gq-linked muscarinic M1 receptor action and its recovery (Gamper et al., 2004). We presume that a similar mechanism is operative for P2Y receptors of the Gq/11-linked class, but this has not as yet been specifically examined.
4. Activation or Inactivation of G Protein-Gated K+ Channels by P2Y Receptors.
An entirely different class of K+ channels is specialized for interaction with GPCRs, i.e., the G protein-activated inward rectifiers (Kir3 or GIRK channels). Activations of Gi/o-linked receptors, but not of those linked to Gq/11, generally open these channels (Fernandez-Fernandez et al., 2001; Stanfield et al., 2002). In line with this, in the rat SCG neurons 2-MeSADP, acting at an introduced rat P2Y12 receptor, strongly activates a GIRK channel (EC50: 0.099 ± 0.008 nM) (Simon et al., 2002).
Very exceptionally, some P2Y receptors, which usually link to Gq/11, have also been found to efficiently activate GIRK channels. This case occurs with P2Y1 receptors (Simon et al., 2002), in which 2-MeSADP, for example, acts to open with high potency a GIRK channel in the SCG cell. Another instance has been seen in Xenopus oocytes, with the P2Y2 receptor (activated by ATP or by UTP) opening a GIRK channel, when both that and the receptor are implanted therein (Mosbacher et al., 1998; Mark et al., 2000). In these cases the normally Gq/11-linked P2Y1 or P2Y2 receptors can display in neurons a channel-based transduction that is highly PTX-sensitive and dependent on Gi. This action is fast and direct, through subunits released in the cell membrane from Gi trimers and binding at the K+ channel (Simon et al., 2002; Filippov et al., 2004), i.e., as found with P2Y12. This mechanism of the P2Y1 receptor signal transduction via a K+ channel could be an indicator of an interaction of that receptor in neurons with a recruited GPCR-regulating component [e.g., a regulator of GPCR signaling (RGS) protein], since reconstituted P2Y1 protein gives no coupling to Gi or Go (Waldo and Harden, 2004) (see section VI.A.).
A second, independent, type of interaction with GIRK channels can occur with P2Y receptors, namely the closing of an open GIRK channel. This slow inactivation phase is common in Gq/11-linked GPCRs, occurring in situ at GIRK channels already opened via endogenous Gi/o-linked receptors. It occurs with P2Y1 but not with P2Y12 receptors (Simon et al., 2002). This inactivation is found also with P2Y4 and P2Y6 in the neuron (Filippov et al., 2004) and with P2Y2 in oocyte expression (Mark et al., 2000). It is unaffected by PTX, with Gq/11 being involved, as shown in the P2Y1 case with evidence that included blockade by sequestering the released q/11 subunits using the coexpressed RGS2 protein, which is specific for that action (Filippov et al., 2004). Again, when investigated for P2Y1 receptors, depletion of PIP2 from the membrane and the liberation of IP3 in the cytosol can be seen to be correlated with the course of the GIRK current inactivation, when visualized by a sensor for them, PLC
-PH, fluorescently tagged (Filippov et al., 2004). There appears to be a role for PIP2 in the GIRK inactivation by P2Y receptors, related to that proposed for it (see above) in the M-current K+ channel inactivation, but this requires further specification.
5. Conclusions on the Interactions with Identified Ion Channels.
Five P2Y subtypes have been compared so far (Table 2). Clear differences are seen between the P2Y subtypes, with only P2Y1 and P2Y2 receptors showing a common behavior in the three transductions. These considerable variations support the conclusion that the channel couplings seen are not a general phenomenon produced by overloading with an exogenous receptor. Likewise, for all of these P2Y receptors, their maximum inhibition of the N-type Ca2+ current is well below 100% of the total N-type Ca2+ current recorded and is less than that attainable by test activations of the native 2-adrenergic or M1 muscarinic receptors in the same cell (see references cited in Table 2). There is also evidence for the coupling of some native P2Y receptors to such ion channels in brain neurons (section III.B.1.) and also in autonomic neurons and related cells (Ennion et al., 2004; Lechner et al., 2004, and references cited therein).
Some promiscuity between PTX-sensitive and PTX-insensitive G proteins is described here for these tranductions (Table 2). Such cross-reaction is already known for P2Y2 and P2Y4 receptors in other signaling pathways (see sections III.A. VI.B, and VI.C.). We noted its occurrence here, however, also with the P2Y1 receptor, in its coupling to a Ca2+ channel and to some but not all of its couplings to K+ channels. Indeed, the P2Y1-linked activation of the GIRK K+ channel described is almost entirely Gi/o-mediated, but none of that component is seen in its PLC-dependent transductions; this finding underlines the independence of the channel-coupling pathways in P2Y receptor signaling. The coupling to that channel is not, however, necessarily linked to a relaxation of selectivity for q/11-containing heterotrimers, since the P2Y6 receptor, in contrast with the others, strictly maintains that selectivity in all of the couplings (Table 2) as well as in other transductions. Again, in the independent inactivation reaction of the GIRK channel all four of these P2Y receptors generally associated with Gq/11 linkage were seen here to fully maintain that selectivity.
Such ion channel responses are usually recorded in whole-cell patch-clamping, but this may permit diffusible intracellular cofactors to dialyze out. Most of the analyses of P2Y receptor coupling in the SCG cell discussed above were made instead in the more-laborious perforated-patch mode (using amphotericin B to form small membrane pores), to avoid that possibility. In the case of the P2Y6-mediated Ca2+ channel closure (Filippov et al., 1999), use of this configuration abolished a partial block by PTX (although for other subtypes it can remain). Hence a soluble component, possibly an RGS protein, may cooperate in the membrane-delimited reactions of the P2Y6 receptor to direct its G protein coupling.
The P2Y4 receptor has not so far been shown to occur in neurons, unlike P2Y1 and P2Y6 receptors; its mRNA is prominent in the brain ventricular system, cardiac and skeletal muscles, some smooth muscles, and some other peripheral sites but is undetectable in neurons (Webb et al., 1998). This may be why its coupling to the N-type Ca2+ channel of a neuron can be weak and readily lost (Table 2). Adaptation of such a P2Y subtype may evolve for a different signaling environment than that in neurons.
C. Other Potential Interactions with Ion Channels
Little is known of these as yet. One clue to some other interactions of the P2Y1 receptor comes from the recent finding (Fam et al., 2005) that it can bind strongly to the Na+/H+ exchanger regulatory factor 2 (NHERF-2) through the extreme C-terminal motif DTSL, which is specific to P2Y1 in this family. For comparison, binding of the related NHERF-1 to P2Y1 receptors (Hall et al., 1998) is negligible (Fam et al., 2005). The three membrane-located NHERF subtypes either activate or inhibit various Na+/H+ exchangers, but these actions are indirect since it is now known that NHERFs are actually scaffolding proteins, which can localize various exchangers in membrane microdomains with selected receptors and signaling intermediates, e.g., Gq, Src, and certain isoenzymes of PLC and of PKC (Donowitz et al., 2005). For example, the NHERF-2 scaffold can link a cGMP-dependent protein kinase or a protein kinase A anchoring protein and hence protein kinase A, to modulate a tethered Na+/H+ exchanger by specific phosphorylations thereon (Cha et al., 2005; Donowitz et al., 2005).
When the endogenous P2Y1 receptor (as studied in C6 glioma cells) is linked through its tail to the NHERF-2 scaffold, the Ca2+ transients produced by its activation by 2-MeSADP become strongly prolonged (Fam et al., 2005). This will change the P2Y1 selectivity for the various calcium-sensitive signaling cascades and for ion channel interactions. Another interaction of the P2Y1 receptor is with the chloride channel of the cystic fibrosis transmembrane conductance regulator (CFTR); in renal epithelial cells, 2-MeSADP activation of native P2Y1 receptors stimulates the chloride channel activity of the CFTR. This is again an indirect action arising from the NHERF-2 colocalization of this P2Y subtype and the CFTR; expression of a dominant negative NHERF-2 mutant blocks the CFTR regulation through P2Y1, as does a blocker of the protein kinase A anchoring protein binding of PKA (Guerra et al., 2004). The evidence suggested that P2Y1 receptor-mediated PKC activation leads to potentiation of PKA and its action on the associated CFTR channels.
Additionally, a highly unusual mode of GPCR interaction with an ion channel has been suggested for several P2Y receptors by Lee et al. (2003b). A novel, unidentified, voltage-gated channel of the Xenopus oocyte, Tin, was reported to be activated and modulated after expression of human P2Y1,2,6,11, but not by P2Y4 nor by other Gq/11-linked GPCRs. It was deduced that this does involve a direct receptor binding to the channel. However, expression of G also activates this channel and the mechanism and physiological significance are at present unclear.
IV. Principles of P2Y Receptor Classification
As already outlined above, eight distinct mammalian P2Y receptors have been cloned and recognized so far: the P2Y1,2,4,6,11,12,13 and the recently reclassified P2Y14 receptors (Abbracchio et al., 2003). The missing numbers in the P2Y1-14 sequence represent GPCRs cloned from nonmammalian vertebrates or receptors for which a functional response to nucleotides has not yet been convincingly demonstrated.
Pharmacologically (Table 1) P2Y receptors can be broadly subdivided into 1) adenine nucleotide-preferring receptors mainly responding to ADP and ATP. This group includes human and rodent P2Y1, P2Y12, and P2Y13, and human P2Y11 (which has, however, been recently reported to also respond to UTP) (White et al., 2003); 2) uracil nucleotide-preferring receptors. This group includes human P2Y4 and P2Y6 responding to either UTP or UDP; 3) receptors of mixed selectivity (human and rodent P2Y2, rodent P2Y4 and, possibly, P2Y11); and 4) receptors responding solely to the sugar nucleotides UDP-glucose and UDP-galactose (P2Y14).
From a phylogenetic and structural (i.e., protein sequence) point of view, two distinct P2Y receptor subgroups characterized by a relatively high level of sequence divergence have been identified (Jacobson et al., 2002; Abbracchio et al., 2003). The first subgroup includes P2Y1,2,4,6,11 subtypes and the second subgroup encompasses the P2Y12,13,14 subtypes (see dendrogram in Fig. 1 reproduced from Abbracchio et al., 2003). Alignment of the deduced amino acid sequences of the cloned P2Y receptors has shown that the human members of this family are 21 to 48% identical (Table 3). The highest degree of sequence identity is found among the second subgroup of P2Y12,13,14. Interestingly, despite clear phylogenetic relationships with the first subgroup, the P2Y11 subtype seems to differ from all the others, for both sequence and pharmacological differences between species (e.g., canine versus human) and also based on its absence in the murine and rat genomes (Table 1). Thus, it might be hypothesized that this receptor differentiated from P2Y1 and subsequently underwent many modifications and insertions that led to a dissimilar receptor, despite the common origin. Cotranscription and intergenic splicing of the P2Y11 gene might be another evolutionary event accounting for its dissimilarity from the other P2Y receptors.
|
The two P2Y receptor subgroups highlighted above also differ in several other features. In particular, specific amino acid motifs in TM6 and TM7 have been previously proposed to be important for binding to extracellular nucleotides (Erb et al., 1995; Jiang et al., 1997b; Boarder and Webb, 2001; Jacobson et al., 2002). All human P2Y receptors share the typical TM6 H-X-X-R/K motif that might be important for agonist activity (Erb et al., 1995; Jiang et al., 1997b; Boarder and Webb, 2001; Jacobson et al., 2002) (Table 3; see also section V.). A Q/K-X-X-R defining motif in TM7 has also been proposed to participate in ligand binding for the P2Y1,2,4,6 and P2Y11 receptors. In P2Y12,13,14 receptors, this motif is substituted with K-E-X-X-L, which might affect ligand binding characteristics (Table 3). In humans, the genes of P2Y12,13,14 receptors cluster in the same region of chromosome 3, together with the gene encoding for the P2Y1 receptor; in this region; three additional still unidentified "orphan" GPCRs structurally related to the known P2Y receptors can also be found (Table 1). The mapping of the known genes of the P2Y receptors and the structurally related orphans in the human genome is detailed in Simon and Barnard (2003). Interestingly, two of these orphan receptors (i.e., GPR87 and H963) also show full conservation of the defining motifs in TM6 and TM7 typically found in P2Y12,13,14. Their functional characterization may eventually lead to inclusion in this P2Y receptor subgroup.
Finally, these two P2Y receptor subgroups also differ in their primary coupling to transductional G proteins. In particular, receptors in the first subgroup (i.e., P2Y1,2,4,6,11) all principally use Gq/G11 to activate the PLC/IP3 pathway and release intracellular calcium, whereas receptors in the second subgroup (i.e., P2Y12,13,14) almost exclusively couple to members of the Gi/o family of G proteins (Table 3; see also section III. and individual receptor subsections and individual receptor summary tables that appear at the end of the article). Secondary couplings have been also reported, especially for receptors of the first subgroup in heterologous expression systems (Simon et al., 2002; Burnstock, 2003; King and Townsend-Nicholson, 2003; Köttgen et al., 2003; White et al., 2003). For receptors of the second group, P2Y13 has been also reported to couple to G16 and stimulate PLC in recombinant systems overexpressing this G protein (Communi et al., 2001a; Marteau et al., 2003), whereas activation of the native P2Y14 receptor in astrocytes and microglia has been shown to increase intracellular calcium levels (Fumagalli et al., 2003; Bianco et al., 2005). Such "promiscuity" of G protein-coupling may depend on the indirect activation of additional G protein subtypes within protein complexes containing the P2Y receptor.
Thus, a division into two subgroups could be considered, based on 1) phylogenetic (i.e., sequence) similarity (Fig. 1; Table 3), the 2) presence of amino acid defining motifs proposed to be important for ligand binding (Table 3; see also section V.), and 3) selectivity of primary G protein coupling (Table 3). However, we prefer to wait to formally implement this subdivision until there is a more complete knowledge of these receptors, with some of the orphan P2Y-like receptors still waiting for deorphanization, the possibility of new receptors still to be discovered, and the place of the P2Y11 receptor still to be clearly resolved.
A. Chemical Structure of Agonist and Antagonist Ligands
Most of the P2Y receptor subtypes are still lacking potent and selective synthetic agonists and antagonists. However, considerable progress in exploring structure-activity relationships (SARs) has been made for P2Y1 and P2Y12 receptors and to a lesser extent for the P2Y2 receptor. Radioligand binding studies have been successfully carried out at the P2Y1 and P2Y12 receptors, but so far not at any other P2Y receptor subtype. Here we describe the current state of molecular probes known for the P2Y receptors, categorized by the chemical class of the endogenous agonists.
|
; Boeynaems et al., 2003; Marteau et al., 2003). At P2Y1 receptors, derivatives of ADP tend to be full agonists (the EC50 of ADP is 100 nM), whereas ATP appears to be a partial agonist. At P2Y12 receptors, ADP derivatives activate (the EC50 of ADP is 100 nM) and 5'-triphosphate derivatives antagonize (Gachet and Hechler, 2005). At P2Y13 receptors, both ADP and ATP might act as full agonists, with EC50 values of 100 nM. However, under some conditions, ATP can behave as a weak partial agonist, suggesting that, as described for the P2Y1 receptor, the activity of ATP in recombinant systems may vary according to the level of expression of the P2Y13 receptor (see also below).
Phosphate modifications among P2Y receptor ligands often serve to increase their stability toward ecto-nucleotidases. For example, the added stability of a terminal thiophosphate group resulted in its incorporation in some useful P2Y nucleotide agonists. One such analog, ADPS (2), is a potent agonist of both P2Y1 (EC50 = 96 nM), P2Y12 (EC50 = 82 nM) (Jacobson et al., 2002), and P2Y13 receptors (EC50 = 42 nM) (Communi et al., 2001a). Although terminal thiophosphates are enzymatically more stable than the oxygen equivalents, they are subject to chemical oxidation reactions; thus, solutions of these compounds are prone to instability.
Figure 3 shows the structures of adenine-derived nucleotide agonists of P2Y receptors. The structure of the nucleobase of adenine nucleotides has been extensively probed for effects at P2Y receptors. 8-Aza and 1-deaza modifications are generally tolerated. A fluorescent adenine-modified derivative of ATP (5) behaved as a potent P2Y1 receptor agonist (Sharon et al., 2004). The 2-position of the adenine ring can accommodate a wide variety of substituents, with resultant activation of both P2Y1 and P2Y12 receptors. Long-chain and sterically bulky groups may be accommodated at the 2-position. In particular, 2-alkylthio ethers (Fischer et al., 1993; Brown et al., 2000b) appear to provide high potency at these subtypes when bonded to a variety of alkyl or alkylaryl groups. Notably, the smallest member of this class is 2-MeSADP [6, which is a potent agonist (EC50) at P2Y1 (6 nM), P2Y12 (1 nM), and P2Y13 (1 nM) receptors (Jacobson et al., 2002; Marteau et al., 2003); however, see also above and individual receptor subsection] and is highly selective in comparison with other P2Y receptor subtypes. For example, at the P2Y2 receptor compound 6 is inactive at 100 µM. The corresponding triphosphate, 2-MeSATP (7), is less potent and selective as a P2Y receptor agonist, since it also activates P2X receptors (King, 1998). The sterically bulky p-aminophenylethylthio analog (2-[2-(4-aminophenyl)ethylhio]adenosine 5'-triphosphate) (8), potently activated the P2Y1 receptor (EC50 = 1 nM) (Fischer et al., 1993). The SAR of alkynyl substitutions at the 2-position of P2Y1 receptor agonists has been explored (Cristalli et al., 2005).
Although AMP is inactive at the P2Y1 receptor, addition of a 2-thioether substituent as a receptor "anchor" causes AMP analogs to activate P2Y1 receptors. Among these derivatives, 2-hextylthioadenosine 5'monophosphate (9) was especially potent, with an EC50 of 59 nM at the turkey P2Y1 receptor (Boyer et al., 1996b). Certain 2-thioether derivatives of AMP derivatives also activate the P2Y12 receptor in C6 glioma cells in the micromolar range. The -thio modification of AMP analogs (e.g., 10) increases potency at the P2Y1 receptor (Fischer et al., 1999). Such monophosphate derivatives have also been reported to inhibit ecto-nucleotidases, which complicates their use as P2Y receptor agonists.
A BH2 moiety may serve as a substitute for an ionized oxygen atom of the -phosphate of ATP derivatives in promoting binding to the P2Y1 receptor binding site. Thus, 5'-(1-boranotriphosphate) derivatives such as the 2-methylthio derivative (11) have been found to potently activate the P2Y1 receptor (Nahum et al., 2002). Because the 1-boranotriphosphate moiety is chiral, it was possible to separate two stable isomers in this series. The more potent isomer of 11 displayed an EC50 of 2.6 nM at the rat P2Y1 receptor.
The ribose moiety of nucleotide derivatives was also modified, resulting in enhanced potency at the P2Y1 and P2Y12 receptors. Simple carbocyclic (cyclopentyl) analogs of ATP were found to enhance antagonist affinity at the P2Y12 receptor (see below) (van Giezen and Humphries, 2005). Similarly, at the P2Y1 receptor, carbocyclic and even acyclic substitutions of ribose were studied. In general, carbocyclics and ring-constrained nucleotide analogs were able to maintain agonism at the P2Y1 receptor, whereas acyclic derivatives proved to be exclusively antagonists (see below). A ring-expanded, yet nonglycosidic, dehydroanhydrohexitol analog MRS2255 (12) activated the P2Y1 receptor with an EC50 of 3.0 µM (Nandanan et al., 2000).
Among the more successful examples of the use of carbocyclic or sterically constrained carbocyclic substitution of the ribose moiety for P2Y receptor interactions are the "methanocarba" analogs (Nandanan et al., 2000; Kim et al., 2002). These analogs incorporate a conformationally fixed bicyclic ring system, consisting of fused cyclopentane and cyclopropane rings, in place of the ribose moiety. Depending on the position of fusion, the resulting nucleotides may adopt one of two conformations: (N), north, or (S), south. Correlation of ring geometry with the biological activities helped define the conformational requirements of the ribose moiety in P2Y receptor binding and led to pharmacological probes of unusual selectivity and affinity. For example, the two isomeric methanocarba equivalents of ATP indicated a strong preference (ratio of potency >100-fold) at the P2Y1 receptor for the (N)-isomer 13 over the (S)-isomer 14 (Kim et al., 2002).
Combination of this ring system with other favorable modifications of ADP or ATP resulted in large qualitative differences from the native nucleotides in receptor activation. For example, whereas ,-methylene ATP (,-meATP) is a weak partial agonist at the human P2Y1 receptor, the corresponding (N)-methanocarba-,-meATP (15) was a full agonist with an EC50 of 158 nM (Ravi et al., 2002). MRS2365 (16), the most potent known agonist of the human P2Y1 receptor, with an EC50 of 0.4 nM, induces platelet shape change without aggregation (Chhatriwala et al., 2004). In addition, the high selectivity of 16 for the P2Y1 receptor in comparison to its inactivity at P2Y12 and P2Y13 receptors was striking, in contrast to the relatively nonselective 2-MeSADP (2). Thus, the P2Y12 and P2Y13 receptors have very different conformational preferences within the ribose-binding region than does the P2Y1 receptor. At P2Y13 receptors, under optimal experimental conditions, ATP (2) and 2-MeSATP (5) are equipotent as agonists.
Figure 4 shows the structures of nucleotide-based antagonists of P2Y receptors. A successful approach to the development of potent and selective P2Y1 receptor antagonists was made possible by the observation by Boyer et al. (1996a) that naturally occurring adenosine bisphosphate derivatives such as A3P5P (17) act as partial agonists or antagonists of the receptor. Thus, the splitting and repositioning of the phosphate groups of the 5'-diphosphate group of ADP to separate ribose positions (5'- and either 3'- or 2'-) reduces efficacy at the P2Y1 receptor. Removal of the 2'-hydroxyl group and addition of the potency-enhancing N6-methyl group resulted in MRS2179 (18), and the corresponding 2-chloro analog MRS2216 (19), which became full antagonists at the P2Y1 receptor with IC50 values of 300 and 100 nM, respectively (Nandanan et al., 1999; Brown et al., 2000b). The SAR of alkyl, thioether, and other substitutions at the 2-position of bisphosphate antagonists has been explored (Nandanan et al., 1999, Raboisson et al., 2002a). Raboission et al. (2002b) synthesized a C-nucleotide bisphosphate derivative (20) that antagonized P2Y1 receptors. In addition, the adenine N9 nitrogen is not essential in P2Y1 receptor interaction, and similarly the N1 nitrogen was found to be unnecessary through the evaluation of 1-deaza analogs (Nandanan et al., 1999).
|
Upon introduction of the conformationally preferred (N)-methanocarbo ring system into this series of bisphosphate nucleotide antagonists, the P2Y1 receptor affinity was further enhanced. Thus, MRS2279 (21), the (N)-methanocarbo equivalent of the riboside (19), and the corresponding 2-iodo derivative MRS2500 (22) demonstrated high affinity in competitive antagonism at the human, turkey, rat, and mouse P2Y1 receptors (Nandanan et al., 2000; Boyer et al., 2002; Waldo et al., 2002; Cattaneo et al., 2004). The Ki value of MRS
