Molecular Structure of P2Y Receptors: Mutagenesis, Modeling, and Chemical Probes

Abstract

There are eight subtypes of P2Y receptors (P2YRs) that are activated, and in some cases inhibited, by a range of extracellular nucleotides. These nucleotides are ubiquitous, but their extracellular concentration can rise dramatically in response to hypoxia, ischemia, or mechanical stress, injury, and release through channels and from vesicles. Two subclasses of P2YRs were defined based on clustering of sequences, second messengers, and receptor sequence analysis. The numbering system for P2YR subtypes is discontinuous; i.e., P2Y_1–14Rs have been defined, but six of the intermediate-numbered cloned receptor sequences (e.g., P2y₃, P2y₅, P2y_7–10) are not functional mammalian nucleotide receptors. Of these two clusters, the P2Y_12–14 subtypes couple via Gα_i to inhibit adenylate cyclase, while the remaining subtypes couple through Gα_q to activate phospholipase C. Collectively, the P2YRs respond to both purine and pyrimidine nucleotides, in the form of 5′-mono- and dinucleotides and nucleoside-5′-diphosphosugars. In recent years, the medicinal chemistry of P2Y receptors has advanced significantly, to provide selective agonists and antagonists for many but not all of the subtypes. Ligand design has been aided by insights from structural probing using molecular modelling and mutagenesis. Currently, the molecular modelling of the receptors is effectively based on the X-ray structure of the CXCR4 receptor, which is the closest to the P2Y receptors among all the currently crystallized receptors in terms of sequence similarity. It is now a challenge to develop novel and selective P2YR ligands for disease treatment (although antagonists of the P2Y₁₂R are already widely used as antithrombotics).

The P2Y receptors (P2YRs) are a family of eight G protein-coupled receptors (GPCRs) that belong to the rhodopsin-like branch GPCRs, also known as class A or Family 1 GPCRs.^1,2 There are activated, and in some cases inhibited, by a range of naturally occurring extracellular nucleotides. These nucleotides are ubiquitous, but their extracellular concentration can rise dramatically in response to hypoxia, ischemia, or mechanical stress, injury, and release through channels and from vesicles. The P2Y family can be further divided into a subfamily of five P2Y₁, P2Y₂, P2Y₄, P2Y₆, and P2Y₁₁Rs (“P2Y₁-like”) that stimulate phospholipase C (PLC) through G_q protein and a second subfamily of P2Y₁₂, P2Y₁₃, and P2Y₁₄Rs (“P2Y₁₂-like”) that inhibit adenylate cyclase through G_i protein (Table 1).³ Other effector pathways have been documented, such as coupling of the P2Y₁₁R to G_s as well as to G_q in some cells to induce stimulation of cyclic AMP production.⁴

Table 1.

Properties P2YRs and key agonist antagonist ligands.

Group	Subtype, gene symbol	G protein	Ligand agonist/ antagonist	pEC50	Other interactions, Subtype
P2Y1-like	P2Y1, P2RY1	Gq	MRS2365 8	9.40	selective
			2-MeSADP 3	8.18	P2Y1, P2Y12
			ADPβS 6	5.62	P2Y1, P2Y12
			ADP 1	5.09	P2Y1, P2Y12
			MRS2500 31d	9.02	selective
			MRS2279 30d	8.10	selective
			MRS2179 29d	6.48	selective
	P2Y2, P2RY2	Gq, Gi	UTP 12	8.10	P2Y4
			MRS2698 20b	8.10	selective
			ATP 2	7.07	P2Y4,11 + others
			PSB1114 17	6.87	selective
			MRS2768 16	5.72	selective
			PSB-716 40	5.01
	P2Y4, P2RY4	Gq, Gi	MRS4062 21	7.64	selective
			UTP 12	6.26	P2Y4
			GTP 5	5.18	nonselective
			ATP 2a	6.15	P2Y2,11 + others
	P2Y6, P2RY6	Gq	MRS2957 19b	7.92	selective
			5-iodo-UDP 23	7.83	selective
			PSB-0474 22	7.15	selective
			MRS2578 49 [noncompetitive]	7.43	selective
	P2Y11, P2RY11	Gq, Gs	NF546 43c	6.27
			ATP 2	4.77	P2Y2,4 + others
			ATPγS 7	4.62	nonselective
			NF157 42	7.35
			NF340 44	7.14	selective
P2Y12-like	P2Y12, P2RY12	Gi	2-MeSADP 3	8.3	P2Y1,13
			ADP 1d	7.22	P2Y1,13
			PSB-0739 39	9.8	selective
			AR-C69931MX 34	9.40	P2Y11
			AZ11931285 36d	~9	selective
			AR-C67085 33bd	8.3	P2Y11
			AZD6140 37	7.90	selective
	P2Y13,b P2RY13	Gi	ADP 1	7.94	P2Y1,12
			2-MeSADP 3	7.72	P2Y1,12
			MRS2211 48	5.97	selective
	P2Y14, P2RY14	Gi	MRS2905 26	8.70	selective
			MRS2690 29	7.31	selective
			MRS2802 25	7.20	selective
			UDP 11d	6.80	P2Y6
			UDP-glucose 27	6.45	P2Y2
			PPTN 56	8.7	selective

^aAt the human P2Y₄ receptor, but ATP acts as an agonist at the rat or mouse P2Y₄ receptor.

^cNF546 activates the P2Y₁₁R, although it belongs to a structural class of antagonists.

^dUsed as a radioligand, when labelled with [³H], [³²P], [³³P] or [¹²⁵I], as appropriate.]

There are species differences in the occurrence, function and ligand preference of the P2YRs, e.g. mice lack a P2Y₁₁R gene in their genome.⁵ The structures of representative ligands used as tools for defining the role of the P2YRs are shown in Figures 1 (agonists) and 2 (antagonists). The principal native agonists for these receptors are: ADP 1 for P2Y₁, P2Y₁₂, and P2Y₁₃; ATP 2 for P2Y₂, P2Y₄ (rat, but not human) and P2Y₁₁; UDP 11 for P2Y₆ and P2Y₁₄; UTP 12 for P2Y₂ and P2Y₄; UDP-glucose 27 and other UDP-sugars for P2Y₁₄. In some cases, the same nucleotide that activates one P2YR subtype may antagonize another subtype, for example ATP and other 5′-triphosphate analogues act as antagonists at the P2Y₁₂R in platelets.⁶

Figure 1.

Nonselective and selective P2YR agonists derived from nucleotides. A. Purine derivatives. B. and C. Pyrimidine derivatives. The P2Y potencies of ligands selected from these figures are provided in Table 1.

STRUCTURAL CHARACTERIZATION OF P2Y RECEPTORS: MOLECULAR MODELING AND MUTAGENESIS

As for other G protein-coupled receptors (GPCRs), the P2YRs display the characteristic structural motif of a single polypeptide with seven α-helical transmembrane domains (TMs).³ The ligand binding site is thought to be primarily within a cleft surrounded by the TM domains. However, there is evidence that the extracellular loops (ELs) are also involved in coordinating the nucleotide ligands. This is especially applicable to EL2, which often extends into the binding cleft. On the cytosolic side, the three intracellular loops (ILs) and the C-terminus are involved to varying degrees in signalling functions within proximity of the docked G protein. None of the P2YRs are solved structurally using X-ray crystallography. However, extensive site-directed mutagenesis of residues that are predicted to coordinate the ligands,^7–18 structure activity studies of ligand analogues,^19–30 and molecular modelling^7,9.31–34 have provided insight into molecular recognition of the P2YRs.

Molecular modeling of the P2YRs and ligand docking have identified putative recognition functions for some of the conserved TM residues. Early in this progression, positively charged residues were predicted to coordinate the negatively charged phosphate moieties that are a required feature of known P2YR agonists.^1,3 The division of the P2YRs into two subfamilies (P2Y₁R-like and P2Y₁₂R-like) corresponds not only to second messenger pathways, but also to sequence alignments, phylogenetic analysis, and the identification of residues with key functions for each of the subtypes. The sequence identity between the P2Y₁R-like and P2Y₁₂R-like subfamilies is only approximately 20%. Therefore, the modelling and mode of ligand docking must be established independently for each of the subfamilies. The sequence identity within each subfamily is higher, typically in the 40–50% range.

Specific residues within the TM domains of the P2Y₁R appear to be involved in ligand recognition or receptor functionality according to mutagenesis data and can be subdivided in two categories. Positively charged residues (using the numbering convention of Ballesteros and Weinstein³⁶) are likely implicated in the coordination of the phosphate moiety of the ligands, namely R128^3.29, K280^6.55, and R310^7.39. Neutral residues, namely Y273^6.48, H277^6.52, Q307^7.36 and Ser314^7.43 were also essential or modulatory for the action of nucleotides at the P2Y₁R.³⁵ According to mutagenesis data, charged residues of the EL domains were also found to be involved in ligand recognition or receptor functionality. Specifically, in the P2Y₁R mutagenesis of D204, E209, and R287 interfered with activation by nucleotides, while in the P2Y₂R, similar effects were found for R177, R180, and R272.^12,13 Furthermore, P2Y₁R mutagenesis has demonstrated the presence of two putative disulfide bridges necessary for proper activation: one between TM3 and EL2 that is found in most class A GPCRs, and a second less common disulfide bridge formed between the N-terminus and the junction between EL3 and TM7.¹³ Beyond the P2Y₁R, the following residues within the TMs domain were found to be involved in ligand recognition for other members of the first subfamily: Y118^3.37 and Y268^6.58 for the P2Y₂R¹²; Y261^6.48, R268^6.55, and R307^7.39 for the P2Y₁₁R.¹⁰ Moreover, within the second subfamily the positively charged residues R256^6.55 and K280^7.35, and neutral residues, H253^6.48 and Y259^6.52, were found to be involved in ligand recognition for the P2Y₁₂R.¹⁴

P2YR molecular modeling, initially based on highly speculative grounds and using templates of low resolution and little sequence homology^7,31, has progressed dramatically in the past 20 years. The accuracy of the modelling appears to have improved as judged by the ability to correlate features of the docked ligands with the receptor structure. Various GPCR structures have been used as templates for P2YR homology modeling. Initially, the high resolution structure of bovine rhodopsin revealed in 2000 was used to build models,³⁷ and homology modeling has been applied to all of the P2YRs at various times.³ The first P2Y subtype to which extensive molecular modelling was applied in conjunction with site-directed mutagenesis was the P2Y₁R. The putative nucleotide binding pocket of this subtype was proposed to be located within the upper third (from the extracellular surface) of TMs 3, 6, and 7. The positively charged residues occurring in each of these TMs of the P2Y₁R, which were conserved among the P2Y₁-like receptors, clustered around the putative position of the negatively charged phosphate group of the native agonist ADP. The β₂ adrenergic receptor structure³⁸ was used as an improved modelling template for the P2YR family, but that was eventually superseded. Recently, an alternative template was shown to be more consistent with the characteristics of the P2YRs with respect to modelling.^24,34 X-ray structures of the antagonist-bound form of the human CXCR4, a peptide receptor, were reported in 2010 by Stevens and coworkers.³⁹ The sequence identity of the human P2Y₁₂R and the CXCR4 chemokine receptor is 22% overall and 26% for the TM domains. Additional structural features of CXCR4 were shared with P2YRs, making the crystal structure of CXCR4 a more suitable modeling template for both subfamilies of the P2YRs over other available GPCR crystal structures.³⁴

The impact of using different templates for P2YR modeling is illustrated with two cases: the P2Y₄R and the P2Y₁₄R (Figure 3). One of the first models of the P2Y₁₄R in complex with the natural agonist UDP-glucose is shown in Figure 3A.³ The homology model was constructed on the basis of the crystal structure of rhodopsin, which at the time was the only GPCR with a crystallographically solved structure.³⁷ Consequently, EL2 overlays and occludes the extracellular opening of the binding cavity, as in rhodopsin. This situation most likely does not reflect the actual conformation of the loop in the P2Y receptors as it would impede the access of nucleotides to the binding cavity. The greater difficulty of modeling loop regions in comparison to the TMs has been a long standing problem in GPCR modeling, and is not particular to the P2YRs.

Figure 3.

A. Previous model of the P2Y₁₄R, based on the structures of rhodopsin, in complex with the natural agonist UDP-glucose 27. Three cationic residues, shown in sticks representation with gray carbons, coordinate the phosphate moieties of the ligand, shown in space filling representation with yellow carbons. The backbone of the receptor is schematically rendered as a ribbon colored with a continuum spectrum going from blue, at the N-terminus, to dark red, at the C-terminus. B. Recent model of the P2Y₄R based on the structures of the CXCR4 receptor, in complex with a synthetic agonist, MRS4062 21. Three cationic residues, shown in sticks representation with gray carbons, coordinate the phosphate moieties of the ligand, shown in space filling representation with yellow carbons. The backbone of the receptor is schematically rendered as a ribbon colored with a continuum spectrum going from blue, at the N-terminus, to dark red, at the C-terminus.

Three cationic residues, i.e. K171^EL2, R253^6.55, and K277^7.35, coordinate the phosphate moieties of the ligands in the model. These residues are conserved as cationic residues in all the subtypes belonging to the subfamily of the G_i coupled P2YRs, which encompasses the P2Y₁₂, P2Y₁₃ and P2Y₁₄ subtypes.³ Figure 3B shows a recent model of the P2Y₄R in complex with a synthetic agonist, namely a CTP derivative substituted with a phenylalkoxyimino group at position 4 of the pyrimidine ring, MRS4062 21.²⁴ In this case, the homology model was constructed on the basis of the crystal structure of CXCR4.³⁹ Moreover, the CXCR4 receptor shares a disulfide bridge with the P2YRs to connect the N-terminus with the junction of EL3 and TM7. The EL2 is solvent exposed and leaves ample room for the nucleotides to diffuse into and out of the receptor. Also in this case, three cationic residues, i.e. R112^3.29, R265^6.55, and R292^7.39, coordinate the phosphate moieties of the ligands in the model. These residues are conserved as cationic residues in all the subtypes belonging to the P2Y₁-like subfamily of the G_q-coupled P2YRs, which encompasses the P2Y₁, P2Y₂, P2Y₄, P2Y₆ and P2Y₁₁ subtypes,³ and are thought to coordinate the phosphate moiety in each case.

CHEMICAL PROBES FOR STUDYING P2Y RECEPTORS

There are inherent difficulties in the medicinal chemistry of P2YRs. Nucleotide analogues are difficult to synthesize in reproducible yield, and purification and characterization are more complicated than for many other classes of small molecule ligands of GPCRs. Furthermore, in pharmacological studies, especially in vivo, interconversion of nucleotides by ectonucleotidases and other enzymes is a serious complication.⁴¹ Endogenous nucleotides that are produced by all cells in culture often complicate the interpretation. Efforts to introduce chemical modifications that will stabilize the compounds also may reduce their potency at the P2YRs. Despite these limitations, there are now selective agonist and antagonist ligands for many but not all of the P2YRs (Table 1), and this effort has accelerated in recent years. Selective nucleotide agonists have been reported for P2Y₁, P2Y₂, P2Y₄, P2Y₆ and P2Y₁₄ subtypes, and selective nucleotide antagonists are now known for P2Y₁ and P2Y₁₂ subtypes. Moreover, the identification of novel chemotypes to antagonize various P2Y receptors is progressing. For example, the screening of chemically diverse libraries and the use of in silico screening to greatly narrow the list of candidates have had some success in finding atypical antagonists for P2YRs.

A summary of many of the most useful agonists and antagonists at each of the P2Y subtypes is provided below (structures in Figures 1 and 2). It is now a challenge to develop novel P2YR ligands for disease treatment, in addition to the widespread use of antagonists of the P2Y₁₂R as antithrombotics⁴². This requires overcoming various complicating factors such as the inherent instability and lack of bioavailability of nucleotide derivatives. Also, in vivo results emphasize the widespread occurrence of these receptors in the body, with multiple effects, both positive and detrimental, associated with the activation of each subtype. Nevertheless, promising results suggest the possible use of such agents in endocrine, gastrointestinal, inflammatory, cardiovascular, ischemic and neurological diseases.

Figure 2.

Nonselective and selective P2YR antagonists derived either from: A. nucleotides and nucleosides, or B. nonnucleotide derivatives. The P2Y potencies of ligands selected from these figures are provided in Table 1.

P2Y₁R

C2-Alkylthio (and arylalkylthio) modifications of adenine nucleotides are often well-tolerated at the P2Y₁R. The endogenous nucleotide ADP 1 and its monosubstituted more potent derivative 2-MeSADP 3 are full agonists at the P2Y₁R. However, they are nonselective due to activation of the P2Y₁₂ and P2Y₁₃Rs.⁴³ The corresponding 5′-triphosphate derivatives, i.e. ATP 2 and 2-MeSATP 4 are reported to activate the P2Y₁R, but in some models demonstrate lower efficacy.¹

The N⁶ position of adenine nucleotides that act as P2Y₁R ligands can be substituted only with small alkyl groups, with the order of potency Me > Et ≫ Pr. N⁶-Arylalkyl analogues are inactive at the P2Y₁R. Thus, the N⁶ group occupies a small hydrophobic pocket in the binding site. The ribose and phosphate moieties have also been extensively modified in studies of P2Y₁R SAR. Thiophosphate modifications, which introduce a new chiral center if present in a nonterminal phosphate moiety, have been useful in SAR studies and increase the biological stability, e.g. ADPβS 6 is a potent P2Y₁R agonist. One result of this effort was the ability to effectively convert potent P2Y₁R agonists into potent antagonists. The seminal discovery by Boyer and Harden and colleagues that 3′,5′-bisphosphate derivatives of adenosine tend to antagonize the P2Y₁R, made the introduction of many nucleotide antagonists of this receptor possible.⁴⁴ This led to later generations of more potent antagonists of the P2Y₁R that display no residual efficacy at the receptor, such as MRS2179 29, which is used widely as a pharmacological probe.

The ribose moiety of nucleotide derivatives in nature freely changes conformation, and these conformations have been described mathematically as a “pseudorotational cycle”.⁴⁰ Conformational scrutiny of the ribose moiety has facilitated the introduction of more potent and selective ligands, both nucleotide agonists and antagonists of the P2Y₁R. Potent antagonists, i.e. 2-iodo derivative MRS2500 31 and its 2-chloro analogue MRS2279 30, contain a rigid bicyclic ribose substitution, i.e. a bicycle[3.1.0]hexane, that maintain a conformation of the ribose-like ring that is favored at the P2Y₁R.⁴⁵ Therefore, the entropic barrier for the ligand to bind to the receptor is diminished resulting in a gain in affinity. This type of modification is termed North (N)-methanocarba, because it maintains a (N)-envelope conformation. The introduction of potent antagonists of the P2Y₁R has led to exploration of such compounds as antithrombotic agents.

While introducing rigidity to the ribose-like moiety is capable of greatly enhancing the P2Y₁R potency, the application of greater flexibility has also been explored. Certain acyclic nucleotide derivatives (bisphosphates and bisphosphonates) have found application as P2Y₁R antagonists of moderate potency, but no P2Y₁R agonists were discovered in that series. An example of one such acyclic antagonist is a symmetric bisphosphate MRS 2298 32.⁴⁵

Chemical diversity was extensively explored in the search for P2Y₁R antagonists. A representative heterocyclic compound discovered through optimization of a high throughput screening hit was a substituted 1-phenyl-3-methyl pyrazol-5-one 50, was found to be an orally-bioavailable antagonist of the P2Y₁R with a K_i of 90 nM.²⁷ Other novel antagonists of the P2Y₁R have been reported.^65–68

Also in the agonist series, the North (N)-methanocarba analog of 2-MeSADP, i.e. MRS 2365 8, selectively activates the P2Y₁R with great selectivity over other ADP-preferring subtypes, i.e. P2Y₁₂ and P2Y₁₃Rs.⁴³ A further benefit of the (N)-methanocarba modification is that the 5′-phosphate esters is more stable toward the hydrolytic action of 5′-nucleotidase (CD73). Dinucleoside triphosphates and tetraphosphates tend to activate the P2Y₁R and display greater stability than the corresponding mononucleotide analogues. A means of further stabilizing the phosphate linkage is the introduction of borano analogues of the phosphate, such as in the dinucleotides that act as P2Y₁R agonists, e.g. boronate-containing diadenosine triphosphate derivative 10 introduced by Fischer and co-workers, are being explored for application to Type 2 diabetes.²⁵

P2Y₂R

UTP 4 as a native P2YR agonist does not distinguish between P2Y₂R and P2Y₄R. ATP 2 is a second endogenous agonist of the P2Y₂R, and its SAR at this receptor is much less explored than UTP. UTPγS 14 and 2-thioUTP 20a are potent agonists of the P2Y₂R. A β,γ-difluoromethylene derivative of 4-thio-UTP PSB1114 17 is ~60-fold selective for the P2Y₂R.³⁰ A 2′-amino derivative MRS 2698 20b was shown to be 300-fold selective for the P2Y₂R in comparison to the P2Y₄R.¹⁹

Dinucleoside tetraphosphates with either uracil or mixed uracil/adenine nucleobases activate the P2Y₂R and display enhanced stability in pharmacological preparations. Several such derivatives have been explored for therapeutics. For example, INS 365 (Diquafosol) 18 is used for treating dry eye disease in Japan by virtue of its action as a moderately potent P2Y₂R agonist, but it is nonselective in comparison to P2Y₄R.^46,47 A simplified analogue similar to the dinucleotide series, MRS 2768 16 (uridine tetraphosphate δ-phenyl ester) is somewhat selective for the P2Y₂R, but achieves only moderate potency.⁴⁸

A phosphonate analogue of 5-aryl UMP, SVP333 15, acts as a partial agonist of the P2Y₂R, but it is not competitive with other agonists and has therefore been described as an allosteric agonist.⁴⁹ Compound 15 displayed an EC₅₀ of 400 nM and reached a maximal efficacy of 43% of full agonist. The compound is more stable than other P2Y₂R agonists, not being subject to cleavage by nucleotidases. It is also unusual that a monophosphate equivalent interacts potently with a P2YR.

A uracil nonnucleotide derivative AR-C11892510 45 was reported in abstract form to be a P2Y₂R antagonist.¹ Suramin 38 and Reactive blue 2 41, discovered early as P2 antagonists, have been chemically modified to increase receptor subtype selectivity. One such simplified derivative that maintains potency at the P2Y₂R is PSB-716 40.⁵⁰

P2Y₄R

UTP 4 activates this receptor universally, but the action of ATP 2 at the P2Y₄R is species-dependent, i.e. acting as an antagonist at the human P2Y₄R and an agonist at the rat P2Y₄R. It is useful to note that GTP activates this receptor with only one order of magnitude less potency than UTP. A hydrophobic pocket around the uracil 4 position was identified by SAR studies, and modifications or alkyl extensions at this site alter the P2Y₂/P2Y₄R selectivity of UTP analogues.²⁴ For example, an N⁴-phenylpropyloxyimino derivative of CTP, MRS 4062 21, is a P2Y₄R agonist that ass ~30-fold selectivity in comparison to both P2Y₂R and P2Y₆R. P2Y₄R agonists might prove useful in treating chronic constipation by inducing chloride and water secretion in the intestinal epithelia.⁶⁴

P2Y₆R

UDP 3 was previously thought to be a native agonist exclusively of the P2Y₆R, but it also activates the P2Y₁₄R at similar concentrations.⁵¹ Modifications of UDP that favor the P2Y₆R have resulted in UDPβS 13, 3-phenacyl UDP (PSB 0474) 22, 5-iodo-UDP (MRS 2693) 23 and 5-methoxy-UDP 24.^23,29,54 Furthermore, uracil dinucleoside triphosphates, such as Up₃U 19a and the more potent analogue MRS2957 19b, have been used as moderately selective P2Y₆R agonists.²³ P2Y₆R agonists such as MRS2957 both promote insulin release and prevent TNF-induced apoptosis in pancreatic islets cells in culture, which might prove useful in the treatment of diabetes.⁶²

The P2Y₆R-preferred conformation of the ribose ring is the opposite twist from that preferred by the P2Y₁R. Initially, the locked South (S)-conformation was predicted to be favored by Monte Carlo computer simulation of docking of various nucleotide conformations to the P2Y₆R, and this prediction was later confirmed by chemical synthesis and pharmacological comparison with UDP.⁴⁰ In fact, the methanocarba-UDP isomer that maintains a South (S) conformation 25 is roughly twice as potent as UDP, and the corresponding (N)-methanocarba analogue of UDP (not shown) is inactive.

Few antagonists of the P2Y₆R have been reported. MRS 2578 49 is a non-competitive P2Y₆R antagonist that is predicted to react irreversibly with the receptor because of the essential two isothiocyanate groups.⁵² However, it has disadvantages as a pharmacological probe, such as limited aqueous stability and low solubility.

P2Y₁₁R

ATPγS 7 is among the most potent P2Y₁₁R agonists, but selective nucleotide activators of this receptor have not been reported. However, a derivative of suramin, NF546 43, was found to activate the P2Y₁₁R selectively.⁵³ Curiously, other suramin derivatives such as NF 157 42, antagonize the P2Y₁₁R, but not in a fully selective fashion. The suramin antagonist derivative NF340 44 appears to be selective for the P2Y₁₁R.

Some potent adenine nucleotide antagonists of the P2Y₁₂R, e.g. 5′-triphosphates, were found to activate the P2Y₁₁R, and therefore should be used with caution in pharmacological experiments. For example, 2-propylthio-β,γ-dichloromethylene-ATP (AR-C 67085 33b) activates the P2Y₁₁R with an EC₅₀ value of 1.5 μM.

P2Y₁₂R

ADP 1 is the native agonist of the P2Y₁₂R, and 2-MeSADP 3 is more potent at this subtype. 5′-Triphosphates such as AR-C 69931MX 34 (Cangrelor), act as competitive antagonists of the P2Y₁₂R and consequently display potent antithrombotic activity.⁵⁵ Many novel nucleotide and nonnucleotide P2Y₁₂R antagonists have been explored.^56–60 Several thienopyridine derivatives that are widely used antithrombotic agents act as prodrugs of potent and irreversibly binding P2Y₁₂R antagonists, i.e. Clopidgrel 53a and Prasugrel 54a and their active metabolites 53b and 54b, respectively. AZD 6140 37 (Ticagrelor) is a nucleoside derivative that acts as a potent, reversible P2Y₁₂R antagonist and is now in clinical use. PSB-0739 39, is a potent nonnucleotide P2Y₁₂R antagonist derived from Reactive blue 2 38.¹⁵

P2Y₁₃R

The native agonist of the P2Y₁₃R is ADP 3, and unlike at P2Y₁ and P2Y₁₂Rs, 2-MeSADP 3 is less active. ATP 2 is also much weaker and less efficacious. P2Y₁₃R agonists have been suggested for the treatment of dyslipidemia and atherosclerosis.⁶¹ Various pyridoxal phosphates were screened for antagonist potency at the human P2Y₁₃R, and MRS 2211 48 proved to be more potent and selective than the parent antagonists PPADS 46 and iso-PPADS 47.

P2Y₁₄R

UDP-glucose 27 and UDP 13 are native agonists of the P2Y₁₄R, and other UDP-sugars also activate the receptor but at higher concentrations. MRS 2690 28, the 2-thio derivative of UDP-glucose, is a potent and selective agonist at the P2Y₁₄R. UDP-glucose but not MRS2690 can activate the P2Y₂R. The most potent agonist of the P2Y₁₄R yet reported, α,β-methylene-2-thio analogue MRS2905 26, and its related analogue α,β-Difluoromethylene-UDP, MRS 2802 25, are highly selective agonists of the P2Y₁₄R in comparison to the P2Y₆R. The P2Y₁₄R facilitates release of mediators from mast cells, and thus might be a target for asthma.⁶³ Several classes of P2Y₁₄R antagonists of nanomolar affinity, including PPTN 56, were recently reported.⁶⁴ PPTN was shown to decrease the release from human LAD2 mast cells that was facilitated by a P2Y₁₄R agonist.⁶⁹

Conclusion

The development of selective ligands of the two subclasses of P2YRs has progressed in recent years to provide selective agonists and antagonists for many but not all of the subtypes. Ligand design has been aided by insights from structural probing using molecular modelling and mutagenesis. The current molecular modelling is often based on the X-ray structure of the CXCR4 receptor, which is the closest to the P2Y receptors among all the currently crystallized receptors in terms of sequence similarity and key structural features. It is now a challenge to develop novel and selective P2YR ligands for disease treatment (although antagonists of the P2Y₁₂R are already widely used as antithrombotics).

There have been significant recent advances in the structural probing of P2YRs and in the medicinal chemistry of selective ligands and their pharmacology. Potent purine and pyrimidine nucleotide analogues that selectively interact with P2YRs have aided in the characterization of regulation of many physiological and pathophysiological processes. It is apparent that this ubiquitous cell signaling system has implications for understanding and treating many diseases. Thus, this field has provided fertile ground for pharmaceutical development.