Structure-Guided Design of A₃ Adenosine Receptor-Selective Nucleosides: Combination of 2-Arylethynyl and Bicyclo[3.1.0]hexane Substitutions

Abstract

(N)-Methanocarba adenosine 5′-methyluronamides containing known A₃ AR (adenosine receptor)-enhancing modifications, i.e. 2-(arylethynyl)adenine and N⁶-methyl or N⁶-(3-substituted-benzyl), were nanomolar full agonists of human (h) A₃AR and highly selective (K_i ~0.6 nM, N⁶-methyl 2-(halophenylethynyl) analogues 13, 14). Combined 2-arylethynyl-N⁶-3-chlorobenzyl substitutions preserved A₃AR affinity/selectivity in the (N)-methanocarba series (e.g. 3,4-difluoro full agonist MRS5698 31, K_i 3 nM, human and mouse A₃) better than for ribosides. Polyaromatic 2-ethynyl N⁶-3-chlorobenzyl analogues, such as potent linearly extended 2-p-biphenylethynyl MRS5679 34 (K_i hA₃ 3.1 nM; A₁, A_2A: inactive) and fluorescent 1-pyrene adduct MRS5704 35 (K_i hA₃ 68.3 nM) were conformationally rigid; receptor docking identified a large, mainly hydrophobic binding region. The vicinity of receptor-bound C2 groups was probed by homology modeling based on recent X-ray structure of an agonist-bound A_2AAR, with a predicted helical rearrangement requiring an agonist-specific outward displacement of TM2 resembling opsin. Thus, X-ray structure of related A_2AAR is useful in guiding design of new A₃AR agonists.

Introduction

The structure activity relationship (SAR) of nucleoside agonists at the A₃ adenosine receptor (AR), a G protein-coupled receptor (GPCR) of the rhodopsin-like family A, has been extensively explored.¹ Several A₃AR-selective agonists are currently in advanced clinical trials for the treatment of hepatocellular carcinoma, autoimmune inflammatory diseases,^2–4 such as rheumatoid arthritis, psoriasis, and dry eye disease, and other conditions.^5–8 Other potential applications of such agonists could be osteoarthritis, Crohn’s disease, ischemia, and other inflammatory disorders. As applied to cancer models, there is evidence that multiple modes of benefit from A₃AR agonists are obtained: myeloprotection and reduction of neuropathic pain, in addition to the antiproliferative effects on tumors in vivo. A₃AR-selective antagonists are also of interest for therapeutic applications in asthma, glaucoma, septic shock and other conditions.^9–12 Thus, this subtype of AR (others are A₁AR, A_2AAR, and A_2BAR) has provided numerous opportunities for translation to therapeutics.

In order to achieve selectivity of adenosine derivatives for the A₃AR, modifications are typically introduced on the N⁶ position (often m-substituted benzyl groups) and on the ribose moiety (often 5′-N-methyluronamide), such as occur on prototypical agonists 1 and 2 (Chart 1).¹³ Replacement of the conformationally flexible ribose tetrahydrofuryl group with a rigid bicyclo[3.1.0]hexane (methanocarba) ring system in an isomeric form that enforces a North (N) envelope conformation increases both the binding affinity and selectivity of adenosine derivatives as A₃AR agonists.^14–16 For example, analogues that combined N⁶-(3-halobenzyl) and (N)-methanocarba modifications, such as 3 and 4, display anti-inflammatory activity, protection in a model of lung injury, and protection against chemotherapy-induced neuropathic pain.^17–19 Cristalli and coworkers have explored the SAR leading to enhancement of A_2AAR and A₃AR affinity by alkyn-2-yl groups, such as hexyn-2-yl, at the adenine C2 position of adenine 9-ribosides.²⁰ The enhancement of human (h) A₃AR affinity by such groups was also observed in the (N)-methanocarba series.²¹ Because ligand affinity at the A₃AR is often dependent on species, it was particularly important to establish the effect in species other than human. The combination of N⁶-(3-halobenzyl), 2-alkyn-2-yl, and 5′-N-methyluronamide modifications was found to be particularly well suited for application to agonists displaying A₃AR selectivity across species.

Chart 1.

Derivatives of adenosine as A₃AR-selective agonists in the ribose (1, 2) and (N)-methanocarba (3, 4) series.

There is a precedent for the lack of compatibility in AR recognition of multiple modifications of nucleoside derivatives at the N⁶ (i.e. 3-halobenzyl) and C2 positions (i.e. aryl or large hydrophobic groups), each of which may promote affinity at A₃AR and other subtypes.^13,22 There also is indication that N⁶-3-(substituted benzyl) and 2-arylethynyl (introduced by Cristalli and colleagues²³) modifications are not additive in their effect on A₃AR selectivity of 9-ribosides.²⁴ The present study establishes that these two substitutions are fully additive and conducive to high selectivity when applied to (N)-methanocarba adenosine derivatives. We have also used molecular modeling based on a recent X-ray structure of an A_2AAR complex with a 5′,2,N⁶ trisubstituted agonist^25.26 and with insights from other GPCRs²⁶ to probe the vicinity of the 2-arylethynyl groups when docked to the A₃AR. Thus, the crystallographic structure of one AR subtype is used to guide the design of new analogues at another AR subtype following necessary customization for this ligand set.

Results

Chemical synthesis

We have explored 2-arylethynyl substitution of the adenine ring of previously reported (N)-methanocarba 5′-N-methyluronamide nucleoside derivatives that act as selective A₃AR agonists.^14,15,21 Two series, depending on the N⁶ substitution (N⁶-methyl: 8 – 26 and N⁶-3-chlorobenzyl: 27 – 36), that are closely patterned on known potent agonists are included (Scheme 1). N⁶-Methyladenosines are typically significantly more potent at the hA₃AR than at rat (r) A₃AR, while N⁶-(3-halobenzyl)adenosines tend to display greater species-independent selectivity.

Scheme 1.

Synthesis of (N)-methanocarba nucleoside analogues. R₁ = Me or 3-Cl-benzyl. Reagents: (i) MeNH₂ or 3-Cl-benzylamine, Et₃N, MeOH, rt; (ii) 40% MeNH₂, MeOH, rt; (iii) HC≡CAr, Pd(PPh₃)₂Cl₂, CuI, Et₃N, DMF, rt; (iv) 10% TFA, MeOH, 70°C; (v) 1-ethynylpyrene or 4-ethynylpyrene 51, Pd(PPh₃)₂Cl₂, CuI, Et₃N, DMF, rt; (vi) HCl, 1 N, dioxane, 60°C.

The synthetic route used to prepare (N)-methanocarba 5′-N-methyluronamido derivatives containing a 2-arylethynyl group involved a key Sonogashira reaction²⁷ at a 2-iodoadenine moiety of intermediate 43 or 44 (Scheme 1). L-Ribose 38 was converted as previously reported into the 2′,3′-protected intermediate 39 containing a 5-ethyl ester, which was then subjected to a Mitsunobu condensation with 2-iodo-6-chloropurine to give 40.^21,28 A N⁶-methyl or N⁶-(3-chlorobenzyl) group was added by nucleophilic substitution of 6-chloro at room temperature to provide intermediates 41 and 42, respectively, followed by aminolysis of the ester at room temperature leading to 5′-N-methyluronamides 43 and 44. A Sonogashira reaction was then carried out with a variety of commercially available arylacetylenes to give protected nucleosides 45 – 46 (N⁶-methyl) and 47a - f (N⁶-3-chlorobenzyl). Finally, acid hydrolysis of the isopropylidene protecting group provided N⁶-methyl 9 – 26 and N⁶-3-chlorobenzyl 27 – 36 nucleosides. The associated synthesis of a polyaromatic ethynyl intermediate 51 is shown in Scheme 2.

Scheme 2.

Synthesis of an intermediate arylalkynyl derivative. (i) HC≡C-TMS, Pd(PPh₃)₂Cl₂, CuI, Et₃N, DMF, rt; (ii) TBAF, THF.

Pharmacological activity

Radioligand binding assays at three hAR subtypes were carried out using standard ³H (52, 53) and ¹²⁵I-labeled (54) nucleosides (Table 1).^29–31 The membrane preparations were obtained from Chinese hamster ovary (CHO) cells (A₁ and A₃) or human embryonic kidney (HEK293) cells (A_2A) stably expressing a hAR subtype.^28,29,32,33 A₃AR binding curves generally showed a Hill coefficient of ~1. The nucleoside analogues were not screened for activity at the hA_2BAR, because other (N)-methanocarba nucleosides were previously noted to be very weak or inactive at that subtype.¹⁵ For example, in cyclic AMP assays, agonist 3 displayed 30,000-fold selectivity for the hA₃AR in comparison to the hA_2BAR.¹⁷ Some previously reported 2-chloro and 2-alkynyl (N)-methanocarba agonists (3 – 8) were used for comparison in the biological assays.²⁰ For exploring species differences in AR ligand recognition, binding assays were also performed on selected nucleoside derivatives at three mouse (m) ARs (Table 2) expressed in HEK293 cells.^21,33b

Table 1.

Binding affinity of a series of (N)-methanocarba adenosine derivatives at three subtypes of hARs and the functional efficacy at the hA₃AR.

Compd	Structure		Affinity (Ki, nM) or % inhibitiona			%Efficacyb
	R1 or R3	R2	hA1	hA2A	hA3	hA3
3c	Cl	3-Cl-Bn	260±60	2300±100	0.29±0.04	103±7
4c,d	Cl	3-I-Bn	136±22	784±97	1.5±0.2	100
5c	H	3-I-Bn	700±270	6200±100	2.4±0.5	100
6d	C≡CH	3-Cl-Bn	174±23	(48%)	1.30±0.38	ND
7d	C≡C(CH2)2CH3	3-Cl-Bn	1040±83	(80%)	0.82±0.20	ND
8c,d	Cl	CH3	2100±1700	(6%)	2.2±0.6	ND
9		CH3	(13%±6%)	(14%±7%)	0.85±0.22	89.3±7.7
10		CH3	(11%±4%)	(13%±4%)	1.01±0.36	86.8±9.2
11		CH3	(21%±4%)	(17%±2%)	0.97±0.38	97.7±9.1
12		CH3	(12%±2%)	(10%±5%)	0.97±0.24	95.8±6.7
13		CH3	(21%±11%)	(19%±3%)	0.53±0.09	80.3±5.8
14		CH3	(27%±7%)	(30%±5%)	0.58±0.04	84.2±6.2
15		CH3	(10%±2%)	1270±300	1.60±0.60	90.9±1.9
16		CH3	(14%±1%)	(30%±1%)	1.22±0.31	97.4±9.1
17		CH3	(13%±7%)	(26%±1%)	0.91±0.06	97.5±12.3
18		CH3	(10%±5%)	(19%±14%)	1.07±0.14	109±4.1
19		CH3	(6%±3%)	(6%±6%)	1.65±0.08	108±1.9
20		CH3	(8%±3%)	(47%±4%)	1.66±0.36	99.6±3.3
21		CH3	(13%±3%)	(38%±5%)	3.78±1.16	110±4.9
22		CH3	(23%±5%)	(7%±5%)	10.1±1.9	78.4±6.5
23		CH3	(15%±7%)	5300±600	2.57±0.78	87.4±9.9
24		CH3	(21%±6%)	(29%±8%)	3.10±1.26	110±2.9
25		CH3	(25%±1%)	(34%±8%)	1.67±0.18	94.6±4.4
26		CH3	(15%±5%)	(52%±1%)	3.48±1.36	108±3.5
27		3-Cl-Bn	(20%±3%)	(27%±3%)	1.34±0.30	101±5.9
28		3-Cl-Bn	(20%±6%)	(42%±2%)	2.16±0.34	102±1.4
29		3-Cl-Bn	(19%±2%)	(52%±12%)	1.92±0.57	103±1.5
30		3-Cl-Bn	(4%±4%)	1740±590	4.45±1.39	91.5±11.4
31		3-Cl-Bn	(6%±4%)	(41%±10%)	3.49±1.84	95.7±6.4
32		3-Cl-Bn	1520±300	(44%±4%)	2.27±0.70	76.6±13.1
33		3-Cl-Bn	(6%±5%)	(38%±5%)	6.75±2.78	86.7±5.4
34		3-Cl-Bn	(2%±2%)	(0%±0%)	3.06±1.35	89.0±4.5
35		3-Cl-Bn	(8%±2%)	3110±530	68.3±12.5	77.8±11.6
36		3-Cl-Bn	(11±5%)	(4±3%)	660±170	97.1±3.3

^aAll experiments were done on CHO or HEK293 (A_2A only) cells stably expressing one of three subtypes of the four hARs. The binding affinity for A₁, A_2A and A₃ARs was expressed as K_i values (n = 3–5) and was determined by using agonist radioligands ([³H]N⁶-R-phenylisopropyladenosine 52 (R-PIA), [³H]2-[p-(2-carboxyethyl)phenyl-ethylamino]-5′-N-ethylcarboxamidoadenosine 53 (CGS21680), or [¹²⁵I]N⁶-(4-amino-3-iodobenzyl)adenosine-5′-N-methyl-uronamide 54 (I-AB-MECA), respectively), unless noted.^29–31 A percent in parentheses refers to inhibition of radioligand binding at 10 μM (n = 3).

^bUnless noted, the efficacy at the hA₃AR was determined by inhibition of forskolin-stimulated cAMP production in AR-transfected CHO cells.^32–34 At a concentration of 10 μM, in comparison to the maximal effect of 5′-N-ethylcarboxamidoadenosine 48 (=100%) at 10 μM. Data are expressed as mean±standard error (n = 3).

^cValues from Lee et al.; Tchilibon et al.^15,16

^dValues from Melman et al.²¹

ND, not determined.

Table 2.

Binding affinity of a series of (N)-methanocarba adenosine derivatives at three subtypes of mARs.

Compd	Affinity (Ki, nM) or % inhibitiona
	mA1	mA2A	mA3
3b	15.3±5.8	10,400±1,700	1.49±0.46
4b	7.32±1.5	5,350±860	0.80±0.14
6b	45.6±7.9	(41%)i	0.85±0.08
7b	1390±430	(42%)i	6.06±1.21
8b	55.3±6.0	20,400±3,200	49.0±3.9
13	(29±2%)	(0%)	37.7±1.1
14	(55±5%)	(2±1%)	37.2±2.0
27	(50±5%)	(2±1%)	1.23±0.14
28	(65±3%)	(7±2%)	2.38±0.04
29	(51±12%)	(19±3%)	2.64±0.22
30	(35±3%)	(55%)	2.39±0.38
31	(14±3%)	(27±2%)	3.08±0.23
32	261±19	(5±2%)	0.82±0.06
33	(18±3%)	(22%)	3.66±0.25
34	(41±6%)	(6±1%)	10.8±0.93
35	(8±2%)	(64%)	47.6±4.6

^aCompetition radioligand binding assays using [¹²⁵I]54 (A₁ and A₃ARs) and [³H]53 (A_2AAR) were conducted with membranes prepared from HEK293 cells expressing recombinant mA₁, A_2A, or A₃ARs. The data (n = 3–4) are expressed as K_i values. A percent in parentheses refers to inhibition of radioligand binding at 10 μM.

^bValues from Melman et al.²¹

The simplest N⁶-methyl 2-phenylethynyl analogue 9 displayed a subnanomolar K_i value at the hA₃AR and was nearly inactive at the hA₁AR and hA_2AAR, with <20% inhibition of binding at 10 μM. Therefore, the degree of A₃AR selectivity of 9 was estimated to be >10,000-fold. The effects of phenyl modification of the C2-arylethynyl group at the C2 position were explored initially in the N⁶-methyl 5′-N-methyluronamide series. These derivatives displayed great freedom of substitution of the arylethynyl moiety, i.e. aza substitution of a CH of phenyl (10), monohalo (11 – 17) and dihalo (19, 20) or other (18, 21 – 23) substituent groups on the phenyl ring, and the presence of additional aryl rings (24 – 26). The 3,4-difluorophenyl analogue 19 was particularly A₃AR-selective with insignificant inhibition at A₁ and A_2AARs. Planar polyaromatic groups, such as α-naphthyl 25 and the larger phenanthrene 26, in the N⁶-methyl series did not interfere with the binding to the hA₃AR. The presence of a branched p-t-Bu group in 22 reduced A₃AR binding affinity by 12-fold compared to the H analogue 9. The most potent A₃AR ligands in the N⁶-methyl series were halo-substituted compounds 13 and 14 with K_i values of 0.5 – 0.6 nM. A p-acetyl substitution in 23 was lower in hA₃AR affinity and higher in hA_2AAR affinity than most other ring substitutions, unlike in the riboside series of Cristalli and coworkers,^23a in which a 2(-p-acetylphenylethynyl) group favored A₃AR affinity and selectivity.

The variability of binding affinity upon substitution of the arylethynyl moiety was greater at the hA_2AAR than at the hA₁AR. For example, the hA_2AAR affinity increased substantially to a K_i value of 1.3 μM upon replacement of 3-F 12 with 3-Cl in 15. In contrast, the same change slightly decreased hA₃AR affinity. At the hA₁AR, only 10 to 30% of binding inhibition was typically seen at 10 μM for a variety of substitutions.

The SAR was extended by the analysis of eleven analogues prepared in the N⁶-(3-chlorobenzyl) series (27 – 36). The effect in 27 of adding a phenyl group to the ethynyl substituent of the simpler acetylene derivative 6 was complete retention of the affinity at hA₃AR and a reduction of the hA₁AR affinity from K_i = 174 nM to >10 μM, thus providing high selectivity. Furthermore, with only marginal binding of 27 at the hA_2AAR, the hA₃AR selectivity was roughly 10,000-fold in comparison to both hA₁AR and hA_2AAR. For comparison, by adding a n-propyl chain in 7 rather than a phenyl ring, the hA₁AR affinity was reduced only 6-fold in comparison to 6, and the hA_2AAR affinity appeared to be increased. Thus, a flat aryl ring rather than a flexible alkyl chain provides the appropriate geometry for hA₃AR selectivity. A potent A₃AR ligand 34 in the N⁶-(3-chlorobenzyl) series containing a rigid, linearly extended 2-p-biphenylethynyl group served as a model ligand for receptor docking due to its steric and conformational constraints. It displayed a K_i value at the hA₃AR of 3.06 nM and exceptionally high selectivity in comparison to the hA₁ and hA_2AARs (no inhibition observed). The hA₃AR affinities of 34 and the corresponding N⁶-methyl analogue 24 were equal. Thus, the combination of 2-arylethynyl, 5′-N-methyluronamide, and N⁶-3-chlorobenzyl substitution preserved the hA₃AR selectivity of (N)-methanocarba nucleosides, even for large 2-arylethynyl moieties.

Species differences in the binding affinity of the nucleoside derivatives were explored. Nanomolar A₃AR affinity was maintained in a murine species only for the N⁶-(substituted benzyl) series (Figure 1). The N⁶-methyl-2-arylethynyl derivatives 13 and 14 were determined to be ~70-fold weaker at the mA₃AR than at the hA₃AR, consistent with the previously reported reduced affinity at mA₃AR of N⁶-methyl-2-Cl derivative 8.²¹ The mA₃AR/mA₁AR selectivity in the N⁶-methyl series was enhanced with the elongation and rigidification of the C2 substituent. However, with N⁶-3-chlorobenzyl substitution, elongation at C2 did not significantly reduce affinity at the mA₃AR compared to hA₃AR (except for a 3-fold reduction for the elongated biaryl derivative 34). In the N⁶-3-chlorobenzyl series, particularly high selectivity compared to mA₁AR and mA_2AAR was generally present. There was greater variability in the degree of binding inhibition at mA₁AR than at mA_2AAR. Elongation of a 2-ethynyl group in 6 with a straight alkyl chain in 7 reduced mA₃AR affinity by 7-fold, but elongation with a phenyl ring in 27 only slightly reduced mA₃AR affinity. The most potent derivative at the mA₃AR was the p-aminophenyl analogue 32 with a K_i value of 0.82 nM. Curiously, although this derivative remained selective, its A₁AR affinity was enhanced over other members of the series at both in mouse (K_i 261 nM) and human (K_i 1520 nM).

Figure 1.

Inhibition of binding of the radioligand [¹²⁵I]54 (0.3 nM) at the mA₃AR by compounds 13, 31, 32 and 34. K_i values are found in Table 2.

Functional data were determined in an assay consisting of hA₃AR-induced inhibition of the production of adenosine 3′,5′-cyclic phosphate (cAMP) in membranes of CHO cells expressing the hA₃AR (Table 1).³⁴ Inhibition by 10 μM NECA (48, 5′-N-ethylcarboxamidoadenosine) was set at 100% relative efficacy. The novel (N)-methanocarba 5′-N-methyluronamide derivatives at 10 μM were predominantly full agonists at the A₃AR, with a few analogues showing relative efficacy of 80% or less (p-fluoro 13, t-butyl-phenyl 22, and 1-pyrene 35 derivatives). A full concentration response curve in hA₃AR-mediated inhibition of adenylate cyclase for 3,4-difluoro analogue 31 provided an EC₅₀ value of 1.2±0.7 nM (Figure 2), i.e. slightly more potent than its binding affinity.

Figure 2.

Functional agonism by the tested in an assay of adenylate cyclase in membranes of CHO cells expressing the hA₃AR. Activity of the 2-(3,4-difluorophenylethynyl)-N⁶-3-chlorobenzyl-5′-N-methyluronamide-(N)-methanocarba analogue 31 (EC₅₀ value 1.2±0.7 nM). The full agonist 5′-N-ethyluronamidoadenosine 48 was included for comparison (representing 100% efficacy). The experiment was repeated three times, and the average is shown.

Molecular modeling

A homology model of the hA₃AR based on an agonist-bound hA_2AAR X-ray structure (PDB code 3QAK)²⁵ was used to study the putative interactions between the C2-substituted agonist 34 and the A₃AR. The binding mode of 34 obtained after Induced Fit docking (IFD)³⁵ revealed the following binding interactions of the ligand (Figure 3). The key residues embedding the adenosine moiety of the agonists in the receptor binding site are mostly conserved among AR subtypes. The 3′- and 2′-hydroxyl groups were located in proximity to Ser271 (7.42) and His272 (7.43), respectively, and could form H-bonds with these residues. The NH group of the 5′-N-methylcarboxamido moiety was involved in H-bonding with the side chain hydroxyl group of Thr94 (3.36). Both the 6-amino group and the N7 atom of the adenine ring H-bonds with Asn250 (6.55). A π–π interaction was observed between the adenine ring of 34 and Phe168 (EL2), while the side chains of Leu246 (6.51) and Ile268 (7.39) offered CH-π interactions to the adenine moiety of 34. The ligand-receptor interactions observed in the present model were in good agreement with the data of site-directed mutagenesis and with our previously published models of ARs, including the studies of AR agonists docked to the A_2AAR crystal structure.^36–38 In the model obtained, the 3-chlorobenzyl ring of docked 34 was located in the hydrophobic pocket formed by Val169 (EL2), Met174 (5.35), Met172 (5.33), Ile253 (6.58), and Leu264 (7.35). More precisely, the N⁶-3-chlorobenzyl substituent of 34 was locked in the hydrophobic pocket of the A₃AR among TM5, TM6, TM7, and EL2 by CH-π interactions with the side chains of Val169 (EL2) and Ile253 (6.58). Ile253 (6.58) is not conserved among the ARs: in A₁, A_2A and A_2B the residue at position 6.58 is a smaller threonine. Moreover, favorable interactions between the backbone group of Met172 (5.33) and the chloro atom stabilized the compound in the binding site.

Figure 3.

The binding mode (two different views with A. wire helices and B. tube helices) of linearly extended N⁶-3-chlorobenzyl analogue 34 obtained after IFD to the A₃AR homology model. The template was built based on the crystal structure of an agonist-bound A_2AAR (purple helices)²⁵ and adjusted for movement of TM2 based on other templates (β₂-adrenergic receptor, brown helices; opsin, green helices).^40,41 The most successful template for TM2 was the structure of opsin.

The arylethynyl substituent at the C2 position of 34 was oriented toward the extracellular part of the A₃AR in close proximity to TM2. A_2AAR is characterized by four constraining disulfide bridges in the extracellular domains, unlike the A₃AR, which has only one disulfide bond in that region. Thus, we expected the A₃AR to be more subject to reorganization of the TMs. The disulfide bond between Cys77 (3.25) and Cys166 (EL2) is conserved not only among the AR subtypes but also among the family A GPCRs, and it is crucial for the expression and the function of the receptors. This disulfide bond involving Cys166 holds in place the backbone of two neighboring EL2 residues that are important in ligand coordination, i.e. Gln167 and Phe168. Two other disulfide bridges of A_2AAR are between EL1 and EL2, namely between Cys71 and Cys159 and between Cys74 and Cys146. Another disulfide bond involves the two cysteines in EL3, Cys259 and Cys262. It has been speculated that the presence within the extracellular domains of the three disulfide bridges unique to A_2AAR is not crucial for the tertiary structure and the stability of the receptor, but is indeed important for ligand recognition.³⁹ From the structural point of view, from a comparison between the A_2AAR structure and other GPCR structures, the presence of the two disulfide bridges between EL1 and EL2 forced the extracellular terminal of TM2 toward the TM bundle, thus reducing the size of the pocket embedding the agonist C2-substituents.

The IFD procedure applied to these rigid, extended A₃AR agonists at the A_2AAR-based model of A₃AR was unable to accommodate the long and straight arylethynyl substituent at the C2 position due to a steric clash with the TM2 residues. In particular, the phenyl ring of the N⁶-methyl 2-phenylethynyl derivative 9, a high affinity agonist at A₃AR, was clearly clashing with Ser73 in TM2, suggesting that the orientations of the extracellular terminal of TM2 and the EL1 in the A_2AAR-based model of A₃AR were not optimal for the accessibility of these methanocarba derivatives to the A₃AR binding site.

New hybrid models of the A₃AR (Figure 3) were built using different templates for the homology modeling of the extracellular part of TM2, namely an agonist-bound human β₂ adrenergic receptor crystallographic structure (designated A₃AR-β₂adr)⁴⁰ and the structure of the opsin in the activated state (designated A₃AR-ops).⁴¹ In the hybrid A₃AR model based on both the A_2AAR and the β₂-adrenergic receptor structures, the extracellular extremity of TM2 moved outward by about 4 Å at the Cα atom of Ser73. This movement created a larger pocket for the C2 arylethynyl substituents of the A₃AR agonists, thereby allowing derivatives such as agonist 9 to dock in the binding cavity without steric clashes with TM2. Nevertheless, the docking of the N⁶-methyl 2-biphenylethynyl agonist 24 to the binding site of the A₃AR-β₂adr model did not produce any reasonable pose due to the steric clash between the distal phenyl ring of the compound and the residues in TM2. The unfit docked pose of 24 to the A₃AR-β₂adr model of A₃AR suggested that a larger pocket for the longer C2-substituted agonists was needed.

The outward movement of the extracellular terminal of TM2 in the hybrid A₃AR-ops model was by approximately 7 Å at the Cα of Ser73, with the creation of a larger pocket for the accessibility of C2 substituents of the present rigid chain-extended A₃AR agonists, such as compounds 24 and 34. The docking poses of 24 and 34 in the A₃AR-ops model showed the biphenylethynyl moiety of the C2 chain pointing toward the extracellular environment. The dihedral angle between the two phenyl rings was approximately 30–40°, making the C2 chain non-coplanar with the ethynyl moiety. The biphenylethynyl side chain was stabilized by favorable hydrophobic interactions with residues in TM2, EL2, and TM7, namely Ile268 (7.39), Tyr265 (7.36), Val72 (2.64), Leu264 (7.35), Phe168 (EL2), and the carbon chain of Gln167 (EL2). The side chain of Gln167, after the IFD optimization, pointed away from the binding cavity, opening the cavity to the agonists. The docking pose of 34 showed the distal phenyl ring of the C2 chain between the carbon chain of Gln167 and the aromatic ring of Tyr265. Tyr265 is in the wall of the binding pocket, but π interactions with the ligand were not evident. An analogue 22 that deviated from planarity because of the branched t-Bu group was significantly less potent in hA₃AR binding. The presence of a halogen atom at the o-, m- or p-positions of the phenyl ring in the C2 chain was studied in the binding sites of A₃AR models and the A_2AAR crystal structure in order to understand the gain of affinity of compounds 15 and 30 at the A_2AAR. The m-Cl atoms of the C2 side chain of 15 and 30 in the predicted orientations in the A₃AR-ops model were located between Tyr271 and Gln167 (EL2), locked in the pocket by interactions between the halogen atom of the compounds and the OH group of Tyr271 and the side chain CO group of Gln167. The correspondent residue of Gln167 in human A_2AAR is the hydrophobic and bulky Leu167. Nevertheless, in the docked poses of compounds 15 and 30 in the binding site of A_2AAR, the m-Cl atom was able to generate strong interactions with the hydroxyl groups and the aromatic moieties of Tyr271 (7.35) and Tyr9 (1.35) of A_2AAR. Instead, the orientation of the o-Cl atom of the docked agonists 14 and 29 was not optimal to create interactions with the side chain of Tyr271 (7.35) of A_2AAR. On the other hand, in the docked complexes of 14 and 29 with the A₃AR-ops model, the o-Cl atom was located between Tyr265 (7.35) and Val169 (EL2), creating favorable interactions with the OH group of Tyr265 and the backbone NH group of Val169. The residue corresponding to Val169 in the A_2AAR is Glu169, of which the side chain was oriented such that the γ-carboxyl group interacted with the backbone NH group, i.e. no longer available to interact with the o-Cl atom of 14 or 29.

Discussion

In previous studies,¹³ the SAR in AR binding was studied for two series of ribonucleosides that are particularly suited for selectivity at the hA₃AR: N⁶-methyl and N⁶-3-halobenzyl. N⁶-3-Halobenzyl substitution is greatly favored for maintaining selectivity of ribosides at the A₃AR in mouse and rat. Upon addition of the (N)-methanocarba modification,²¹ a rise in affinity was observed at the mA₁AR for certain 2-chloroadenine analogues, leading to a reduction of AR selectivity in murine species. However, straight chain alkynyl groups at the C2 position, such as pentynyl 7, alleviated this problem by reducing affinity at the mA₁AR. Substitution with 2-arylethynyl groups was not examined in the (N)-methanocarba series of A₃AR agonists by Melman et al.,²¹ but we have introduced this modification in the present study.

Previously, in the ribose-5′-uronamide series, combination of 2-phenylalkynyl groups with various bulky N⁶ substituents was not additive in its effect on A₃AR affinity. In a series of A₃AR-selective N⁶-arylurea derivatives, a 2-phenylalkynyl group reduced the rA₃AR affinity in comparison to a 2-chloro analogue containing the same bulky N⁶-arylurea, although selectivity was increased.⁴² The combination of N⁶-(substituted benzyl) groups and a 2-phenylethynyl modification produced much lower affinity in hA₃AR binding than the same structure with H at the C2 position.²⁴ A study of 2-pyrazolyl-N⁶-substituted adenosine derivatives concluded that bulky substitutions at those two positions generally did not benefit from additivity in hA₃AR binding affinity.^13c

We have now explored the effects of various 2-(arylethynyl) groups at the adenine C2 position on AR affinity of (N)-methanocarba 5′-N-methyluronamido nucleosides. The resulting analogues (both N⁶-methyl and N⁶-3-halobenzyl) were full agonists of the hA₃AR of nanomolar affinity that were consistently highly selective (typically >1000-fold vs. hA₁AR and hA_2AAR). The most potent and selective N⁶-methyl compounds were p-F 13 and o-Cl 14 analogues. At the mA₃AR, selectivity generally remained, but for N⁶-methyl derivatives the high affinities achievable were somewhat lower than at hA₃AR. Thus, the combination of a 2-(arylethynyl) group and large N⁶ substitutions was better tolerated at the A₃AR in the methanocarba series than in the 9-riboside series, as characterized in earlier reports. There is also a broad flexibility of substitution of the 2-(arylethynyl) moiety, with halo, hydrophobic, and hydrophilic substitutions without losing A₃AR selectivity. This feature could also benefit specific pharmacological characteristics, such as pharmacokinetic properties. The 3,4-difluoro substitution of 31 might impede possible in vivo metabolic transformation of this ring (cf. P2Y₁₂ receptor antagonist Ticagrelor^1b).

The pharmacokinetic properties of these analogues have not been measured. However, they represent an increase in hydrophobicity, as well as selectivity, in comparison to most widely used A₃AR agonists. For example, the cLog P values of potent 2-chlorophenylethynyl 29, 4-fluorophenylethynyl 30 and 3,4-fluorophenylethynyl 31 derivatives are 4.65, 4.08 and 4.15, respectively. The cLog P values of more polar ribosides 1 and 2 are 0.48 and 1.20, respectively, which is possibly less desirable for full bioavailability. The total polar surface area of 29 is 121.9 Ų, which is a more favorable value for one of the physical parameters that predicts drug-likeness compared to 131.1 Ų for both 1 and 2.⁴³

Even large, planar polyaromatic groups at C2 (separated by an acetylene moiety) were tolerated and retained nanomolar affinity. In binding to the A_2AAR, a 2-(2-(naphth-1-yl)-ethyloxy) analogue bound with high affinity, and the location of the naphthyl group was predicted by docking.⁴⁴ In the present study, the size of the polycyclic C2 groups, i.e. in the potent phenanthrene-ethynyl derivative 26, exceeded that of all previous AR ligands. The 1-pyrene derivative 35 and its less potent 4-pyrene isomer 36 represent different sites of attachment of the 2-ethynyl moiety to the same large polyaromatic group. In receptor binding experiments, differences were observed that could provide insight into the geometry of the binding site considering the spatial and conformational constraints of these agonists. The 4-pyrene derivative 36 (with one additional ring added) has a close analogue in the N⁶-methyl series, i.e. 26.

These observations were subjected to molecular modeling analysis, which also predicted useful analogues to be synthesized. The ribose moiety is tightly anchored through a H-bond network involving TM3 and TM7, and the geometry of the unnatural glycosidic bond to the nucleobase is restricted. Thus, there is little angular flexibility of the extended, rigid C2 substituents. Thus, anchoring of the adenosine moiety to conserved amino acid residues provides a framework for exploring the region surrounding the linearly extended C2 substituent. The placement of a 2-arylethynyl substituent in the hA₃AR binding site was predicted by Dal Ben et al. using the inactive structure of the hA_2AAR as template, and the orientation is similar to that found here.^23b However, the present study predicts the docking mode with greater detail and confidence, because the homology model is based on the agonist-bound A_2AAR structure.²⁵ With the present set of rigidified analogues, we have located a very large hydophobic pocket on the receptor that depends on an outward diplacement of TM2 in order to accommodate multiple fused rings. Thus, the binding would be allowed by the plasticity of the receptor structure to attain a ligand-specific reorganization. The reduction of hA₃AR affinity for the larger C2 substituents in 35 and 36 suggests that although TM2 could shift position to accommodate such groups, there is an energetic cost for these more bulky planar polyaromatic groups. As the movement of TM2 increases, some stabilizing interactions with other regions of the receptor would be progressively lost.

The A₃AR homology model built based exclusively on the crystal structure of an agonist-bound A_2AAR required an adjustment of the position of TM2, which was based on the orientation of TM2 in other agonist-bound or activated GPCR templates, i.e. the β₂-adrenergic receptor or opsin. The most successful template for TM2 was the structure of opsin, which displaced the upper part of TM2 relative to the agonist-bound A_2AAR by ~7 Å. Thus, we predict an outward displacement of TM2 similar to the opsin structure, which is specific for agonists containing rigid C2 extensions. The inability of the A_2AAR to rearrangement might contribute to the A₃AR selectivity of these compounds.

The displacement of one or more TMs to accommodate a sterically bulky ligand, assuming that there are no other constraints on the helix such as a proximal disulfide bond, might be a general phenomenon in GPCRs. The ability of the helices to readjust position or “breathe” to enable a larger ligand to bind has already been proposed, specifically with respect to “multi-conformational space of the antagonist-like state of the human A₃ receptor”.⁴⁵ The example of altered A_2AAR conformation specific to binding of a sterically extended agonist 6-(2,2-diphenylethylamino)-9-((2R,3R,4S,5S)-5-(ethylcarbamoyl)-3,4-dihydroxytetrahydrofuran-2-yl)-N-(2-(3-(1-(pyridin-2-yl)piperidin-4-yl)ureido)ethyl)-9H-purine-2-carboxamide (59, UK-432097) was already documented in the X-ray structure.²⁵ In particular, outward movements of EL3 and extracellular portion of TM7 were associated exclusively with this bulky, multifunctionalized agonist.

The implications for receptor coupling and signaling of this type of displacement, such as the agonist-dependent movement of TM2 that we predict here, are unknown. If this conformational change associated with a family of ligands is propagated to the intracellular regions, we speculate that there could differential effects on multiple signaling pathways, and might eventually provide a rational basis for the design of biased GPCR agonists.

The enhancement of A_2AAR affinity in the 2-(3-chlorophenylethynyl) analogues 15 and 30 indicated a consistent interaction with a specific site on the hA_2AAR. Modeling analysis suggested halogen-π interactions⁴⁶ with two tyrosine residues in the C2 binding cavity of A_2AAR, namely Tyr271 (7.36) and Tyr9 (1.35), to explain the activity of the 2-(3-chlorophenylethynyl) methanocarba adenosine derivatives 15 and 30 at the A_2AAR.

The freedom to insert polyaromatic ring systems on the 2-ethynyl group suggests inclusion of reporter groups, such as fluorescent dye moieties at this site. In fact, the pyrene analogue 35 is highly fluorescent and could be explored as spectroscopic probes of moderate affinity for receptor binding experiments. The fluorescent properties of pyrene are sensitive to the environment. The p-amino derivative 32 has a high affinity at both the mA₃AR and hA₃AR, and could potentially be derivatized for radioisotope incorporation or affinity cross-linking. We are also interested in the design of radiofluorinated nucleosides for positron emission tomographic (PET) imaging of the A₃AR in vivo.⁴⁷ Several fluorinated species in this study could provide the basis for incorporation of ¹⁸F into a high affinity A₃AR agonist.

In conclusion, we have found a new series of highly potent and selective series of A₃AR agonists containing combined substitution of the (N)-methanocarba ring system and arylethynyl groups at the adenine C2 position. The binding affinities demonstrated high tolerance of steric bulk and ring substitution by extension at the C2 position, with many aryl groups providing K_i values in the nM range. Molecular modeling suggests that this reflects conformational plasticity of the A₃AR. Potent agonist ligands should be useful in future pharmacological studies of the antiinflammatory, anticancer and antiischemic properties of A₃AR agonists.

Experimental Section

Chemical synthesis

Materials and instrumentation

L-ribose, and other reagents and solvents were purchased from Sigma-Aldrich (St. Louis, MO). Alcohol derivative 39 was prepared as reported.^{28 1}H NMR spectra were obtained with a Brucker 400 spectrometer using CDCl₃ and CD₃OD as solvents. Chemical shifts are expressed in δ values (ppm) with tetramethylsilane (δ 0.00) for CDCl₃ and water (δ3.30) for CD₃OD. TLC analysis was carried out on glass sheets precoated with silica gel F₂₅₄ (0.2 mm) from Aldrich. The purity of final nucleoside derivatives was checked using a Hewlett–Packard 1100 HPLC equipped with a Zorbax SB-Aq 5 μm analytical column (50 × 4.6 mm; Agilent Technologies Inc, Palo Alto, CA). Mobile phase: linear gradient solvent system: 5 mM TBAP (tetrabutylammonium dihydrogenphosphate)-CH₃CN from 80:20 to 0:100 in 13 min; the flow rate was 0.5 mL/min. Peaks were detected by UV absorption with a diode array detector at 230, 254, and 280 nm. All derivatives tested for biological activity showed >95% purity by HPLC analysis (detection at 254 nm). Low-resolution mass spectrometry was performed with a JEOL SX102 spectrometer with 6-kV Xe atoms following desorption from a glycerol matrix or on an Agilent LC/MS 1100 MSD, with a Waters (Milford, MA) Atlantis C18 column. High resolution mass spectroscopic (HRMS) measurements were performed on a proteomics optimized Q-TOF-2 (Micromass-Waters) using external calibration with polyalanine, unless noted. Observed mass accuracies are those expected based on known performance of the instrument as well as trends in masses of standard compounds observed at intervals during the series of measurements. Reported masses are observed masses uncorrected for this time-dependent drift in mass accuracy. Synthetic procedures for 45 – 47 are in Supporting information. LogP and total polar surface area values were calculated using ChemBioDraw Ultra (Version 12.0.3, PerkinElmer, Boston, MA).

(1S,2R,3S,4R,5S)-2,3-Dihydroxy-N-methyl-4-(6-(methylamino)-2-(phenylethynyl)-9H-purin-9-yl)bicyclo[3.1.0]hexane-1-carboxamide (9)

A solution of compound 45a (29 mg, 0.06 mmol) in methanol (2 mL) and 10% trifluoromethane sulfonic acid (2 mL) was heated at 70 °C for 5 h. Solvent was evaporated under vacuum, and the residue was purified on flash silica gel column chromatography (CH₂Cl₂:MeOH = 25:1) to give compound 9 (21 mg, 81%) as a syrup. ¹H NMR (CD₃OD) δ 8.12 (s, 1H), 7.67-7.65 (m, 2H), 7.47-7.43 (m, 2H), 5.06 (d, J = 5.2 Hz, 1H), 4.08 (d, J = 6.4 Hz, 1H), 3.15 (br s, 3H), 2.84 (s, 3H), 2.13-2.10 (m, 1H), 1.88 (t, J = 5.2 Hz, 1H), 1.41-1.39 (m, 1H). HRMS calculated for C₂₂H₂₃N₆O₃ (M + H)⁺: 419.1832; found 419.1818.

(1S,2R,3S,4R,5S)-2,3-Dihydroxy-N-methyl-4-(6-(methylamino)-2-(pyridin-2-ylethynyl)-9H-purin-9-yl)bicyclo[3.1.0]hexane-1-carboxamide (10)

Compound 10 (80%) was prepared from compound 46 following the same method as used for compound 9. ¹H NMR (CD₃OD) δ 8.65 (s, 1H), 8.01 (s, 1H), 7.91–7.97 (m, 1H), 7.87 (d, J = 8.8 Hz, 1H), 7.53-7.48 (m, 1H), 5.15 (d, J = 5.6 Hz, 1H), 4.09 (d, J = 8.4 Hz, 1H), 3.14 (br s, 3H), 2.83 (s, 3H), 2.13-2.06 (m, 1H), 1.85 (t, J = 5.2 Hz, 1H), 1.42-1.40 (m, 1H). HRMS calculated for C₂₁H₂₂N₇O₃ (M + H)⁺: 420.1784; found 420.1797.

(1S,2R,3S,4R,5S)-4-(2-((2-Fluorophenyl)ethynyl)-6-(methylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (11)

Compound 11 (82%) was prepared from compound 45b following the same method as used for compound 9.¹H NMR (CD₃OD) δ 8.11 (s, 1H), 7.49-7.39 (m, 3H), 7.25-7.20 (m, 1H), 5.06 (d, J = 5.2 Hz, 1H), 4.9 (s, 1H), 4.02 (d, J = 6.8 Hz, 1H), 3.15 (br s, 1H), 2.84 (s, 3H), 2.13-2.10 (m, 1H), 1.88 (t, J = 4.8 Hz, 1H), 1.41-1.38 (m, 1H). HRMS calculated for C₂₂H₂₂FN₆O₃ (M + H)⁺: 437.1737; found 437.1753.

(1S,2R,3S,4R,5S)-4-(2-((3-Fluorophenyl)ethynyl)-6-(methylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (12)

Compound 12 (78%) was prepared from compound 45c following the same method as used for compound 9. ¹H NMR (CD₃OD) δ 8.11 (s, 1H), 7.67 (t, J = 6.0 Hz, 1H), 7.52-7.47 (m, 1H), 7.28-7.22 (m, 2H), 5.07 (d, J = 6.8 Hz, 1H), 4.03 (d, J = 6.8 Hz, 1H), 3.13 (br s, 1H), 2.84 (s, 3H), 2.13-2.10 (m, 1H), 1.88 (t, J = 4.8 Hz, 1H), 1.41-1.40 (m, 1H). HRMS calculated for C₂₂H₂₂FN₆O₃ (M + H)⁺: 437.1737; found 437.1718.

(1S,2R,3S,4R,5S)-4-(2-((4-Fluorophenyl)ethynyl)-6-(methylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (13)

Compound 13 (81%) was prepared from compound 45d following the same method as used for compound 9. ¹H NMR (CD₃OD) δ 8.11 (s, 1H), 7.75-7.68 (m, 2H), 7.23-7.18 (m, 2H), 5.05 (d, J = 6.0 Hz, 1H), 4.02 (d, J = 6.4 Hz), 1H), 3.14 (br s, 3H), 2.84 (s, 3H), 2.13-2.10 (m, 1H), 1.89 (t, J = 4.8 Hz, 1H), 1.41-1.39 (m, 1H). HRMS calculated for C₂₂H₂₂FN₆O₃ (M + H)⁺: 437.1737; found 437.1722.

(1S,2R,3S,4R,5S)-4-(2-((2-Chlorophenyl)ethynyl)-6-(methylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (14)

Compound 14 (85%) was prepared from compound 45e following the same method as used for compound 9. ¹H NMR (CD₃OD) δ 8.11 (s, 1H), 7.72 (d, J = 7.2 Hz, 1H), 7.54 (d, J = 7.2 Hz, 1H), 7.46-7.36 (m, 2H), 5.10 (d, J = 6.4 Hz, 1H), 4.04 (d, J = 6.8 Hz, 1H), 3.15 (br s, 3H), 2.83 (s, 3H), 2.12-2.08 (m, 1H), 1.86 (t, J = 4.8 Hz, 1H), 1.41-1.39 (m, 1H). HRMS calculated for C₂₂H₂₂ClN₆O₃ (M + H)⁺: 453.1442; found 453.1449.

(1S,2R,3S,4R,5S)-4-(2-((3-Chlorophenyl)ethynyl)-6-(methylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (15)

Compound 15 (82%) was prepared from compound 45f following the same method as used for compound 9. ¹H NMR (CD₃OD) δ 8.11 (s, 1H), 7.67 (s, 1H), 7.59 (d, J = 7.2 Hz, 1H), 7.49-7.42 (m, 2H), 5.06 (d, J = 6.4 Hz, 1H), 4.02 (d, J = 6.8 Hz, 1H), 3.14 (br s, 3H), 2.84 (s, 3H), 2.13-2.10 (m, 1H), 1.88 (t, J = 4.8 Hz, 1H), 1.41-1.39 (m, 1H). HRMS calculated for C₂₂H₂₂ClN₆O₃ (M + H)⁺: 453.1442; found 453.1442.

(1S,2R,3S,4R,5S)-4-(2-((4-Chlorophenyl)ethynyl)-6-(methylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (16)

Compound 16 (84%) was prepared from compound 45g following the same method as used for compound 9. ¹H NMR (CD₃OD) δ 8.10 (s, 1H), 7.64 (d, J = 8.4 Hz, 2H), 7.46 (d, J = 8.4 Hz, 2H), 5.06 (d, J = 5.6 Hz, 1H), 4.88 (s, 1H), 4.02 (d, J = 6.4 Hz, 1H), 3.15 (br s, 3H), 2.84 (s, 3H), 2.13-2.09 (m, 1H), 1.88 (t, J = 4.8 Hz, 1H), 1.41-1.38 (m, 1H). HRMS calculated for C₂₂H₂₂ClN₆O₃ (M + H)⁺: 453.1460; found 453.1454.

(1S,2R,3S,4R,5S)-4-(2-((4-Bromophenyl)ethynyl)-6-(methylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (17)

Compound 17 (76%) was prepared from compound 45h following the same method as used for compound 9. ¹H NMR (CD₃OD) δ 8.10 (s, 1H), 7.62 (d, J = 8.8 Hz, 2H), 7.56 (d, J = 8.4 Hz, 2H), 5.05 (d, J = 6.8 Hz, 1H), 4.88 (s, 1H), 4.02 (d, J = 6.4 Hz, 1H), 3.14 (br s, 3H), 2.84 (s, 3H), 2.13-2.09 (m, 1H), 1.88 (t, J = 5.2 Hz, 1H), 1.41-1.38 (m, 1H). HRMS calculated for C₂₂H₂₂BrN₆O₃ (M + H)⁺: 497.0937; found 497.0948.

(1S,2R,3S,4R,5S)-4-(2-((3-Aminophenyl)ethynyl)-6-(methylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (18)

Compound 18 (67%) was prepared from compound 45i following the same method as used for compound 9.¹H NMR (CD₃OD) δ 8.09 (s, 1H), 7.14 (t, J = 8.0 Hz, 1H), 6.98-6.94 (m, 2H), 6.80-6.78 (m, 1H), 5.06 (d, J = 5.6 Hz, 1H), 4.02 (d, J = 6.4 Hz, 1H), 3.15 (br s, 3H), 2.85 (s, 3H), 2.12-2.09 (m, 1H), 1.87 (t, J = 4.8 Hz, 1H), 1.41-1.39 (m, 1H). HRMS calculated for C₂₂H₂₂N₇O₃ (M + H)⁺: 432.1784; found 432.1799.

(1S,2R,3S,4R,5S)-4-(2-((3,4-Difluorophenyl)ethynyl)-6-(methylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (19)

Compound 19 (83%) was prepared from compound 45j following the same method as used for compound 9. ¹H NMR (CD₃OD) δ 8.12 (s, 1H), 7.58-7.50 (m, 1H), 7.50-7.48 (m, 1H), 7.40-7.34 (m, 1H), 5.06 (d, J = 6.4 Hz, 1H), 4.02 (d, J = 6.4 Hz, 1H), 3.14 (br s, 3H), 2.84 (s, 3H), 2.13-2.09 (m, 1H), 1.88 (t, J = 4.8 Hz, 1H), 1.41-1.40 (m, 1H). HRMS calculated for C₂₂H₂₁F₂N₆O₃ (M + H)⁺: 455.1643; found 455.1639.

(1S,2R,3S,4R,5S)-4-(2-((3,5-Difluorophenyl)ethynyl)-6-(methylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (20)

Compound 20 (84%) was prepared from compound 45k following the same method as used for compound 9. ¹H NMR (CD₃OD) δ 8.12 (s, 1H), 7.31-7.26 (s, 2H), 7.14-7.09 (m, 1H), 7.49-7.42 (m, 2H), 5.06 (d, J = 5.2 Hz, 1H), 4.89 (s, 1H), 4.03 (d, J = 6.4 Hz, 1H), 3.14 (br s, 3H), 2.84 (s, 3H), 2.12-2.09 (m, 1H), 1.88 (t, J = 4.8 Hz, 1H), 1.41-1.39 (m, 1H). HRMS calculated for C₂₂H₂₁F₂N₆O₃ (M + H)⁺: 455.1643; found 455.1630.

(1S,2R,3S,4R,5S)-4-(2-((4-Ethylphenyl)ethynyl)-6-(methylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (21)

Compound 21 (79%) was prepared from compound 45l following the same method as used for compound 9. ¹H NMR (CD₃OD) δ 8.09 (s, 1H), 7.57 (d, J = 8.0 Hz, 2H), 7.29 (d, J = 8.0 Hz, 2H), 5.06 (d, J = 5.2 Hz, 1H), 4.91 (s, 1H), 4.02 (d, J = 6.4 Hz, 1H), 3.15 (br s, 3H), 2.84 (s, 3H), 2.74-2.68 (m, 2H), 2.13-2.09 (m, 1H), 1.88 (t, J = 4.8 Hz, 1H), 1.41-1.38 (m, 1H), 1.30 (t, J=7.6 Hz, 3H). HRMS calculated for C₂₄H₂₇N₆O₃ (M + H)⁺: 447.2145; found 447.2130.

(1S,2R,3S,4R,5S)-4-(2-((4-tert-Butylphenyl)ethynyl)-6-(methylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (22)

Compound 22 (83%) was prepared from compound 45m following the same method as used for compound 9. ¹H NMR (CD₃OD) δ 8.10 (s, 1H), 7.59 (d, J = 8.4 Hz, 2H), 7.50 (d, J = 8.4 Hz, 2H), 5.06 (d, J = 6.4 Hz, 1H), 4.02(d, J = 6.4 Hz, 1H), 3.15 (br s, 3H), 2.85 (s, 3H), 2.13-2.10 (m, 1H), 188(t, J = 4.8 Hz, 1H), 1.41-1.39 (m, 10H). HRMS calculated for C₂₆H₃₁N₆O₃ (M + H)⁺: 475.2458; found 475.2450.

(1S,2R,3S,4R,5S)-4-(2-((4-Acetylphenyl)ethynyl)-6-(methylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (23)

Compound 23 (80%) was prepared from compound 45n following the same method as used for compound 9. ¹H NMR (CD₃OD) δ 8.11 (s, 1H), 8.06 (d, J = 8.4 Hz, 2H), 7.77 (d, J = 8.0 Hz, 2H), 5.06 (d, J = 6.4 Hz, 1H), 4.03 (d, J = 6.4 Hz, 1H), 3.15 (br s, 3H), 2.84 (s, 3H), 2.64 (s, 3H), 2.14-2.10 (m, 1H), 1.88 (t, J = 4.8 Hz, 1H), 1.41-1.39 (m, 1H). HRMS calculated for C₂₄H₂₅N₆O₄ (M + H)⁺: 461.1937; found 461.1937.

(1S,2R,3S,4R,5S)-4-(2-(Biphenyl-4-ylethynyl)-6-(methylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (24)

Compound 24 (74%) was prepared from compound 45o following the same method as used for compound 9. ¹H NMR (CD₃OD) δ 8.11 (s, 1H), 7.75-7.67 (m, 6H), 7.47 (t, J = 7.6 Hz, 2H), 7.38 (t, J = 7.2Hz, 1H), 5.06 (d, J = 6.0 Hz, 1H), 4.89 (s, 1H), 4.03 (d, J = 6.4 Hz, 1H), 3.16 (br s, 3H), 2.84 (s, 3H), 2.13-2.10 (m, 1H), 189 (t, J = 4.8 Hz, 1H), 1.42-1.39 (m, 1H). HRMS calculated for C₂₈H₂₇N₆O₃ (M + H)⁺: 495.2145; found 495.2141.

(1S,2R,3S,4R,5S)-2,3-Dihydroxy-N-methyl-4-(6-(methylamino)-2-(naphthalen-1-ylethynyl)-9H-purin-9-yl)bicyclo[3.1.0]hexane-1-carboxamide (25)

Compound 25 (73%) was prepared from compound 45p following the same method as used for compound 9. ¹H NMR (CD₃OD) δ 8.56 (d, J = 7.6 Hz, 1H), 8.11 (s, 1H), 8.00-7.90 (m, 3H), 7.68 (t, J = 5.6 Hz, 1H), 7.62-7.53 (m, 2H), 5.10 (d, J = 5.2 Hz, 1H), 4.93 (s, 1H), 4.06 (d, J = 6.4 Hz, 1H), 3.19 (br s, 3H), 2.79 (s, 3H), 2.15-2.12 (m, 1H), 189 (t, J = 4.8 Hz, 1H), 1.41-1.38 (m, 1H). HRMS calculated for C₂₆H₂₅N₆O₃ (M + H)⁺: 469.1988; found 469.2005.

(1S,2R,3S,4R,5S)-2,3-Dihydroxy-N-methyl-4-(6-(methylamino)-2-(phenanthren-9-ylethynyl)-9H-purin-9-yl)bicyclo[3.1.0]hexane-1-carboxamide (26)

Compound 26 (65%) was prepared from compound 45q following the same method as used for compound 9. ¹H NMR (CD₃OD) δ 8.84-8.77 (m, 2H), 8.66-8.63 (m, 1H), 8.26 (s, 1H), 8.12 (s, 1H), 7.97 (d, J = 7.6 Hz, 1H), 7.80-7.73 (m, 3H), 7.66 (t, J = 7.2 Hz, 1H), 5.10 (d, J = 6.0 Hz, 1H), 4.92 (s, 1H), 4.07 (d, J = 6.4 Hz, 1H), 3.20 (br s, 3H), 2.81 (s, 3H), 2.15-2.12 (m, 1H), 1.90 (t, J = 4.8 Hz, 1H), 1.42-1.38 (m, 1H). HRMS calculated for C₃₀H₂₇N₆O₃ (M + H)⁺: 519.2145; found 519.2137.

(1S,2R,3S,4R,5S)-4-(6-(3-Chlorobenzylamino)-2-(phenylethynyl)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (27)

PdCl₂(PPh₃)₂ (6.13 mg, 0.008 mmol), CuI (1.2 mg, 0.004 mmol), phenylacetylene (30 μL, 0.26 mmol) and triethylamine (60 μL, 0.4 mmol) was added to a solution of compound 44 (26 mg, 0.04 mmol) in anhydrous DMF (1 mL), and stirred at room temperature overnight. Solvent was evaporated under vacuum, and the residue was roughly purified on flash silica gel column chromatography. The resulting compound was dissolved in methanol (2 mL) and 10% trifluoromethane sulfonic acid (2 mL) and heated at 70 °C for 5 h. Solvent was evaporated under vacuum, and the residue was purified on flash silica gel column chromatography (CH₂Cl₂:MeOH = 25:1) to give compound 27 (17 mg, 76%) as a syrup. ¹H NMR (CD₃OD) δ 8.13 (s, 1H), 7.66-7.63 (m, 2H), 7.46-7.42 (m, 4H), 7.37-7.26 (m, 3H), 5.06 (d, J = 5.6 Hz, 1H), 4.9 (br s, 2H) 4.04 (d, J = 6.4 Hz, 1H), 2.84 (s, 3H), 2.14-2.11 (m, 1H), 1.88 (t, J = 4.8 Hz, 1H), 1.41-1.37 (m, 1H). HRMS calculated for C₂₈H₂₆ClN₆O₃ (M + H)⁺: 529.1755; found 529.1740.

(1S,2R,3S,4R,5S)-4-(6-(3-Chlorobenzylamino)-2-((4-fluorophenyl)ethynyl)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (28)

Compound 28 (68%) was prepared from compound 44 following the same method as used for compound 27. ¹H NMR (CD₃OD) δ 8.13 (s, 1H), 7.70-7.63 (m, 2H), 7.59-7.55 (m, 1H), 7.45 (s, 1H), 7.37-7.21 (m, 3H), 7.19-7.16 (m, 1H), 5.06 (d, J = 5.6 Hz, 1H), 4.9 (s, 1H), 4.58 (br s, 2H), 4.04 (d, J = 7.6 Hz, 1H), 2.84 (s, 3H), 2.17-2.10 (m, 1H), 1.88 (t, J = 4.8 Hz, 1H), 1.41-1.38 (m, 1H). HRMS calculated for C₂₈H₂₅ClFN₆O₃ (M + H)⁺: 547.1661; found 547.1652.

(1S,2R,3S,4R,5S)-4-(6-(3-Chlorobenzylamino)-2-((2-chlorophenyl)ethynyl)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (29)

Compound 29 (65%) was prepared from compound 44 following the same method as used for compound 27. ¹H NMR (CD₃OD) δ 8.14 (s, 1H), 7.72-7.70 (m, 1H), 7.53 (d, J = 8.0 Hz, 1H), 7.47-7.25 (m, 6H), 5.11 (d, J = 6.8 Hz, 1H), 4.90 (s, 1H), 4.05 (d, J = 6.8 Hz, 1H), 2.82 (s, 3H), 2.12-2.09 (m, 1H), 1.86 (t, J = 4.8 Hz, 1H), 1.40-1.38 (m, 1H). HRMS calculated for C₂₈H₂₄Cl₂N₆O₃Na (M + Na)⁺: 585.1185; found 585.1167.

(1S,2R,3S,4R,5S)-4-(6-(3-Chlorobenzylamino)-2-((3-chlorophenyl)ethynyl)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (30)

Compound 30 (66%) was prepared from compound 44 following the same method as used for compound 27. ¹H NMR (CD₃OD) δ 8.15 (s, 1H), 7.66-7.57 (m, 2H), 7.48-7.26 (m, 6H), 5.07 (d, J = 6.4 Hz, 1H), 4.85 (s, 1H), 4.04 (d, J = 6.8 Hz, 1H), 2.84 (s, 3H), 2.14-2.10 (m, 1H), 1.88 (t, J = 4.8 Hz, 1H), 1.41-1.39 (m, 1H). HRMS calculated for C₂₈H₂₅Cl₂N₆O₃ (M + H)⁺: 563.1365; found 563.1359.

(1S,2R,3S,4R,5S)-4-(6-(3-Chlorobenzylamino)-2-((3,4-difluorophenyl)ethynyl)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (31)

Compound 31 (63%) was prepared from compound 44 following the same method as used for compound 27. ¹H NMR (CD₃OD) δ 8.14 (s, 1H), 7.58 (t, J =8.8 Hz, 1H), 7.47-7.44 (m, 2H), 7.39-7.25 (m, 4H), 5.06 (d, J = 6.4 Hz, 1H), 4.89 (s, 1H), 4.04 (d, J = 6.4 Hz, 1H), 2.84 (s, 3H), 2.13-2.10 (m, 1H), 1.88 (t, J = 4.8 Hz, 1H), 1.41-1.38 (m, 1H). HRMS calculated for C₂₈H₂₄F₂ClN₆O₃ (M + H)⁺: 565.1566; found 565.1559.

(1S,2R,3S,4R,5S)-4-(2-((4-Aminophenyl)ethynyl)-6-(3-chlorobenzylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (32)

Compound 32 (59%) was prepared from compound 44 following the same method as used for compound 27. ¹H NMR (CD₃OD) δ 8.02 (s, 1H), 7.84 (d, J =8.8 Hz, 2H), 7.33-7.23 (m, 4H), 6.60 (d, J = 8.4 Hz, 2H), 5.17 (d, J = 6.4 Hz, 1H), 4.85 (s, 1H), 4.73 (br s, 2H), 4.02 (d, J = 6.8 Hz, 1H), 2.83 (s, 3H), 2.08-2.05 (m, 1H), 1.81 (t, J = 4.8 Hz, 1H), 1.41-1.37 (m, 1H). HRMS calculated for C₂₈H₂₇ClN₇O₃ (M + H)⁺: 545.1041; found 545.1045.

3-((6-(3-Chlorobenzylamino)-9-((1S,2R,3S,4R,5S)-3,4-dihydroxy-5-(methyl-carbamoyl)bicyclo[3.1.0]hexan-2-yl)-9H-purin-2-yl)ethynyl)benzoic acid (33)

PdCl₂(PPh₃)₂ (3.0 mg, 0.004 mmol), CuI (1.0 mg, 0.004 mmol), phenylacetylene (18.7 mg, 0.12 mmol) and triethylamine (20 μL, 0.2 mmol) was added to a solution of compound 44 (12.68 mg, 0.02 mmol) in anhydrous DMF (1 mL), and stirred at room temperature overnight. Solvent was evaporated under vacuum, and the residue was roughly purified on flash silica gel column chromatography. The resulting compound was dissolved in dioxane (2 mL) and 1N HCl (1.5 mL) and heated at 60 °C for 2 h. After completion of starting material, solvent was evaporated under vacuum, and the residue was purified on flash silica gel column chromatography (CH₂Cl₂:MeOH:TFA = 25:1:0.1) to give compound 33 (7 mg, 61%) as a syrup. ¹H NMR (CD₃OD) δ 8.28 (s, 1H), 8.18-8.16 (m, 1H), 8.11-8.07 (m, 1H), 7.87 (d, J =7.6 Hz, 1H), 7.59-7.54 (m, 2H), 7.47 (s, 1H), 7.41-7.26 (m, 2H), 5.09 (d, J = 6.4 Hz, 1H), 4.91 (s, 1H), 4.05 (d, J = 6.4 Hz, 1H), 2.85 (s, 3H), 2.14-2.11 (m, 1H), 1.88 (t, J = 4.8 Hz, 1H), 1.42-1.38 (m, 1H). HRMS calculated for C₂₉H₂₆ClN₆O₅ (M + H)⁺: 573.1653; found 573.1646.

(1S,2R,3S,4R,5S)-4-(2-(Biphenyl-4-ylethynyl)-6-(3-chlorobenzylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (34)

Compound 34 (68%) was prepared from compound 44 following the same method as used for compound 27. ¹H NMR (CD₃OD) δ 8.13 (s, 1H), 7.74-7.66 (m, 7H), 7.49-7.45 (m, 2H), 7.40-7.26 (m, 4H), 5.07 (d, J = 5.6 Hz, 1H), 4.9 (s, 1H), 4.60 (br s, 2H), 4.05 (d, J = 6.4 Hz, 1H), 2.85 (s, 3H), 2.14-2.11 (m, 1H), 1.89 (t, J = 4.8 Hz, 1H), 1.42-1.40 (m, 1H). HRMS calculated for C₃₄H₃₀ClN₆O₃ (M + H)⁺: 605.2068; found 605.2083.

(1S,2R,3S,4R,5S)-4-(6-(3-Chlorobenzylamino)-2-(pyren-1-ylethynyl)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (35)

Compound 35 (91%) was prepared from compound 44 following the same method as used for compound 33. ¹H NMR (CD₃OD) δ 8.71 (d, J = 9.2 Hz, 1H), 8.26-8.23 (m, 4H), 8.16-8.13 (m, 2H), 8.08-8.03 (m, 3H), 7.54 (s, 1H), 7.43 (d, J = 7.6 Hz, 1H), 7.34 (t, J = 8.0 Hz, 1H), 7.27 (d, J = 8.0 Hz, 1H), 5.06 (d, J = 6.4 Hz, 1H), 4.83 (s, 1H), 4.04 (d, J = 6.4 Hz, 1H), 2.81 (s, 3H), 2.10-2.07 (m, 1H), 1.89 (t, J = 4.8 Hz, 1H), 1.41-1.37 (m, 1H). HRMS calculated for C₃₈H₃₀ClN₆O₃ (M + H)⁺: 653.2068; found 653.2078.

(1S,2R,3S,4R,5S)-4-(6-(3-Chlorobenzylamino)-2-(pyren-4-ylethynyl)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide (36)

Compound 36 (69%) was prepared from compound 44 following the same method as used for compound 33. ¹H NMR (CD₃OD) δ 8.82 (d, J = 7.6 Hz, 1H), 8.51 (s, 1H), 8.29-8.21 (m, 3H), 8.16-8.12 (m, 4H), 8.04 (t, J = 7.6 Hz, 1H), 7.54 (s, 1H), 7.43 (d, J = 7.6 Hz, 1H), 7.35 (t, J = 8.0 Hz, 1H), 7.28 (d, J = 8.4 Hz, 1H), 5.1 (d, J = 6.0 Hz, 1H), 4.06 (d, J = 6.0 Hz, 1H), 2.80 (s, 3H), 2.18-2.09 (m, 1H), 1.89 (t, J = 4.8 Hz, 1H), 1.42-1.39 (m, 1H). HRMS calculated for C₃₈H₃₀ClN₆O₃ (M + H)⁺: 653.2068; found 653.2056.

(1S,2R,3S,4R,5S)-Ethyl-(2,3-O-isopropylidene)-4-(2-iodo-6-(methylamino)-9H-purin-9-yl)bicyclo[3.1.0]hexane-1-carboxylate (41)

Methylamine hydrochloride (0.353g, 5.23 mmol) and triethylamine (1.4 mL, 16.6 mmol) was added to a solution of compound 40 (0.528g, 1.04 mmol) in anhydrous methanol (15 mL) and stirred at room temperature overnight. Solvent was evaporated under vacuum, and the residue was purified on flash silica gel column chromatography (hexane:ethylacetate=1:1) to give compound 41 (0.470 g, 94%) as a foamy solid. ¹H NMR (CD₃OD) δ 7.94 (s, 1H), 5.83 (d, J = 7.2 Hz, 1H), 4.94 (s, 1H), 4.80 (d, J = 6.0 Hz, 1H), 4.33-4.27 (m, 2H), 3.05 (br s, 3H), 2.25-2.21 (m, 1H), 1.65-1.61 (m, 1H), 1.53-1.49 (m, 4H), 1.34 (t, J = 7.2 Hz, 3H), 1.29 (s, 3H). HRMS calculated for C₁₈H₂₃IN₅O₄ (M + H)⁺: 500.1072; found 500.1075.

(1S,2R,3S,4R,5S)-(2,3-O-Isopropylidene)-4-(2-iodo-6-(methylamino)-9H-purin-9-yl)-N-methylbicyclo[3.1.0]hexane-1-carboxamide (43)

40% Methylamine solution (10 mL) was added to a solution of compound 41 (0.470 g, 0.94 mmol) in methanol (15 mL) and stirred at room temperature for 48 h. Solvent was evaporated under vacuum, and the residue was purified on flash silica gel column chromatography (CH₂Cl₂:MeOH=40:1) to give compound 43 (0.360 g, 79%) as a syrup. ¹H NMR (CD₃OD) δ 7.95 (s, 1H), 5.72 (d, J = 7.2 Hz, 1H), 4.93 (s, 1H), 4.84 (d, J = 7.2 Hz, 1H), 3.05 (br s, 3H), 2.90 (s, 3H), 2.17-2.11 (m, 1H), 1.54-1.49 (m, 4H), 1.39 (t, J = 5.2 Hz, 1H), 1.30 (s, 3H). HRMS calculated for C₁₇H₂₂IN₆O₃ (M + H)⁺: 485.0798; found 485.0803.

Pharmacological characterization

The nucleoside derivatives, dissolved as stock solutions in DMSO (5 mM) and stored frozen, were evaluated in binding^29–31 and a functional assay³⁴ at the A₃AR, binding assays at the A₁AR and A_2AAR, and a functional assay at the hA₃AR (details in Supporting Information). Use of heterologously expressed mouse ARs was as reported.^33b,49 Protein content was determined⁴⁸ and IC₅₀ values in binding inhibition transformed to K_i values⁵¹ as reported.

Molecular modeling

The Homology Model module of MOE was utilized to build a new molecular model of the hA₃AR (details in Supporting Information). The recently reported model of the complex of the nonselective AR agonist 59 docked to the crystal structure of the A_2AAR was used as a template for modeling of the A₃AR.^25,51 The sequence alignment of the A_2A and A₃ARs was performed with MOE, taking into account positions of highly conserved amino acid residues. The position of full agonist 59 inside the A_2AAR was taken into account during the modeling. In the resulting initial homology model of the A₃AR (prior to movement of TM2), the complex 5′,2,N⁶ trisubstituted agonist 59 used to crystallize the A_2AAR was removed and the simpler 5′-uronamide 58 docked inside the receptor. The Glide program of the Schrodinger package⁵² was used to dock the other agonists to the A₃AR model obtained. The receptor grid generation was performed for the box with a center in the centroid of 58 in its initial position. The size of the box was determined automatically. The extra precision mode (XP) of Glide was used for the docking. The binding site was defined as 58 and all amino acid residues located within 5Å from 58. All A₃AR residues located within 2Å from the binding site were used as a shell. The following parameters of energy minimization were used: OPLS2005 force field, water was used as an implicit solvent, a maximum of 5000 iterations of the Polak-Ribier conjugate gradient minimization method was used with a convergence threshold of 0.01 kJ·mol⁻¹·Å⁻¹. Another reference nucleoside 58 was also docked in this hA₃AR homology model.

Hybrid A₃AR models involving movement of TM2, as explained above, were utilized to study the binding mode of analogue 34 and all other novel analogues in Table 1 using InducedFit docking implemented in the Schrödinger package. The grid generation was performed for a cubic box with sides of 26 Å and having a center in the centroid of the agonist molecules. The default values were used for other parameters.

Abbreviations

AR

adenosine receptor

cAMP

adenosine 3′,5′-cyclic phosphate

CHO

Chinese hamster ovary

Cl-IB-MECA

2-chloro-N⁶-(3-iodobenzyl)-5′-N-methylcarboxamidoadenosine

DIPEA

diisopropylethylamine

DCM

dichloromethane

DMF

N,N-dimethylformamide

DMEM

Dulbecco’s modified Eagle’s medium

EDTA

ethylenediaminetetraacetic acid

EL

extracellular loop

GPCR

G protein-coupled receptor

HEK

human embryonic kidney

I-AB-MECA

N⁶-(4-amino-3-iodobenzyl)adenosine-5′-N-methyl-uronamide

IFD

Induced Fit docking

NECA

5′-N-ethylcarboxamidoadenosine

HEPES

4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

HRMS

high resolution mass spectroscopy

NMR

nuclear magnetic resonance

R-PIA

N⁶-R-phenylisopropyladenosine

TEA

triethylamine

TLC

thin layer chromatography

TM

transmembrane domain