Functionalized Congeners of A₃ Adenosine Receptor-Selective Nucleosides Containing a Bicyclo[3.1.0]hexane Ring System

Abstract

(N)-Methanocarba nucleosides containing bicyclo[3.1.0]hexane replacement of the ribose ring previously demonstrated selectivity as A₃ adenosine receptor (AR) agonists (5′-uronamides) or antagonists (5′-truncated). Here, these two series were modified in parallel at the adenine C2 position. N⁶-3-Chlorobenzyl-5′-N-methyluronamides derivatives with functionalized 2-alkynyl chains of varying length terminating in a reactive carboxylate, ester, or amine group were full, potent human A₃AR agonists. Flexibility of chain substitution allowed the conjugation with a fluorescent cyanine dye (Cy5) and biotin, resulting in binding K_i values of 17 and 36 nM, respectively. The distal end of the chain was predicted by homology modeling to bind at the A₃AR extracellular regions. Corresponding l-nucleosides were nearly inactive in AR binding. In the 5′-truncated nucleoside series, 2-Cl analogues were more potent at A₃AR than 2-H and 2-F, functional efficacy in adenylate cyclase inhibition varied, and introduction of a 2-alkynyl chain greatly reduced affinity. SAR parallels between the two series lost stringency at distal positions. The most potent and selective novel compounds were amine congener 15 (K_i = 2.1 nM) and truncated partial agonist 22 (K_i = 4.9 nM).

Introduction

Selective agonist and antagonist ligands of the A₃ adenosine receptor (AR^a), a G-protein-coupled receptor (GPCR), have been reported.¹ A₃AR antagonists are proposed for the treatment of cancer, inflammation, glaucoma, and asthma.^2–5 A₃AR agonists are already in clinical trials for liver cancer, rheumatoid arthritis, psoriasis, and dry eye disease,^6,7 with additional envisioned applications for bowel inflammation, ischemia, and other autoimmune inflammatory disorders.^8–10

Selective A₃AR antagonists encompass both non-purine heterocyclic antagonists^1,11 and nucleoside^12,13 antagonists. Nucleoside-derived A₃AR antagonists have the advantage of a species-independent pharmacological profile;⁴ i.e., they tend to maintain A₃AR selectivity in both human and murine species, while most of the non-purine antagonists are selective for the human (h) but not rat (r) A₃AR. Modification of the ribose moiety of adenosine derivatives, and to a lesser extent N⁶ and C2 substituents, has been shown to reduce agonist efficacy and thus be useful for converting A₃AR agonists into antagonists.^12,13

Both the binding and selectivity of adenosine derivatives as prototypical A₃AR agonists can be substantially increased by replacing the flexible ribose scaffold with a rigid bicyclo[3.1.0] hexane ring system resulting in a class of carbocyclic nucleosides related to 1 (Chart 1).^14–16 These (N)-methanocarba adenosine agonists have been found to possess high potency and enhanced selectivity for the A₃AR,¹¹ and an agonist of this class (MRS3558, 1a) has been shown to display antiarthritic activity and protection in a model of lung ischemia.^17,18 The N⁶-3-iodobenzyl derivative 1b (MRS1898) was recently radioiodinated to form an A₃AR agonist radioligand that can be used selectively in murine and in hA₃AR binding studies.¹⁹

Chart 1.

(N)-Methanocarba Derivatives of Adenosine as A₃AR-Selective Agonists (1a, 1b) and Antagonists (2a, 2b)^a

^aGeneral formulas I and II represent two parallel structural series included in the present study. The agonist corresponding to X = Cl, n = 3, m = 2, R = H in series II was already reported.²¹

We describe in this work two new series of selective A₃AR ligands that extend recent reports, i.e., agonists and antagonists of the (N)-methanocarba family, derived through largely parallel modifications at the adenine C2 position.^20,21 Thus, the new agonists contained a 5′-N-methyluronamide group, and this amide group was removed entirely in the truncated series. The principle of lowering the relative efficacy of full agonists by removing a flexible 5′-amide group to produce, in some cases, hA₃AR antagonists was shown for 4′-thionucleosides by Jeong et al. and for (N)-methanocarba analogues by Melman et al. (i.e., 2a–c).^13,20 The stereospecificity of binding to the ARs was also probed with the preparation of corresponding l-nucleosides. In both series, the conformationally restricted (N)-methanocarba scaffold was retained, and most of the analogues contained a substituted 6-(benzylamino) moiety. The most structurally diverse modifications were made at the adenine C2 position, with the addition of extended alkynyl chains terminating in a carboxylic, ester, or amino group, following a functionalized congener approach.^22,23 This terminal functionalized group was remote from the minimal pharmacophore required for recognition at the primary binding site of the receptor. The intended application of the functional groups is for covalent coupling to various other moieties, including spectroscopic reporter groups, without losing biological activity.

Results

Chemical Synthesis

In this study, we have elaborated and extended functionalization at the C2 position of previously reported (N)-methanocarba nucleoside derivatives, which were shown to be selective A₃AR agonists^15,16,21 and antagonists.²⁰ The series of 5′-N-methyluronamides (Chart 1, II) are similar to the reported agonists 1, and the series of truncated nucleosides (I) are similar to the reported antagonists 2. Here, these two series were modified in parallel at the adenine C2 position.

The synthetic route to the 5′-N-methyluronamide (N)-methanocarba derivatives 5–16 is shown in Scheme 1A. The synthesis of the 2′,3′-protected intermediate 27, containing an N⁶-(3-chlorobenzyl) group, from l-ribose was previously reported.²⁴ This 2-iodo intermediate was then subjected to Sonogashira coupling²⁵ with the corresponding alkyne esters. The products 28–30 of varying alkynyl chain length were deprotected to liberate the 2′,3′-hydroxyl groups to provide nucleosides 8–10, followed by hydrolysis of the methyl ester to afford agonists 5–7. Alternatively, compounds 28–30 were aminolyzed using the appropriate alkyldiamine in excess, and the product amines were deprotected to yield the amine functionalized agonists 11–16 of varying chain length.

Scheme 1^a.

^a (A) (i) HC≡C(CH₂)_nCOOMe, Pd(PPh₃)₂Cl₂, CuI, Et₃N, DMF, room temp; (ii) 10% TFA, MeOH, 70 °C; (iii) appropriate alkyldiamine, MeOH, room temp; (iv) KOH, MeOH, room temp. (B) (v) For 17a: biotin, HATU, DIEA, DMF, room temp. For 17b: biotin-ε-aminocaproyl N-succinimidyl ester, Et₃N, DMF, room temp. (vi) N-succinimidyl ester of chain-derivatized Cy5 cyanine dye, bicarbonate buffer, DMF.

The amino and carboxylic derivatives contained chemically functionalized, extended alkynyl chains to serve as functionalized congeners for conjugation to other biologically active moieties or to carriers. ^23,26 Several representative biologically active conjugates for the detection and characterization of the A₃AR were prepared as shown in Scheme 1B. Biotin conjugates of varying length 17a and 17b were synthesized from the amine 11, which was the analogue having the shortest chain in the homologous series of amine congeners. The biotin conjugate 17a was prepared by a HATU coupling with biotin, and the extended chain analogue 17b was prepared from the corresponding biotin-ε-aminocaproyl active ester. The fluorescent derivative 18 similarly was prepared by treating 11 with the active ester of a chain-derivatized Cy5 cyanine dye.²⁷

Selected l-enantiomers 19–21 of the 5′-N-methyluronamide methanocarba agonist series were also prepared, as shown in Scheme 2. The isopropylidene intermediate 31 was prepared from d-ribose as previously reported. A Mitsunobu condensation²¹ with 2-iodo-6-chloropurine provided 32, which was first converted to the N⁶-(3-chlorobenzyl) derivative 33 and then treated with excess methylamine to yield 34. This 2-iodo intermediate was subjected to a Sonogashira coupling²⁵ with the corresponding alkyne esters to provide compounds 35–37 of varying alkynyl chain length. Esters 35 and 37 were aminolyzed with ethylenediamine followed by hydrolysis of the acetonide group with 10% TFA to provide compounds 20 and 21. In the case of compound 36, the 2′,3′-isopropylidene group was first deprotected, and hydrolysis of the ester afforded the carboxylic acid derivative 19.

Scheme 2 ^a.

^a (i) 2-Iodo-6-chloropurine, Ph₃P, DIAD, THF, room temp; (ii) 3-chlorobenzylamine, Et₃N, MeOH, room temp; (iii) 10% methylamine, MeOH, room temp; (iv) HC≡C(CH₂)_nCOOMe, Pd(PPh₃)₂Cl₂, Cul, Et₃N, DMF, room temp; (v) KOH, MeOH; (vi) ethylenediamine, MeOH, room temp; (vii) 10% TFA, MeOH, 70 °C.

Scheme 3 shows the new synthetic route used to prepare the truncated (N)-methanocarba derivatives 22–26. While the previous study of truncated (N)-methanocarba nucleosides as antagonists of the hA₃AR depended on a low yielding decarboxyation step, an alternative simple synthetic approach was designed. The protected cyclopentenone 38 was stereoselectively reduced to the corresponding alcohol 39,²⁸ which was then subjected to cyclopropanation to yield the glycosyl donor 40 for the truncated nucleosides and therefore converged on the previously reported route.²⁰ Compound 40 was subjected to a Mitsunobu condensation²⁰ with the appropriate C2 substituted 6-chloropurine derivative to give the 2-H (41) and 2-halopurine (42, 43) derivatives. The 2-1 derivative 44 was prepared in a similar manner and was substituted with 3-chlorobenzylamine at the 6 position to give 45, which serves as a precursor for 2-alkynyl derivatives in the truncated series. This route afforded the isopropylidene-protected truncated nucleosides 46–49, which were subsequently deprotected to provide compounds 22–25 for biological testing.

Scheme 3^a.

^a (i) NaBH₄, CeCl₃ ·7H₂O, MeOH, 0 °C; (ii) Et₂Zn, CH₂I₂, CH₂Cl₂, room temp; (iii) 2-substituted-6-chloropurine, Ph₃P, DIAD, THF, room temp; (iv) 3-iodo- or 3-chlorobenzylamine or cyclopentylamine, Et₃N, MeOH, room temp; (v) 10% TFA, MeOH, 70 °C; (vi) HC≡C(CH₂)₂COOMe, Pd(PPh₃)₂Cl₂, CuI, Et₃N, DMF, room temp; (vii) ethylenediamine, MeOH, room temp.

A truncated 2-iodo intermediate 45 was subjected to Sonogashira coupling²⁵ with methyl pentynate to give compound 50, which upon amination with ethylenediamine and subsequent acid hydrolysis provided the truncated nucleoside 26 containing an amine functionalized chain.

Pharmacological Activity

Several previously reported 2-chloro-(N)-methanocarba agonists (1a, 1b, 3, 4, 6, 9, and 12) and antagonists (2a–c) were used for comparison in the biological assays (Table 1). Binding assays at three hAR subtypes were carried out using standard radioligands^29–31 and membrane preparations from Chinese hamster ovary (CHO) cells (A₁ and A₃) or human embryonic kidney (HEK293) cells (A_2A) stably expressing a hAR subtype (Table 1).^32,33 Since activity within the class of (N)-methanocarba nucleosides was previously noted to be very weak or absent at the hA_2BAR,¹⁶ we did not include this receptor in the screening protocol. Functional data were determined in an assay of adenylate cyclase (A₃AR-induced inhibition).³⁴ Functional data for selected compounds in a functional assay consisting of A₃AR-induced stimulation of guanine nucleotide binding are reported in Table 1.³⁵

Table 1. Potency of a Series of (N)-Methanocarba Adenosine Derivatives at Three Subtypes of Human ARs and the Functional Efficacy at the A₃AR.

compd	structure		affinity (Ki, nM) or % inhibitiona			% efficacy,b A3
	R1	R2	A1	A2A	A3
Series A
1ac	Cl	3-Cl-Bn	260 ± 60h	2300 ±100	0.29 ± 0.04	103 ± 7
1b	Cl	3-I-Bn	136 ± 22d,h	784 ± 97d	1.5 ± 0.2c	100c
3c	H	3-I-Bn	700 ± 270h	6200 ± 100	2.4 ± 0.5	100
4c	Cl	cyclopentyl	18.3 ± 6.3h	3250 ± 300	3.7 ± 0.9	101
5	C≡C(CH2)2COOH	3-Cl-Bn	(17 ± 4%)	(24 ± 12%)	61.1 ± 35.8	106 ±18
6c,e	C≡C(CH2)3COOH	3-Cl-Bn	14,900 ± 3500h	(43%)	2.38 ± 0.56	ND
7	C≡C(CH2)4COOH	3-Cl-Bn	(11 ± 5%)	(38 ± 2%)	12.4 ± 1.8	74.4 ± 5.1
8	C≡C(CH2)2COOCH3	3-Cl-Bn	(41 ± 0%)	(43 ± 5%)	5.5 ± 0.3	90.0 ± 6.2
9c	C≡C(CH2)3COOCH3	3-Cl-Bn	482 ± 23h	(49%)	1.17 ± 0.27	ND
10	C≡C(CH2)4COOCH3	3-Cl-Bn	(20 ± 7%)	3540 ± 1370	11.1 ± 2.3	85.0 ± 10.2
11	C≡C(CH2)2CONH(CH2)2NH2	3-Cl-Bn	2320 ± 520	550 ± 83	2.5 ± 0.5	96.9 ± 6.8
12c	C≡C(CH2)3CONH(CH2)2NH2	3-Cl-Bn	454 ± 44h	(81%)	2.17 ± 0.51	ND
13	C≡C(CH2)4CONH(CH2)2NH2	3-Cl-Bn	(47 ± 3%)	277 ± 33	3.4 ± 0.7	100 ± 2
14	C≡C(CH2)2CONH(CH2)3NH2	3-Cl-Bn	(31 ± 9%)	979 ± 181	3.1 ± 0.4	97.9 ± 18.4
15	C≡C(CH2)2CONH(CH2)4NH2	3-Cl-Bn	(30 ± 2%)	766 ± 109	2.1 ± 0.4	102 ± 13
16	C≡C(CH2)2CO[NH(CH2)2]2NH2	3-Cl-Bn	(33 ± 2%)	890 ± 110	15.4 ± 4.4	87.6 ± 21.2
17a	C≡C(CH2)2CONH(CH2)2NH-biotin	3-Cl-Bn	(1 ± 1%)	(51 ± 2%)	36.4 ± 5.6	84.5 ± 12.0
17b	C≡C(CH2)2CONH(CH2)2NH-CO(CH2)5NH-biotin	3-Cl-Bn	(12 ± 4%)	(47 ± 11%)	57.7 ± 16.2	107 ± 18
18	C≡C(CH2)2CONH(CH2)2NH-CO-(CH2)5Cy5g	3-Cl-Bn	(36 ± 3%)	4730 ± 1020	17.2 ± 3.1	94.4 ± 9.6
Series B
19	C≡C(CH2)3COOH	3-Cl-Bn	(0%)	(9 ± 2%)	5270 ± 1090	ND
20	C≡C(CH2)2CONH(CH2)2NH2	3-Cl-Bn	(33 ± 7%)h	(8 ± 1%)	(30 ± 1%)	25.6 ± 10.9
21	C≡C(CH2)4CONH(CH2)2NH2	3-Cl-Bn	(36 ± 5%)h	(5 ± 1%)	4630 ± 1060	52.0 ± 9.4
Series C
2ad	Cl	3-Cl-Bn	3070 ± 1500h	4510 ± 910	1.06 ± 0.36	2.9 ± 3.7f
2bd,e	Cl	3-Br-Bn	1760 ± 1010h	1600 ± 480	0.73 ± 0.30	46 ± 4,5.8 ± 0.8f
2cd,e	Cl	3-I-Bn	3040 ± 610h	1080 ± 310	1.44 ± 0.60	44 ± 6, 1.0 ± 3.2f
22	H	3-Cl-Bn	1600 ± 150	4520 ± 830	4.9 ± 0.7	58.3 ± 8.0
23	H	3-I-Bn	839 ± 68	2330 ± 680	11.0 ± 2.2	80.8 ± 19.0
24	F	3-I-Bn	513 ± 4	4000 ± 780	10.7 ± 0.9	19.0 ± 3.0
25	Cl	cyclopentyl	109 ± 16	1640 ± 360	120 ± 31	95.1 ± 4.6
26	C≡C(CH2)2CONH(CH2)2NH2	3-Cl-Bn	(15 ± 2%)	(35 ± 6%)	404 ± 67	0.9 ± 8.5

^aAll experiments were done on CHO or HEK293 (A_2A only) cells stably expressing one of four subtypes of human ARs. 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]51, [³H]54, or [¹²⁵I]53, respectively), unless otherwise noted. A percent in parentheses refers to inhibition of radioligand binding at 10 μM.

^bUnless otherwise noted, the efficacy at the human A₃AR was determined by inhibition of forskolin-stimulated cAMP production in AR-transfected CHO cells. At a concentration of 10 μM, in comparison to the maximal effect of 51 (= 100%) at 10 μM. Data are expressed as mean ± standard error (n = 3). ND, not determined.

^cValues from Melman et al.; Lee et al.; Tchilibon et al.^15,16,21

^dValues from Melman et al.²⁰

^e2b, MRS5147; 2c, MRS5127; 6, MRS5151.^21,33

^fA₃AR functional assay consisted of stimulation of [³⁵S]GTPγS binding at 10 μM, expressed as a percentage of the full effect induced by 10 μM 51 (100 ± 5%).

^gStructure given in Scheme 1B.

^hA₁AR binding determined using [³H]52.

Compounds 5–15 in the (N)-methanocarba 5′-N-methyluronamide series (series A), which had functionalized 2-alkynyl chains terminating in a carboxylic acid, ester, or amino group, were potent A₃AR ligands. These derivatives having charged, polar, and chemically reactive groups at the end of a chain were intended for structural probing of distal regions of the ligand binding site of the receptor and for the design of specialized high-affinity ligands, such as fluorescent probes.

The optimal chain length for A₃AR affinity among carboxylic acid derivatives occurred with the previously reported compound 6,²¹ which had three methylenes in the chain. The corresponding ester derivative 9²¹ was also the most potent in the series of homologous esters 8–10. A shorter chain carboxylic acid derivative 5 was considerably less potent than other members of the series in binding to the hA₃AR.

The carboxylic acid derivatives were then extended by amide linkage to diaminoalkyl groups. In the series of resulting terminal amino derivatives 11–13, the length of the carboxyl alkyl chain was varied, and in the series of 11 and 14 and 15 the length of the diaminoalkyl chain was varied. While 12 displayed an atypically high A ₁AR affinity (K_i = 454 nM) and 13 an atypically high A_2AAR affinity (K_i = 270 nM), most of the members of these series were highly selective for the A₃AR. The K_i values at the hA₃AR of the primary amino derivatives 11–15 of varying chain length were in the range of 2–3 nM. The most potent and selective novel member of this series, an amine derivative 15 (Figure 1, Table 1), was 360 and > 5000-fold selective in binding to the hA₃AR in comparison to hA₁ and A_2AARs, respectively. Compound 16, containing two free amino groups in the chain, one primary and another secondary, was slightly reduced in affinity at the A₃AR with a K_i value of 15.4 nM.

Figure 1.

Inhibition of radioligand binding by (N)-methanocarba nucleoside analogues in membranes of CHO cells expressing the human A₃AR. Inhibition curves are shown for the agonist analogues containing a 5′-N-methyluronamide group and compounds 15 (amine functionalized congener), 17b (biotinylated probe), and 18 (fluoresecent probe) and for the truncated partial agonist 22. All of the analogues shown were highly selective for the A₃AR in comparison to the A₁ and A_2aARs.

As functionalized congeners, these long chain derivatives were intended for conjugation to carrier moieties for receptor detection and characterization, in a manner that did not prevent receptor recognition.^22,23 A biotin conjugate 17b derived from amine 11 in the series of 5′-N-methyluronamide derivatives displayed a K_i value at the hA₃AR of 58 nM (Figure 1) and selectivity of > 170-fold and 66-fold in comparison to the A₁ and A_2AAR, respectively. A shorter analogue 17a lacking the ε-aminocaproyl spacer chain was slightly more potent in hA₃AR binding with a K_i value of 36 nM. A fluorescent conjugate 18 containing the sulfonated carbocyanine dye Cy5²⁷ was potent and selective in A₃AR binding. The fluorescent excitation and emission maxima of 18 in aqueous medium, 646 and 660 nm, respectively, were characteristic of Cy5 dye conjugates.

Various compounds were examined for the ability to inhibit 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 (51, 5′-N-ethylcarboxamidoadenosine) was set at 100% relative efficacy. The (N)-methanocarba 5′-N-methyluronamide derivatives (series A) were consistently full agonists. Full concentration response curves for compounds 17a and 18 provided IC₅₀ values in hA₃AR-mediated inhibition of adenylate cyclase of 2.50 ± 0.42 and 1.09 ± 0.28 nM, respectively (Figure 2A). Thus, this series provided many new potent and selective A₃AR agonists, and alkynyl chain derivatization at the C2 position appears to be an effective strategy for covalent attaching sterically bulky groups while maintaining A₃AR recognition and activation.

Figure 2.

Functional agonism by the tested in an assay of adenylate cyclase in membranes of CHO cells expressing the hA₃AR. Each experiment was repeated three times. (A) Activity of the 5′-N-methyluronamide (N)-methanocarba analogues 17a and 18. The EC₅₀ values of 17a and 18 were 2.50 ±0.42 and 1.09 ±0.28 nM, respectively. (B) Activity of the truncated (N)-methanocarba analogues 2b and 2c at the hA₃AR. The full agonist 51 was included for comparison (representing 100% efficacy). The percent relative efficacy of 2b and 2c was 46 ± 4% and 44 ± 6%, respectively. The EC₅₀ values of 2b and 2c were 4.2± 0.6 and 12± 1 nM, respectively.

A series of l-nucleosides was prepared as enantiomers of members of the potent agonist functionalized congeners (series B). Compounds 19, 20, and 21 were the enantiomers of the potent A₃AR agonists 5, 11, and 13, respectively. We confirmed that at submicromolar concentrations no substantial AR binding activity was seen with these l-enantiomers, confirming stereoselectivity of binding at the ARs. At 10 μM, 19–21 were weak or inactive in binding at the A₁ and A_2AARs, and at the A₃AR they tended to be slightly more potent in binding. These l-nucleosides induced partial activation of the A₃AR at 10 μM, as confirmed in the cAMP functional assay in which compounds 20 and 21 activated to the degree of 26% and 52%, respectively. Nevertheless, these l-nucleosides can be used at nanomolar concentrations in pharmacological studies as inactive control compounds of similar physicochemical properties, by analogy to the use of inactive enantiomers of other GPCRs.³⁶ At these concentrations, interactions with ARs were lacking. In the truncated 4′-thionucleoside class of A₃AR antagonists, stereospecificity of binding was similarly demonstrated.⁴⁵

The effects of modification at the C2 position were explored in the truncated series C. In the 5′-N-methyluronamide derivatives (series A), even the presence of a long chain at the C2 position did not interfere with the binding and activation of the A₃AR. However, in the series of nucleosides that were truncated at the 4′-carbon, variability in both binding affinity and efficacy at the hA₃AR was observed, depending on the nature of the C2 substitution. Compounds 22–24, lacking a functionalized chain at the C2 position, were potent and selective A₃AR ligands with binding K_i values of 5–11 nM. Removal of the 2-C1 substitution in N⁶-benzyl-type 4′-truncated derivatives (e.g., 22, 23) or its replacement with 2-F (e.g., 24) maintained A₃AR selectivity but reduced the A₃ AR affinity by 5- to 10-fold. Replacement of 2-C1 with 2-H either had no effect (cf. 2a and 22) on affinity at the A_2AAR or moderately reduced it (cf. 2c and 23), and affinity at the A₁AR was moderately increased. The most A₃AR selective truncated compound was the 2-unsubstituted N⁶-(3-chlorobenzyl) derivative 22 with 330-fold and 940-fold selectivity in comparison to the A₁ and A_2AAR, respectively.

An amine functionalized congener 26 in the truncated series, with a K_i value of 403 nM, proved to be weaker in binding to the A₃AR than those analogues with less sterically bulky substitution at the C2 position. However, it nevertheless displayed moderate selectivity of > 20-fold in comparison to both the A₁ and A_2AARs. This derivative was analogous to compound 11 in the 5′-N-methyluronamide series, which was 160-fold more potent in binding to the A₃AR.

A 5′-truncated N⁶-cyclopentyl derivative 25, by analogy to compound 4,³⁷ was intended to display enhanced affinity at both A₁ and A₃ARs but not at the A_2AAR. However, the loss of the 5′-N-methyluronamide group reduced affinity at A₁ and A₃ARS by 6- and 32-fold, respectively. Thus, N⁶-benzyl substitution of the (N)-methanocarba nucleosides preserved the AR selectivity profile upon truncation better than an N⁶-cyclopentyl group.

Selected compounds in the present series were tested in binding at the rA₃AR. The K_i values (nM) were determined to be the following: 6, 66.5 ± 3.9; 15, 8.68 ± 0.85; 18, 9.62 ± 1.50; 22, 20.3 ± 2.0. Thus, the high A₃AR affinity was maintained in a murine species.

The A₃AR efficacy in the cAMP functional assay varied greatly in this series of truncated (N)-methanocarba derivatives. Compounds 2b and 2c, which were reported as A₃AR antagonists based on an assay of [³⁵S]GTPγS binding,²⁰ were partial agonists in the assay of adenylate cyclase (Figure 2B). Both partial (e.g., 22) and nearly full agonism (e.g., 23 and 25) were observed for the novel truncated nucleosides in the cyclase assay. Replacement of the 2-Cl of the truncated analogues 2b with hydrogen in 22 resulted in a 58% relative efficacy, which was comparable to 46% for 2b. The 2-H 23 and 2-F 24 N⁶-(3-iodobenzyl) analogues displayed relative efficacy of 81% and 19%, respectively. Compound 25 inhibited the production of cAMP via the hA₃AR to the same extent as 51 and was therefore a full agonist as judged by adenylate cyclase inhibition.

A homology model of the hA₃AR based on the A_2AAR X-ray structure^38,39 was used to study the putative interactions between the chain-functionalized agonist 15 and the A₃AR. The binding mode of 15 obtained after InducedFit docking revealed the following binding mode of the ligand (Figure 3). The 2′- and 3′-hydroxyl groups were located in proximity to Ser271 (7.42) and His272 (7.43) and could form H-bonds with these residues. The oxygen atom of the 5′-N-methylcarboxamido moiety was involved in H-bonding with the side chain hydroxyl group of Thr94 (3.36). Both N⁶-amino group and the N7 atom of the adenine ring formed H-bonds with Asn250 (6.55). A π−π interaction was observed between the adenine ring of 15 and Phe168 (EL2). 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.^38,40 In the model obtained, the 3-chlorobenzyl ring was located in the hydrophobic pocket formed by Val169 (EL2). Met172(EL2), Met174(EL2), Met177(5.38), and Ile253(EL3).

Figure 3.

Binding mode of amine congener 15 in the 5′-N-methyluronamide series obtained after InducedFit docking to the A₃AR homology model built based on the crystal structure of the A_2AAR.

The long chain substituent at the C2 position of 15 was oriented toward the extracellular part of the receptor, which is a prerequisite in functionalized congeners.²³ A similar orientation of the 2-propynyl fragment of 15 was previously suggested for C2 substituted agonists of the A_2B receptor.^41,42 The carbonyl oxygen of the amide group of the C2 substituent of 15 was H-bonded to the side chain NH₂-group of Gln 167 (EL2), and consequently the NH-group of the C2 substituent of 15 was oriented toward the hydroxyl group of Tyr265 (7.36) (Figure 3). The terminal amine of the ligand interacted with negatively charged Glu258 (EL3), and it was also H-bonded to the backbone oxygen atom of Val259 (EL3).

Discussion

In the present work and in a previous study,²⁰ parallel patterns of SAR in receptor binding were observed in the two series of 5′-CONHCH₃ (full agonist) and truncated (N)-methanocarba nucleosides, in cases with substituted N⁶-benzyl groups and small substituents present at the adenine C2 position. This parallel behavior was not maintained in a derivative bearing a chain at the C2 position.

The replacement of the adenine 2-Cl with H in the series of substituted N⁶-benzyl-(N)-methanocarba agonists containing the 5′-CONHCH₃ group was previously shown to have only minor effects on AR binding.¹⁶ Specifically, it moderately reduced affinity at the A₁ and A_2AARs without a significant reduction at the A₃AR (cf. 1a and 3). In the present 5′-N-methyluronamide series with more elaborate modification at the C2 position, the preservation of potent and selective binding to the hA₃AR was demonstrated.

The flexibility of chain substitution at C2 allowed the conjugation with sterically bulky moieties. This distal region of the nucleosides was predicted by homology modeling to be at A₃AR extracellular regions and consequently free from the steric constraints present at the pharmacophore binding region. Thus, this series of highly potent and selective A₃AR agonist functionalized congeners is suitable for conjugation to prosthetic groups for radiolabeling, fluorescent detection, avidin complexation, and polymeric coupling with retention of biological activity. The functionalized congener approach was already applied to 1,4-dihydropyridine derivatives as antagonists of the hA₃AR.²⁶

Specifically, agonists in this series were conjugated with biotin and a fluorescent dye with retention of high A₃AR affinity and selectivity. The fluorescent conjugate 18 was a highly potent, full agonist with a K_i values at the hA₃AR and the rA₃AR of 17.2 and 9.6 nM, respectively. The fluorescent dye Cy5, which was incorporated covalently in 18, has favorable properties for use in biological systems including near-infrared imaging, with excitation and emission maxima at 649 and 670 nm, respectively, and with a relatively high quantum yield of 0.28.²⁷ Another fluorescent agonist of the hA₃AR having lower affinity than 18 was used in fluorescence correlation spectroscopy to demonstrate that the receptor exists in heterogeneous microdomains of individual living cells.⁴³

In our previous study of truncated (N)-methanocarba nucleosides,²⁰ the docking of two representative structures in a rhodopsin-based homology model of the hA₃AR indicated that both (N)-methanocarba agonists and nucleoside-derived antagonists were able to bind in a very similar mode in the receptor binding site, which is common to other A₃AR-selective nucleosides, such as the prototypical agonist 2-chloro-N⁶-(3-iodobenzyl)-5′-N-methylcarboxamidoadenosine (Cl-IB-MECA).²⁰ We have shown that the SAR parallels between these two series were less stringent with respect to distal modification, e.g., when a long functionalized chain was attached at the C2 position. Thus, in the truncated series the attachment of a functionalized amino chain at the C2 position greatly reduced A₃AR affinity. Therefore, among the 2-alkynyl nucleosides in this study, only those containing a 5′-N-methyluronamide group, i.e., full agonists, were suitable as functionalized congeners.

Truncation of the 5′-CONHCH₃ moiety of A₃AR agonists in the (N)-methanocarba class was again demonstrated as a means of reducing the relative efficacy of A₃AR-selective agonists to produce partial agonism, as indicated in the cyclase assay. A previous study of 2-triazolo derivatives of adenosine demonstrated modulation of the relative efficacy at the A₃AR by structural changes at the 2 position.^12c In the truncated series, the remaining efficacy at the hA₃AR, as was evident in a cAMP functional assay, ranged from small (e.g., 24 and 26) to substantial (e.g., 23).

We compared selected compounds in the ability to stimulate the A₃AR using multiple functional criteria. When stimulation of [³⁵S]GTPγS binding was used as a criterion of receptor activation, truncated (N)-methanocarba nucleosides (e.g., 2a–c) reported by Melman et al.²⁰ did not induce a substantial activation of the A₃AR, with < 10% of the full agonism. However, in the current study using a different functional assay that measured inhibition of the production of cAMP to detect antagonism, considerable activity in the truncated series was seen. The efficacy in the cyclase assay generally exceeded the efficacy in [³⁵S]GTPγ S binding. Other truncated nucleosides in the 4′-thio series were A₃ AR antagonists in both [³⁵S]GTPγS and cAMP assays.⁴⁴ Recently, truncated nucleosides in the 4′-oxo series were demonstrated to be A₃ AR antagonists in the cAMP assay.⁴⁵

In the truncated series, removal of the 2-Cl substitution or replacement with 2-F in 24 moderately reduced the A₃AR affinity but retained high selectivity. The introduction of fluorine in 24 was intended for possible radiofluorination as a means of preparing tracers for in vivo PET imaging of the A₃AR.⁴⁷ One member of the previously reported truncated (N)-methanocarba nucleoside series 2b, (1R,2R,3S,4R,5S)-4-(6-(3-bromobenzylamino)-2-chloro-9H-purin-9-yl)bicyclo-[3.1.0]hexane-2,3-diol (MRS5147), was recently labeled with the short-lived positron-emitter ⁷⁶Br to form a selective hA₃AR radioligand of high affinity for use in positron emission tomography (PET). This ligand was intended for imaging of the A₃ AR in murine as well as primate species, because high affinity binding was maintained at the rA₃AR. Binding of these truncated (N)-methanocarba nucleosides at the A₃AR was competitive,²⁰ and therefore, the related derivatives in this study should be useful as receptor probes.

In conclusion, two highly potent and selective series of nucleoside ligands containing the (N)-methanocarba ring system have been synthesized and characterized as A₃AR agonists, partial agonists, and antagonists. Furthermore, in the 5′-N-methyluronamide series but not in the truncated series, the functionalized congener approach was found to maintain potency and selectivity at the hA₃AR and therefore provides an entry into high affinity molecular probes. Thus, the binding and activation profiles of the truncated analogues indicated the biological interdependence of substitution at the C2, 5′, and N⁶ positions. Potent agonist ligands containing biotin and fluorescent dye moieties should be useful in future pharmacological studies of the localization, association, regulation, and internalization of the A₃AR and might provide an alternative to the use of radioligands for this receptor.

Experimental Section

Chemical Synthesis. Materials and Instrumentation

d-Ri-bose, l-ribose, and other reagents and solvents were purchased from Sigma-Aldrich (St. Louis, MO). Alcohol derivative 31 was prepared as reported.²⁰¹ H NMR spectra were obtained with a Varian Gemini 300 spectrometer using CDCl₃ and CD₃OD as solvents. Chemical shifts are expressed in δ values (ppm) with tetramethylsilane (δ 0.00) for CDCl₃ and water (δ 3.30) for CD₃OD. TLC analysis was carried out on aluminum sheets precoated with silica gel F₂₅₄ (0.2 mm) from Aldrich. All target compounds are ≥95% pure as determined by HPLC. HPLC mobile phases consisted of the following: for system A, linear gradient solvent system of CH₃CN/triethyl ammonium acetate from 5/95 to 60/40 in 20 min and flow rate of 1.0 mL/min; for ystem B, linear gradient solvent system of CH₃CN/tetrabutyl ammonium phosphate from 20/80 to 60/40 in 20 min and flow rate of 1.0 mL/min. 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 otherwise noted. Observed mass accuracies are those expected on the basis of 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.

(1′S,2′R,3′S,4′S,5′S)-4′-[6-(3-Chlorobenzylamino)-2-(4-methoxycarbonyl-l-butynyl)-9-yl]-2′,3′-0-isopropylidenebicyclo[3.1.0]-hexane-1′-carboxylic Acid N′-Methylamide (28)

To a solution of compound 27 (52 mg, 0.087 mmol) in anhydrous DMF (1.5 mL), PdCl₂(PPh₃)₂ (12 mg, 0.017 mmol), CuI (2 mg, 0.010 mmol), methyl-ω-pentynate (39 mg, 0.347 mmol), and then triethylamine (0.12 mL, 0.86 mmol) were added. The reaction mixture was stirred at room temperature overnight. Solvent was evaporated under vacuum, and the residue was purified on flash silica gel column chromatography (CH₂Cl₂/MeOH = 30:1) to give compound 28 (40 mg, 78%) as a foamy syrup. ¹H NMR (CD₃OD, 300 MHz) δ 8.16 (s, 1H), 7.42–7.44 (m, 4H), 5.81 (d, J =6.7 Hz, 1H), 5.02 (s, 1H), 4.82 (m, 2H), 3.78 (s, 3H), 2.85 (s, 3H), 2.72–281 (m, 5H), 2.09–2.14 (m, 1H), 1.55 (s, 3H), 1.41 (t, J=5.1 Hz, 1H), 1.29 (s, 3H), 0.82–0.96 (m, 1H). HRMS calculated for C₂₉H₃₂ClN₆O₅ (M + H)⁺: 579.2105: found 579.2123.

(1′S,2′R,3′S,4′S,5′S)-4′-[6-(3-Chlorobenzylamino)-2-(4-(β-aminoethyl aminocarbonyl)-1-butynyl)-9-yl]-2′,3′-dihydroxybicyclo [3.1.0]hexane-1′-carboxylic Acid N-Methylamide (11)

To a solution of compound 28 (20 mg, 0.034 mmol) in methanol (0.3 mL), ethylenediamine (1.5 mL) was added. The mixture was stirred overnight at room temperature. Solvent was evaporated, and the residue was roughly purified on flash silica gel column chromatography. The aminated product was dissolved in methanol (1.5 mL) and 10% trifluoroacetic acid (1.5 mL) and heated at 70 °C for 15 h. Solvent was evaporated and the residue was purified on flash silica gel column chromatography (CH₂Cl₂/MeOH/NH₄OH = 10:1:0.1) to give compound 11 (15 mg, 79%). ¹H NMR (CD₃OD, 300 MHz) δ 8.09 (s, 1H), 7.39 (s, 1H), 7.27–7.32 (m, 3H), 4.99 (d, J = 6.6 Hz, 1H), 4.78– 4.91 (m, 2H), 3.97 (d, J = 6.6 Hz, 1H), 3.33–3.36 (m, 4H), 2.87 (s, 3H), 2.75–2.80 (m, 4H), 2.53 (t, J = 7.2 Hz, 2H), 2.59–3.41 (m, 1H), 1.88 (t, J = 4.5 Hz, 1H), 1.34–1.43 (m, 1H). HRMS calculated for C₂₇H₃₁ClN₆O₂Na (M + Na)⁺: 589.2068; found 589.2054.

(1′S,2′R,3′S,4′S,5′S)-4′-[6-(3-Chlorobenzylamino)-2-(4-methoxycarbonyl)-1 -butyn yl)-9-yl]- 2′,3′ -dihydroxybicyclo [3.1.0] hexane-1′-carboxylic Acid N-Methylamide (8)

A solution of compound 28 (67 mg, 0.115 mmol) in methanol (3 mL) and 10% triflromethanesulfonic acid (2 mL) was heated at 70 °C overnight. Solvent was evaporated and the residue was purified on flash silica gel column chromatography (CH₂Cl₂/MeOH = 20:1) to give compound 8 (46 mg, 75%). ¹H NMR (CD₃OD, 300 MHz) δ 8.13 (s, 1H), 7.40 (s, 1H), 7.22–7.45 (m, 3H), 5.10 (d, J = 6.3 Hz, 1H), 4.80–4.86 (m, 2H), 4.03 (d, J = 6.6 Hz, 1H), 3.63 (s, 3H), 2.86 (s, 3H), 2.76–2.85 (m, 2H), 2.46–2.58 (m, 2H), 2.05–2.10 (m, 1H), 1.77–1.83 (m, 1H), 1.35–1.40 (m, 1H), 0.82–0.96 (m, 1H). HRMS calculated for C₂₆H₂₈ClN₆O₅ (M + H)⁺: 539.1731; found 539.1743.

(1′S,2′R,3′S,4′S,5′S)-4′-[6-(3-Chlorobenzylamino)-2-(4-hydroxycarbonyl)-l-butynyl)-9-yl]-2′,3′-dihydroxybicyclo[3.1.0]hexane-1′-carboxylic Acid N′-Methylamide (5)

To a solution of ester 8 (30 mg, 0.055 mmol) in methanol (1.5 mL), 1 M solution of potassium hydroxide (1 mL) was added. The mixture was stirred at room temperature overnight. The reaction mixture was neutralized with acetic acid, and solvent was evaporated, and the residue was purified on flash silica gel column chromatography (CH₂Cl₂/MeOH = 10:1) to give compound 5 (23 mg, 80%). ¹H NMR (CD₃OD, 300 MHz) δ 8.06 (s, 1H), 7.38 (s, 1H), 7.24–7.35 (m, 3H), 5.14 (d, J= 5.8 Hz, 1H), 4.79–4.86 (m, 2H), 4.03 (d, J = 6.3 Hz, 1H), 2.87 (s, 3H), 2.42–2.80 (m, 4H), 2.04– 2.32 (m, 1H), 1.82 (t, J =4.8 Hz, 1H), 1.36–1.40 (m, 1H), 0.84– 0.97 (m, 1H). HRMS calculated for C₂₅H₂₆ClN₆O₅ (M + H)⁺: 525.9562; found 525.9583.

(1′S,2′R,3′S,4′S,5′S)-4′-[6-(3-Chiorobenzylamino)-2-(N-biotinyl(β-aminoethylaminocarbonyl)-1-butynyl)-9-yl]-2′, 3′-dihydroxybicyclo[3.1.0]hexane-1′-carboxylic Acid N-Methylamide (17a)

To a solution of compound 11 (4 mg, 0.007 mmol) in dry DMF (0.5 mL), biotin (1.89 mg, 0.0077 mmol), HATU (3.2 mg, 0.0084 mmol), and DIEA (1.6 μL, 0.009 mmol) were added. The mixture was stirred at room temperature overnight. Solvent was evaporated and the residue was purified on flash silica gel column chromatography (CH₂Cl₂/MeOH/NH₄OH =5:1:0.1) to give compound 17a (3.5 mg, 62%). ¹H NMR (CD₃OD, 300 MHz) δ 8.14 (s, 1H), 7.43 (s, 1H), 7.29–7.39 (m, 3H), 5.07 (d, J = 6.9 Hz, 1H), 4.78–4.86 (m, 2H), 4.48–4.52 (m, 1H), 4.28–4.32 (m, 1H), 4.04 (d, J =6.3 Hz, 1H), 3.70–3.82 (m, 2H), 3.24–3.31 (m, 1H), 3.08–3.23 (m, 1H), 2.91 (s, 3H), 2.70–2.83 (m, 4H), 2.56 (t, J =7.5 Hz, 2H), 2.04–2.15 (m, 2H), 1.89 (t, J =5.1 Hz, 1H), 1.33–1.70 (m, 10H), 0.84–1.02 (m, 1H). HRMS calculated for C₃₇H₄₆ClN₁₀O₆S (M + H)⁺: 793.3011; found 793.3030.

(1′S,2′R,3′S,4′S,5′S)4′-[6-(3-Chlorobenzylamino)-2-(N-biotinyl{5-aminopentanyl}(β-aminoethylaminocarbonyl)-1-butynyl)-9-yl]-2′,3′-dihydroxybicyclo[3.1.0]hexane-1′-carboxylic Acid N-Methylamide (17b)

To a solution of compound 11 (3.4 mg, 0.0059 mmol) in DMF (0.5 mL), sulfo-NHS-LC-biotin (10 mg, 0.017 mmol, Pierce) and 10 μL of triethyl amine were added. The mixture was stirred overnight at room temperature. Solvent was evaporated and the residue was purified on flash silica gel column chromatography (CH₂Cl₂/MeOH/NH₄OH = 7:1:0.1) to give compound 17b (2.9 mg, 55%). ¹H NMR (CD₃OD, 300 MHz) δ 7.99 (s, 1H), 7.29 (s, 1H), 7.19–7.32 (m, 3H), 4.93 (d, J = 5.4 Hz, 1H), 4.70–4.80 (m, 2H), 4.36–4.42 (m, 1H), 4.16–4.22 (m, lH), 3.88 (d,J = 6.9 Hz, 1H), 3.42–3.62 (m, 2H), 3.03–3.15 (m, 4H), 2.89 (s, 3H), 2.56–2.86 (m, 6H), 2.41 (t, J = 6.6 Hz, 1H), 1.96–2.13 (m, 5H), 1.15–1.16 (m, 14H), 0.74–0.88 (m, 1H). C₄₃H₅₇C1N₁₁O₇S (M+H)⁺: 906.3852; found 906.3878.

(1′S,2′R,3′S,4′S,5′S)-4′-[6-(3-Chlorobenzylamino)-2-(N-cyaiune(β-aminoethylaminocarbonyl)-1-butynyl)-9-yl]-2′,3′-dihydroxybicyclo[3.1.0]hexane-1′-carboxylic Acid N-Methylamide (18)

To a solution of compound 11 (1.69 mg, 0.0029 mmol) in DMF (0.3 mL), Cy5 fluorescent dye (2.36 mg, 0.0029 mmol) and bicarbonate buffer (60 μL) were added. The mixture was stirred at room temperature overnight. The reaction mixture was covered with aluminum foil in order to protect from light. Solvent was evaporated and the residue was purified on flash silica gel column chromatography (CH₂Cl₂/MeOH/NH₄OH = 3:1:0.1) to give compound 18 (3.2 mg, 89%) as a dark-blue syrup. ¹H NMR (CD₃OD, 300 MHz) δ 8.21–8.42 (m, 1H), 8.08 (s, 1H), 7.85–7.93 (m, 2H), 7.29–7.46 (m, 6H), 6.28–6.45 (m, 1H), 5.07 (d, J = 6.9 Hz, 1H), 4.84–4.86 (m, 2H), 3.97–4.20 (m, 3H), 3.53 (t,J = 6.0Hz, 2H), 3.12 (t,J = 6.0Hz, 2H), 2.89 (s, 3H), 2.49 (t,J =7.5 Hz, 2H), 2.52–2.64 (m, 2H), 2.01–2.17 (m, 2H), 1.56–1.89 (m, 9H), 1.22–1.49 (m, 10H), 0.86–0.96 (m, 1H). HRMS calculated for C₆₀H₆₈ClIN₁₀ O₁₁S₂ (M ⁺): 1203.4199; found 1203.4175.

(1′R,2′S,3′R,4′R,5′R)-4′-(6-Chloro-2-iodopurin-9-yl)-2′,3′-isopropylidenebicyclo[3.1.0]hexane-1′-carboxylic Acid Ethyl Ester (32)

To a solution of triphenylphosphine (0.357 g, 1.361 mmol) and 6-chloro-2-iodopurine (0.286 g, 1.019 mmol) in dry THF (4 mL), DIAD (0.26 mL, 1.359 mmol) was added. The mixture was stirred at room temperature for 10 min. A solution of compound 31 (0.165 g, 0.681 mmol) in THF (2 mL) was added to the reaction mixture, and the mixture was further stirred overnight at room temperature. Solvent was evaporated and the residue was purified on flash silica gel column chromatography (hexane/ethyl acetate = 3:1) to give the product 32 (0.2 g, 58%) as a foam. ¹H NMR (CDCl₃, 300 MHz) δ 1.31–1.37 (m, 4H), 1.56 (s,6H), 1.77–1.80 (m, 1H), 2.10–2.22 (m, 1H), 4.19–4.27 (m, 2H), 4.77 (d, J = 14.2 Hz, 1H), 4.91 (s, 1H), 5.83 (d, J = 15.3 Hz, 1H),7.97(s, 1H). HRMS calculated for C₁₇H₁₉ClIN₄O₄ (M + H)⁺: 505.0061; found 505.0073.

(1′R,2′S,3′R,4′R,5′R)-4′-[6-(3-Chlorobenzylamino)-2-iodopurin-9- yl]-2′,3′-isopropylidenebicyclo[3.1.0]he xane-1′-carboxylic Acid Ethyl Ester (33)

To a solution of compound 32 (0.273 g, 0.54 mmol) in anhydrous methanol (5 mL), triethylamine (1 mL, 7.569 mmol) and 3-chlorobenzylamine (0.33 mL, 2.697 mmol) were added. The mixture was stirred overnight at room temperature. Solvent was evaporated and the residue was purified on flash silica gel column chromatography (hexane/ethyl acetate = 1:1) to give the product 33 (0.255 g, 78%) as a foamy solid. ¹H NMR (CDCl₃, 300 MHz) δ 7.57 (s, 1H), 7.20–7.25 (m, 4H), 6.03–6.15 (m, 1H), 5.83 (d, J = 4.45 Hz, 1H), 5.28 (s, 1H), 4.65– 4.88 (m, 3H), 4.11–4.22 (m, 2H), 2.18–2.23 (m, 1H), 1.22–1.80 (m, 7H), 1.10–1.21 (m, 4H). HRMS calculated for C₂₄H₂₆ClI-N₅O₄ (M + H)⁺: 610.0639; found 610.0649

(1′R,2′S,3′R,4′R,5′R)-4′-[6-(3-Chlorobenzylamino)-2-iodopurin-9-yl]-2′-,3′-isopropylidenebicyclo[3.1.0]hexane-1′-carboxylic Acid-N-Methylamide (34)

To a solution of compound 33 (0.258 g, 0.423 mmol) in methanol (5 mL), 40% methylamine (4 mL) was added. The mixture was stirred at room temperature for 2 days. Solvent was evaporated and the residue was purified on flash silica gel column chromatography (CH₂Cl₂/MeOH = 40:1) to give the product 34 (0.165 g, 66%) as a foam. ¹H NMR(CDCl₃, 300 MHz) δ 7.58 (s, 1H), 7.25–7.37 (m, 4H), 6.52 (bs, 1H), 5.65 (d, J = 6.9 Hz, 1H), 4.73–4.83 (m, 3H), 2.97 (s, 3H), 2.04–2.07 (m,1H), 1.62–1.82 (m, 1H), 1.55 (s, 3H), 1.22–1.33 (m, 5H). HRMS calculated for C₂₃H₂₅ClIN₆O₃ (M + H)⁺: 595.0643: found 595.0649.

(1′R,2′S,3′R,4′R,5′R)-4′-[6-(3-Chlorobenzylamino)-2-(4-hydroxycarbonyl)-1-butynyl)-9-yl]-2′,3′-dihydroxybicyclo[3.1.0]hexane-1′-carboxylic Acid N-Methylamide (19)

To a solution of compound 34 (32 mg, 0.053 mmol) in anhydrous DMF (1.0 mL), PdCl₂(PPh₃)₂ (7 mg, 0.01 mmol), CuI (2 mg, 0.01 mmol), methyl-ω-hexynate (27 mg, 0.214 mmol), and then triethylamine (0.014 mL, 0.10 mmol) were added. The reaction mixture was stirred at room temperature overnight. Solvent was evaporated under vacuum, and the residue was roughly purified on flash silica gel column chromatography. To a solution of the product in methanol (1 mL), 10% TFA (1 mL) was added, and the mixture was heated at 70 °C for 7 h. Solvent was evaporated under vacuum, the crude reaction mixture was further dissolved in methanol (1 mL), and 1 M solution of KOH (1 mL) was added, and the mixture was stirred at room temperature overnight. Reaction mixture was neutralized with acetic acid, solvent was evaporated under vacuum, and the residue was purified on flash silica gel column chromatography (CH₂Cl₂/MeOH = 20:1) to give compound 19 (18 mg, 62%) as a colorless powder. ¹H NMR (CD₃OD, 300 MHz) δ 7.97 (s, 1H), 7.30 (s, 1H), 7.14– 7.22 (m, 3H), 4.97 (d, J = 6.7 Hz, 1H), 4.73 (brs, 3H), 3.90 (d, J = 6.6 Hz, 1H), 2.75 (s, 3H), 2.42 (t, J = 7.2 Hz, 2H), 2.28 (t, J = 7.2 Hz, 2H), 1.96 (m, 1H), 1.83 (m, 3H), 1.75 (m, 1H), 1.25 (m, 1H). HRMS calculated for C₂₆H₂₈ClN₆O₅ (M + H)⁺: 539.1731: found 539.1724.

(1R,2R,3S,4S,5S)-2,3-O-Isopropylidene-2,3,4-trihydroxybicyclo[3.1.0]hexane (40)

To a solution of compound 39 (0.773 g, 4.95 mmol) in dry CH₂Cl₂ (25 mL), Et₂Zn (19.8 mL, 1 M solution in hexane) was added at 0 °C. The mixture was stirred for 15 min. After 15min, CH₂I₂ (3.2 mL, 39.57 mmol) was added at the same conditions and then the reaction mixture was stirred at room temperature overnight. After completion of reaction, it was quenched with saturated ammonium chloride solution and the aqueous layer was extracted with CH₂Cl₂. The combined organic layer was dried (Na₂SO₄), filtered, and evaporated. The residue was purified on flash silica gel column chromatography (hexane/ethyl acetate = 3:1) to give compound 40 (0.710 g, 85 %) as a colorless syrup. ¹H NMR (CDCl₃, 300 MHz) δ 4.88 (t, J = 6.3 Hz, 1H), 4.46 –4.54 (m, 2H), 2.37 (d, J = 9.3 Hz, 1H), 1.83 – 1.86 (m, 1H), 1.61–1.66 (m, 1H), 1.55 (s,3H), 1.29 (s,3H), 0.95– 0.99 (m, 1H), 0.62–0.68 (m, 1H). HRMS calculated for C₉H₁₅O₃ (M + H)⁺: 171.0943; found 171.0954.

(1′R,2′R,3′S,4′S,5′S)-4′-(6-Chloro-2-iodopurine)-2′,3′-O-isopropylidene-2′,3′-dihydroxybicyclo[3.1.0]hexane (44)

To a solution of triphenylphosphine (132 mg, 0.5 mmol) and 6-chloro-2-iodopurine (141 mg, 0.50 mmol) in dry THF (2 mL), DIAD (0.1 mL, 0.50 mmol) was added. The mixture was stirred at room temperature for 10 min. A solution of compound 31 (43 mg, 0.252 mmol) in THF (2 mL) was added to the reaction mixture, and the mixture was further stirred overnight at room temperature. Solvent was evaporated and the residue was purified on flash silica gel column chromatography (hexane/ethyl acetate = 3:1) to give the product 44 (65 mg, 61%) as a foam. ¹H NMR (CDCl₃, 300 MHz) d 8.07 (s, 1H), 5.37 (t, J = 2.1 Hz, 1H), 4.67 (d, J = 5.7Hz, 1H), 2.13–2.18 (m, 1H), 1.65–1.69 (m, 2H), 1.27 (s, 3H), 1.26 (s, 3H), 0.96–1.01 (m, 2H). HRMS calculated for C₁₄H₁₅ClIN₄O₂ (M + H)⁺: 432.9927; found 432.9928.

(1′R,2′R,3′S,4′S,5′S)-4′-[6-(3-Chlorobenzylamino)-2-iodopurine]-2′,3′- O-isopropylidene-2′,3′-dihydroxybicyclo[3.1.0]hexane (45)

To a solution of compound 44 (0.105 g, 0.243 mmol) in anhydrous methanol (5 mL), triethylamine (0.4 mL, 3.38 mmol) and 3-chlorobenzylamine (0.15 mL, 1.21 mmol) were added. The mixture was stirred overnight at room temperature. Solvent was evaporated and the residue was purified on flash silica gel column chromatography (hexane/ethyl acetate = 1:1) to give the product 45 (0.096 g, 74%) as a foam. ¹H NMR (CDCl₃, 300 MHz) δ 7.63 (s, 1H), 7.37 (s, 1H), 7.27–7.26 (m, 3H), 6.25 (bs, 1H), 5.34 (t, J = 5.7Hz, 1H), 4.94 (s,2H), 4.64 (d, J =6.9Hz, 1H), 2.06–2.09 (m, 1H), 1.61–1.69 (m,2H), 1.54(s,3H), 1.26(s,3H), 0.88–0.96 (m, 2H). HRMS calculated for C₂₁H₂₂ClIN₅O₂ (M + H)⁺: 538.0497; found 538.0507.

(1′R,2′R,3′S,4′ S,5′ S)-4′-[6-(3-Iodobenzylamino)purine]-2′,3′-dihydroxybicyclo[3.1.0]hexane (23)

A solution of compound 46 (32 mg, 0.063 mmol) in MeOH (2 mL) and 10% trifluoroacetic acid (1.5 mL) was heated at 70 °C overnight. Solvent was evaporated and the residue was purified on flash silica gel column chromatography (CH₂Cl₂/MeOH =30:1) to give compound 23 (23 mg, 79%). ¹H NMR (CD₃OD, 300 MHz) δ 8.34 (s, 1H), 8.32 (s, 1H), 7.8 (s, 1H), 7.64 (d,J = 8.1 Hz, 1H), 7.42 (d, J =7.5 Hz, 1H), 7.13 (t, J= 7.8 Hz, 1H), 4.90 (s, 2H), 4.72 (t, J= 6.3 Hz, 2H), 3.96 (d, J = 6.6 Hz, 1H), 1.96–2.04 (m, 1H), 1.67–1.72 (m, 1H), 1.30–1.34 (m, 1H), 0.75–0.83 (m, 1H). HRMS calculated for C₁₈H₁₉IlN₅O₂ (M +H)⁺: 464.0596; found 464.0584.

(1′R,2′R,3′S,4′S,5_′S)-4′-[6-(3-Iodobenzylamino)-2-fluoro-purine]-2′,3′-dihydroxybicyclo[3.1.0]hexane (24)

To a solution of compound 42 (10 mg, 0.03 mmol) in anhydrous methanol (0.5 mL), triethylamine (0.06 mL, 0.42 mmol) and 3-iodobenzylamine (0.02 mL, 0.15 mmol) were added. The mixture was stirred overnight at room temperature. Solvent was evaporated, and the residue was roughly purified on flash silica gel column chromatography. The product was dissolved in methanol (1 mL) and 10% TFA (0.5 mL), and the reaction mixture was heated at 70 °C for 6 h. Solvent was evaporated and the residue was purified on flash silica gel column chromatography (hexane/ ethyl acetae = 1:1) to give compound 24 (8 mg, 55%). ¹H NMR (CD₃OD, 300 MHz) δ 8.13 (s, 1H), 7.78 (s, 1H), 7.63 (dd, J = 7.8 Hz, 1H), 7.41 (d, J = 7.8 Hz, 1H), 7.11 (t, J = 7.8 Hz, 1H), 4.76 (s, 1H), 4.71–4.76 (m, 2H), 3.92–3.96 (m, 2H), 1.96–2.01 (m, 1H), 1.64–1.67 (m, 1H), 0.73–0.91 (m, 2H). HRMS calculated for C₁₈H₁₈FIN₅O₂ (M + H)⁺: 482.0473; found 482.0489.

(1′R,2′R,3′S,4′S,5′S)-4′-[6-(3-Chlorobenzylamino)-2-(4-methoxycarbonyl-1-butynyl)-9-yl]-2′,3′-O-isopropylidenebicyclo[3.1.0]-hexane (50)

To a solution of compound 45 (43 mg, 0.08 mmol) in anhydrous DMF (1.5mL),PdCl₂(PPh₃)₂ (11 mg,0.015mmol), CuI (1.5 mg, 0.007 mmol), methyl-ω-pentynate (35 mg, 0.31 mmol), and then triethylamine (0.11 mL, 0.80 mmol) were added. The reaction mixture was stirred at room temperature overnight. Solvent was evaporated under vacuum and the residue was purified on flash silica gel column chromatography (hexane/ ethyl acetate = 1:1) to give compound 50 (30 mg, 74%) as a foamy syrup. ¹H NMR (CDCl₃, 300 MHz) 7.98 (s, 1H), 7.32– 7.21 (m, 4H), 6.19 (bs, 1H), 5.31 (t, J = 5.24, 1H), 5.12 (s, 2H), 4.60 (d, J= 7.2 Hz, 1H), 3.78 (s, 3H), 2.69 –2.84 (m, 4H), 2.06– 2.14 (m, 1H), 1.62–1.72 (m, 1H), 1.56 (s,3H), 1.25 (s,3H), 0.88– 1.04 (m, 2H). HRMS calculated for C₂₇H₂₉ClN₅O₂ (M+ H)⁺: 522.9953; found 522.9967.

(1′ R,2′ R,3′S,4′S,5′S)-4′ -[6-(3-Chlorobenzylamino)-2-(4-β-aminoethylaminocarbonyl)-1-butynyl)-9-yl]-2′,3′-dihydroxybicyclo [3.1.0]hexane (26)

To a solution of compound 50 (25 mg, 0.047 mmol) in methanol (0.4 mL), ethylenediamine (1.5 mL) was added. The mixture was stirred overnight at room temperature. Solvent was evaporated, and the residue was roughly purified on flash silica gel column chromatography. The aminated product was dissolved in methanol (1.5 mL) and 10% trifluoroacetic acid (1.0 mL), and the mixture was heated at 70 °C for 15 h. Solvent was evaporated and the residue was purified on flash silica gel column chromatography (CH₂Cl₂/MeOH = 5:1) to give compound 26 (14 mg, 58%). ¹H NMR (CD₃OD, 300 MHz) δ 8.12 (s, 1H), 7.25–7.39 (m, 4H), 5.97 (s, 1H), 4.70 (t, J = 6.3 Hz, 1H), 3.95 (d, J = 6.6 Hz, 1H), 3.77 (t, J = 6.3 Hz, 2H), 2.94 (m, 2H), 2.52 (t, J = 7.8 Hz, 2H), 1.97–2.01 (m, 1H), 1.68–1.72 (m, 2H), 1.30–1.35 (m, 2H), 0.89–0.91 (m, 2H), 0.78–0.82 (m, 1H). HRMS calculated for C₂₅H₂₉ClN₇O₃ (M + H)⁺: 510.2000; found 510.2020.

UV–Visible Spectra of Biotinyl (17) and Fluorescent (18) Probes and Fluorescence Spectrum of 18

UV–visible spectra were measured on an HP diode array spectrophotometer, model no. 8452A. The concentration of each was 40 μM in pH 7.5 aqueous buffer (50 mM Tris-HCl containing 10 mM MgCl₂). The parameters for the compounds are as follows: 17a, λ 276 nm (ε = 9890 L/(mol·cm)); 17b, λ 276 nm (ε = 7400 L/(mol·cm)); 18, λ 274 nm (ε = 16 100 L/(mol·cm)), λ 650 nm (ε = 44900 L/(mol·cm)). Fluorescence spectra were measured on a PerkinElmer LS503: excitation λ_max = 646 nm, emission λ_max ⁼ 660 nm.

Receptor Binding and Functional Assays

[³H]2-Chloro-N⁶-cyclopentyladenosine (52, [³H]CCPA, 42.6 Ci/mmol) was a custom synthesis product (Perkin-Elmer). [¹²⁵I]N⁶-(4-Amino-3-iodobenzyl)adenosine-5′-N′-methyluronamide (53, [¹²⁵I]I-AB-MECA, 2200 Ci/mmol) and [³H](2-[p-(2-carboxyethyl)-phenylethylamino]-5′-N′-ethylcarboxamidoadenosine) (54, [³H] CGS21680, 40.5 Ci/mmol) were purchased from Perkin-Elmer Life and Analytical Sciences (Boston, MA). Test compounds were prepared as 5 mM stock solutions in DMSO and stored frozen.

Cell Culture and Membrane Preparation

CHO cells stably expressing the recombinant hA₁, hA₃, and rA₃Rs and HEK-293 cells stably expressing the hA_2AAR were cultured in Dulbecco's modified Eagle medium (DMEM) and F12 (1:1) supplemented with 10% fetal bovine serum, 100 units/mL penicillin, 100μg/mL streptomycin, and 2 μmol/mL glutamine. In addition, 500 μg/mL Geneticin was added to the A_2A media, while 800 μg/mL hygromycin was added to the A₁ and A₃ media. After being harvested, cells were homogenized and suspended in PBS. Cells were then centrifuged at 240g for 5 min, and the pellet was resuspended in 50 mM Tris-HCl buffer (pH 7.5) containing 10 mM MgCl₂. The suspension was homogenized and was then ultracentrifuged at 14,330 g for 30 min at 4 °C. The resultant pellets were resuspended in Tris buffer and incubated with adenosine deaminase (3 units/mL) for 30 min at 37 °C. The suspension was homogenized with an electric homogenizer for 10 s, pipetted into 1 mL vials, and then stored at −80 °C until the binding experiments. The protein concentration was measured using the BCA protein assay kit from Pierce Biotechnology, Inc. (Rockford, IL).⁴⁹

Binding Assays

Into each tube in the binding assay was added 50 μL of increasing concentrations of the test ligand in Tris-HCl buffer (50 mM, pH 7.5) containing 10 mM MgCl₂, 50 μL of the appropriate agonist radioligand, and finally 100 μL of membrane suspension. For the A₁AR (22 μg of protein/tube) the radioligand used was [³H]51 (final concentration of 3.5 nM) or [³H]52 (final concentration of 1.0 nM). For the A_2AAR (20 μg/tube) the radioligand used was [³H]53 (10 nM). For the A₃AR (21 μg/tube) the radioligand used was [¹²⁵I]54 (0.34 nM). Nonspecific binding was determined using a final concentration of 10 μM 51 diluted with the buffer. The mixtures were incubated at 25 °C for 60 min in a shaking water bath. Binding reactions were terminated by filtration through Brandel GF/B filters under reduced pressure using a M-24 cell harvester (Brandel, Gaithersburg, MD). Filters were washed three times with 3 mL of 50 mM ice-cold Tris-HCl buffer (pH 7.5). Filters for A₁ and A_2AAR binding were placed in scintillation vials containing 5 mL of Hydrofluor scintillation buffer and counted using a Perkin-Elmer liquid scintillation analyzer (Tri-Carb 2810TR). Filters for A₃AR binding were counted using a Packard Cobra II γ-counter. The K_i values were determined using GraphPad Prism for all assays.

cAMP Accumulation Assay

Intracellular cAMP levels were measured with a competitive protein binding method.³⁴ CHO cells that expressed the recombinant hA₃AR were harvested by trypsinization. After centrifugation and resuspended in medium, cells were planted in 24-well plates in 1.0 mL medium. After 24 h, the medium was removed and cells were washed three times with 1 mL of DMEM, containing 50 mM HEPES, pH 7.4. Cells were then treated with the agonist 51 or test compound in the presence of rolipram (10 μM) and adenosine deaminase (3 units/mL). After 45 min forskolin (10 μM) was added to the medium, and incubation was continued for an additional 15 min. The reaction was terminated by removing the supernatant, and cells were lysed upon the addition of 200 μL of 0.1 M ice-cold HCl. The cell lysate was resuspended and stored at −20 °C. For determination of cAMP production, 100 μL of the HCl solution was used in the Sigma direct cAMP enzyme immunoassay following the instructions provided with the kit. The results were interpreted using a Bio-Tek Instruments ELx808 Ultra microplate reader at 405 nm.

Molecular Modeling

The Prime program of the Schrodinger package⁴⁸ was utilized to build a new molecular model of the hA₃AR. The recently reported model of the complex of the nonselective AR agonist 51 docked to the crystal structure of the A_2AAR was used as a template for modeling of the A₃AR.³⁸ The sequence alignment of the A_2A and A₃ARs was performed with Prime, taking into account positions of highly conserved amino acid residues. The default values of all Prime parameters were used. The position of full agonist 51 inside the A_2A receptor was taken into account during the modeling. Thus, the resulting model of the A₃AR contained 51 prepositioned inside the receptor. The Glide program of the Schrodinger package was used to redock 51 to the A₃AR model obtained. The receptor grid generation was performed for the box with a center in the centroid of 51 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 ligand scaling factor was set to 1.0. The geometry of the ligand binding site of the complex of A₃AR with 51 docked was optimized. The binding site was defined as 51 and all amino acid residues located within 5 Å from 51. All A₃AR residues located within 2 Å 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 ·A⁻¹. Another reference nucleoside, 5′-N′-methylcarboxamidoadenosine, was also docked in this hA₃AR homology model (Supporting Information).

The A₃AR model obtained was utilized to study a binding mode of amine congener 15 using InducedFit docking implemented in the Schrödinger package. The grid generation was performed for the box with a side of 32 Å with a center in the centroid of 51. The default values were used for other parameters.