Synthesis and anti-renal fibrosis activity of conformationally locked truncated 2-hexynyl-N(6)-substituted-(N)-methanocarba-nucleosides as A3 adenosine receptor antagonists and partial agonists.

Truncated N(6)-substituted-(N)-methanocarba-adenosine derivatives with 2-hexynyl substitution were synthesized to examine parallels with corresponding 4'-thioadenosines. Hydrophobic N(6) and/or C2 substituents were tolerated in A3AR binding, but only an unsubstituted 6-amino group with a C2-hexynyl group promoted high hA2AAR affinity. A small hydrophobic alkyl (4b and 4c) or N(6)-cycloalkyl group (4d) showed excellent binding affinity at the hA3AR and was better than an unsubstituted free amino group (4a). A3AR affinities of 3-halobenzylamine derivatives 4f-4i did not differ significantly, with Ki values of 7.8-16.0 nM. N(6)-Methyl derivative 4b (Ki = 4.9 nM) was a highly selective, low efficacy partial A3AR agonist. All compounds were screened for renoprotective effects in human TGF-β1-stimulated mProx tubular cells, a kidney fibrosis model. Most compounds strongly inhibited TGF-β1-induced collagen I upregulation, and their A3AR binding affinities were proportional to antifibrotic effects; 4b was most potent (IC50 = 0.83 μM), indicating its potential as a good therapeutic candidate for treating renal fibrosis.

Introduction

Extracellular adenosine acts as a signaling molecule with a generally cytoprotective function in the body. Adenosine mediates cell signaling through binding to four known subtypes (A1, A2A, A2B, and A3) of adenosine receptors (ARs).[1−4] A1, A2A, and A3ARs are activated by low levels of adenosine (EC50 = 0.01–1.0 μM) similar to physiological levels of adenosine, whereas A2BAR is activated by high levels of adenosine (EC50 = 24 μM).[5] A1 and A3ARs are Gi-coupled G protein-coupled receptors (GPCRs), and A2A and A2BARs are Gs-coupled GPCRs. Binding of adenosine to the ARs modulates second messengers such as adenosine 3′,5′-cyclic phosphate (cAMP), inositol triphosphate (IP3), and diacylglycerol (DAG).[1−5] For example, the Gi-coupled A3AR inhibits adenylate cyclase (AC), resulting in cAMP down-regulation, while it stimulates phospholipase C (PLC), which increases the levels of IP3 and DAG. Therefore, ARs have been attractive targets for the development of new therapeutic agents related to cell signaling.

Chronic kidney disease (CKD) is characterized by kidney fibrosis and is becoming a major health problem worldwide,[6] and the use of renin–angiotensin–aldosterone system (RAAS) inhibitors[7,8] is one of a few therapeutic options for the treatment of CKD. However, the efficacy of RAAS inhibitors is limited;[9] it is, therefore, highly desirable to develop new therapeutic agents to improve the prognosis of CKD patients. Extracellular adenosine in the kidney dramatically increases in response to renal hypoxia and ischemia, and increased adenosine has been reported to be associated with CKD.[10] ARs were upregulated in unilateral ureteral obstructed rat kidneys, which is a well-characterized model of CKD,[11] and A3AR knockout mice were protected against ischemia- and myoglobinuria-induced kidney failure.[10] Therefore, A3AR antagonists may become effective renoprotective agents for the treatment of CKD.

Adenosine as a natural ligand has served as a good lead for the development of new AR ligands.[5] Extensive modifications on the N6 and/or 4′-CH2OH of adenosine have been explored, giving several potent and selective A3AR agonists[12,13] such as N6-(3-iodobenzyl)-5′-N-methylcarbamoyladenosine (IB-MECA),[14] 2-chloro-N6-(3-iodobenzyl)-5′-N-methylcarbamoyladenosine (Cl-IB-MECA),[15]N6-(3-iodobenzyl)-5′-N-methylcarbamoyl-4′-thioadenosine (thio-IB-MECA),[16] 2-chloro-N6-(3-iodobenzyl)-5′-N-methylcarbamoyl-4′-thioadenosine (thio-Cl-IB-MECA),[17] and 3′-amino-N6-{5-chloro-2-(3-methylisoxazol-5-ylmethoxy)benzyl}-5′-N-methylcarbamoyladenosine (CP-608039).[18] These compounds contain the potency- and efficacy-enhancing 5′-methyluronamide moiety and the N6-hydrophobic moiety. Also, AR agonists that combined N6-alkyl and 2-alkynyl substitutions proved useful in the identification of A3 or A2B AR agonists with various selectivity profiles, depending on the type of 2-alkynyl substitution.[19] On the other hand, the truncated nucleosides where the 5′-methyluronamide of the A3AR agonists was deleted were converted into potent and selective A3AR antagonists, because there was no 5′-uronamide, which serves as the hydrogen bonding donor required for receptor activation.[20] Among these, compound 1 showed potent antiglaucoma[21] activity (Chart 1). Introduction of the 2-hexynyl group on the C2-position of 1 but no substitution on the N6-position converted 1 into dually acting A2AAR agonist and A3AR antagonist 2.[22] Molecular modeling and empirical structure activity studies in both the ribose and the 4′-thioribose series indicated that the C2 binding sites of A2AAR and A3AR were spacious enough to accommodate the bulky substituent.

Chart 1

Design Strategy for Truncated (N)-Methanocarba-Nucleosides in This Studya

Ki values (nM) or % inhibition at 10 μM in binding to human A1, A2A, and A3 adenosine receptors.

Truncated (N)-methanocarba-nucleosides 3(20a) were also reported to be selective and potent A3AR antagonists, indicating that compound 3 can also serve as a good template for the development of A3AR ligands. Thus, we designed and synthesized the truncated C2-hexynyl-(N)-methanocarba-nucleosides 4, which hybridize the structure of C2-hexynyl derivative 2 with that of (N)-methanocarba-nucleoside 3 to determine if similar biological trends between 2 and 4 were observed. For the synthesis of the target nucleoside 4, copper-catalyzed[23] and palladium-catalyzed[24] cross-coupling reactions were employed as key steps for functionalization of the C2-position of 6-chloropurine nucleosides. Herein, we report the synthesis of truncated C2-hexynyl-N6-substituted-(N)-methanocarba-nucleosides 4 as potent and selective A3AR antagonists and their renoprotective effects using TGF-β1-stimulated mProx cells, a cell culture model for kidney fibrosis.[25]

Results and Discussion

Chemistry

The desired C2-hexynyl-methanocarba-adenosine derivatives 4a–4i were synthesized from our known cyclopentenone intermediate 5(26) using a palladium-catalyzed cross-coupling reaction as a key step (Scheme 1).

Scheme 1

Synthesis of Truncated (N)-Methanocarba-Nucleosides

Reagents and conditions: (a) NaBH4, CeCl3–7H2O, methanol, 0 °C, 2 h; (b) Et2Zn, CH2I2, CH2Cl2, rt, 5 h; (c) 2-iodo-6-chloropurine, Ph3P, DIAD, THF, rt, 18 h; (d) 1-hexyne, (Ph3P)4Pd, Cs2CO3, CuI, DMF, 50 °C, 6 h; (e) 2 N HCl/THF (1/1), 40 °C, 18 h; (f) R–NH2, Et3N, ethanol, 90 °C, 18 h.

The cyclopentenone derivative 5 was converted to the glycosyl donor 7 according to the reported procedure[27] developed by our laboratory. Direct condensation of 7 with 6-chloro-2-iodopurine[28] under the standard Mitsunobu conditions in THF afforded the β-anomer 8 in 67% yield, similar to a literature report.[29] The anomeric β-configuration of 8 was readily assigned by the diagnostic coupling constants typical of the boat conformation of the bicyclo[3.1.0]hexane system, which has been extensively confirmed by X-ray crystallography and NMR analysis.[30] The coupling constants of the JH1′,H2′ and JH1′,H5′ should be zero, because both H1–C–C–H2 and H1–C–C–H5 dihedral angles with trans relationships are close to 90°,[30] indicating that 1′-H of 8 should appear as a singlet. Indeed, 1H NMR of 8 showed that 1′-H appeared as a singlet at 5.03 ppm, confirming the structure of 8. Sonogashira[31] coupling of 8 with 1-hexyne in the presence of palladium catalyst yielded the C2-hexynyl derivative 9 (56%). Treatment of 9 with 2 N HCl gave the 6-chloro derivative 10. Substitution of the 6-position of 10 with ammonia and various primary alkyl-, cycloalkyl-, and arylalkyl-amines afforded the final nucleosides 4a–4i.

The target nucleosides were also synthesized using a lithiation-mediated stannyl transfer reaction[28a] and a copper-catalyzed cross-coupling reaction[23] as key steps for functionalization of the C2-position (Scheme 2). The glycosyl donor 7 was condensed with 6-chloropurine under the same conditions used in Scheme 1 to give 6-chloropurine derivative 11. Treatment of 11 with LiTMP followed by reacting the resulting anion with tri-n-butyltin chloride afforded the C2-stannyl derivative 12 exclusively.[28a] Copper-catalyzed coupling[23] of 12 with 1-iodohexyne yielded the 2-hexynyl derivative 9, which was converted to the same final nucleosides 4a–4i according the same procedure used in Scheme 1.

Scheme 2

Alternative Synthesis of Truncated (N)-Methanocarba-Nucleosides

Reagents and conditions: (a) 6-chloropurine, Ph3P, DIAD, THF, rt, 18 h; (b) LiTMP, Bu3SnCl, THF, −78 °C, 5 h; (c) 1-iodohexyne, CuI, DMF, 50 °C, 16 h.

Binding Affinity

Binding assays were carried out using standard radioligands and membrane preparations from Chinese hamster ovary (CHO) cells stably expressing the human (h) A1 or A3AR, RBL-2H3 basophilic leukemia cells expressing rat (r) A3AR, or human embryonic kidney (HEK)-293 cells expressing the hA2AAR.[32] Binding at the hA3AR or rA3AR in this study was carried out using [125I]N6-(3-iodo-4-aminobenzyl)-5′-N-methylcarboxamidoadenosine (I-AB-MECA, 13) as a radioligand. Binding at the hA1AR using [3H] (-)-N6-2-phenylisopropyl adenosine (R-PIA, 14) or hA2AAR using [3H]CGS21680 (2-[p-(2-carboxyethyl)phenylethylamino]-5′-N-ethylcarboxamidoadenosine, 15) was carried out. In cases of weak binding, the percent inhibition of radioligand binding to the hA1AR and hA2AAR was determined at 10 μM. Nonspecific binding was defined using 10 μM of 5′-N-ethylcarboxamidoadenosine (NECA, 16).

Because binding affinity of similar (N)-methanocarba compounds was reported to be very weak or absent at the hA2BAR subtype,[33] we did not include this receptor in the radioligand binding assays. To confirm that activity of the present chemical series is weak at the A2BAR, we performed a functional assay in CHO cells expressing the hA2BAR. Compound 4b at 10 μM produced only 15.7 ± 12.6% of the activation of cAMP production seen with full agonist 16.

As shown in Table 1, a variety of N6-alkyl, cycloalkyl, and arylalkyl substituents in truncated (N)-methanocarba-nucleoside derivatives have produced nanomolar binding affinity at the hA3AR subtype, indicating that bulky C2 and N6 substituents could be tolerable in the binding site of A3AR. However, a hydrophobic substituent at the N6-position reduced the binding affinity greatly at the hA2A AR subtype in the presence of a hydrophobic C2-hexynyl group, and only an unsubstituted 6-amino group showed good binding affinity (Ki = 100 nM) at the hA2AAR, indicating that the N6 binding site of hA2AAR is small. This trend is similar to that of truncated 2-hexynyl-4′-thioadenosine (2),[22] but truncated carbanucleoside derivative 4a was 14-fold less potent than truncated 4′-thioadenosine derivative 2. This result may be due to the fixed conformation of (N)-methanocarba-nucleosides unlike the flexible conformation of 4′-thioadenosine derivatives, hindering them from forming a favorable hydrophobic interaction in the binding site of A2AAR. However, all compounds showed very weak binding affinity at the hA1AR, suggesting that the binding sites may not be large enough to accommodate the bulky C2 and/or N6 substituent. Among compounds tested, 4b (R = CH3) exhibited the highest binding affinity (Ki = 4.9 nM) at the hA3AR subtype with high selectivity for the hA1 and hA2AARs. The primary amine-substituted N6-alkyl- and N6-cycloalkyl- derivatives (4b–4e) generally exhibited better binding affinity at the hA3AR than the free amino derivative 4a, except cyclopentyl derivative 4e. The order of compounds showing high binding affinity at the hA3AR is as follows: 4b (R = CH3, Ki = 4.6 nM) > 4c (R = ethyl, Ki = 6.7 nM) > 4d (R = cyclopropyl, Ki = 9.2 nM) > 4a (R = H, Ki = 16.2 nM). The binding affinities of 3-halobenzylamine derivatives 4f–4i at the hA3AR did not differ significantly, with Ki values of 7.8–16.0 nM. The binding affinity at the hA3AR in this series decreased in the following order: 3-Cl derivative 4h, 3-Br derivative 4g > 3-I derivative 4f > 3-F derivative 4i. All synthesized compounds 4a–4i have also produced nanomolar binding affinity at the rA3AR, but they showed weaker binding affinity than that at the hA3AR. The N6-alkyl derivatives 4b and 4c exhibited lower binding affinity at the rA3AR than the free amino derivative 4a, the N6-cycloalkyl derivatives 4d and 4e, and the 3-halobenzylamine derivatives 4f–4i, which showed similar binding affinities at the rA3AR, with Ki values in the range of 10.7–65 nM. The 3-chlorobenzyl derivative 4h exhibited the highest binding affinity (Ki = 10.7 nM) at the rA3AR, unlike the N6-methyl derivative 4b showing the highest affinity (Ki = 4.9 nM) at the hA3AR.

Table 1

Binding Affinities and Anti-Renal Fibrosis Activity of Truncated 2-Hexynyl-N6-Substituted Derivatives 4a–4i and Reference Nucleosides 2 and 3 at hARs and rA3AR

		Ki (nM) or % inhibition at 10 μMa
compd no.	R	hA1AR	hA2AAR	hA3AR	rA3AR	IC50 (μM)d
2b		39 ± 10%	7.19 ± 0.6	11.8 ± 1.3	NDe	NDe
3c		3040 ± 610	1080 ± 310	1.44 ± 0.6	NDe	18.6
4a	H	29% ± 6%	100 ± 10	16.2 ± 6.7	65 ± 18	6.12
4b	methyl	14% ± 4%	7490 ± 590	4.90 ± 1.30	231 ± 81	0.83
4c	ethyl	31% ± 7%	2860 ± 1060	6.70 ± 1.80	176 ± 47	0.84
4d	cyclopropyl	2170 ± 510	2200 ± 660	9.20 ± 0.40	39 ± 19	11.8
4e	cyclopentyl	1580 ± 240	1760 ± 410	160 ± 50	58 ± 39	>50
4f	3-iodobenzyl	48% ± 5%	2530 ± 170	12.0 ± 6.0	26 ± 22	7.88
4g	3-bromobenzyl	38% ± 6%	3150 ± 170	8.60 ± 4.80	59 ± 37	10.4
4h	3-chlorobenzyl	19% ± 8%	3310 ± 1220	7.80 ± 1.70	10.7 ± 1.6	2.87
4i	3-fluorobenzyl	21% ± 4%	27% ± 5%	16.0 ± 10.0	43 ± 30	3.17

All binding experiments were performed using adherent mammalian cells stably transfected with cDNA encoding the appropriate hAR (A1AR and A3AR in CHO cells and A2AAR in HEK-293 cells) or rA3AR expressed endogenously in RBL-2H3 cells. Binding was carried out using 1 nM [3H]14, 10 nM [3H]15, or 0.5 nM [125I]13 as radioligands for A1, A2A, and A3ARs, respectively. Values expressed as a percentage in italics refer to percent inhibition of specific radioligand binding at 10 μM for 3 – 5 duplicate determinations, with nonspecific binding defined using 10 μM 16.

ref (22).

ref (20a).

Concentration to inhibit the TGF-β1-induced collagen I mRNA expression by 50%.

Not determined.

In a cAMP functional assay[34] at the hA3AR expressed in CHO cells, the most potent compound 4b behaved as a partial agonist, in contrast to full antagonists 2 and 3 (Figure 1). Compound 4b at 10 μM displayed an EC50 of 45.8 nM and a maximal stimulation of cAMP formation of 29.1 ± 5.0% relative to the full agonist 16 (= 100%). Similarly, other compounds proved to be partial agonists of the hA3AR (% activation relative to 16, triplicate determination): 4c, 15.5 ± 6.7; 4d, 19.8 ± 4.6; 4e, 27.1 ± 14.6; 4i, 18.9 ± 7.5. Compounds 4f, 4g, and 4h induced <5% of the activation seen with 16 and were therefore antagonists.

Figure 1

Effect of compound 4b on forskolin-induced stimulation of cAMP production at the hA3AR expressed in CHO cells, compared to 16 as reference full agonist (= 100%). A representative curve from three determinations is shown.

Renoprotective Effects

All synthesized compounds were tested for an antifibrotic effect in murine proximal (mProx) cells, a cell line of mouse proximal tubular epithelial cells.[25] As shown in Table 1, most of the tested compounds strongly inhibited transforming growth factor (TGF)-β1-induced collagen I upregulation. Compound 4b showed the most potent inhibitory activity (IC50 = 0.83 μM) against TGF-β1-induced collagen I mRNA expression (Figure 2). The binding affinity at the A3AR was almost proportional to the antifibrotic activity, which indicates that the small N6-hydrophobic substituent is also favored for renoprotective effects.

Figure 2

Inhibition of TGF-β1-induced COL1A1 gene expression in mProx24 cells, a cell line of mouse proximal tubular epithelial cells, by 4b. Data are mean ± SE of three experiments. *p < 0.05 vs TGF-β1-stimulated mProx24 cells: arelative increase in COL1A1 gene expression (1.0 is the effect of 5 ng/mL TGF-β1), bat the concentration of 4b in μM indicated in column 1.

Molecular Docking Study

The truncated C2-substituted thio-ribose compound 2 (A2AKi = 7.19 nM) exhibited excellent binding affinity, and the methanocarba analogue 4a (A2AKi = 100 nM) showed ≈14-fold less binding affinity at the hA2A AR. In addition, the presence of the 3-iodobenzyl group at the N6-position in 4f led to a substantial decrease in its binding affinity at the hA2AAR with a Ki of 2530 nM. In view of the observed variations in the hA2AAR binding affinities among these compounds, molecular docking and binding free energy calculations were carried out considering the X-ray structure of the hA2AAR complexed with an agonist, 16 (PDB code 2YDV[35]). The common interactions among N6-unsubstituted compounds 2 and 4a at the hA2AAR includes: (i) the adenine ring stabilized through π–π stacking interaction with Phe168 (extracellular loop 2) and a H-bonding interaction with Asn2536.55, (ii) the exocyclic 6-amino group H-bonded with Asn2536.55 and Glu169, and (iii) the projection of C2-hexynyl group toward the extracellular side exhibiting hydrophobic interaction with Phe168, Ile662.55, Leu2677.32, Met2707.35, Ile2747.39, and Tyr2717.36 residues (Figure 3).

Figure 3

Predicted binding modes of N6-unsubstituted nucleosides 2 (A) and 4a (B) in the hA2AAR agonist-bound crystal structure. Compounds 2 and 4a are depicted in ball-and-stick, with carbon atoms in magenta and purple, respectively. The key amino acid residues are shown as capped-stick, with carbon atoms in white. The Connolly surface of the receptor was generated by MOLCAD with green color and z-clipped for visual convenience. The hydrogen bonds are shown as black dashed lines, and the nonpolar hydrogen atoms are not displayed for clarity.

In contrast, they exhibited different binding modes at the ribose binding site formed by Val843.32, Leu853.33, Trp2466.48, Leu2496.51 and Ile2747.39, Ser2777.42, and His2787.43. The 2′- and 3′-hydroxyl groups of 2 formed H-bonds with the two key residues His2787.43 and Ser2777.42, respectively (Figure 3A), whereas 4a lost one of the key H-bond interactions with Ser2777.42 (Figure 3B). This residue Ser277 is a key residue reported to be important for hA2AAR agonistic activity and potency using site-directed mutagenesis.[36−38] It appears that a decrease in the binding affinity of 4a at the hA2AAR could be due to the loss of H-bonding with Ser2777.42 at the ribose binding site. The loss of H-bonding with Ser2777.42 may particularly be attributed to the methanocarba ring (4a), being less flexible than the thio-ribose ring (2). Furthermore, the calculated prime MM-GBSA binding free energies (ΔGbind) for 2 and 4a were −104.16 and −90.37 kcal/mol, respectively, which are in good agreement with their observed binding affinities at the hA2AAR. However, 4f with a bulky group at the N6-position did not fit well at the binding site of the hA2A AR. These results show that the H-bond interactions with both Ser2777.42 and His2787.43 at the ribose binding site are important for high affinity and potency, and the bulky group at the N6-position is unfavorable toward high binding affinity at the hA2AAR.

In addition, we also performed the molecular docking studies of the analogues 2 and 4a to hA3AR (see Figure 1S in Supporting Information). Because the X-ray crystal structure of hA3AR is not available yet, the homology model available in the Protein Data Bank (PDB code 1OEA) was used. The docking results showed that the binding modes of the analogues in hA3AR are flipped compared to those in hA2AAR. In hA3AR, the bulky C2-hexynyl group positions toward the middle of the trans-membrane region exhibited hydrophobic interactions. However, in hA2AAR, there is limited space at the bottom of the pocket, making the bulky hexynyl group face toward the extracellular side. The NH2 group at N6-position forms the hydrogen bonding with Asn6.55 in both hA2AAR (Asn253) and hA3AR (Asn250). Interestingly, there is a relatively bigger space near this region in hA3AR, whereas the NH2 group binds tightly in the pocket of hA2AAR. It appears according to this docking mode that this is why the N6-substituted derivatives (4b–4i) maintained their binding affinity at the hA3AR, but not at the hA2AAR.

Conclusions

The series of truncated N6-substituted-(N)-methanocarba-adenosine derivatives, 4a–4i with 2-hexynyl group were synthesized in order to examine if this class of nucleosides behaves as the corresponding 4′-thioadenosine derivatives. The functionalization at the C2-position of 6-chloropurine derivatives was achieved using lithiation-mediated stannyl transfer and copper- or palladium-catalyzed cross-coupling reactions. It was revealed that all synthesized nucleosides showed very high binding affinity at the hA3AR as well as at the rA3AR, as in the case of the corresponding 4′-thioadenosine derivatives, indicating that the hydrophobic N6 and/or C2 substituent could be tolerable in the binding site of the A3AR. However, only an unsubstituted 6-amino group in the presence of a bulky C2-hexynyl group was associated with high binding affinity at the hA2AAR (compound 4a). This trend is similar to that of the corresponding 4′-thioadenosine derivatives, serving as dual acting A2A and A3AR ligands. However, the binding affinity at the hA2AAR of the truncated (N)-methanocarba-nucleoside 4a is 14-fold less potent than the truncated 4′-thioadenosine derivative 2. It is attributed to the loss of key hydrogen bonding due to the rigid structure, which was confirmed by a hA2AAR molecular docking study.

The specific structure–activity relationship for this series of conformationally constrained nucleosides might arise from the molecule of lacking in the flexibility required for optimal interaction in the binding site because of the rigidity of (N)-methanocarba-nucleosides. From this study, N6-methyl derivative 4b was discovered as a preferred hA3AR ligand (low efficacy partial agonist) with high selectivity, whereas 3-chlorobenzyl derivative 4h was discovered as the most potent/selective rA3AR ligand in this series.

The nature of the N6 substituent in this chemical series modulates the level of hA3AR agonist efficacy (ranging from nearly 0% to 29% of full agonist). For these assays, we used a CHO cell with a high level of stable expression of the hA3AR, which would tend to amplify partial agonist action. Because even these partial agonists have a relatively low efficacy, they can be expected to behave similarly to full antagonists in some pharmacological models, especially in cases of low receptor expression.[39]

A3AR antagonist 1 was recently shown to inhibit unilateral ureteral obstruction-induced renal fibrosis and collagen I upregulation.[40] This suggests that A3AR antagonists might be useful therapeutically to block the development and attenuate the progression of renal fibrosis. All of the compounds synthesized here were screened for renoprotective activity. Among compounds tested, 4b exhibited the most potent inhibitory activity (IC50 = 0.83 μM) against TGF-β1-induced collagen I upregulation. These findings indicate that this series of truncated (N)-methanocarba-nucleoside derivatives acting as partial agonists of low efficacy or as antagonists, which show high binding affinity at the human A3AR, can serve as a good lead for the development of antirenal fibrosis agents.

Experimental Section

Chemical Synthesis

General Methods

1H NMR spectra (CDCl3, CD3OD, or DMSO-d) were recorded on a Varian Unity Invoa 400 MHz instrument. The 1H NMR data are reported as peak multiplicities: s for singlet, d for doublet, dd for doublet of doublets, t for triplet, q for quartet, brs for broad singlet, and m for multiplet. Coupling constants are reported in hertz. 13C NMR spectra (CDCl3, CD3OD, or DMSO-d) were recorded on a Varian Unity Inova 100 MHz instrument. 19F NMR spectra (CDCl3, CD3OD) were recorded on a Varian Unity Inova 376 MHz instrument. The chemical shifts were reported as parts per million (δ) relative to the solvent peak. Optical rotations were determined on Jasco III in appropriate solvent. UV spectra were recorded on U-3000 made by Hitachi in methanol or water. Infrared spectra were recorded on FT-IR (FTS-135) made by Bio-Rad. Melting points were determined on a Buchan B-540 instrument and are uncorrected. Elemental analyses (C, H, and N) were used to determine the purity of all synthesized compounds, and the results were within ±0.4% of the calculated values, confirming ≥95% purity. Reactions were checked with TLC (Merck precoated 60F254 plates). Flash column chromatography was performed on silica gel 60 (230–400 mesh, Merck). Reagents were purchased from Aldrich Chemical Co. Solvents were obtained from local suppliers. All the anhydrous solvents used were redistilled over CaH, P2O5 or sodium/benzophenone prior to the reaction.

6-Chloro-9-((3aR,3bR,4aS,5R,5aS)-hexahydro-2,2dimethylbicyclo[3.1.0]hex-1(5)-eno[3,2-d] [1,3]dioxol-5-yl)-2-iodo-9H-purine (8)

To a stirred solution of 2-iodo-6-chloropurine (1.23 g, 4.4 mmol) and triphenylphosphine (Ph3P) (1.90 g, 4.4 mmol) in anhydrous THF (20 mL) was added diisopropyl azodicarboxylate (DIAD) (1.44 mL, 9.16 mmol) in THF (10 mL) under N2 at 0 °C, and the mixture was stirred at the same temperature for 15 min. To this solution was added a solution of compound 7(19) (0.5 g, 2.93 mmol) in THF (10 mL) at 0 °C, and the reaction mixture was stirred at room temperature for 16 h. The reaction mixture was concentrated under reduced pressure, and the crude residue was purified by flash silica gel column chromatography (hexane: EtOAc = 3:1) to give 8 (0.85 g, 67%) as a white solid: mp 94–96 °C; UV (MeOH) λmax 282 nm. MS (ESI): [M + H]+ calcd for C14H15ClIN4O2, 432.9923; found, 432.9931; [α]25D = −10.4 (c 0.2, MeOH); 1H NMR (CDCl3) δ 0.95–1.01 (m, 2 H), 1.26 (s, 3 H), 1.55 (s, 3 H), 1.63–1.68 (m, 1 H), 2.12–2.18 (m, 1 H), 4.65–4.68 (m, 1 H), 5.03 (s, 1 H), 5.35–5.38 (m, 1 H), 8.12 (s, 1 H); 13C NMR (CDCl3) δ 9.4, 24.4, 25.6, 26.1, 26.5, 61.5, 81.6, 89.1, 112.8, 116.9, 132.1, 143.9, 150.9, 152.1. Anal. (C14H14ClIN4O2) C, H, N.

6-Chloro-2-(hex-1-ynyl)-9-((3aR,3bR,4aS,5R,5aS)-hexahydro-2,2-dimethylbicyclo [3.1.0]hex-1(5)-eno[3,2-d][1,3]dioxol-5-yl)-9H-purine (9)

To a stirred solution of 8 (0.30 g, 0.69 mmol) in anhydrous DMF (10 mL) were added tetrakis(triphenylphosphine)palladium ((Ph3P)4Pd) (0.20 g, 0.17 mmol), copper iodide (0.016 g, 0.08 mmol), cesium carbonate (0.226 g, 0.69 mmol), and 1-hexyne (0.07 mL, 0.62 mmol) at room temperature and the reaction mixture was stirred at 50 °C for 5 h. The reaction mixture was cooled to room temperature, quenched with saturated NaHCO3 (5 mL) solution, and diluted with EtOAc (10 mL). The organic layer was separated and aqueous layer was further extracted with EtOAc (3 × 5 mL). The combined organic layers were washed with brine (10 mL) and water (10 mL), dried over anhydrous MgSO4, and filtered. The solvent was evaporated under reduced pressure and the crude residue was purified by flash silica gel column chromatography (hexane: EtOAc = 3:1) to give 9 (0.15 g, 56%) as a white foam: mp 105–107 °C; UV (MeOH) λmax 284 nm. MS (ESI+): [M + H]+ calcd for C20H24ClIN4O2, 387.1582; found, 387.1586; [α]25D = +2.5 (c 0.2, MeOH). 1H NMR (CDCl3, 400 MHz) δ: 0.94–1.03 (m, 5 H), 1.25 (s, 3 H), 1.47–1.54 (m, 2 H), 1.55 (s, 3 H), 1.63–1.70 (m, 3 H), 2.12–2.17 (m, 1 H), 2.48–2.51 (t, 2 H, J = 7.2 Hz), 4.63–4.65 (d, 1 H, J = 5.1 Hz), 5.13 (s, 1 H), 5.35–5.38 (t, 1 H, J = 6.00 Hz), 8.13 (s, 1 H). 13C NMR (CDCl3, 100 MHz) δ: 9.3, 13.7, 19.2, 22.3, 24.4, 25.3, 26.0, 26.5, 30.2, 60.7, 79.7, 81.4, 89.1, 90.9, 112.6, 130.8, 144.2, 146.3, 151.1, 151.3. Anal. (C20H23ClIN4O2) C, H, N.

(1R,2R,3S,4R,5S)-4-(6-Chloro-2-(hex-1-ynyl)-9H-purin-9yl)bicyclo[3.1.0]hexane-2,3-diol (10)

To a stirred ice-cooled solution of 9 (0.30 g, 0.77 mmol) in THF (3 mL) was added 2 N HCl (3 mL), and the mixture was stirred at 40 °C for 16 h. The reaction mixture was neutralized with saturated NaHCO3 (2 mL) solution and then diluted with EtOAc (10 mL). The organic layer was separated, and the aqueous layer was further extracted with EtOAc (2 × 5 mL). The combined organic layers were washed with brine (10 mL), dried over anhydrous MgSO4, filtered, and evaporated under reduced pressure. The crude residue was purified by flash silica gel column chromatography (hexane: EtOAc = 2:1) to give 10 (0.22 g, 73%) as a white solid: mp 120–122 °C; UV (MeOH) λmax 285 nm. MS (ESI+): [M + H]+ calcd for C17H20ClN4O2, 347.1269; found, 347.1274; [α]25D = +27.0 (c 0.2, MeOH). 1H NMR (CDCl3, 400 MHz) δ: 0.84–0.87 (m, 1 H), 0.94–0.97 (t, 3 H, J = 7.2 Hz), 1.29–1.32 (m, 1 H), 1.47–1.53 (m, 2 H), 1.62–1.70 (m, 3 H), 2.10–2.14 (m, 1 H), 2.47–2.51 (t, 2 H, J = 7.2 Hz), 4.05–4.06 (d, 1 H, J = 6.0 Hz), 4.86–4.89 (t, 1 H, J = 6.0 Hz), 5.04 (s, 1 H), 8.20 (s, 1 H). 13C NMR (CDCl3, 100 MHz) δ: 7.8, 14.1, 19.6, 19.9, 22.7, 24.6, 30.6, 63.4, 72.3, 77.1, 79.9, 91.8, 131.3, 144.4, 146.5, 151.6, 151.7. Anal. (C17H19ClN4O2) C, H, N.

(1R,2R,3S,4R,5S)-4-(6-Amino-2-(hex-1-ynyl)-9H-purin-9-yl)bicyclo[3.1.0]hexane-2, 3-diol (4a)

A solution of 10 (0.05 g, 0.42 mmol) in saturated NH3 in t-BuOH (5 mL) was stirred at 110 °C for 16 h. The reaction mixture was evaporated, and the residue was purified by flash silica gel column chromatography (CH2Cl2: MeOH = 9: 1) to give 4a (0.103 g, 87%) as a white solid: mp 122–124 °C; UV (MeOH) λmax 271 nm. MS (ESI+): [M + H]+ calcd for C17H22N5O2, 329.1796; found, 329.1797; [α]25D = +14.7 (c 1.75, MeOH). 1H NMR (CD3OD, 400 MHz) δ: 0.76–0.78 (m, 1 H), 0.96–1.00 (t, 3 H, J = 7.2 Hz), 1.34–1.37 (m, 1 H), 1.50–1.70 (m, 5 H), 1.98–2.01 (m, 1 H), 2.45–2.48 (t, 2 H, J = 7.2 Hz), 3.86–3.88 (d, 1 H, J = 6.8 Hz), 4.66–4.69 (t, 1 H, J = 5.6 Hz), 4.83 (s, 1 H), 8.24 (s, 1 H). 13C NMR (CD3OD) δ: 8.2, 14.1, 19.6, 19.7, 23.2, 24.7, 31.6, 64.0, 73.0, 77.4, 81.3, 88.6, 120.3, 141.4, 147.9, 157.2, 167.2. Anal. (C17H21N5O2) C, H, N.

General Procedure for the Synthesis of 4b–4i

To a solution of 10 (1 equiv) in EtOH (10 mL) were added Et3N (3 equiv) and the appropriate amine (1.5 equiv) at room temperature, and the mixture was stirred at 90 °C for 18 h in a steel bomb. The reaction mixture was evaporated and the residue was purified by flash silica gel column chromatography (CH2Cl2/MeOH = 12:1) to give 4b–4i.

(1R,2R,3S,4R,5S)-4-(2-(Hex-1-ynyl)-6-(methylamino)-9H-purin-9-bicyclo[3.1.0]hexane-2,3-diol (4b)

Yield: 75%; white solid; mp 118–120 °C; UV (MeOH) λmax 273 nm. MS (ESI+): [M + H]+ calcd for C18H24N5O2, 342.1925; found, 342.1925; [α]25D = +35.5 (c 0.2, MeOH). 1H NMR (CD3OD, 400 MHz) δ: 0.73–0.79 (m, 1 H), 0.96–1.00 (t, 3 H, J = 7.2 Hz), 1.34–1.38 (m, 1 H), 1.49–1.71 (m, 5 H), 1.95–2.01 (m, 1 H), 2.45–2.49 (t, 2 H, J = 7.2 Hz), 3.11 (brs, 3 H), 3.84–3.86 (d, 1 H, J = 6.8 Hz), 4.64–4.67 (t, 1 H, J = 5.6 Hz), 4.82 (s, 1 H), 8.16 (s, 1 H). 13C NMR (CD3OD, 100 MHz) δ: 7.9, 13.9, 19.6, 19.6, 23.2, 24.6, 27.8, 31.6, 63.8, 73.0, 77.2, 81.6, 87.9, 120.2, 140.3, 148.1, 149.4, 156.6. Anal. (C18H23N5O2) C, H, N.

(1R,2R,3S,4R,5S)-4-(6-(ethylamino)-2-(hex-1-ynyl)-9H-purin-9-yl)bicyclo[3.1.0]hexane-2,3-diol (4c)

Yield: 76%; white solid; mp 98–100 °C; UV (MeOH) λmax 273 nm. MS (ESI+): [M + H]+ calcd for C19H26N5O2, 356.2081; found, 356.2083; [α]25D = +8.5 (c 0.2, MeOH); 1H NMR (CD3OD, 400 MHz) δ: 0.74–0.77 (m, 1 H), 0.96–1.00 (t, 3 H, J = 7.2 Hz), 1.27–1.31 (t, 3 H, J = 7.2 Hz), 1.34–1.37 (m, 1 H), 1.50–1.70 (m, 5 H), 1.96–2.00 (m, 1 H), 2.45–2.50 (t, 2 H, J = 7.2 Hz), 3.62 (brs, 2 H), 3.84–3.85 (d, 1 H, J = 6.8 Hz), 4.64–4.67 (t, 1 H, J = 5.6 Hz), 4.81 (s, 1 H), 8.16 (s, 1 H). 13C NMR (CD3OD, 100 MHz) δ: 7.9, 14.0, 15.1, 19.5, 19.2, 23.2, 24.6, 31.6, 36.6, 63.8, 73.1, 77.1, 81.6, 87.9, 119.9, 140.3, 148.1, 149.5, 155.9. Anal. (C19H25N5O2) C, H, N.

(1R,2R,3S,4R,5S)-4-(6-(Cyclopropylamino)-2-(hex-1-ynyl)-9H-purin-9-yl)bicyclo[3.1.0]hexane-2,3-diol (4d)

Yield: 68%; white solid; mp 94–96 °C; UV (MeOH) λmax 275.0 nm. MS (ESI+): [M + H]+ calcd for C20H26N5O2, 356.2081; found, 368.2078; [α]25D = +20.2 (c 0.2, MeOH). 1H NMR (CD3OD, 400 MHz) δ: 0.61–0.65 (m, 2 H), 0.73–0.79 (m, 1 H), 0.86–0.91 (m, 2 H), 0.96–1.00 (t, 3 H, J = 7.2 Hz), 1.33–1.38 (m, 1 H), 1.48–1.71 (m, 5 H), 1.95–2.01 (m, 1 H), 2.46–2.50 (t, 2 H, J = 7.2 Hz), 3.05 (brs, 1 H), 3.84–3.86 (d, 1 H, J = 6.8 Hz), 4.65–4.67 (t, 1 H, J = 5.6 Hz), 4.83 (s, 1 H), 8.18 (s, 1 H). 13C NMR (CD3OD, 100 MHz) δ: 7.7, 7.9, 14.0, 19.6, 19.7, 23.2, 24.6, 24.8, 31.6, 63.8, 73.0, 77.2, 81.6, 88.2, 120.1, 140.7, 148.0, 149.8, 157.1. Anal. (C20H25N5O2) C, H, N.

(1R,2R,3S,4R,5S)-4-(6-(Cyclopentylamino)-2-(hex-1-ynyl)-9H-purin-9-yl)bicyclo[3.1.0]hexane-2,3-diol (4e)

Yield: 65%; white solid; mp 95–97 °C; UV (MeOH) λmax 275 nm. MS (ESI+): [M+H]+ calcd for C22H30N5O2, 396.2394; found, 396.2400; [α]25D = +21.4 (c 0.2, MeOH). 1H NMR (CD3OD, 400 MHz) δ: 0.74–0.78 (m, 1 H), 0.96–1.00 (t, 3 H, J = 7.2 Hz), 1.34–1.37 (m, 1 H), 1.50–1.83 (m, 10 H), 1.96–1.99 (m, 2 H), 2.06–2.12 (m, 2 H), 2.45–2.48 (t, 2 H, J = 7.2 Hz), 3.83–3.85 (d, 1 H, J = 6.4 Hz), 4.60 (brs, 1 H), 4.64–4.67 (t, 1 H, J = 5.6 Hz), 4.81(s, 1 H), 8.18 (s, 1 H). 13C NMR (CD3OD, 100 MHz) δ: 7.9, 13.9, 19.55, 19.60, 23.2, 24.6, 24.7, 31.6, 34.0, 53.5, 63.8, 73.0, 77.1, 81.7, 87.9, 119.8, 140.3, 148.1, 149.5, 155.5. Anal. (C22H29N5O2) C, H, N.

(1R,2R,3S,4R,5S)-4-(6-(3-Iodobenzylamino)-2-(hex-1-ynyl)-9H-purin-9-yl)bicyclo[3.1.0]hexane-2,3-diol (4f)

Yield: 88%; white solid; mp 128–130 °C; UV (MeOH) λmax 274 nm. MS (ESI+): [M + H]+ calcd for C24H27IN5O2, 544.1204; found, 544.1212; [α]25D = +30.3 (c 0.2, MeOH). 1H NMR (CD3OD, 400 MHz) δ: 0.73–0.79 (m, 1 H), 0.96–1.00 (t, 3 H, J = 7.2 Hz), 1.34–1.37 (m, 1 H), 1.50–1.58 (m, 2 H), 1.60–1.72 (m, 3 H), 1.96–2.01 (m, 1 H), 2.45–2.48 (t, 2 H, J = 7.2 Hz), 3.85–3.87 (d, 1 H, J = 6.4 Hz), 4.65–4.67 (t, 1 H, J = 5.6 Hz), 4.76 (brs, 2 H), 4.83 (s, 1 H), 7.07–7.11 (t, 1 H, J = 7.6 Hz), 7.39–7.41 (m, 1 H), 7.59–7.61 (m, 1 H), 7.80 (s, 1 H), 8.19 (s, 1 H). 13C NMR (CD3OD, 100 MHz) δ: 7.9, 14.0, 19.5, 19.6, 23.2, 24.6, 31.6, 44.4, 63.8, 73.1, 77.1, 81.7, 88.1, 94.9, 120.1, 128.2, 131.4, 137.4, 137.9, 140.7, 143.1, 147.9, 149.9, 155.7. Anal. (C24H26IN5O2) C, H, N.

(1R,2R,3S,4R,5S)-4-(6-(3-Bromobenzylamino)-2-(hex-1-ynyl)-9H-purin-9-yl)bicyclo[3.1.0]hexane-2,3-diol (4g)

Yield: 90%; white solid; mp 124–126 °C; UV (MeOH) λmax 275 nm. MS (ESI+): [M + H]+ calcd for C24H27BrN5O2, 496.1343; found, 496.1350; [α]25D = +21.2 (c 0.2, MeOH). 1H NMR (CD3OD, 400 MHz) δ: 0.75–0.77 (m, 1 H), 0.96–1.00 (t, 3 H, J = 7.2 Hz), 1.34–1.37 (m, 1 H), 1.50–1.56 (m, 2 H), 1.60–1.71 (m, 3 H), 1.97–1.99 (m, 1 H), 2.45–2.48 (t, 2 H, J = 7.2 Hz), 3.85–3.87 (d, 1 H, J = 6.8 Hz), 4.65–4.68 (t, 1 H, J = 5.6 Hz), 4.80 (brs, 2 H), 4.83 (s, 1 H), 7.22–7.26 (m, 1 H), 7.36–7.41 (m, 2 H), 7.59 (s, 1 H), 8.19 (s, 1 H). 13C NMR (CD3OD, 100 MHz) δ: 7.9, 14.0, 19.2, 19.6, 23.2, 24.6, 31.6, 44.5, 63.9, 73.1, 77.2, 81.6, 88.0, 120.1, 123.5, 127.6, 131.3, 131.4, 131.8, 140.7, 143.2, 148.0, 150.0, 155.8. Anal. (C24H26BrN5O2) C, H, N.

(1R,2R,3S,4R,5S)-4-(6-(3-Chlorobenzylamino)-2-(hex-1-ynyl)-9H-purin-9-yl)bicyclo[3.1.0] hexane-2,3-diol (4h)

Yield: 90%; white solid; mp 109–111 °C; UV (MeOH) λmax 274 nm. MS (ESI+): [M + H]+ calcd for C24H27ClN5O2, 452.1848; found, 452.1842; [α]25D = +13.1 (c 0.2, MeOH). 1H NMR (CD3OD, 400 MHz) δ: 0.73–0.79 (m, 1 H), 0.90–1.00 (t, 3 H, J = 7.2 Hz), 1.34–1.37 (m, 1 H), 1.50–1.72 (m, 5 H), 1.96–2.04 (m, 1 H), 2.44–2.48 (t, 2 H, J = 7.2 Hz), 3.85–3.87 (d, 1 H, J = 6.8 Hz), 4.65–4.68 (t, 1 H, J = 5.6 Hz), 4.80 (brs, 2 H), 4.83 (s, 1 H), 7.24–7.34 (m, 3 H), 7.42–7.43 (m, 1 H), 8.19 (s, 1 H). 13C NMR (CD3OD, 100 MHz) δ: 7.9, 14.0, 19.5, 19.6, 23.1, 24.6, 31.6, 44.6, 63.8, 73.0, 77.1, 81.7, 88.1, 120.5, 127.1, 128.3, 128.8, 131.1, 135.4, 140.7, 142.9, 147.9, 149.9, 155.8. Anal. (C24H26ClN5O2) C, H, N.

(1R,2R,3S,4R,5S)-4-(6-(3-Fluorobenzylamino)-2-(hex-1-ynyl)-9H-purin-9-yl)bicyclo[3.1.0] hexane-2,3-diol (4i)

Yield: 70%; white solid; mp 99–101 °C; UV (MeOH) λmax 273 nm. MS (ESI+): [M + H]+ calcd for C24H27FN5O2, 436.2143; found, 436.2141; [α]25D = +2.5 (c 0.2, MeOH). 1H NMR (CD3OD, 400 MHz) δ: 0.73–0.79 (m, 1 H), 0.96–1.00 (t, 3 H, J = 7.2 Hz), 1.35–1.37 (m, 1 H), 1.49–1.71 (m, 5 H), 1.95–2.01 (m, 1 H), 2.44–2.47 (t, 2 H, J = 7.2 Hz), 3.85–3.87 (d, 1 H, J = 6.4 Hz), 4.65–4.67 (t, 1 H, J = 5.6 Hz), 4.80 (brs, 2 H), 4.83 (s, 1 H), 6.94–6.99 (m, 1 H), 7.12–7.15 (m, 1 H), 7.20–7.22 (m, 1 H), 7.30–7.35 (m, 1 H), 8.18 (s, 1 H). 13C NMR (CD3OD, 100 MHz) δ: 7.9, 13.9, 19.5, 19.6, 23.1, 24.6, 31.6, 44.6, 63.8, 73.0, 77.1, 81.7, 88.0, 115.3, 115.5, 120.0, 124.5, 131.2, 131.3, 140.7, 143.5, 148.0, 150.0, 155.8. Anal. (C24H26FN5O2) C, H, N.

6-Chloro-9-((3aR,3bR,4aS,5R,5aS)-Hexahydro-2,2-dimethylbicyclo[3.1.0]hex-1(5)-eno [3,2-d][1,3] dioxol-5-yl)-9H-purine (11)

Compound 6 (0.50 g, 2.93 mmol) was converted to 11 (0.63 g, 70%) as a white solid according to the same procedure used in the preparation of 8: mp 84–86 °C; UV(CH2Cl2) λmax 265 nm. MS (ESI+): [M + H]+ calcd for C14H16ClN4O2, 307.0956; found, 307.0951; [α]25D = −40.5 (c 0.2, MeOH). 1H NMR (CDCl3, 400 MHz) δ: 0.96–1.04 (m, 2 H), 1.24 (s, 3 H), 1.56 (s, 3 H), 1.70–1.75 (m, 1 H), 2.13–2.19 (m, 1 H), 4.68–4.70 (m, 1 H), 5.08 (s, 1 H), 5.36–5.39 (t, 1 H, J = 6.8 Hz), 8.18 (s, 1 H), 8.78 (s, 1 H). 13C NMR (CDCl3, 100 MHz) δ: 9.3, 24.3, 25.5, 26.0, 26.1, 61.4, 81.4, 89.1, 112.6, 132.1, 143.9, 151.3, 151.4, 152.3. Anal. (C14H15ClN4O2) C, H, N.

2-(Tributylstannyl)-6-chloro-9-((3aR,3bR,4aS,5R,5aS)-hexahydro-2,2-dimethylbicyclo[3.1.0] hex-1(5)-eno[3,2-d][1,3]dioxol-5-yl)-9H-purine (12)

To a stirred solution of 2,2,6,6-tetramethylpiperidine (TMP, 1.36 mL, 8.00 mmol) in dry hexane (5 mL) and dry THF (10 mL) was added n-butyllithium (5.6 mL, 1.5 M solution in hexanes, 8.47 mmol) dropwise at −78 °C over 30 min, and the mixture was stirred at the same temperature for 1 h. To this mixture, a solution of 11 (0.50 g, 1.60 mmol) in dry THF (10 mL) was added dropwise, and the mixture was stirred at −78 °C for 30 min. Tributyltin chloride (1.74 mL, 8.0 mmol) was successively added dropwise to the dark reaction mixture, and the mixture was stirred at the same temperature for another 2 h. The resulting dark solution was quenched by dropwise addition of a saturated aqueous NH4Cl solution (15 mL). After the mixture was stirred at room temperature for 15 h, the mixture was diluted with CH2Cl2 (15 mL). The organic layer was washed with saturated NaHCO3 solution, dried over anhydrous MgSO4, and filtered. The solvent was evaporated under reduced pressure. The crude syrup was purified by flash silica gel column chromatography (hexane/EtOAc = 5:1) to give 12 (0.68 g, 70%) as a colorless syrup: UV (MeOH) λmax 269 nm. MS (ESI+): [M + H]+ calcd for C26H42ClN4O2Sn, 597.2011; found, 595.2020; [α]25D = −36.5 (c 0.2, MeOH). 1H NMR (CDCl3) δ: 0.85–0.97 (m, 11 H), 1.13–1.38 (m, 15 H), 1.52–1.67 (m, 10 H), 2.03–2.13(m, 1 H), 4.79–4.81 (d, 1 H, J = 7.2 Hz), 4.96 (s, 1 H), 5.43- 5.47 (m, 1 H), 8.01 (s, 1 H). 13C NMR (CD3OD) δ: 9.4, 10.9, 24.3, 25.8, 26.1, 26.7, 27.4, 29.1, 62.1, 81.9, 89.1, 112.4, 131.0, 142.9, 149.8, 150.5, 182.1.

6-Chloro-2-(hex-1-ynyl)-9-((3aR,3bR,4aS,5R,5aS)-hexahydro-2,2-dimethylbicyclo[3.1.0]hex-1(5) -eno[3,2-d][1,3]dioxol-5-yl)-9H-purine (9)

To a stirred solution of 12 (0.40 g, 0.60 mmol) and copper iodide (0.013 g, 0.06 mmol) in anhydrous DMF (5 mL) was added 1-iodohexyne (0.124 mL, 0.6 mmol) in DMF (4 mL) dropwise via syringe pump over a period of 1 h at room temperature, and the reaction mixture was stirred at 50 °C for 5 h. The reaction mixture was cooled to room temperature, quenched with saturated NaHCO3 (5 mL) solution, and diluted with EtOAc (10 mL). The organic layer was separated, and the aqueous layer was further extracted with EtOAc (3 × 5 mL). The combined organic layers were washed with brine (5 mL) and water (5 mL), dried over anhydrous MgSO4, and filtered. The solvent was evaporated under reduced pressure, and the residue was purified by flash silica gel column chromatography (hexane/EtOAc = 3: 1) to give 9 (0.155 g, 60%) as a white foam, whose spectral data were identical to those of authentic sample.

Biological Assays

Cell Culture and Membrane Preparation

CHO cells expressing the recombinant hA1 or A3R and HEK-293 cells expressing the hA2AAR were cultured in Dulbecco’s modified Eagle’s medium (DMEM) and F12 (1:1) supplemented with 10% fetal bovine serum, 100 U/ml penicillin, 100 μg/mL streptomycin, and 2 μmol/mL glutamine. RBL-2H3 cells endogenously expressing rA3AR were cultured as described.[41] Cells were harvested by trypsinization. After homogenization and suspension, cells were centrifuged at 500g for 10 min, and the pellet was resuspended in 50 mM Tris·HCl buffer (pH 7.4) containing 10 mM MgCl2. The suspension was homogenized with an electric homogenizer for 10 s and was then recentrifuged at 20 000g for 20 min at 4 °C. The resultant pellets were resuspended in buffer containing 3 U/mL adenosine deaminase, and the suspension was stored at −80 °C until the binding experiments. The protein concentration was measured using the Bradford assay.[42]

Binding Assays at the hA1 and hA2AARs

For binding to the hA1AR, 50 μL of increasing concentrations of a test ligand and 50 μL of [3H]14 (2 nM, PerkinElmer, Boston, MA) were incubated with membranes (40 μg/tube) from CHO cells stably expressing the hA1 AR at 25 °C for 60 min in 50 mM Tris·HCl buffer (pH 7.4; MgCl2, 10 mM) in a total assay volume of 200 μL.[32] Nonspecific binding was determined using 10 μM of N6-cyclopentyladenosine (CPA, 17). For hA2AAR binding, membranes (20 μg/tube) from HEK-293 cells stably expressing the hA2AAR were incubated at 25 °C for 60 min with a final concentration of 15 nM [3H]15 (American Radiolabeled Chemicals, Inc., St. Louis, MO) in a mixture containing 50 μL of increasing concentrations of a test ligand and 200 μL of 50 mM Tris·HCl, pH 7.4, containing 10 mM MgCl2. Compound 16 (10 μM) was used to define nonspecific binding. The reaction was terminated by filtration with GF/B filters. Filters for A1 and A2AAR binding were placed in scintillation vials containing 5 mL of Hydrofluor scintillation buffer and counted using a PerkinElmer Tricarb 2810TR Liquid Scintillation Analyzer.

Binding Assay at the hA3AR and rA3AR

Each tube in the competitive binding assay contained 100 μL membrane suspension (20 μg protein), 50 μL [125I]13 (1.0 nM, PerkinElmer, Boston, MA), and 50 μL of increasing concentrations of the test ligands in Tris·HCl buffer (50 mM, pH 8.0) containing 10 mM MgCl2, 1 mM EDTA.[32] Nonspecific binding was determined using 10 μM of 16 in the buffer. The mixtures were incubated at 25 °C for 60 min. Binding reactions were terminated by filtration through Whatman GF/B filters under reduced pressure using a MT-24 cell harvester (Brandell, Gaithersburgh, MD, USA). Filters were washed three times with 9 mL ice-cold buffer. Filters for A3AR binding were counted using a PerkinElmer Cobra II γ-counter.

Cyclic AMP Accumulation Assay

Intracellular cAMP levels were measured with a competitive protein binding method.[34] CHO cells that expressed the recombinant hA2BAR or hA3AR were harvested by trypsinization. After centrifugation and resuspension in medium, cells were plated in 24-well plates in 0.5 mL medium. After 24 h, the medium was removed, and cells were washed three times with 1 mL DMEM, containing 50 mM N-(2-hydroxyethyl)piperazine-N′-2-ethanesulfonic acid (HEPES), pH 7.4. Cells were then treated with agonists and/or test compounds in the presence of rolipram (10 μM) and adenosine deaminase (3 units/mL). For assay of the hA3AR but not the hA2BAR, forskolin (10 μM) was added to the medium after 45 min. After the addition of forskolin, the incubation was continued an additional 15 min. The reaction was terminated upon removal of 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, protein kinase A (PKA) was incubated with [3H]cAMP (2 nM) in K2HPO4/EDTA buffer (K2HPO4, 150 mM; EDTA, 10 mM), 20 μL of the cell lysate, and 30 μL 0.1 M HCl or 50 μL of cAMP solution (0–16 pmol/200 μL for standard curve). Bound radioactivity was separated by rapid filtration through Whatman GF/C filters and washed once with cold buffer. Bound radioactivity was measured by liquid scintillation spectrometry.

Statistical Analysis

Binding and functional parameters were calculated using Prism 5.0 software (GraphPAD, San Diego, CA, USA). IC50 values obtained from competition curves were converted to Ki values using the Cheng–Prusoff equation.[43] Data were expressed as mean ± standard error of the mean.

Antifibrosis Assay

Immortalized murine proximal tubular cells (mProx24) derived from microdissected proximal tubular segments of C57BL6/J adult mouse kidneys were supplied from Dr. Sugaya at St. Marianna University School of Medicine, Kanagawa, Japan. mProx24 were maintained in DMEM supplemented with 10% fetal calf serum (FCS; Gibco), 100 U/ml penicillin, 100 μg/mL streptomycin, and 44 mM NaHCO3 under 5% CO2 environment at 37 °C. Cells were cultured in 6-well plate for mRNA analysis. At next day after seeding cell on 6-well plate, the cultured cells were growth-arrested with a DMEM medium containing 0.15% FCS for 24h. Each synthesized compound was dissolved in DMSO to 50 mM and it was diluted to 20 mM, 10 mM, and 1 mM. After cells were pretreated with the synthesized compound dissolved in DMEM containing 0.15% FCS for 1 h, treated with recombinant human transforming growth factor-β1 (hTGF β1, R&D Systems) 5 ng/mL for 6 h. Total RNA was extracted from mProx24 using Trizol (Invitrogen) according to the standard protocol. mRNA expressions were measured by real-time PCR using StepOnePlus (Applied Biosystems) with 20 μL reaction volume consisting of cDNA transcripts, primer pairs, and SYBR Green PCR Master Mix (Applied Biosystems). Quantifications were normalized to 18S. The sequences of mouse collagen Iα1 primer pairs are 5′-GAACATCACCTACCA CTGCA-3′ and 5′-GTTGGGATGGAGGGAGTTTA-3′.

Molecular Modeling

The X-ray crystal structure of the human A2A AR in complex with an agonist, 16 (PDB ID: 2YDV)[34] was retrieved from the protein data bank (PDB) and prepared using the Protein Preparation Wizard in Maestro v9.2 (Schrödinger, LLC, NY, U.S.A.), where water and ions were removed, hydrogen atoms were added and optimized, and then the protein was minimized using the Optimized Potentials for Liquid Simulations-all atom (OPLS-AA) 2005 force field. The structures of the molecules were sketched in the Maestro and energy minimized using Impact v5.7 (Schrödinger, LLC, NY, U.S.A.) considering conjugant gradient algorithm with the maximum minimization cycles of 1000 and convergence gradient of 0.001 kJ/mol-Å. The four docking programs Glide-SP (standard precision), Glide-XP (extra precision), GOLD, and Surflex-dock showed consistent results, and the Glide-XP docking results are presented. The receptor grid box with 10 Å around the centroid of the cocrystallized NECA was generated. The best binding poses of 2, 4a, and 4f were selected for the calculation of the receptor–ligand binding free energy (ΔGbind) using Prime molecular mechanics-generalized Born surface area (MM-GBSA) module (Schrödinger, LLC, NY, U.S.A.).

The Ballesteros–Weinstein double-numbering system[44] is used to describe the transmembrane (TM) location of the amino acids. Along with numbering their positions in the primary amino acid sequence, the residues have numbers in parentheses (X.YZ) that indicate their position in each transmembrane (TM) helix (X), relative to a conserved reference residue in that TM helix (YZ).