Adenine Nucleotide Analogues Locked in a Northern Methanocarba Conformation: Enhanced Stability and Potency as P2Y₁ Receptor Agonists

Abstract

Preference for the Northern (N) ring conformation of the ribose moiety of nucleotide 5′-triphosphate agonists at P2Y₁, P2Y₂, P2Y₄, and P2Y₁₁ receptors, but not P2Y₆ receptors, was established using a ring-constrained methanocarba (a 3.1.0-bicyclohexane) ring as a ribose substitute (Kim et al. J. Med. Chem. 2002, 45, 208–218.). We have now combined the ring-constrained (N)-methanocarba modification of adenine nucleotides with other functionalities known to enhance potency at P2 receptors. The potency of the newly synthesized analogues was determined in the stimulation of phospholipase C through activation of turkey erythrocyte P2Y₁ or human P2Y₁ and P2Y₂ receptors stably expressed in astrocytoma cells. An (N)-methanocarba-2-methylthio-ADP analogue displayed an EC₅₀ at the hP2Y₁ receptor of 0.40 nM and was 55-fold more potent than the corresponding triphosphate and 16-fold more potent than the riboside 5′-diphosphate. 2-Cl–(N)-methanocarba-ATP and its N⁶-Me analogue were also highly selective, full agonists at P2Y₁ receptors. The (N)-methanocarba-2-methylthio and 2-chloromonophosphate analogues were full agonists exhibiting micromolar potency at P2Y₁ receptors, while the corresponding ribosides were inactive. Although β,γ-methylene-ATP was inactive at P2Y receptors, β,γ-methylene-(N)-methanocarba-ATP was a potent hP2Y₁ receptor agonist with an EC₅₀ of 160 nM and was selective versus hP2Y₂ and hP2Y₄ receptors. The rates of hydrolysis of Northern (N) and Southern (S) methanocarba analogues of AMP by rat 5′-ectonucleotidase were negligible. The rates of hydrolysis of the corresponding triphosphates by recombinant rat NTPDase1 and 2 were studied. Both isomers were hydrolyzed by NTPDase 1 at about half the rate of ATP hydrolysis. The (N) isomer was hardly hydrolyzed by NTPDase 2, while the (S) isomer was hydrolyzed at one-third of the rate of ATP hydrolysis. This suggests that new, more stable and selective nucleotide agonists may be designed on the basis of the (N)-conformation, which greatly enhanced potency at P2Y₁ receptors.

Introduction

The P2 nucleotide receptors consist of two families: G-protein-coupled receptors (GPCR) termed P2Y, of which seven mammalian subtypes (P2Y_{1,2,4,6,11,12,13}) have been cloned, and ligand-gated cation channels termed P2X, of which seven mammalian subtypes (P2X_1–7) have been cloned.^1–3 Adenine nucleotides are required for activation of P2Y₁, P2Y₁₁, P2Y₁₂, and P2Y₁₃ subtypes, while uracil nucleotides activate P2Y₂, P2Y₄, and P2Y₆ subtypes. P2Y receptors may couple to multiple second messengers, but all except the recently cloned^3–5 P2Y₁₂ and P2Y₁₃ receptors lead to activation of phospholipase C (PLC) via G_q, resulting in a rise in intracellular calcium. The G_i-coupled P2Y₁₂ and P2Y₁₃ receptors occur on a more distant branch of the sequence dendrogram than the other members of the P2Y family⁶ and are closer in sequence to a newly characterized GPCR that is selective for sugar nucleotides (e.g. UDP-glucose).⁷

The medicinal chemistry of P2 receptors is underdeveloped compared to many other GPCRs.^8–10 Reasons for the lack of progress in this field include nonselectivity, chemical instability, heterogeneity, and nonbioavailability of known ligands. Other difficulties in characterizing the receptors include the enzymatic instability of the phosphate groups of nucleotides to form biologically active metabolites. Indeed, most of the known P2 antagonists also inhibit ectonucleotidase activity, thus augmenting the effects of the endogenous agonist(s). Receptor heterogeneity,¹¹ oligomerization,¹² and desensitization¹³ are further complications.

Despite the above-mentioned difficulties, the SAR of ATP analogues (e.g., 1–5; Chart 1) at P2 receptors has been probed.^8–10 Selective agonists (P2Y₁, P2Y₆)14,15 and antagonists (P2Y₁, P2Y₁₂, P2X₁, P2X₇)14,16–¹⁹ have been reported for only a few of the subtypes. 2-Methylthio-ATP (1, 2-MeSATP) is one of the most potent agonists at a number of P2Y and P2X receptors.¹⁰ It was formerly regarded as selective for P2Y receptors; however, by adjustment for its lability in classical smooth muscle assays in which P2X receptors have been assayed, it is a highly potent agonist within the P2X superfamily.²⁰ At the P2Y₁ receptor, which is present in the heart, skeletal and various smooth muscles, prostate, ovary, and brain,²¹ the potency order for activation is 2-methylthioadenosine 5′-diphosphate (2-MeSADP) > 2-Me-SATP > ADP > ATP, and AMP and UTP are inactive. 2-Chloro-ATP and N⁶-methyl-ATP (2 and 3, respectively) are moderately potent agonists at P2Y₁ receptors.²² MRS 2055 4 (N⁶-methyl-2-(5-hexenyl)thioadenosine 5′-triphosphate) is an agonist with micromolar potency at a glial adenylate cyclase coupled P2Y receptor (now identified as identical to P2Y₁₂)^23,24 but is inactive at P2Y₁ receptors. β,γ-Methylene ATP 5 activates P2X₁, but not P2Y₁, receptors.^21,22

Chart 1.

P2Y receptor ligands are being explored for therapeutic applications in the cardiovascular, endocrine, and other systems. A selective P2Y₁ receptor agonist may have potential as an antihypertensive or antidiabetic agent.^25–27 A selective antagonist of the P2Y₁ receptor may be potentially useful in antithrombotic therapy,^28,29 since activation of this subtype is associated with the initial shape change during platelet aggregation.²⁹ ATP analogues have been shown to be effective insulin secretagogues.³⁰ A P2Y₁ receptor in osteoclasts has been linked to bone disintegration.³¹ First cloned from chick brain,³² the distribution of the P2Y₁ receptor in the human brain has been described³³ and abnormalities in P2Y₁ receptor levels have been detected in the Alzheimer’s brain.³⁴ A linkage between P2Y₁ receptors and the MAP kinase/apoptosis pathway has been proposed.³⁵ Selective agonists and antagonists at the P2Y₁ receptor have been reported.^10,14

3′,5′-Bisphosphate nucleotides have been explored as selective antagonists of the P2Y₁ receptor, and nearly nanomolar affinities have been achieved.¹⁴ The ribose moiety of P2Y₁ receptor antagonists may be replaced with carbocylics, smaller and larger rings, acyclics, and conformationally constrained rings,¹⁴ resulting in retention of affinity for the receptor. The ribose rings of nucleosides and nucleotides can adopt a range of conformations, although their target receptors likely prefer specific conformations. For P2Y₁ antagonists we have replaced the ribose moiety with a carbocyclic ring locked in a preferred conformation, a fusion of cyclopropane and cyclopentane rings known as the methanocarba modification.^36–38 Two structural variations, depending on the position of the cyclopropane ring, restrict the ring pucker, i.e., hold the ribose-like ring (pseudosugar) in either a Northern (N, 2′-exo) 6a or Southern (S, 2′-endo) 6c envelope conformation (Chart 1), as defined in the pseudorotational cycle. Preference for the Northern (N) ring conformation of the ribose moiety of nucleoside 5′-triphosphate agonists at P2Y₁, P2Y₂, P2Y₄, and P2Y₁₁ receptors, but not P2Y₆ receptors, was established using a constrained methanocarba (bicyclic) ring.³⁹

In the present study we combined the ring-constrained (N)-methanocarba modification of adenine nucleotides with other functionalities known to enhance potency at P2 receptors. In several cases, the constrained ring conferred dramatic increases in agonist potencies at P2Y₁ receptors, and generally the modification preserved the potencies at P2Y₂ receptors.

Results

Chemical Synthesis

We prepared methanocarbocyclic analogues of various synthetic adenine nucleotides (Table 1) in which fused cyclopropane and cyclopentane rings fixed the pseudoribose moiety in a rigid (N)-envelope conformation. The unsubstituted ATP 6a and AMP 7a analogues were reported in a previous study.³⁹ The new nucleotide analogues 8a–15a and 17a were prepared in the ammonium salt form according to the methods shown in Schemes 1 and 2 and were tested biologically as agonists at various P2Y receptors (Table 1). Identity was confirmed using NMR (¹H and ³¹P) and high-resolution mass spectrometry, and purity was demonstrated using high-pressure liquid chromatography (HPLC) in two different solvent systems (Table 2).

Table 1.

Biological Potencies of Methanocarba Derivatives of Adenine Nucleotides and Their Corresponding Riboside Analogues (Stimulation of Phospholipase C through Activation of P2Y Receptors)

compd	analogue	subtype	a, methanacarba EC50 (nM) of (N)-methanocarba analoguea	b, riboside EC50 (nM) of corresponding ribose analoguea	ref
6	ATP	tP2Y1	14 ± 3	2800 ± 700
		hP2Y1	52 ± 22	1500 ± 200
		hP2Y2	91 ± 5	85 ± 12
		hP2Y4	32% at 10 μM	c
		hP2Y11	35 ± 5% at 10 μM	17300 ± 2800
7	AMP	tP2Y1	d	d
8	2-MeSATP	tP2Y1	3.7 ± 1.7	8 ± 2	40
		hP2Y1	23 ± 11	34 ± 11
9	2-MeSADP	tP2Y1	0.88 ± 0.37	6 ± 3	40
		hP2Y1	0.40 ± 0.23	6.4 ± 3.5
		hP2Y2	54 ± 8% at 10 μM	d
		hP2Y4	d	d
		hP2Y11	d	d
10	2-MeSAMP	tP2Y1	1190 ± 500	15% at 100 μM
		hP2Y1	590 ± 20	d
		hP2Y11	d	d
11	2-MeS(O)AMP	tP2Y1	36.6 ± 12.5	d
		hP2Y1		d
12	2-C1-ATP	tP2Y1	3 ± 1	720 ± 20	22
		hP2Y1	7 ± 1	770 ± 120	46
		hP2Y2	1900b	2300 ± 300
13	2-C1-AMP	tP2Y1	3900 ± 2900	d
		hP2Y1	1890 ± 770	d
14	N6-Me-ATP	tP2Y1	92 ± 32	19000 ± 6000	22
		hP2Y1	33b
		hP2Y2	197 ± 2
15	2-C1-N6-MeATP	tP2Y1	6 ± 3
		hP2Y1	9 ± 1
		hP2Y2	8400 ± 3600
16	2-C1-N6-MeAdo-5′-CONH-Me	tP2Y1	d
17	β,γ-Me-ATP	tP2Y1	11,300 ± 1700	d	22
		hP2Y1	158b	d
		hP2Y2	d	d
		hP2Y4	d

^aAverage ± SEM data for 6a and 7a reported in ref 39. A percentage refers to percent of maximal effect.

^bAverage of two determinations.

^cATP is an antagonist at hP2Y₄ receptors.50

^dNo effect as either agonist or antagonist at 10 μM.

Scheme 1.

Synthesis of (N)-Methanocarbaadenosine 5′-Phosphate Derivatives^a

^a Reagents: (i) DEAD, Ph₃P, THF; (ii) NH₃ or CH₃NH₂, THF; (iii) palladium black, HCO₂H, methanol; (iv) NaSCH₃, DMF, 90–100 °C; (v) di-tert-butyl N,N-diethylphosphoramidite, tetrazole, THF, room temp, 20 min and then m-CPBA, −78 °C to room temp; (vi) DOWEX 50×8–200, methanol, 60–70 °C, ~80%; (vii) carbonyldiimidazole, DMF, room temp, 6 h; (viii) 5% triethylamine/H₂O/MeOH, 2 h; (ix) tributylammonium pyrophosphate or tributylammonium phosphate, DMF, 3 days.

Scheme 2.

Synthesis of the (N)-Methanocarba Analogue of β,γ-Me-ATP^a

^a Reagents: (i) carbonyldiimidazole, DMF, room temp, 6 h; (ii) triethylamine/H₂O/MeOH (1:10:10), 2 h; (iii) methylenediphosphonic acid, triethylamine, DMF, 3 days.

Table 2.

Synthetic Data for Nucleotide Derivatives, Including Structural Verification Using High-Resolution Mass Spectrometry and Purity Verification Using HPLC

compd	formula	FAB (M – H+) m/z		HPLC retention time in min (purity in %a)		yieldb(%)
		calcd	found	system A	system B unless otherwise noted
8a	C13H20N5O12P3S	561.9964	561.9966	11.0 (99)	19.7 (99)	35
9a	C13H19N5O9P2S	482.0301	482.0291	10.5 (98)	19.4 (98)	40
10a	C13H18N5O6PS	402.0637	402.0650	9.0 (98)	15.6 (99)	60
10b	C11H16N5O7PS	394.0586	394.0591	14.7 (98)	16.2 (99)	16
11a	C13H18N5O7PS	418.0586	418.0607	6.3 (97)	12.7 (97)	65
11b	C11H16N5O8PS	410.0535	410.0545	11.2 (97)	17.1 (97)	16
12a	C12H17N5O12P3Cl	551.9667	551.9673	7.4 (97)	19.1 (98)	30
13a	C12H15N5O6PCl	390.0370	390.0384	7.7 (99)	11.1(99)	66
13b	C10H13N5O7PS	382.0319	382.0322	9.2 (99)	17.5 (99)	42
14a	C13H20N5O12P3	530.0243	530.0247	6.0 (98)	8.8 (98)	25
15a	C13H19N5O12P3Cl	563.9853	563.9866	11.2 (98)	6.6 (98)	32
16a	C14H17N6O3Cl	353.1129	353.1139	10.4 (98)	10.3c (99)	d
17a	C13H20N5O11P3	514.0294	514.0306	3.2 (98)	15.6 (97)	32

^aPurity of each derivative was ≥97%, as determined using HPLC with two different mobile phases. System A: gradient of 0.1 M TEAA/CH₃CN from 95/5 to 40/60. System B: gradient of 5 mM TBAP/CH₃CN from 80/20 to 40/60. System C: gradient of 0.1 M TEAB/ CH₃CN from 100/0 to 90/10 in 20 min.

^bThe percent yields refer to the overall yield for each phosphorylation sequence.

^cSystem C.

^dSynthesis described in ref 45.

As in the previous study,³⁹ since the phosphorus oxychloride method^40,41 was unsatisfactory for phosphorylation of (N)-methanocarba nucleosides at the 5′-hydroxyl group, an alternative multistep process was used. This consisted of the formation of a monophosphate using the phosphoramidite method followed by condensation with additional phosphate or pyrophosphate groups using carbonyldiimidazole. By this method, the monophosphate analogues were cleanly converted to either di- or triphosphate via a phosphorimidazolidate intermediate.⁴¹ During the sequence, it was necessary to hydrolyze the cyclic carbonates⁴² 38–41, which were formed at the 2′,3′-dihydroxy groups, using a 5% triethylamine solution.

The (N)-methanocarba analogues of various ATP derivatives substituted at the 2- and/or 6-positions of the purine base moiety or at the β,γ position of the triphosphate moiety were prepared by this approach. To construct the pseudoribose ring leading to the (N)-methanocarbaadenosine precursors, a 3.1.0-bicyclohex-ane structure was prepared by the general approach of Lee et al.,⁴³ which utilized ring closure metathesis. An adenine precursor 18 or 19 was condensed with the protected bicycloheptane derivative 20⁴³ using Mitsunobu conditions to give 21³⁹ or 22 (Scheme 1). To introduce an amino or methylamino group at the 6-position, it was necessary to substitute 6-Cl with the corresponding amine following the Mitsunobu reaction to yield 23–26. Deprotection of the 5′-benzyl group of 23–26 was carried out smoothly using Pd black/formic acid, even in the presence of 2-chloro substitution, to give 27, 29, and 30. This was followed by 5′-phosphorylation using a phosphoramidite method⁴⁴ to give the di-tert-butyl protected monophosphates 31, 34, and 35. Alternatively, the 2-chloro group of 27 was replaced with methylthio to give the nucleoside 28, leading to the protected monophosphate 32. In the process of synthesis of the 2-methylthio-AMP derivative 32, the corresponding sulfoxide 33 was isolated as a byproduct. The monophosphates 31–35 were deprotected and either tested biologically or carried further to the di- and triphosphates by the method shown in Scheme 1. The (N)-methanocarba β,γ-methylene triphosphate analogue 17a was prepared from the unsubstituted AMP equivalent 7a³⁹ using methylene diphosphonate instead of pyrophosphate (Scheme 2).

Three riboside 5′-monophosphates 10b, 11b, and 13b were prepared using methods shown in Scheme 3 for comparison to the corresponding (N)-methanocarba derivatives. The isopropylidene group was used as protection for the 2′- and 3′-hydroxyl groups, and phosphorylation was carried out using the phosphoramidite method, as above.

Scheme 3.

Synthesis of Adenosine 5′-Monophosphate Analogues^a

^a Reagents: (i) tetrazole, di-tert-butyl N,N-diethylphosphoramidite, THF, then m-CPBA, CH₂Cl₂; (ii) DOWEX 50W×8–200, H₂O/ MeOH; (iii) NaSCH₃, DMF, 90 °C.

Biological Activity

(N)-Methanocarba analogues (denoted “a”, Table 1) were tested for agonist activity by measuring P2Y₁-receptor-promoted phospholipase C activity in turkey erythrocyte membranes and by measuring inositol phosphate accumulation in 1321N1 human astrocytoma cells stably expressing the human P2Y₁, P2Y₂, P2Y₄, or P2Y₁₁ receptors. In each case, full concentration effect curves were generated and EC₅₀ values from at least three separate experiments were determined. The agonist potencies of the corresponding ribose-containing nucleotides (denoted “b”) were simultaneously determined for comparison purposes.

Given the high potencies observed with (N)-methanocarba analogues of ATP and UTP and the much greater potency of the (N)-methanocarba as opposed to (S)-methanocarba conformation, we extended our studies to (N)-methanocarba derivatives synthesized containing various functionalities that we and others have shown to produce potent or receptor-selective adenine nucleotide analogues.^8–10 As has been observed for adenine nucleotides, substitution at the 2-position, such as 2-methylthio, in the (N)-methanocarba-nucleotide series enhanced P2Y₁ receptor potency. Thus, the high P2Y₁ potency of 2-MeSATP 8b (Figure 1) was preserved in the corresponding (N)-methanocarba analogue 8a. The most potent P2Y₁ agonist in this series was the 2-methylthio diphosphate analogue 9a (Figure 2), which displayed an EC₅₀ of 0.88 nM at the turkey P2Y₁ receptor and was 7-fold more potent than the corresponding ribose analogue 9b and 26-fold more potent than 8a. At other receptors 9a was inactive (P2Y₄, P2Y₁₁) or weakly active (P2Y₂). The (N)-methanocarba-2-methylthio monophosphate analogue 10a and the (N)-methanocarba-2-chloro monophosphate analogue 13a were full agonists exhibiting micromolar potency at the turkey and human P2Y₁ receptors. Oxidation of the thioether to a sulfoxide, in 11a, reduced tP2Y₁ receptor potency by 30-fold. The corresponding 5′-monophosphate ribosides 10b, 11b, and 13b were inactive at P2Y₁ receptors. The (N)-methanocarba-2-chloro 5′-triphosphate analogue 12a was 240-fold more potent at the turkey P2Y₁ receptor than the corresponding riboside 12b.

Figure 1.

Effects of the (N)-methanocarba modification of 2-MeS-ATP on phospholipase C activity in 1321N1 astrocytoma cells expressing the hP2Y₁ receptor. Concentration-dependent stimulation of inositol phosphate formation by compounds 8a (□) and 8b (■) is shown. The data shown are typical curves for at least three experiments.

Figure 2.

Effects of the (N)-methanocarba modification of 2-MeS-ADP on phospholipase C activity in 1321N1 astrocytoma cells expressing the hP2Y₁ receptor. Concentration-dependent stimulation of inositol phosphate formation by compounds 9a (○) and 9b (●) is shown. The data shown are typical curves for at least three experiments.

As for the riboside series, substitution at the exocyclic amine of adenine was tolerated for (N)-methanocarba analogues acting as P2Y receptor agonists. The N⁶-methyl 5′-triphosphate analogue 14a was a potent agonist at the turkey and human P2Y₁ receptors and was 5-fold more potent as a full agonist at the hP2Y₁ than at the hP2Y₂ receptor. Combination of N⁶-methyl and 2-chloro modifications in 15a resulted in a very potent agonist for the turkey and human P2Y₁ receptors and selectivity for the hP2Y₁ receptor versus hP2Y₂ receptor of 930-fold. None of the (N)-methanocarba analogues of adenine nucleotide derivatives studied here were agonists at the human P2Y₄ receptor (data not shown).

Since the 2-chloro-N⁶-methyl 5′-triphosphate analogue 15a was a potent P2Y₁ receptor agonist, compound 16a⁴⁵ was designed to evaluate the feasibility of replacing the triphosphate moiety with an uncharged moiety. However, it was inactive either as an agonist or as an antagonist at the turkey P2Y₁ receptor. β;,γ-Me-ATP 17b was inactive as an agonist at the P2Y₁ receptor.46,47 In contrast, and consistent with the potency enhancement observed for other nucleotide analogues, the (N)-methanocarba-β,γ-methylene analogue 17a was a relatively potent (EC₅₀ = 11 μM) full agonist at the turkey P2Y₁ receptor and a very potent (EC₅₀ = 160 nM) agonist at the hP2Y₁ receptor (Figure 3). In contrast, this analogue was inactive at the P2Y₂ and P2Y₄ receptors.

Figure 3.

Effects of the (N)-methanocarba modification in triphosphate-modified adenine nucleotides on activation of phospholipase C in 1321N1 astrocytoma cells expressing hP2Y₁ receptors, showing concentration-dependent stimulation of inositol phosphate formation by compounds 14a (●) and 14b (○). The data shown are typical curves for at least three experiments.

Because of the well-established lability of nucleotide analogues in pharmacological studies, the enzymatic stability of the simple (N)-methanocarba and (S)-methanocarba analogues of AMP and ATP was investigated (Tables 3 and 4). Neither the (N)-methanocarba nor (S)-methanocarba 5′-monophosphate derivatives 7a and 7c, respectively, were substrates for recombinant rat ecto- 5′-nucleotidase (Table 3).⁴⁸ Hydrolysis of the (N)-methanocarba and (S)-methanocarba derivatives of ATP 6a and 6c, respectively, by recombinant rat ectoapyrase (NTPDase1) and ecto-ATPase (NTPDase2)⁴⁹ stably expressed in CHO cells was also examined (Table 4). The (N)-methanocarba-ATP analogue was resistant to hydrolysis by NTPDase2 (4% as effective as ATP as a substrate), whereas the corresponding (S)-methanocarba-ATP was hydrolyzed by NTPDase2 at one-third (34%) the rate of ATP hydrolysis. In contrast, both molecules were good substrates for NTPDase1, at about half the rate of ATP hydrolysis (52% and 57%, respectively).

Table 3.

Hydrolysis Rates of Monophosphate Derivatives as Substrates of Purified Recombinant Rat Ecto-5′-nucleotidase

substrate	activitya (%)
AMP	100
7a, (N)-methanocarba-AMP39	0.14
7c, (S)-(±)-methanocarba-AMP39	0.0

^aValues are means of two experiments with duplicate determinations in each. The catalytic activity in the presence of AMP (100%) corresponded to 5.6 nmol/(min μg) protein.

Table 4.

Hydrolysis of the Triphosphate Derivatives as Substrates of Rat NTPDase1 (Ectoapyrase) and NTPDase2 (Ecto-ATPase) Stably Expressed in CHO Cells

substrate	activitya (%)
	NTPDase1	NTPDase2
ATP	100	100
6a, (N)-methanocarba-ATP39	52	4
6c, (S)-(±)-methanocarba-ATP39	57	34

^aThe hydrolysis rate with the two derivatives as substrate is compared to that of ATP (100%) for each enzyme. Values are means of two experiments with duplicate determinations in each. Nontransfected CHO cells were analyzed as control. They contained no significant activity for the extracellular hydrolysis of ATP. 100% values correspond to the following: NTPDase1, 42.1 nmol/(min 10⁶ cells); NTPDase2, 13.5 nmol/(min 10⁶ cells).

Discussion

In the previous study,³⁹ we concluded that substitution of the ribose moiety with fused cyclopentane and cyclopropane rings in a pseudorotational (N)-conformation produced agonists at certain adenine and uracil nucleotide activated P2Y receptors. At P2Y₁, P2Y₂, P2Y₄, and P2Y₁₁ receptors, but not P2Y₆ receptors, an (N)-methanocarba analogue was much more potent than the corresponding molecule constrained in the Southern conformation. Here, we have extended the synthesis to multiple substituted analogues of ring-constrained (N)-methanocarba-ATP, ADP, and AMP and illustrated that these derivatives exhibit as high or even higher potency for activation than the corresponding riboside, depending on the P2Y subtype involved. In some cases, alreadypotent ATP analogues were increased in potency upon (N)-methanocarba modification. In other cases, an inactive compound became biologically active upon such substitution.

A key finding in this study is that remarkably potent agonists, especially at the P2Y₁ receptor, can be produced by introducing the (N)-methanocarba modification into molecules having other substitutions known to increase P2Y receptor potency. This was very notable in the (N)-methanocarba analogues of 2-Cl-ATP 13a and N⁶-Me-ATP 14a, which exhibited increases in potency in the (N)-methanocarba form versus ribosides of >200-fold. However, perhaps the most remarkable enhancement of activity occurred with β,γ-Me-ATP, although a full agonist at P2X receptors is not an agonist (or antagonist) at the P2Y₁ and other P2Y receptors. However, the (N)-methanocarba analogue of β,γ-Me-ATP was a full agonist at the turkey and human P2Y₁ receptors, exhibiting an EC₅₀ for the human receptor of 160 nM. We have not yet extended this analysis to a direct comparison of relative activities of (N)- versus (S)-conformers to molecules that also possess other functional groups, e.g., 2-MeS or N⁶-CH₃ substitution, known to increase binding affinity and selectivity for P2Y receptor agonists.

As we have observed with many different molecules that were synthesized and tested previously, results in human and turkey P2Y₁ receptors were similar. The potencies determined in assays of phospholipase C in turkey erythrocyte membranes corresponded closely to the EC₅₀ values determined for the same molecules at the recombinant human P2Y₁ receptor stably expressed in 1321N1 human astrocytoma cells. This correlation was closest with a series of analogues containing the (N)-methanocarba substitution with various functional groups, e.g., 2-Cl, 2-MeS, N⁶-CH₃, that we and others had previously demonstrated to increase P2Y receptor agonist potency in the riboside series. An explanation for the relatively large difference in apparent potency of the (N)-methanocarba analogue of β,γ-Me-ATP at the turkey versus human P2Y₁ receptor is not readily apparent.

Although not studied in detail here, the (N)-methanocarba analogues exhibit resistance to hydrolysis relative to naturally occurring nucleotides. This observation may be particularly significant for 5′-monophosphate derivatives, since we have shown that they potently and selectively activate P2Y₁ receptors but may be some-what stable to dephosphorylation by 5′-nucleotidase. However, it should be pointed out that this resistance to hydrolysis would at most only make a minor contribution to the very large difference in potencies seen between certain (N)-methanocarba compounds and their natural ribose counterpart. That is, the K_m/V_max values of ecto-nucleotidases on 1321N1 cells do not result in sufficient nucleotide hydrolysis to result in large shifts in the activation curves of, for example, ATP or UTP. More work will be required to obtain a clearer view on the capacity of the (N)-methanocarba molecules to serve as substrates for nucleotide-metabolizing ectoenzymes.

Thus, we established that constraining the ribose-like ring as methanocarba analogues was an important principle for producing P2Y receptor-subtype selectivity, similar to our previous demonstration of this phenomenon for P2Y₁ receptor antagonists. The (N)-methanocarba analogues generally displayed increased potency over the corresponding nucleotide agonists at the P2Y₁ and P2Y₂ receptors. The high potency of several of the derivatives (e.g., the selective P2Y₁ receptor agonist 9a) suggested the possibility of preparation of high-affinity radioligands in the (N)-methanocarba series. The potent and selective P2Y₁ agonists (EC₅₀ in nM at the tP2Y₁ receptor unless otherwise noted) found in the present study were the following: 9a (0.88), 12a (3), 15a (9), 8a (23), 17a (160 at human), and 10a (590). Compound 14a was roughly equipotent at P2Y₁ and P2Y₂ receptors.

In conclusion, remarkable P2Y₁ potency was generated in derivatives that combined the (N)-methanocarba modification with other functional groups known to enhance potency and/or selectivity in P2 receptor agonists. These molecules provide a new armamentarium for the study of P2Y receptors. Moreover, they provide an attractive new chemical backbone on which to build highly potent and selective P2Y receptor agonists and antagonists that, in lacking a ribose moiety, are not nucleotides, and therefore, they are less likely to bind to the myriad of nucleotide binding proteins that are not P2Y receptors. Furthermore, enhanced stability to some nucleotidase enzymes is indicated in the methanocarba series. These findings promise to be useful for defining the microscopic determinants of the binding sites on these receptors and for designing novel pharmacological probes and/or therapeutic agents.

Experimental Section

1. Chemical Synthesis

Nucleosides and synthetic reagents were purchased from Sigma Chemical Co. (St. Louis, MO) and Aldrich (Milwaukee, WI). 2,6-Dichloropurine was obtained from Sigma. The protected intermediate 21 was synthesized in our laboratory as described.^43,45 (N)-Methanocarba analogues of ATP 6a and AMP 7a and the corresponding (±)-(S)-methanocarba analogues 6c and 7c, respectively, were prepared as described.³⁹ Compound 16a was prepared as described.⁴⁵

¹H NMR spectra were obtained with a Varian Gemini-300 spectrometer (300 MHz) using D₂O/CDCl₃ and CD₃OD as a solvent. ³¹P NMR spectra were recorded at room temperature by use of a Varian XL-300 spectrometer (121.42 MHz); orthophosphoric acid (85%) was used as an external standard.

Purity of compounds was checked using a Hewlett-Packard 1090 HPLC apparatus equipped with an SMT OD-5-60 RP-C18 analytical column (250 mm × 4.6 mm; Separation Methods Technologies, Inc., Newark, DE) in two solvent systems.

System A was a linear gradient solvent system consisting of 0.1 M TEAA/CH₃CN from 95/5 to 40/60 in 20 min, and the flow rate was 1 mL/min. System B was a linear gradient solvent system consisting of 5 mM TBAP/CH₃CN from 80/20 to 40/60 in 20 min, and the flow rate was 1 mL/min. System C was a linear gradient solvent system consisting of 0.1 M TEAB/CH₃CN from 100/0 to 90/10 in 20 min, and the flow rate was 1 mL/min.

Peaks were detected by UV absorption using a diode array detector. All derivatives tested for biological activity showed ≥97% purity in the HPLC systems.

Low-resolution CI(NH₃) (chemical ionization) mass spectra were carried out with Finnigan 4600 mass spectrometer, and high-resolution EI (electron impact) mass spectrometry was carried out with a VG7070F mass spectrometer at 6 kV. High-resolution FAB (fast-atom bombardment) mass spectrometry was performed with a JEOL SX102 spectrometer using 6 kV Xe atoms following desorption from a glycerol matrix.

Purification of the nucleotide analogues for biological testing was carried out on DEAE-A25 Sephadex columns as described above.

General Phosphorylation Procedure: Synthesis of (N)-Methanocarbaadenosine Derivatives (23–26). (1′S,2′ R,3′S,4′R,5′S)-4-(6-Amino-9H-purin-9-yl)-1-[(phenylmethoxy)methyl]bicyclo[3.1.0]hexane-2,3-(O-isopropylidene) (23)

To a solution of 21 (0.2 g, 0.47 mmol) in i-PrOH (3 mL) was added NH₃ (2 M solution in i-PrOH, 5 mL, 10 mmol), and the reaction mixture was heated at 90 °C in a closed tube for 15 h for complete reaction. The resulting mixture was concentrated under reduced pressure, and the residue obtained was purified by flash chromatography using 9/1 CHCl₃/MeOH to furnish 0.182 g of 23 (95%). 1H NMR (CDCl₃) δ 8.35 (s, 1H), 8.30 (s, 1H), 7.37 (s, 5H), 5.87 (bs, 2H), 5.31 (d, 1H, J = 6.8 Hz), 5.13 (s, 1H), 4.64 (qAB, 2H, J = 11.7, 20.5 Hz), 4.51 (d, 1H, J = 6.8 Hz), 3.97 (d, 1H, J = 10.7 Hz), 3.35 (d, 1H, J = 10.7 Hz), 1.72–1.62 (m, 1H), 1.55 (s, 3H), 1.32–1.26 (m, 1H), 1.23 (s, 3H), 1.00–0.93 (m, 1H).

Compounds 24 and 25 were synthesized from 21, and 26 was synthesized from 22. All were produced in 90–95% yields.

(1′S,2′R,3′S,4′R,5′S)-4-(6-Methylamino-9H-purin-9-yl)-1-[(phenylmethoxy)methyl]bicyclo[3.1.0]hexane-2,3-(O-isopropylidene) (24)

¹H NMR (CDCl₃) δ 8.41 (s, 1H), 8.22 (s, 1H), 7.40–7.30 (m, 5H), 6.01 (bs, 1H), 5.30 (d, 1H, J = 7.2 Hz), 5.12 (s, 1H), 4.62 (qAB 2H, J = 12.1, 20.9 Hz), 4.51 (d, 1H, J = 7.2 Hz), 3.94 (d, 1H, J = 9.9 Hz), 3.35 (d, 1H, J = 9.9 Hz), 3.21 (s, 3H), 1.70–1.62 (m, 1H), 1.55 (s, 3H), 1.29–1.25 (m, 1H), 1.22 (s, 3H), 0.98–0.92 (m, 1H).

(1′S,2′R,3′S,4′R,5′S)-4-(6-Amino-2-chloro-9H-purin-9-yl)-1-[(phenylmethoxy)methyl]bicyclo[3.1.0]hexane-2,3-(O-isopropylidene) (25)

¹H NMR (CDCl₃) δ 8.22 (s, 1H), 7.40–7.33 (m, 5H), 6.48 (bs, 2H), 5.32 (d, 1H, J = 7.15 Hz), 5.07 (s, 1H), 4.61 (qAB, 2H, J = 12.1, 19.2 Hz), 4.51 (d, 1H, J = 7.2 Hz), 3.94 (d, 1H, J = 9.9 Hz), 3.42 (d, 1H, J = 9.9 Hz), 1.64–1.56 (m, 1H), 1.55 (s, 3H), 1.30–1.26 (m, 1H), 1.24 (s, 3H), 0.98–0.933 (m, 1H).

(1′S,2′R,3′S,4′R,5′S)-4-(6-Methylamino-2-chloro-9H-pu-rin-9-yl)-1-[(phenylmethoxy)methyl]bicyclo[3.1.0]hexane-2,3-(O-isopropylidene) (26)

¹H NMR (CDCl₃) δ 8.15 (s, 1H), 7.40–7.28 (m, 5H), 6.29 (bs, 1H), 5.32 (d, 1H, J = 7.1 Hz), 5.05 (s, 1H), 4.61 (qAB, 2H, J = 12.4, 17.6 Hz), 3.90 (d, 1H, J = 10.2 Hz), 3.42 (d, 1H, J = 10.2 Hz), 3.17 (bs, 3H), 1.60–1.56 (m, 1H), 1.54 (s, 3H), 1.30–1.26 (m, 1H), 1.24 (s, 3H), 0.96–0.92 (m, 1H).

Procedure for Debenzylation of 24–26

See ref 39.

(1′S,2′R,3′S,4′R,5′S)-4-(6-Amino-2-chloro-9H-purin-9-yl)-1-[hydroxymethyl]bicyclo[3.1.0]hexane-2,3-(O-isopropylidene) (27)

¹H NMR (CDCl₃) δ 7.90 (s, 1H), 5.56 (d, 1H, J = 7.4 Hz), 4.79 (s, 1H), 4.66 (d, 1H, J = 7.4 Hz), 4.26 (d, 1H, J = 11.5 Hz), 3.37 (d, 1H, J = 11.5 Hz), 1.78–1.70 (m, 1H), 1.55 (s, 3H), 1.26 (s, 3H), 1.20–1.14 (m, 1H), 1.20–0.96 (m, 1H).

(1′S,2′R,3′S,4′R,5′S)-4-(6-Methylamino-9H-purin-9-yl)-1-[hydroxymethyl]bicyclo[3.1.0]hexane-2,3-(O-isopropylidene) (29)

¹H NMR (CDCl₃) δ 8.36 (s, 1H), 7.77 (s, 1H), 5.61 (d, 1H, J = 7.4 Hz), 4.78 (s, 1H), 4.65 (d, 1H, J = 7.4 Hz), 4.33 (d, 1H, J = 11.5 Hz), 3.25 (d, 1H, J = 11.5 Hz), 3.20 (bs, 3H), 1.79–1.74 (m, 1H), 1.55 (s, 3H), 1.25 (s, 3H), 1.17–1.13 (m, 1H), 1.01–0.96 (m, 1H).

(1′S,2′R,3′S,4′R,5′S)-4-(6-Methylamino-2-chloro-9H-pu-rin-9-yl)-1-[hydroxymethyl]bicyclo[3.1.0]hexane-2,3-(O-isopropylidene) (30)

¹H NMR (CD₃OD) δ 8.14 (s, 1H), 5.30 (d, 1H, J = 7.1 Hz), 4.87 (s, 1H), 4.61 (d, 1H, J = 7.1 Hz), 3.91 (d, 1H, J = 11.5 Hz), 3.52 (d, 1H, J = 11.5 Hz), 3.00 (bs, 3H), 1.64–1.60 (m, 1H), 1.44 (s, 3H), 1.18 (s, 3H), 1.08–1.04 (m, 1H), 0.92–0.87 (m, 1H).

(1′S,2′R,3′S,4′R,5′S)-4-(6-Amino-2-methylthio-9H-purin-9-yl)-1-[hydroxymethyl]bicyclo[3.1.0]hexane-2,3-(O-isopropylidene) (28)

Compound 27 (30 mg, 0.086 mmol) was dissolved in DMF (1.5 mL) and treated with sodium thiomethoxide (90 mg, 1.3 mmol). The reaction mixture was heated in a sealed tube at 90 °C for 1.5 h. The solvent was removed under vacuum, and the product was purified by preparative thin-layer chromatography (9/1 CHCl₃/MeOH) to furnish 25 mg of 28 (80%). ¹H NMR (CDCl₃) δ 7.76 (s, 1H), 5.82 (bs, 2H), 5.56 (d, 1H, J = 7.1 Hz), 4.78 (s, 1H), 4.71 (d, 1H, J = 7.1 Hz), 4.25 (d, 1H, J = 11.5 Hz), 3.35 (d, 1H, J = 11.5 Hz), 2.59 (s, 3H), 1.71–1.66 (m, 1H), 1.55 (s, 3H), 1.26 (s, 3H), 1.18–1.14 (m, 1H), 0.99–0.93 (m, 1H).

Protected (N)-Methanocarbaadenosine 5′-Monophosphate Derivatives (31–35)

These monophosphate derivatives were synthesized by procedures similar to those used to prepare the unsubstituted adenine analogue.³⁹

(1′S,2′R,3′S,4′R,5′S)-4-(6-Amino-2-chloro-9H-purin-9-yl)-1-[(di-tert-butyl phosphate)methyl]bicyclo[3.1.0]hex-ane-2,3-(O-isopropylidene) (31)

¹H NMR (CDCl₃) δ 8.23 (s, 1H), 6.91 (bs, 2H), 5.41 (d, 1H, J = 7.1 Hz), 5.16 (s, 1H), 4.64 (d, 1H, J = 7.1 Hz), 4.59 (dd, 1H, J = 5.5, 11.0 Hz), 3.98 (dd, 1H, J = 6.3, 11.0 Hz), 1.84–1.76 (m, 1H), 1.63 (s, 3H), 1.61 (s, 9H), 1.59 (s, 9H), 1.42–1.37 (m, 1H), 1.33 (s, 3H), 1.18–1.13 (m, 1H).

(1′S,2′R,3′S,4′R,5′S)-4-(6-Amino-2-methylthio-9H-purin-9-yl)-1-[(di-tert-butyl phosphate)methyl]bicyclo[3.1.0]-hexane-2,3-(O-isopropylidene)(32)and(1′S,2′R,3′S,4′R,5′S)-4-(6-Amino-2-methylsulfoxy-9H-purin-9-yl)-1-[(di-tert-butyl phosphate)methyl]bicyclo[3.1.0]hexane-2,3-(O-isopropylidene) (33)

To a mixture of 28 (20 mg, 0.055 mmol) and tetrazole (0.012 g, 0.165 mmol) in anhydrous THF (2 mL) was added di-tert-butyl-N,N-diethylphosphoramidite (0.023 mL, 0.083 mmol), and the mixture was stirred at room temperature overnight. The reaction mixture was cooled to −78 °C, and a solution of 70% mCPBA (0.012 g, 0.066 mmol) in 1 mL of CH₂Cl₂ was warmed to room temperature. A total of 1 mL MeOH was added, and the mixture was concentrated and purified by preparative TLC using 10% MeOH in CHCl₃ to furnish 32 (12 mg) and 33 (12 mg).

32: ¹H NMR (CDCl₃) δ 7.98 (s, 1H), 5.87 (bs, 2H), 5.31 (d, 1H, J = 7.2 Hz), 5.07 (s, 1H), 4.59 (d, 1H, J = 7.2 Hz), 4.49 (dd, 1H, J = 5.8, 11.0 Hz), 3.83 (dd, 1H, J = 6.3, 11.0 Hz), 2.59 (s, 3H), 1.75–1.70 (m, 1H), 1.55 (s, 3H), 1.49 (s, 18H), 1.30–1.27 (m, 1H), 1.23 (s, 3H), 1.05–1.00 (m, 1H).

33: ¹H NMR (CDCl₃) δ 8.26 (s, 1H), 6.66 (bs, 2H), 5.33 (d, J = 7.14 Hz, 1H), 5.22 (s, 1H), 4.60–4.50 (m, 2H), 3.81–3.71 (m, 2H), 2.93 (s, 3H), 1.76–1.70 (m, 1H), 1.54 (s, 3H), 1.50 (s, 9H), 1.34–1.30 (m, 1H), 1.25 (s, 3H), 1.23 (s, 9H), 1.07–1.02 (m, 1H).

(1′S,2′R,3′S,4′R,5′S)-4-(6-Methylamino-9H-purin-9-yl)-1-[(di-tert-butyl phosphate)methyl]bicyclo[3.1.0]hexane-2,3-(O-isopropylidene) (34)

¹H NMR (CDCl₃) δ 8.45 (s, 1H), 8.05 (s, 1H), 6.04 (bs, 1H), 5.30 (d, 1H, J = 7.2 Hz), 5.08 (s, 1H), 4.57 (d, 1H, J = 7.2 Hz), 4.47 (dd, 1H, J = 5.8, 11.0 Hz), 3.90 (dd, 1H, J = 6.3, 11.0 Hz), 3.20 (bs, 3H), 1.79–1.71 (m, 1H), 1.54 (s, 3H), 1.51 (s, 9H), 1.50 (s, 9H), 1.29–1.26 (m, 1H), 1.22 (s, 3H), 1.08–1.03 (m, 1H).

(1′S,2′R,3′S,4′R,5′S)-4-(6-Methylamino-2-chloro-9H-pu-rin-9-yl)-1-[(di-tert-butyl phosphate)methyl]bicyclo[3.1.0]-hexane-2,3-(O-isopropylidene) (35)

¹H NMR (CDCl₃) δ 8.04 (s, 1H), 5.38 (d, 1H, J = 7.1 Hz, 1H), 5.04 (s, 1H), 4.58 (d, 1H, J = 7.1 Hz), 4.42–4.52 (m, 1H), 3.90–4.08 (m, 1H), 3.20 (bs 3H), 1.79–1.71 (m, 1H), 1.54 (s, 3H), 1.51 (s, 9H), 1.50 (s, 9H), 1.29–1.26 (m, 1H), 1.22 (s, 3H), 1.08–1.03 (m, 1H).

General Procedure for Deprotection of (N)-Methanocarbaadenosine 5′-Monophosphate Derivatives (Synthesis of 13a, 10a, 11a, 36, and 37)

Deprotection was carried out by procedures similar to those used to prepare the unsubstituted adenine analogue.³⁹

(1′S,2′R,3′S,4′R,5′S)-4-(6-Amino-2-methylthio-9H-purin-9-yl)-1-[phosphoryloxymethyl]bicyclo[3.1.0]hexane-2,3-diol (10a)

¹H NMR (D₂O) δ 8.53 (s, 1H), 4.91 (s, 1H), 4.48 (dd, 1H, J = 4.7, 11.0 Hz), 4.05 (d, 1H, J = 6.6 Hz), 3.69 (dd, 1H, J = 4.7, 11.0 Hz), 2.63 (s, 3H), 1.92–1.84 (m, 1H), 1.58–1.40 (m, 1H), 1.18–0.98 (m, 1H). ³¹P NMR (D₂O) δ 0.613.

(1′S,2′R,3′S,4′R,5′S)-4-(6-Amino-2-methylthiooxy-9H-purin-9-yl)-1-[phosphoryloxymethyl]bicyclo[3.1.0]hexane-2,3-diol (11a)

¹H NMR (D₂O) δ 8.83 (s, 1H), 5.46 (d, J = 6.04 Hz, 1H), 4.52 (dd, J = 4.94, 10.71 Hz, 1H), 4.06 (d, J = 6.04 Hz, 1H), 3.72 (dd, J = 4.94, 10.71 Hz, 1H), 2.99 (s, 3H), 2.04–1.92 (m, 1H), 1.26–1.22 (m, 1H), 1.04–0.98 (m, 1H). ³¹P NMR (D₂O) δ 0.623.

(1′S,2′R,3′S,4′R,5′S)-4-(6-Amino-2-chloro-9H-purin-9-yl)-1-[phosphoryloxymethyl]bicyclo[3.1.0]hexane-2,3-di-ol (13a)

¹H NMR (D₂O) δ 8.60 (s, 1H), 4.72 (s, 1H), 4.48–4.40 (m, 1H), 3.99 (d, 1H, J = 6.6 Hz), 3.62–3.76 (m, 1H), 1.94–1.82 (m, 1H), 1.60–1.52 (m, 1H), 1.08–0.95 (m, 1H).

The imidazolate 42, derived from 13a, displayed a single ³¹P NMR resonance in D₂O at δ −11.824.

(1′S,2′R,3′S,4′R,5′S)-4-(6-Methylamino-9H-purin-9-yl)-1-[phosphoryloxymethyl]bicyclo[3.1.0]hexane-2,3-diol (36)

¹H NMR (D₂O) δ 8.47 (s, 1H), 8.26 (s, 1H), 4.87 (s, 1H), 4.76 (d, 1H, J = 6.6 Hz), 4.46 (dd, 1H, J = 5.0, 11.0 Hz), 3.99 (d, 1H, J = 6.6 Hz), 3.66 (dd, 1H, J = 5.0, 11.0 Hz), 3.12 (bs, 3H), 1.88–1.82 (m, 1H), 1.54–1.48 (m, 1H), 1.00–0.92 (m, 1H).

(1′S,2′R,3′S,4′R,5′S)-4-(6-Methylamino-2-chloro-9H-pu-rin-9-yl)-1-[phosphoryloxymethyl]bicyclo[3.1.0]hexane-2,3-diol (37)

¹H NMR (D₂O) δ 8.42 (s, 1H), 4.71 (s, 1H), 4.52 (m, 1H), 3.97 (d, 1H, J = 6.3 Hz), 3.72–3.62 (m, 1H), 3.01 (bs, 3H), 1.72–1.80 (m, 1H), 1.62–1.48 (m, 1H), 1.04–0.94 (m, 1H).

General Method of Synthesis of (N)-Methanocarba-ATP Analogues (8a, 12a, 14a, and 15a)

These triphosphate derivatives were synthesized by procedures similar to the procedure used to prepare the unsubstituted adenine analogue³⁹ and the procedure for 5′-diphosphate analogue 9a.

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

A mixture of 10a (14 mg, 0.035 mmol) and 1,1′-carbonyldiimidazole (28 mg, 0.175 mmol) in 1 mL of anhydrous DMF was stirred at room temperature for 6 h. Et₃N/H₂O/ MeOH (1:10:10 by volume, 2 mL) was added, and the mixture was stirred at room temperature for an additional 2 h. The mixture was concentrated to dryness under vacuum. To the residue, phosphoric acid (20 mg, 0.21 mmol) was added and dried under high vacuum. This was suspended in anhydrous DMF (2 mL), and Et₃N (0.1 mL) was added. The mixture was stirred at room temperature for 3 days. Triethylammonium bicarbonate (2 mL) was added, all of the solvent was removed under vacuum, and the residue was purified using Sephadex column chromatography, eluting with 150 mL of water and 150 mL of ammonium bicarbonate to furnish 9a (5.0 mg). ¹H NMR (D₂O) δ 8.34 (s, 1H), 4.87 (s, 1H), 4.71 (d, J = 6.2 Hz, 1H), 4.54–4.48 (m, 1H), 4.02 (d, J = 6.2 Hz, 1H), 3.85–3.78 (m, 1H), 2.58 (s, 3H), 1.91–1.86 (m, 1H), 1.54–1.46 (m, 1H), 1.02–0.96 (m, 1H). ³¹P NMR (D₂O) δ −9.63, −10.70.

(1′S,2′R,3′S,4′R,5′S)-4-(6-Amino-2-methylthio-9H-purin-9-yl)-1-[triphosphoryloxymethyl]bicyclo[3.1.0]hexane-2,3-diol (8a)

¹H NMR (D₂O) δ 8.15 (s, 1H), 4.10–3.98 (m, 1H), 3.90–3.74 (m, 1H), 3.67–3.61 (m, 1H), 3.57–3.52 (m, 1H), 2.59 (s, 3H), 2.02–1.84 (m, 1H), 1.62–1.46 (m, 1H), 1.06–0.94 (m, 1H). ³¹P NMR (D₂O) δ 0.02, −9.6, −10.8.

(1′S,2′R,3′S,4′R,5′S)-4-(6-Amino-2-chloro-9H-purin-9-yl)-1-[triphosphoryloxymethyl]bicyclo[3.1.0]hexane-2,3-diol (12a)

¹H NMR (D₂O) δ 8.22 (s, 1H), 4.86 (s, 1H), 4.72 (d, 1H, J = 5.2 Hz), 3.84–3.74 (m, 2H), 2.14–2.02 (m, 1H), 1.62–1.54 (m, 1H), 1.06–0.96 (m, 1H). ³¹P NMR (D₂O) δ −10.55, −10.93, −22.79.

(1′S,2′R,3′S,4′R,5′S)-4-(6-Methylamino-9H-purin-9-yl)-1-[triphosphoryloxymethyl]bicyclo[3.1.0]hexane-2,3-di-ol (14a)

¹H NMR (D₂O) δ 8.44 (s, 1H), 8.24 (s, 1H), 4.85 (d, 1H, J = 5.2 Hz), 4.75 (s, 1H), 4.57 (dd, 1H, J = 5.2, 11.3 Hz), 4.01 (d, 1H, J = 5.2 Hz), 3.83 (dd, 1H, J = 4.7, 11.3 Hz), 3.09 (bs, 3H), 1.96–1.86 (m, 1H), 1.56–1.48 (m, 1H), 1.04–0.96 (m, 1H). ³¹P NMR (D₂O) δ −10.03, −10.67, −22.54.

(1′S,2′R,3′S,4′R,5′S)-4-(6-Methylamino-2-chloro-9H-pu-rin-9-yl)-1-[triphosphoryloxymethyl]bicyclo[3.1.0]hexane-2,3-diol (15a)

¹H NMR (D₂O) δ 8.42 (s, 1H), 4.86 (s, 1H), 4.74 (s, 1H), 4.64–4.55 (m, 1H), 4.00 (s, 1H), 3.95–3.72 (m, 1H), 3.08 (bs, 3H), 1.96–1.82 (m, 1H), 1.58–1.44 (m, 1H), 1.08–0.96 (m, 1H). ³¹P NMR (D₂O) δ −8.05, −10.73, −22.19.

(1′S,2′R,3′S,4′R,5′S)-4-(6-Amino-9H-purin-9-yl)-1-[b,g-methylenetriphosphoryloxymethyl]bicyclo[3.1.0]hexane-2,3-diol (17a)

A mixture of compound 7a (11 mg, 0.03 mmol)³⁹ and 1,1′-carbonyldiimidazole (24 mg, 0.15 mmol) in DMF (2 mL) was stirred at room temperature for 6 h. Et₃N/H₂O/MeOH (1:10:10 by volume, 1 mL) was added, and the mixture was stirred for 2 h. All of the solvent was removed under vacuum to dryness. The residue and methylene diphosphonic acid (0.32 g, 0.18 mmol) were dried under high vacuum for 1 h and suspended in anhydrous DMF (2 mL), 0.1 mL Et₃N was added, and the mixture was stirred at room temperature for 3 days. Triethylammonium bicarbonate (2 mL) was added, and all of the solvent was removed under vacuum and purified using Sephadex column chromatography, eluting with 100 mL of water and 100 mL of ammonium bicarbonate to furnish 17a (4.0 mg). ¹H NMR (D₂O) δ 8.70 (s, 1H), 8.41 (s, 1H), 4.98 (s, 1H), 4.59 (dd, J = 4.95, 10.72 Hz, 1H), 4.00 (d, J = 6.04 Hz, 1H), 3.82 (dd, J = 4.95, 10.72 Hz, 1H), 2.36 (bs, 2H), 1.98–1.86 (m, 1H), 1.58–1.52 (m, 1H), 1.06–0.98 (m, 1H). ³¹P NMR (D₂O) δ 15.454, 9.024, −10.341.

Phosphoric Acid Mono[(2R,3R,4S,5R)-5-(6-amino-2-chloro-9H-purin-9-yl]-3,4-dihydroxytetrahydrofuran-2-ylmethyl]ester (13b)

This monophosphate was synthesized by procedures similar to those used to prepare the unsubstituted adenine analogue.^{39 1}H NMR (D₂O) δ 8.23 (s, 1H), 5.93 (d, J = 6.04 Hz, 1H), 4.39 (s, 1H), 4.27 (s, 1H), 3.92–3.82 (m, 2H).

(2R,3R,4S,5R)-2-(6-Amino-2-methylsulfanyl-9H-purin-9-yl]tetrahydrofuran-3,4-(O-isopropylidene) (47)

Compound 46 (30 mg, 0.088 mmol) was dissolved in DMF (1.5 mL) and treated with sodium thiomethoxide (30 mg, 0.44 mmol). The reaction mixture was heated in a sealed tube at 90 °C for 1.5 h. The solvent was removed under vacuum, and the product was purified using preparative thin-layer chromatography (9/1 CHCl₃/MeOH) to furnish 20 mg of 47 (65%). 1H NMR (CDCl₃) δ 7.7 (s, 1H), 5.81 (d, J = 4.39 Hz, 1H), 5.62 (s, 2H), 5.34–5.28 (m, 1H), 5.12 (d, J = 4.39 Hz, 1H), 4.47 (s, 1H), 3.98–3.76 (m, 2H), 2.56 (s, 3H), 1.63 (s, 3H), 1.38 (s, 3H).

Phosphoric Acid Mono[(2R,3R,4S,5R)-5-(6-amino-2-methylsulfanyl-9H-purin-9-yl]-3,4-dihydroxytetrahydrofuran-2-ylmethyl]ester (10b) and Phosphoric Acid Mono-[(2R,3R,4S,5R)-5-(6-amino-2-methylsulfinyl-9H-purin-9-yl]-3,4-dihydroxytetrahydrofuran-2-ylmethyl]ester (11b)

Di-tert-butyl-N,N-diethylphosphoramidite (0.02 mL, 0.075 mmol) was added to a mixture of 47 (18 mg, 0.05 mmol) and tetrazole (17 mg, 0.05 mmol) in anhydrous THF (2 mL), and the mixture was stirred at room temperature overnight. The reaction mixture was cooled to −78 °C, treated with a solution of 70% mCPBA (15 mg, 0.06 mmol) in 1 mL of CH₂Cl₂, and warmed to room temperature. A total of 1 mL of MeOH was added, and the mixture was concentrated and purified by preparative TLC using 10% MeOH in CHCl₃ to furnish 48 (10 mg) and 49 (10 mg). These were treated with 30 mg of Dowex 50×8–200 in 1/1 H₂O/MeOH at 80 °C for 1 h. The Dowex resin was filtered, and the filtrate was concentrated to furnish 10b (3 mg, 15%) and 11b (3 mg, 15%).

10b: ¹H NMR (D₂O) δ 8.39 (s, 1H), 6.11 (d, J = 5.5 Hz, 1H), 4.82–4.95 (m, 1H), 4.45 (s, 1H), 4.26 (s, 1H), 3.92–3.82 (m, 2H), 2.59 (s, 3H).

11b: ¹H NMR (D₂O) δ 8.49 (s, 1H), 6.11 (d, J = 5.5 Hz, 1H), 4.81–4.94 (m, 1H), 4.45 (s, 1H), 4.26 (s, 1H), 3.92–3.82 (m, 2H), 2.82 (s, 3H).

2. Enzymatic Analyses

2.1. Hydrolysis of 5′-Triphosphate Derivatives by Recombinant Rat Ecto-5′-nucleotidase

Catalytically active recombinant ecto-5′-nucleotidase was expressed in insect cells using the baculovirus system and purified as described in ref 48. Ecto-5′-nucleotidase (0.5 μg) was incubated in Tris-buffered saline (TPS: 10 mM Tris, 150 mM NaCl, 0.4 mM MgCl₂, pH 7.4) with the corresponding substrate (AMP, 7a, or 7c) at a final concentration of 100 μM. After 10 min of incubation at 37 °C, the reaction was terminated by addition of 5% TCA. The sample was centrifuged for 10 min at 12000g (4 °C). Inorganic phosphate was determined in the supernatant fraction according to the method of Lanzetta et al.⁵¹

2.2. Hydrolysis of 5′-Triphosphate Derivatives by Rat NTPDase1 (Ectoapyrase) and NTPDase2 (Ecto-ATPase) Stably Expressed in CHO Cells. Cell Transfection

CHO cells stably transfected with a plasmid DNA containing NTPDase1 (ectoapyrase) or NTPDase2 (ecto-ATPase) have previously been described.⁴⁹ They were cultured in in HAM’s FC-12 medium containing 10% fetal calf serum, 100 U/mL penicillin and 100 μg/mL streptomycin (Gibco BRL, Eggenstein, Germany) as described in ref 52.

2.3. Measurement of Ectonucleotidase Activity

Transfected CHO cells were seeded in multiwell plates (20 000 cells per well, 1.88 cm²). Surface-located enzyme activity of intact cells was determined 24 h later at 37 °C. Cells were washed twice with physiological saline solution (140 mM NaCl, 5 mM KCl, 10 mM glucose, 1 mM MgCl₂, 2 mM CaCl₂, 20 mM Hepes, pH 7.4) and subsequently incubated in 500 μL of the identical saline solution containing 50 mM of nucleotide substrate (ATP, 6a, or 6c). Aliquots of the culture supernatant were collected at varying time points and subjected to high-performance liquid chromatography (HPLC) analysis as previously described.⁴⁹ The absorbance at 260 nm was continuously monitored, and the nucleotide and nucleoside concentrations were determined from the area under each absorbance peak.

4. Pharmacological Analyses

P2Y-receptor-promoted stimulation of inositol phosphate formation was measured at human P2Y₁, P2Y₂, P2Y₄, and P2Y₁₁ receptors stably expressed in 1321N1 human astrocytoma cells as previously described.^46,50 The K_0.5 values were averaged from 3 to 8 independently determined concentration–effect curves for each compound. Turkey erythracytes were incubated in inositol-free medium (DMEM; Gibco, Gaithersburg, MD) with 0.5 mCi of 2-[³H]myo-inositol (20 Ci/mmol; American Radiolabeled Chemicals, Inc., St. Louis, MO) for 18–24 h in a humidified atmosphere of 95%/5% air/CO₂ at 37 °C. tP2Y₁ receptor-promoted phospholipase C activity was measured in 25 μL of [³H]inositol-labeled ghosts (approximately 175 μg of protein, 200–500000 cpm/assay) in a medium containing 424 μM CaCl₂, 0.91 mM MgSO₄, 2 mM EGTA, 115 mM KCl, 5 mM KH₂PO₄, and 10 mM Hepes, pH 7.0. Assays (200 μL final volume) contained 1 μM GTPγS and the indicated concentrations of nucleotide analogues. Membranes were incubated at 30 °C for 5 min, and total [³H]inositol phosphates were quantified by anion exchange chromatography as previously described.^47,53

5. Data Analysis

Agonist potencies were calculated using a four-parameter logistic equation and the GraphPad software package (GraphPad, San Diego, CA). EC₅₀ values (mean ± standard error) represent the concentration at which 50% of the maximal effect is achieved. Relative efficacies (%) were determined by comparison with the effect produced by a maximal effective concentration of UTP or UDP in the same experiment.

All concentration–effect curves were repeated in at least three separate experiments, carried out in duplicate or triplicate using different membrane preparations.