Sulfur-Containing 1,3-Dialkylxanthine Derivatives as Selective Antagonists at A₁-Adenosine Receptors

Abstract

Sulfur-containing analogues of 8-substituted xanthines were prepared in an effort to increase selectivity or potency as antagonists at adenosine receptors. Either cyclopentyl or various aryl substituents were utilized at the 8-position, because of the association of these groups with high potency at A₁-adenosine receptors. Sulfur was incorporated on the purine ring at positions 2 and/or 6, in the 8-position substituent in the form of 2- or 3-thienyl groups, or via thienyl groups separated from an 8-aryl substituent through an amide-containing chain. The feasibility of using the thienyl group as a prosthetic group for selective iodination via its Hg²⁺ derivative was explored. Receptor selectivity was determined in binding assays using membrane homogenates from rat cortex [[³H]-N⁶-(phenylisopropyl) adenosine as radioligand] or striatum [[³H]-5′-(N-ethylcarbamoyl)adenosine as radioligand] for A₁- and A₂-adenosine receptors, respectively. Generally, 2-thio-8-cycloalkylxanthines were at least as A₁ selective as the corresponding oxygen analogue. 2-Thio-8-aryl derivatives tended to be more potent at A₂ receptors than the oxygen analogue. 8-[4-[(Carboxymethyl)oxy]phenyl]-1,3-dipropyl-2-thioxanthine ethyl ester was >740-fold A₁ selective.

1,3-Dialkyl and other xanthine derivatives inhibit many of the pharmacological and physiological effects of adenosine, e.g., the cardic-depressive,¹ hypotensive,¹ antidiuretic,² and antilipolytic effects,³ by acting as competitive antagonists at A₁- and A₂-adenosine receptor subtypes. The naturally occurring caffeine and theophylline (1; Figure 1) are the most widely used xanthine drugs. However, they are nonselective and relatively weak adenosine antagonists (K_i values of 10 μM or greater). Synthetic analogues of theophylline, containing 1,3-dipropyl, 8-aryl, and 8-cycloalkyl substitutions, are more potent as adenosine antagonists.^4–6 The combination of 1-, 3-, and 8-position substitutions has resulted in analogues such as 8-cyclopentyl-1,3-dipropylxanthine^5,6 (CPX; 2) and 8-[4-[[[[(2-aminoethyl)amino]carbonyl]methyl]oxy]phenyl]-1,3-dipropylxanthine⁷ (XAC; 3) which are more than 4 orders of magnitude more potent than theophylline in binding at A₁-adenosine receptors, and which are A₁-selective by factors of 300 and 60, respectively.

Figure 1.

Xanthines having thio substitutions at the 2- and/or 6-position have been reported to act as antagonists at A₂ receptors in human fibroblasts²¹ and as phosphodiesterase inhibitors with potency comparable to or greater than that of theophylline.^8,9a Remarkably, 6-thiocaffeine and 6-thiotheophylline cause cardiac depression rather than stimulation.^9b Recently, 6-thiocaffeine and 6-thiotheophylline were reported to induce tracheal relaxation, without cardiac or behavioral stimulation.^9a Thio substitution of the NH at position 7 of 8-phenyltheophylline reduced activity by 1000-fold at an A₁ receptor and by nearly 100-fold at an A₂ receptor.²⁰

Results

Chemistry

Various 8-substituted xanthine and thioxanthine derivatives were synthesized via 1,3-dialkyl-5,6-diaminouracils as shown in Scheme I. The substituted uracil and 2-thiouracil intermediates were prepared via an optimized Traube synthesis.^6,7a 1,3-Dimethyl- and 1,3-di-n-propyl-5,6-diaminouracil and their 2-thio derivatives were obtained by condensation of the corresponding dialkyl urea (4a and -b) or thiourea (4c and -d) with ethyl cyanoacetate (5). The products after ring closure, substituted 6-aminouracil derivatives 6, were then nitrosated at the 5-position. The nitroso group was reduced through chemical reduction or catalytic hydrogenation to form the intermediate 1,3-dialkyl-5,6-diaminouracil derivatives 8 in good yield.

Scheme I.

The next step of the synthesis was to form the imidazole ring of the purine nucleus, resulting in the xanthine derivatives, as listed in Table I (compounds 1, 2, and 13–28). The more nucleophilic 5-amino group of compound 8 was acylated by using a carboxylic acid chloride 9, forming the 1,3-dialkyl-5-(acylamino)-6-aminouracil derivatives 10. 1,3-Dialkyl-5-(acylamino)-6-aminouracils derived from thiophene-2-carboxylic acid and -3-carboxylic acid chlorides and from cyclopentanecarboxylic acid chloride were isolated and characterized (Table II). The various 1,3-dialkyl-5-(acylamino)-6-aminouracil derivatives were then cyclized to the corresponding xanthine and 2-thioxanthine derivatives (Table III) by treatment with aqueous sodium hydroxide.

Table I.

Potencies of Xanthine Derivatives at Adenosine A₁ and A₂ Receptors in Nanomolar Concentration Units^a,^b

compd	R	R1	X			Ki(A2)/Ki(A1)
				Ki (A1 receptors)	Ki (A2 receptors)
1a	Me	H	O	8470 ± 1490c	25300 ± 2000a	2.99
1b	Pr	H	O	450 ± 25c	5160 ± 590c	11.5
2a	Me	cyclopentyl	O	10.9 ± 0.9c	1440 ± 70c	133
2b	Pr	cyclopentyl	O	0.9 ± 0.1	410 ± 40	455
13	Me	cyclopentyl	S	10.2 ± 1.5	1390 ± 88	136
14	Pr	cyclopentyl	S	0.655 ± 0.058	314 ± 62	479
15a	Me	2-furyl	O	350 ± 20c	2780 ± 50c	7.94
15b	Pr	2-furyl	O	37 ±6	640 ± 100	16.8
15c	Me	2-furyl	S	182 ± 36	4450 ± 420	10.6
15d	Pr	2-furyl	S	32 ± 5	594 ± 71	18.6
16	Me	3-furyl	O	72.4 ± 3.7c	984 ± 70c	13.6
17	Me	2-thienyl	O	233 ± 48.6	1630 ± 179	6.97
18	Me	3-thienyl	O	152 ± 27	841 ± 109	5.53
19	Pr	2-thienyl	O	16.1 ± 1.96	381 ± 27.7	23.6
20	Pr	3-thienyl	O	10.0 ± 0.03	121 ± 18.2	12.1
21	Me	2-thienyl	S	221 ± 43.3	1740 ± 153	7.87
22	Pr	2-thienyl	S	35.1 ± 6.0	>5000	>142
23	Me	phenyl	O	86.0 ± 2.8c	848 ± 115c	9.85
24	Et	phenyl	O	44.5 ± 1.2c	836 ± 73c	19.4
25	Pr	phenyl	O	10.2 ± 2.6c	180 ± 29c	17.8
26a	Me	phenyl	S	38 ± 6	>7000	>184
26b	Pr	phenyl	S	16.1 ± 2	422 ± 33	26.1

compd	R2				Ki (A2)/Ki (A1)
		X	Ki (A1 receptors)	Ki (A2 receptors)
27	COOH	O	58 ± 3	2200 ± 526	37.8
28	COOH	S	53.8 ± 7.1	315 ± 60.8	5.86
29	COOEt	O	42 ± 3	>5000	>119
30	COOEt	S	6.78 ± 0.64	>5000	>740
3	CONH(CH2)2NH2	O	1.2 ± 0.5	63 ± 21	52.5
31	CONH(CH2)2NH2	S	2.69 ± 0.77	26.3 ± 1.76	9.8
32	CONH(CH2)2NHCH3	O	15.1 ± 1.6d	[9.3 ± 2.1]f	[0.62]
33	CONH(CH2)2NHCH3	S	2.4 ± 0.28	6.80 ± 1.36	2.8
34	CONH(CH2)2N(CH3)2	O	2.8 ± 0.19d	5.03 ± 0.54	1.8
35	CONH(CH2)2N(CH3)2	S	2.55 ± 0.60	27.9 ± 7.5	11
36	CON(CH3)(CH2)2N(CH3)2	O	0.93 ± 0.03d	6.26 ± 0.25	6.7
37	CON(CH3)(CH2)2N(CH3)2	S	2.57 ± 0.67	24.5 ± 8.4	9.5
38		O	12	e	e
39		S	84	870	10
40		O	6.4 ± 2.7	191 ± 13	30
41		S	8.9	322 ± 17	36
42		O	0.87 ± 0.09	180	210
43		S	13.0 ± 3.5	46.8 ± 9.4	3.6
44		O	3.69 ± 0.71d	207 ± 57d	56
45		S	33.5	e	e
46		O	18.3 ± 3.0	147 ± 5	8.1
47		O	16.2 ± 2.7	458 ± 34	28.3
48		O	11.3 ± 1.5	116 ± 25	10.3
49		O	7.44 ± 0.98	630 ± 160	85
50		O	17 ± 1.6	e	e
51		O	1.3 ± 0.12	e	e

compd	R	R1			Ki (A2 receptors)	Ki(A2)/Ki(A1)
			X	Ki (A1 receptors)
52	Me	cyclopentyl	S	40.5 ± 6.6	11500 ± 628	285
53	Pr	cyclopentyl	S	4.87 ± 0.82	2780 ± 730	572
54	Me	cyclopentyl	O	202 ± 26	8980 ± 1300	44.4
55	Pr	cyclopentyl	O	15.5 ± 1.5	3360 ± 270	217
56	Me	phenyl	O	1380 ± 74	11300 ± 777	8.18
57	Et	phenyl	O	1010 ± 321	3510 ± 290	3.47

^aK_i value from a single determination run in triplicate or average of three ± SEM.

^bInhibition of binding of [³H](phenylisopropyl)-adenosine to A₁ receptors in rat cortical membranes and binding of [³H]-5′-(N-ethylcarbamoyl)adenosine to A₂-adenosine receptors in rat striatal membranes was measured as described.^18,19

^cValues taken from Bruns et al.¹⁸

^dValues taken from Jacobson et al.¹⁰

^eNot determined.

^fK_b for inhibition of 5′-(N-ethylcarbamoyl)adenosine-stimulated adenylate cyclase, in pheochromocytoma PC12 cell membranes.^10b

Table II.

Synthesis and Characterization of 1,3-Dialkyl-5-(acylamino)-6-aminouracils

compd	R	R1	X	% yield	mp, °C	formula	anal.
10a	Me	cyclopentyl	S	71	253	C12H18N4O2S	C, H, N
10b	Pr	cyclopentyl	S	92	103	C16H26N4O2S·H2O	C, H, N
10c	Me	2-thienyl	O	76	>300	C11H12N4O3S	C, H, N
10d	Me	3-thienyl	O	78	>300	C11H12N4O3S	C, H, N
10e	Pr	2-thienyl	O	88	143	C15H20N4O3S	C, H, N
10f	Pr	3-thienyl	O	88	144	C15H20N4O3S	H, N; Ca
10g	Me	2-thienyl	S	83	>300	C11H12N4O2S2	C, H, N
10h	Pr	2-thienyl	S	80	150	C15H20N4O2S2	C, H, N

^aC: calcd, 53.56; found, 52.95.

Table III.

Synthesis and Characterization of Xanthine Derivatives

compd	% yield	mp, °C	formula	anal.
13	91	278	C12H16N4OS	C, H, N
14	89	217	C16H24N4OS	C, H, N
15a	80	347	C11HI0N4O3	C, H, N
15b	88	252	C15H18N4O3	C, H, N
15c	85	>350	C11H10N4O2S	C, H, N
15d	88	269	C15H18N4O2S	C, H, N
17	82	>300	C11H10N4O2S	C, H, N
18	92	>300	C11H10N4O2S	C, H, N
19	85	259	C15H18N4O2S	C, H, N
20	87	267	C15H18N4O2S	C, H, N
21	71	>340	C11H10N4O2S2	C, H, N
22	92	298	C15H18N4OS2	C, H, N
26a	91	>350	C13H12N4OS	C, H, N
26b	80	253	C17H20N4OS	C, H, N
33	95	206–208	C22H30N6O3S·¼H2O	C, H, N
35	93	238–240	C23H32N6O3S·¾H2O	C, H, N
37	52	172–174	C24H34N6O3S·½H2O	C, H, N
39	92	210–212	C40H54N8O8S	C, H, N
41	84	182–187	C37H47F3N8O8S·½CF3COOH·½H2O	C, H, N
43	97	238–242 dec	C27H40N8O4S·3HBr·3/2H2O	C, H, N
45	68	160–168	C33H43N7O8S	C, H, Nb
48	85	240 dec	C27H31N6O5SI·2.5H2O	C, H, N
52	85	236	C12H16N4S2	C, H, N
53	84	135	C16H24N4S2	C, H, N
54	79	241	C12H16N4OS	C, H, N
55	81	153	C10H24N4OS	C, H, N
56	84	256	C13H12N4OS	C, H, N
57	72	223	C15H16N4OS	H, C, Na
60b	84	230	C11H9NO3SHg	C, H, N
61a	43	71–73	C6H5O2SI·0.5H2O	C, H

^aC: calcd, 59.32; found, 58.52. N: calcd, 18.65; found, 18.12.

^bN: calcd, 14.05; found, 15.42.

For 8-[p-[(carboxymethyl)oxy]phenyl]xanthine derivatives related to a xanthine amine congener, compound 3, an alternate route was used to form the imidazole ring. 1,3-Dipropyl-5,6-diaminouracil (8b) or the corresponding 2-thiouracil (8d) was condensed with [(p-formylphenyl)-oxy]acetic acid, forming the imine 12. Upon oxidation, the carboxylic acid congeners 27 (XCC) and 28 were obtained. The xanthine carboxylic acid derivatives were then esterified, giving the ethyl esters 29 and 30, respectively, which were treated with neat ethylenediamine as previously reported⁷ to give the amine derivatives 3 and 31. Since the A₂ potency of compound 31 was enhanced over the oxygen analogue (see below), compound 3, we synthesized other 8-aryl-2-thioxanthine derivatives in an effort to increase A₂ potency. Other amine derivatives were synthesized through aminolysis reactions (compounds 32–35) or by carbodiimide coupling (compounds 36–39) as reported.^10b Lysyl conjugates 38–43 were prepared as described.^10c

An N-hydroxysuccinimide ester derivative, 44, was reported to be an irreversible inhibitor of A₁-adenosine receptors at concentrations greater than 50 nM.^10a If shown to be a potent and nonselective adenosine antagonist, this xanthine may be a potential inhibitor of both A₁- and A₂-adenosine receptors. Certain isothiocyanate-containing xanthines related to compound 3 also have been shown to be chemically reactive with A₁ receptors.^10a Efforts to synthesize analogous xanthine–isothiocyanates containing the 2-thio substitution were unsuccessful, likely due to side reactions involving the more reactive thio group.

A thiation reaction was used to generate 6-thioxanthine derivatives from the corresponding oxygen analogues. It is known¹¹ that xanthine derivatives are preferentially thiated at the 6-position with P₄S₁₀. Dioxane was the favored reaction medium to give high yields of the anticipated 6-thio- and 2,6-dithioxanthines (compounds 52–57). For example, CPX was converted to 8-cyclopentyl-1,3-dipropyl-6-thioxanthine (55) by using this thiation reaction.

Iodinated xanthine derivatives, synthesized by using a prosthetic group¹² or by classical methodology, have been introduced as high-affinity radioligands for adenosine receptors.^12,13 We have explored the use of a 2-thienyl substituent as a site for selective iodination, via mercuration (Scheme II). These substituted thiophene derivatives, such as 59 and 63, undergo regioselective mono-mercuration at the unsubstituted 2-position, rapidly and at ambient temperature, in the presence of stoichiometric quantities of mercury salts such as mercuric acetate.¹⁴ The 2-mercuriothiophene salt 60 is then exposed to elemental iodine, resulting in the corresponding 2-iodothiophene derivative 61.

Scheme II.

Use of 2-Thienyl Derivatives as Prosthetic Groups for Mercuration and Subsequent Iodination^a

^aL = leaving group, such as N-hydroxysuccinimide.

New prosthetic groups designed for facile radiodination of functionalized drugs and peptides are still being sought.^15,16 We have used thiophene-2-acetic acid (as its reactive N-hydroxysuccinimde ester, 58) and thiophene-2-methanamine (62) as prosthetic groups for iodination, via mercuration.

Compound 58 reacted with XAC (3) to form an amide, compound 46. This xanthine bearing a 2-alkylthienyl prosthetic group was readily mercurated to give 47.

Iodination via 2-mercuriothiophene intermediates as in Scheme II may be carried out selectively in the presence of other susceptible aromatic groups, such as phenols. Compound 58 reacted with L-tyrosylglycine to form an amide [compound 59, in which R³ = CH(CH₂C₆H₄OH)-CONHCH₂COOH]. Upon sequential treatment with mercuric acetate and iodine (1 equiv), the corresponding monoiodinated peptide derivative, 61 [R⁴ = CONHCH-(CH₂C₆H₄OH)CONHCH₂COOH], was obtained in high yield.

The N-succinoyl derivative [63a; R³ = (CH₂)₂COOH] of thiophene-2-methanamine was mercurated to form an internal salt, 60b [R⁴ = NHCO(CH₂)₂COOH] which precipitated from methanol. Upon treatment with iodine an immediate reaction occurred. This reaction was followed by NMR in DMSO-d₆. The complete reaction of the 2-mercuriothiophene derivative was indicated by shifts of the thiophene aromatic signals to 6.68 and 7.13 ppm from TMS, corresponding to the 2-iodo derivative 61b.

Pharmacology

Affinity at A₁- and A₂-adenosine receptors was measured in competitive binding assays, using as radioligands [³H]-N⁶-(phenylisopropyl)adenosine¹⁷ (with rat cerebral cortical membranes) and [³H]-5′-(N-ethylcarbamoyl)adenosine (with rat striatal membranes),¹⁸ respectively.

A sulfur substitution at the 2-position carbonyl group of 1,3-dialkylxanthines usually did not decrease the affinity of the xanthines for A₁- or A₂-adenosine receptors. In the case of the 2-thio analogue of CPX, compound 14, the A₁ affinity was enhanced by the thio substitution. The 2-thioxanthine amine congener, compound 30, bound with greater affinity at A₂ receptors and with less affinity at A₁ receptors than the corresponding oxygen analogue, compound 3. Potency at A₂ receptors was enhanced 7-fold by the 2-thio substitution in the case of a carboxylic acid congener (compounds 27 and 28).

N-Methylated analogues (32–37) of compound 31 were prepared. As in the 2-oxo series, the secondary N-methylamine derivative 33 was the most potent at A₂ receptors with a K_i value of 6.8 nM. Thus, the combination of two modifications of compound 3 enhanced its A₂ affinity 10-fold.

A sulfur substitution at the 6-position carbonyl group of 1,3-dialkylxanthines was not well tolerated at either A₁ or A₂ binding sites. Thus, the 6-thio analogue of CPX 55 was 17-fold less potent than CPX at A₁ receptors. The 6-thio analogue of 1,3-diethyl-8-phenylxanthine (57) was 23-fold less potent than DPX (24) at A₁ receptors and 12-fold less potent at A₂ receptors. 2,6-Dithio analogues, such as 53, were intermediate in potency between the corresponding 2-thio and 6-thio analogues.

Substitutions of thienyl and furyl groups at the 8-position of xanthines have approximately equivalent effects on affinity at both receptor subtypes. Both substitutions are generally slightly less potent than the corresponding 8-phenylxanthine analogues. For both thienyl and furyl derivatives (several of which were reported previously, by Bruns et al.¹⁸), attachment at the 3-position relative to the heteroatom (sulfur or oxygen, respectively) results in greater potency at adenosine receptors than attachment at the 2-position. Preference for 3-thienyl derivatives was evident, particularly at the A₂ subtype (cf. 18 vs 17 and 20 vs 19).

1,3-Dipropyl-8-(2-thienyl)-2-thioxanthine (22) did not bind measurably to A₂ receptors at its limit of aqueous solubility. At pH 7.7 in Tris buffer this concentration was 5.8 μM, determined with a log ε for absorption in methanol at 320 nm of 4.38. At this concentration there was not even a partial displacement of tritiated [³H]-5′-(N-ethylcarbamoyl)adenosine from striatal membranes. Thus, compound 22 was > 142-fold A₁ selective in these binding assays. Compound 30 is > 740-fold A₁ selective, but low solubility may limit its usefulness. The aqueous solubilities of A₁-selective xanthines 26a and 30 are 9.0 and 5.4 μM, respectively.

Discussion

We have found that 2-thioxanthines are very similar in potency to the corresponding oxygen analogues. In certain cases, as for the CPX analogue 14 and an 8-(2-thienyl) derivative, 22, a greater margin of A₁ selectivity may be achieved by using the 2-thio substitution. A 6-thio substitution is not well tolerated at either A₁ or A₂ receptors.

The feasibility of using a 2-thienyl moiety as a prosthetic group for selective iodination via its Hg²⁺ derivative was explored. By facile and selective mercuration at the 2-thienyl ArH, a site for rapid and regioselective (in the presence of phenols) iodination is created. A xanthine conjugate of XAC and thiophene-2-acetic acid was sequentially mercurated and iodinated by this scheme, resulting in an iodinated xanthine with potential use as an antagonist radioligand for adenosine receptors. This scheme may have applicability to other receptor ligands, including tyrosyl peptides, in which iodination in the presence of essential phenolic groups is desired.

Experimental Section

8- [4- [[[[(2–Aminoethyl)amino]carbonyl]methyl]oxy]phenyl]-1,3-dipropylxanthine (XAC; xanthine, amine congener), 2-chloroadenosine, and 8-cyclohexyl-1,3-dipropylxanthine were obtained from Research Biochemicals, Inc. (Natick, MA). Compounds 32, 34, 36, 38,42, 50, and 51 were reported previously.^7c,10b Amino acid derivatives of XAC and the 2-thio analogue were synthesized in the manner previously described,^7c using the water-soluble 1-[3-(dimethylamino)propyl]-3-ethylcarbodiimide hydrochloride (EDAC) in dimethylformamide. [(p-Formylphenyl)oxy]acetic acid was obtained from Eastman Kodak (Rochester, NY). [³H]-N⁶-(Phenylisopropyl)adenosine and [³H]-5′-(N-ethylcarbamoyl)adenosine were from Du Pont NEN Products, Boston, MA. Thiophene-2-acetic acid was from Aldrich.

New compounds were characterized by 300-MHz proton NMR (unless noted, chemical shifts are in DMSO-d₆ in ppm from TMS), chemical ionization mass spectroscopy (CIMS, NH₃, Finnigan 1015 spectrometer), and C, H, and N analysis. UV spectra were measured in methanol, and the results are expressed as peak wavelengths in nanometers with log ε values in parentheses.

General Procedure for Compound 10

A 1,3-disubstituted 5,6-diaminouracil or the corresponding 1-thiouracil (10 mmol) was suspended in 10 mL of absolute pyridine, and then under stirring 11 mmol of the acid chloride (freshly prepared) was added dropwise. After 5-h stirring at room temperature, the reaction mixture was poured slowly into 100 mL of H₂O, and the precipitate was collected by suction filtration. Purification was done by recrystallization from a EtOH/H₂O mixture. Yields ranged from 70 to 90%.

General Procedure for Compounds 13–26

A 1,3-disubstituted 5-(acylamino)-6-aminouracil or the corresponding 2-thiouracil (10 mmol) was heated under reflux in a mixture of 40 mL of 1 N NaOH and 10 mL of EtOH for 1 h. The hot solution was acidified with acetic acid, resulting in the formation of a precipitate upon cooling. The precipitate was collected and recrystallized from a H₂O/EtOH mixture: yield 80–90% of colorless crystals; ¹H NMR spectrum (compound 13) δ 3.86 (3 H, s, CH₃), 3.68 (3 H, s, CH₃), 3.19 (m, 1 H, cyclohex C1), 2.0 (m, 2 H, cyclohex C2 and C5), 1.6–1.8 (m, 6 H, cyclohex). The NMR spectra of the other compounds were consistent with the assigned structures.

8-[4-[[[[N-[2-(Dimethylamino)ethyl]-N-methylamino]-carbonyl]methyl]oxy]phenyl]-2-thio-1,3-dipropylxanthine (37)

8-[4-[(Carboxymethyl)oxy]phenyl]-1,3-dipropyl-2-thioxanthine (compound 28; 21 mg, 52 μmol), N,N,N′-trimethyl-ethylenediamine (Aldrich, 20 mg, 0.20 mmol), EDAC (45 mg, 0.23 mmol), and 1-hydroxybenzotriazole (HOBt; 25 mg, 0.18 mmol) were combined in 1 mL of DMF. After stirring overnight, 0.5 mL of sodium carbonate (pH 10, 0.5 M) and 2 mL of saturated NaCl were added. After cooling overnight, a white precipitate was collected, yield 13 mg (52%). The NMR and mass spectra were consistent with the assigned structure.

8-[4-[[[[[2-[[[6-[(N-Succinimidyloxy)carbonyl]-n-hexyl]carbonyl]amino]ethyl]amino]carbonyl]methyl]-oxy]phenyl]-1,3-dipropyl-2-thioxanthine (45)

Compound 31 (10.4 mg, 0.024 mmol) was added to a solution of disuccinimidyl suberate (13.1 mg, 0.036 mmol; Pierce, Rockford, IL) in DMF (1 mL) and vigorously stirred for 2 h or until complete by TLC (CHCl₃/MeOH/AcOH, 18/1/1). Dry ether (2 mL) was then added to the suspension followed by the addition of petroleum ether until cloudy. The suspension was allowed to stand at 0 °C for 1 h and then filtered to give an off-white powder: yield 11.5 mg (68%); mp 160–168 °C; ¹H NMR (DMSO-d₆) 0.90 (t, J = 7.2 Hz, 3 H), 0.94 (t, J = 7.2 Hz, 3 H), 1.28 (br m, 4 H), 1.44 (m, 2 H), 1.71 (m, 2 H), 1.83 (m, 2 H), 1.83 (t, J = 7.6 Hz, 2 H), 2.65 (t, J = 7.0 Hz, 2 H), 2.79 (s, 4 H), 3.15 (br s, 4 H), 3.32 (s, H₂O), 4.45 (br t, J = 7.2 Hz, 2 H), 4.54 (s, 2 H), 4.59 (br t, J = 7.2 Hz, 2 H), 7.10 (d, J = 8.7 Hz, 2 H), 7.83 (br s, 1 H), 8.11 (d, J = 8.7 Hz, 2 H), 8.18 (br s, 1 H).

General Procedure for 6-Thiation Reaction

The appropriate xanthine or 2-thioxanthine derivative (10 mmol) was heated with 6 g of P₄S₁₀ in 100 mL of dioxane for 3–5 h under reflux. Insoluble material was removed by filtration, and the filtrate was added dropwise to 200 mL of H₂O with stirring. The precipitate was collected and purified by recrystallization from a H₂O/EtOH mixture; yield 80–85%.

N-Succinimidyl Thiophene-2-acetate (58)

Thiophene-2-acetic acid (1.61 g, 11 mmol), dicyclohexylcarbodiimide (2.34 g, 11 mmol), and N-hydroxysuccinimide (1.30 g, 11 mmol) were added to 50 mL of ethyl acetate containing 10% DMF. After stirring for 2 h, the urea was removed by filtration. The filtrate was washed with aqueous acid/base and evaporated. The residue was recrystallized from ethyl acetate/petroleum ether; yield 2.01 g (74%), mp 127–128 °C; C, H, N, S analysis for C₁₀H₉NO₄S.

5-Mercuriothiophene-2-acetate (60a; R⁴ = COO⁻). Mercuration Reaction

Thiophene-2-acetic acid (0.39 g, 2.8 mmol) was dissolved in 8 mL of methanol. Mercuric acetate was added with stirring, and a white precipitate appeared shortly thereafter. After 1 h, 4 mL of ether was added and the solid was collected by filtration; yield (of 60a) 0.88 g (93%). Mass spectrum (CI, NH₃) shows a peak at 360 z/e corresponding to M + 1 + NH₃.

The thiophene–xanthine derivative 46 was prepared as reported previously.^10b Upon mercuration in dimethylformamide by a similar method, a solid product, compound 47, was obtained and characterized by californium plasma desorption mass spectroscopy.

2-[(N-Succinoylamino)methyl]-5-mercuriothiophene [60b; R⁴ = NHCO(CH₂)₂COO⁻]

Compound 63a (60.5 mg, 0.28 mmol) was dissolved in 5 mL of methanol and treated with 100 mg (0.31mmol) of mercuric acetate. First a solution formed, followed by crystallization of product. After 1 h, ether was added, and the precipitate was collected: yield 98 mg (84%); mp 230 °C dec.

This compound was converted to the corresponding 5-iodo derivative upon treatment at room temperature with iodine or iodine monochloride. ¹H NMR of 2-[(N-succinoylamino)-methyl]-5-mercuriothiophene (61b): 8.46 (1 H, t, NH), 7.13 (1 H, d, Ar-4), 6.68 (1 H, d, Ar-3), 4.37 (2 H, d, CH₂N), 2.43 (2 H, CH₂), 2.35 (2 H, CH₂).

5-Iodothiophene-2-acetic Acid (61a; R⁴ = COOH)

Compound 60a (75 mg, 0.22 mmol) was suspended in 5 mL of dimethylformamide containing 5% DMSO. Iodine crystals (77 mg, 0.31 mmol) were added with stirring. A solution formed within 1 min. Hydrochloric acid (1 M) was added, and the mixture was extracted three times with ethyl acetate. The combined organic layers were dried (Na₂SO₄) and evaporated. The product was purified by column chromatography on silica gel. R_f of product (silica, ethyl acetate/petroleum ether) was 0.79.

The identical product was obtained as follows: N-Iodo-succinimide (236.2 mg, 1.05 mmol) was added to a stirred suspension of compound 59a (326.2 mg, 0.95 mmol) in methanol (30 mL). After 16 h the suspension was filtered and the methanol removed in vacuo. The remaining oil was redissolved in ethyl acetate and washed with 0.5 N HCl, and the product was extracted into a 0.5 N NaOH solution. The basic fraction was washed with CH₂Cl₂, acidified to pH 1.0 with 1 N HCl, and extracted with EtOAc. The product was chromatographed on a silica gel column (eluent, 17/2/1 CHCl₃/MeOH/AcOH), and the solvents were removed from the product fractions in vacuo. Acetic acid was removed by azeotropic distillation with petroleum ether. The light yellow oil was redissolved in ethyl acetate, and ether was added, forming a precipitate which was removed by filtration. Evaporation of the solvent gave compound 61a as a waxy yellow solid (110 mg, 43%).

Peptide Derivatives of Thiophene-2-acetic Acid

Compound 58 reacted with L-tyrosylglycine (274 mg, 1.15 mmol) in dimethylformamide to give N-(thiophene-2-acetyl)-L-tyrosylglycine (230 mg, 55% yield). Upon mercuration in dimethylformamide as for compound 60a, N-(5-mercuriothiophene-2-acetyl)-L-tyrosylglycine (40% yield) was obtained.

2-[(N-Succinoylamino)methyl]thiophene [63a; R = (CH₂)₂COOH]

Thiophene-2-methanamine (Aldrich, 2.37 g, 21 mmol) was dissolved in 20 mL of tetrahydrofuran and treated with a solution of succinic anhydride (2.1 g, 21 mmol) in 20 mL of dimethylformamide. After ½ h, ethyl acetate (50 mL) was added, and the mixture was extracted with citric acid (1 M) three times and with water. The organic layer was dried (MgSO₄). Solvent was removed and petroleum ether was added, causing white crystals of 62a to precipitate: Yield 9.4%; mp 130 °C. The NMR and mass spectra were consistent with the assigned structure. UV spectrum λ_max 233 nm (log ε 4.019).

Biochemical Assays

Stock solutions of xanthines were prepared in the millimolar concentration range in dimethyl sulfoxide and stored frozen. Solutions were warmed to 50 °C prior to dilution in aqueous medium. Inhibition of binding of 1 nM [³H]-N⁶-(phenylisopropyl)adenosine to A₁-adenosine receptors in rat cerebral cortical membranes was assayed as described.¹⁷ Inhibition of binding by a range of concentrations of xanthines was assessed in triplicate in at least three separate experiments. IC₅₀ values, computer generated by using a nonlinear regression formula on the Graphpad program, were converted to K_i values by using a K_D value for [³H]PIA of 1.0 nM and the Cheng–Prusoff equation.¹⁹

Inhibition of binding of [³H]-5′-(N-ethylcarbamoyl)adenosine to A₂-adenosine receptors in rat striatal membranes was measured as described,¹⁸ except that 5 mM theophylline was used to define nonspecific binding. N⁶-Cyclopentyladenosine was present at 50 nM to inhibit binding of the ligand at A₁-adenosine receptors. Inhibition of binding by a range of concentrations of xanthines was assessed in triplicate in at least three separate experiments. IC₅₀ values were converted to K_i values by the method of Bruns et al.,¹⁸ using a conversion factor derived from the affinity of [³H]-5′-(N-ethylcarbamoyl)adenosine at A₂ receptors and the Cheng–Prusoff equation.¹⁹