5'-O-alkyl ethers of N,2-substituted adenosine derivatives: partial agonists for the adenosine A1 and A3 receptors.

Article abstract of DOI:10.1021/jm001114o

New N,5'-di- and N,2,5'-trisubstituted adenosine derivatives were synthesized in good overall yields. Appropriate 5-O-alkyl-substituted ribose moieties were coupled to 6-chloropurine or 2,6-dichloropurine via Vorbrueggen's glycosylation method. Subsequent amination and deprotection of the intermediates yielded compounds 18-35. Binding affinities were determined for rat adenosine A1 and A2A receptors and the human A3 receptor. The ability of compounds 18-35 to inhibit forskolin-induced (10 microM) cyclic AMP (cAMP) production and their ability to stimulate guanosine 5'-O-(3-[35S]thio)triphosphate ([35S]GTPgammaS) binding, via either the adenosine A1 receptor or the adenosine A3 receptor, were assessed. N-Cyclopentyl-substituted adenosine derivatives displayed affinities in the low nanomolar range for the adenosine A1 receptor, whereas N-(3-iodobenzyl)-substituted derivatives had high affinity for the adenosine A3 receptor. Compound 22 had the highest affinity for the adenosine A1 receptor (K(i) value of 16 nM), and compounds 20 and 26 had the highest affinities for the adenosine A3 receptor (K(i) values of 4 and 3 nM, respectively). A chlorine substituent at the 2-position either did not affect or slightly increased the adenosine A1 receptor affinity, whereas the A3 receptor affinity was affected differently, depending on the N-substituent. Furthermore, the introduction of chlorine slightly increased the A3/A1 selectivity ratio. At the 5'-position, an O-methyl substituent induced the highest adenosine A1 receptor affinity, whereas an O-ethyl substituent did so for the A3 receptor. All compounds showed partial agonistic effects in both the cAMP and [35S]GTPgammaS assays, although more marked in the latter assay. In general, the 2-chloro derivatives seemed to have lower intrinsic activities compared to the 2-H-substituted compounds on both the adenosine A1 and the adenosine A3 receptors. The compounds with an N-(3-iodobenzyl) substituent displayed the lowest intrinsic activities. Finally, all compounds also showed partially antagonistic behavior in the [35S]GTPgammaS assay.

Full text of DOI:10.1021/jm001114o

2966
J . Med. Chem. 2001, 44, 2966-2975
5′-O-Alk yl Eth er s of N,2-Su bstitu ted Ad en osin e Der iva tives: P a r tia l Agon ists for
th e Ad en osin e A1 a n d A3 Recep tor s
Erica W. van Tilburg,† Pieter A. M. van der Klein,† J acobien von Frijtag Drabbe Ku¨nzel,† Miriam de Groote,†
Christina Stannek,‡ Anna Lorenzen,‡ and Ad P. IJ zerman*,†
Division of Medicinal Chemistry, Leiden/ Amsterdam Center for Drug Research, P.O. Box 9502, 2300 RA Leiden,
The Netherlands, and Institute of Pharmacology, University of Heidelberg, D-69120 Heidelberg, Germany
Received November 10, 2000
New N,5′-di- and N,2,5′-trisubstituted adenosine derivatives were synthesized in good overall
yields. Appropriate 5-O-alkyl-substituted ribose moieties were coupled to 6-chloropurine or 2,6-
dichloropurine via Vorbru¨ggen’s glycosylation method. Subsequent amination and deprotection
of the intermediates yielded compounds 18-35. Binding affinities were determined for rat
adenosine A1 and A2A receptors and the human A3 receptor. The ability of compounds 18-35
to inhibit forskolin-induced (10 µM) cyclic AMP (cAMP) production and their ability to stimulate
guanosine 5′-O-(3-[35S]thio)triphosphate ([35S]GTPγS) binding, via either the adenosine A1
receptor or the adenosine A3 receptor, were assessed. N-Cyclopentyl-substituted adenosine
derivatives displayed affinities in the low nanomolar range for the adenosine A1 receptor,
whereas N-(3-iodobenzyl)-substituted derivatives had high affinity for the adenosine A3 receptor.
Compound 22 had the highest affinity for the adenosine A1 receptor (Ki value of 16 nM), and
compounds 20 and 26 had the highest affinities for the adenosine A3 receptor (Ki values of 4
and 3 nM, respectively). A chlorine substituent at the 2-position either did not affect or slightly
increased the adenosine A1 receptor affinity, whereas the A3 receptor affinity was affected
differently, depending on the N-substituent. Furthermore, the introduction of chlorine slightly
increased the A3/A1 selectivity ratio. At the 5′-position, an O-methyl substituent induced the
highest adenosine A1 receptor affinity, whereas an O-ethyl substituent did so for the A3 receptor.
All compounds showed partial agonistic effects in both the cAMP and [35S]GTPγS assays,
although more marked in the latter assay. In general, the 2-chloro derivatives seemed to have
lower intrinsic activities compared to the 2-H-substituted compounds on both the adenosine
A1 and the adenosine A3 receptors. The compounds with an N-(3-iodobenzyl) substituent
displayed the lowest intrinsic activities. Finally, all compounds also showed partially
antagonistic behavior in the [35S]GTPγS assay.
In tr od u ction
Partial agonists are compounds that can cause a
submaximal stimulation of the receptor at saturating
concentrations and block the effect of full agonists to
some extent (same binding site at the receptor).6,7 With
the term “intrinsic activity”, the ability of a partial
agonist to elicit a stimulus from the activation of a single
receptor is reflected. In this simplified model,8 a ligand
induces a conformational change in the receptor, thereby
conferring affinity to the receptor for the next signaling
protein in the transduction cascade (e.g., a G protein).
The intensity of the stimulus may be related to the
degree of change in conformation, the frequency of
change in conformation, and the duration of the induced
conformation. In a more complex model,9 it has been
suggested that the active conformation of the receptor
(R*) is already one of the variety of conformations that
the receptor can form spontaneously, in the absence of
an agonist (for a review on this matter see De Ligt et
al.10). Hence, by selective affinity the agonist (or inverse
agonist) either increases or decreases the population of
these conformations (R*), respectively, to produce a
functional response. This way, a single ligand may have
multiple intrinsic efficacies.9
Endogenous adenosine exerts its action via specific
membrane-bound receptors called P1-purinoceptors.
There are four adenosine receptor subtypes, the adeno-
sine A1 and A3 receptors, which inhibit the enzyme
adenylate cyclase after activation, and the A2A and A2B
receptors, which stimulate adenylate cyclase. The ad-
enosine receptors are widely distributed in the body, and
selective ligands for the receptors are required. The
potential therapeutic use of adenosine A1 agonists, for
example, in the treatment of type II diabetes1,2 might
be accompanied by a certain extent of cardiotoxicity.
Useful cardio- and cerebroprotective actions of chroni-
cally administered adenosine A3 receptor agonists3,4
often go with unwanted stimulation of the (rat) A3
receptor in the lungs and immune system, causing
bronchoconstriction and the release of allergic media-
tors.3 One way to circumvent these side effects is the
use of partial agonists,5 which may exploit the differ-
ences in receptor-effector coupling in various tissues
and can achieve selectivity of action in vivo.
* To whom correspondence should be addressed. Phone: (+31) 71
5274651. Fax: (+31) 71 5274565. E-mail: ijzerman@chem.leidenuniv.nl.
† Leiden/Amsterdam Center for Drug Research.
Partial agonists selective for the adenosine A1 recep-
tor that fully inhibit lypolysis in vivo while being devoid
of cardiovascular effects have been successfully synthe-
‡ University of Heidelberg.
10.1021/jm001114o CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/26/2001

5′-O-Alkyl Ethers of Adenosine Derivatives
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 18 2967
Sch em e 1a
atoms present in the base moiety (except for the N
nitrogen, which is less reactive). Treatment of either
2′,3′-O-isopropylideneadenosine or 2′,3′-O-isopropylide-
neinosine with KOH, 18-crown-6, and iodomethane20,21
yielded the N1-methylated derivative or a mixture of
the N1-methylated product together with the N1,5′-
dimethylated one, respectively. Refluxing a 5′-chloro-
5′-deoxyadenosine derivative in methanolate provided
only starting material. Therefore, we chose to alkylate
the ribose first, then couple it to the appropriate
heterocyclic base, and subsequently perform other reac-
tions.
Protection of the 1-, 2-, and 3-hydroxyl groups of
D-ribose (1) was performed according to a method
described by Leonard et al.22 with a minor modification.
D-Ribose was dissolved in a mixture of acetone, MeOH,
and 2,2-dimethoxypropane, and HCl(g) was bubbled
directly through the solution, instead of adding MeOH
saturated with HCl(g). The protected â-sugar 2 was
obtained in good yield (95%). Alkylation of the free
hydroxyl group of 2 was performed under standard
conditions, i.e., treatment with NaH in DMF with
cooling (not in the case of cyclopropyl bromide), followed
by the addition of the alkyl halide iodomethane,
iodoethane, or cyclopropyl bromide, respectively. Com-
pounds 3-5 were obtained in 38-97% yield. Subse-
quently, the protecting groups of the alkylated com-
pounds 3-5 were removed by refluxing the material in
aqueous HCl (0.04 M) for 2 h and subsequent neutral-
ization with BaCO3, and compounds 6-8 were obtained
in good yields.23 Our first efforts to remove the protect-
ing groups with aqueous sulfuric acid (0.02 M) and
EtOH failed. This procedure yielded, besides the desired
product, also a significant amount of 1-O-ethyl-substi-
tuted derivatives.
a
Reaction conditions: (a) acetone, MeOH, 2,2-dimethoxypro-
pane, HCl(g), (b) NaH, DMF, alkyl halide, R1 = CH3, C2H5, or
c-C3H5; (c) (i) 0.04 M HCl or (ii) BaCO3; (d) pyridine, DMAP, acetic
anhydride; (e) 1,2-dichloroethane, silylated base, TMSOTf, R2
)
H, Cl; (f) (i) EtOH/NH3 or (ii) absolute EtOH, appropriate amine,
R3 ) H, c-C5H9, or CH2(C6H4)-m-I.
sized in our laboratory,11,12 as have partial agonists for
the adenosine A3 receptor.13 In these studies, selectivity
for the adenosine A1 or A3 receptor was obtained by the
introduction of either a cyclopentyl or a 3-iodobenzyl
substituent at the N-position, respectively. Further-
more, alkylthio substituents at the 5′-position induced
partial agonism on both the adenosine A1 and A3
receptors.13,14
In the present study, an N-cyclopentyl or an N-(3-
iodobenzyl) group was selected to introduce either
adenosine A1 or adensosine A3 receptor selectivity,
respectively. Attention has now been focused on the
synthetically less accessible 5′-O-alkyl ether derivatives
to induce partial agonism. The synthesis of some 5′-O-
methyladenosine derivatives has been published before,
although in combination with other N-substituents.15
The effect of the 5′-O-alkyl substituents on the intrinsic
activity was studied here, in comparison with the 5′-
alkyl thioether adenosine derivatives that have shown
reduced intrinsic activities,13,14 as mentioned earlier.
When preparing 5′-O-alkyl ether adenosine derivatives,
one must be careful to prevent alkylation at the other
ribose hydroxyl groups,16 as well as at the nitrogen
atoms present in the heterocyclic base. A procedure to
prevent these problems may be the glycosylation of
appropriate protected 5-O-alkyl-substituted riboses to
a heterocyclic base, as described by Vorbru¨ggen et al.17,18
However, yields have often been reported to be quite
low. Finally, in this study, the additional effect of
chlorine introduced at the 2-position of the adenosine
derivatives on the affinity and intrinsic activity was
studied.
The hydroxyl groups of 6-8 were then protected. In
general, benzoyl protecting groups at the 2-, 3-, and
5-positions of the sugar moiety have been used to
improve the formation of â-nucleosides over R-nucleo-
sides in the glycosylation reaction, i.e., when the pro-
tected sugar is coupled to a heterocyclic base.17 Treat-
ment of methyl 5-O-methyl-D-ribofuranose (6) with
benzoyl chloride in CH2Cl2 and pyridine yielded the
desired methyl 2,3-di-O-benzoyl-5-O-methyl-D-ribofura-
noside, although it was difficult to remove the excess
benzoyl chloride.24 The conversion of the 1-O-methyl
group into a 1-O-acetyl group succeeded in reasonable
yield by treatment of methyl 2,3-di-O-benzoyl-5-O-
methyl-D-ribofuranoside with acetic acid, acetic anhy-
dride, and sulfuric acid.24 Since this method is quite
laborious, the possibility of using a fully acetyl protected
sugar in the Vorbru¨ggen coupling reaction was explored
as well. The hydroxyl groups of 6-8 were protected with
acetyl groups to obtain compounds 9-11. Then, the
peracylated compounds 9-11 were coupled to the ap-
propriate heterocyclic base, e.g., 6-chloropurine or 2,6-
dichloropurine, according to the Vorbru¨ggen method18
to give compounds 12-17 in good yields (48-84%).
These good yields indicated that peracylated riboses
were a good alternative compared to the benzoyl-
protected riboses. In fact, recently Vorbru¨ggen et al.
have described that coupling can also be performed with
nonprotected riboses, as they can be persilylated prior
to reaction with a heterocyclic base.25
The compounds were tested in radioligand binding
assays for affinity. Their intrinsic activities were de-
termined in both [35S]GTPγ[S] (guanosine 5′-O-(3-[35S]-
thio)triphosphate) binding assays7,19 and cyclic AMP
(cAMP) assays.
Resu lts a n d Discu ssion
Ch em istr y. The synthesis of the substituted adeno-
sine derivatives 18-35 is depicted in Scheme 1. Selec-
tive alkylation at the 5′-position of adenosine, or one of
its derivatives, is difficult since one must be very careful
to prevent alkylation at the more reactive nitrogen

2968 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 18
van Tilburg et al.
Ta ble 1. Affinities of N,2,5′-trisubstituted Adenosine Analogues at Adenosine A1, A2A, and A3 Receptors Expressed as Ki Values (nM
( SEM, n ) 3) or Percentage of Displacement at 10 µM
Ki (nM) or % displacement at 10-5
M
a
b
c
compd
CPA
2-Cl-CPA
IBMECA
R1
R2
R3
c-C5H9
c-C5H9
CH2(C6H4)-m-I
CH2(C6H4)-m-I
H
c-C5H9
CH2(C6H4)-m-I
A1
A2A
A3
A1:A3
H
H
H
Cl
H
Cl
H
H
7.14 ( 2.30
9.47 ( 0.86
1400 ( 240
710 ( 41
45.9%
281 ( 56
204 ( 44
6.9 ( 0.2
7.2 ( 0.9
40.3 ( 8.1
79.0 ( 30.0
4.2 ( 0.5
28.3 ( 13.0
44.9 ( 4.9
7.3 ( 0.2
23.3 ( 3.9
32.3 ( 9.2
3.3 ( 1.0
15.1 ( 6.9
67.0 ( 44.6
7.4 ( 1.0
45.6 ( 7.4
104 ( 20
9.1 ( 1.9
26.7 ( 4.3
86.0 ( 15.5
20.5 ( 8.1
0.02
0.05
208
98
4.1
0.4
42.6
6.0
0.4
15.3
98.7
0.9
60.0
21.7
0.4
53.1%
39.4%
23.5%
2-Cl-IBMECA
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
CH3
CH3
CH3
CH3
CH3
CH3
C2H5
C2H5
C2H5
C2H5
C2H5
C2H5
c-C3H5
c-C3H5
c-C3H5
c-C3H5
c-C3H5
c-C3H5
164 ( 37
261 ( 65
366 ( 109
373 ( 80
284 ( 55
1390 ( 140
527 ( 21
415 ( 69
33.9%
33.7 ( 6.9
179 ( 40
H
Cl
Cl
Cl
H
H
H
Cl
Cl
Cl
H
H
H
H
169 ( 30
c-C5H9
CH2(C6H4)-m-I
H
c-C5H9
CH2(C6H4)-m-I
H
c-C5H9
CH2(C6H4)-m-I
H
c-C5H9
CH2(C6H4)-m-I
H
16.3 ( 1.3
112 ( 10
2300 ( 320
27.8 ( 2.2
198 ( 36
40.4%
327 ( 15
659 ( 168
1730 ( 700
848 ( 60
739 ( 16
24.3%
27.1 ( 1.8
254 ( 28
34.3
12.6
0.4
80.5
22.3
0.9
575 ( 52
40.5 ( 6.4
733 ( 36
43.7%
Cl
Cl
Cl
595 ( 84
885 ( 89
c-C5H9
CH2(C6H4)-m-I
80.8 ( 12.0
1430 ( 200
34.6%
38.7%
69.8
a
b
Displacement of [3H]DPCPX from rat cortical membranes.32
Displacement of [3H]ZM 241385 from rat striatal membranes.33
c Displacement of [125I]ABMECA from the human A3 receptor expressed in HEK 293 cells.35
From NMR spectroscopy it was concluded that only
â-congeners were formed here. The Lewis acid used in
the Vorbru¨ggen method, trimethylsilyl trifluoromethane-
sulfonate (TMS triflate), belongs to a relatively new
family of acids. It was compared in this study with the
more traditional Lewis acid stannic chloride (SnCl4),
described as high-yielding.26 Both Lewis acids led to
similar yields of the coupled product as a result of the
glycosylation of 6-chloropurine and commercially avail-
able 1-O-acetyl-2,3,5-tri-O-benzoyl-â-D-ribosfuranose. Am-
ination of compounds 12-17 with EtOH/NH3, cyclopen-
tylamine, or (3-iodobenzyl)amine gave the unprotected
substituted adenosine derivatives 18-35.13,14
Biologica l Eva lu a tion . All compounds were tested
in radioligand binding assays to determine their affini-
ties for the adenosine A1 receptor in rat brain cortex,
the A2A receptor in rat striatum, and the human A3
receptor as expressed in HEK 293 cells (Table 1). For
the adenosine A1 receptor, the tritiated antagonist [3H]-
1,3-dipropyl-8-cyclopentylxanthine ([3H]DPCPX) and,
for the adenosine A2A receptor, the tritiated antagonist
[3H]ZM241385(7-amino-2-(2-furyl)-5-[2-(4-hydroxyphenyl)-
ethyl]amino[1,2,4]triazolo[1,5-a][1,3,5]triazine) were used.
Since radiolabeled antagonists are not commercially
available for the adenosine A3 receptor, [125I]ABMECA
(N-(4-aminobenzyl)-5′-methylcarboxamidoadenosine), an
A3 receptor agonist, was used. Displacement experi-
ments were performed in the absence of GTP (guanosine
5′-triphosphate).
compounds had lower affinity for the adenosine A2A
receptor than for the adenosine A1 or A3 receptor.
Compounds with an unsubstituted N amino group (18,
21, 24, 27, 30, 33) showed adenosine A3 receptor
selectivity compared to A1, with A1:A3 selectivity ratios
ranging from 4.1 for compound 18 to 98.7 for compound
24. A cyclopentyl group at the N-position induced high
adenosine A1 receptor affinities and also the highest
selectivity for the A1 receptor, whereas the N-(3-iodo-
benzyl)adenosine derivatives had the highest adenosine
A3 receptor affinities and were highly selective for this
receptor, in line with earlier reports on these N-
substituents.13,14,27 For CCPA (2-chloro-N-cyclopentyl-
adenosine) it is known that its affinity for the rat
adenosine A1 receptor is about 1.5-fold lower than its
affinity for the human A1 receptor.28,29 This suggests
that the affinity of our compounds for the human A1
receptor might also be somewhat higher, and therefore
the selectivity for the human A3 receptor might be
somewhat less than stated here.
In general, the adenosine A1 receptor affinities were
unchanged or increased (up to 2-fold) when a chlorine
was introduced at the 2-position, with a concomitant
slight increase in adenosine A1 receptor selectivity (up
to 4-fold). The introduction of chlorine at the 2-position
had more variable effects on the affinity and selectivity
for the adenosine A3 receptor. The adenosine A3 receptor
affinities of the N-unsubstituted derivatives were slightly
increased, whereas for the N-(3-iodobenzyl)-substituted
compounds they were decreased somewhat. The selec-
tivity for the adenosine A3 receptor was decreased up
Table 1 displays radioligand binding data for all
synthesized di- and trisubstituted final products. Most

5′-O-Alkyl Ethers of Adenosine Derivatives
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 18 2969
Ta ble 2. Inhibition of Forskolin (10µM) Induced cAMP
Production of N,5-Di- and N,2,5′-Trisubtstituted Adenosine
Derivatives at Either the Human Adenosine A1 or the Human
A3 Receptor
to approximately 4-fold when chlorine was introduced
at C2. The effect of chlorine at the 2-position on receptor
affinity is difficult to predict in the case of substituted
adenosines. It has been shown that for both CPA (N-
cyclopentyladenosine) and N-(3-iodobenzyl)adenosine
the affinity and selectivity for the adenosine A1 receptor
are increased when compared to those for the A2A
receptor.30 However, when chlorine was introduced at
the 2-position of 5′-N-methylcarboxamido-substituted
N-(3-iodobenzyl)adenosine,31 the affinity for both the
adenosine A1 and A2A receptors was decreased, while
the A3 receptor affinity increased. On the contrary, in
the present study introduction of chlorine at the 2-posi-
tion of 5′-O-alkyl-substituted derivatives slightly de-
creased the adenosine A3 receptor selectivity of the
compounds compared to that of the A1 receptor.
a
b
compd
CPA
R1
R2
R3
% inhib A1 % inhib A3
69 ( 3
83 ( 1
Cl-IBMECA
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
CH3
CH3
CH3
CH3
CH3
CH3
C2H5
C2H5
C2H5
C2H5
C2H5
C2H5
c-C3H5
c-C3H5
c-C3H5
H
H
H
Cl
H
75 ( 3
74 ( 4
53 ( 10
78 ( 4
73 ( 4
76 ( 2
58 ( 6
33 ( 9
69 ( 3
42 ( 6
17 ( 7
65 ( 7
37 ( 6
11 ( 6
51 ( 5
31 ( 9
7 ( 4
c- C5H9
CH2(C6H4)-m-I
The 5′-substituents also influenced the affinity and
selectivity of the compounds. For the adenosine A1
receptor the compounds with a 5′-O-methyl group had
the highest receptor affinities, whereas the 5′-O-cyclo-
propyl-substituted derivatives displayed the lowest af-
finities in most cases. Only within the N,2-unsubstituted
series (18, 24, 30) 5′-O-ethyladenosine had the lowest
affinity for the adenosine A1 receptor. In the case of the
adenosine A3 receptor, the derivatives with a 5′-O-ethyl
substituent had the highest receptor affinities. The
adenosine A3 receptor affinities of the 5′-O-methyl-
substituted derivatives were equal or somewhat higher
than those of the 5′-O-cyclopropyl-substituted ones.
Although a 5′-O-cyclopropyl group is tolerated fairly well
on both the adenosine A1 and A3 receptors (Ki values in
the low nanomolar range), smaller groups seemed to be
preferred, e.g., 5′-O-methyl for the adenosine A1 receptor
and 5′-O-ethyl for the A3 receptor. The 5′-O-methyl
group induced the highest adenosine A1 receptor selec-
tivity in most cases (within the series containing the
same N- and 2-substituents), whereas the 5′-O-cyclo-
propyl induced most often the highest adenosine A3
receptor selectivity compared to that of A1. The com-
pounds with a 5′-O-ethyl substituent had higher affini-
ties than those with a 5′-O-methyl for the adenosine A3
receptor, and had the highest selectivity for this receptor
in most cases as well, compared to the A2A receptor.
MECA (5′-N-methylcarboxamidoadenosine) has been
described to have higher affinity (and selectivity) than
NECA (5′-N-ethylcarboxamidoadenosine) for the ad-
enosine A3 receptor compared to the A2A receptor. The
results in this present study, however, indicate that the
adenosine A3 receptor accommodates the rather large
5′-alkylthio substituents better than the smaller ones
(5′-S-methyl-5′-thio substituent) and that the adenosine
A3 receptor is well capable of accommodating large 5′-
substituents. Finally, compound 26 displayed a higher
affinity for the adenosine A3 receptor (Ki value of 3.3
nM) than both reference compounds IBMECA (N-(3-
iodobenzyl)-5′-methylcarboxamidoadenosine) and 2-Cl-
IBMECA (2-chloro-N-(3-iodobenzyl)-5′-methylcarbox-
amidoadenosine), although its selectivity for the adeno-
sine A3 receptor compared to the A1 receptor was
somewhat less.
H
Cl c-C5H9
Cl CH2(C6H4)-m-I
H
H
H
Cl
Cl c-C5H9
Cl CH2(C6H4)-m-I
H
H
H
37 ( 15
57 ( 11
69 ( 6
H
c-C5H9
CH2(C6H4)-m-I
6 ( 21
H
67 ( 7
62 ( 6
-72 ( 38c
61 ( 2
H
59 ( 6
26 ( 7
10 ( 12
44 ( 8
19 ( 7
c-C5H9
62 ( 4
CH2(C6H4)-m-I -104 ( 22c
c-C3H5 Cl
c-C3H5 Cl c-C5H9
H
49 ( 9
43 ( 9
c-C3H5 Cl CH2(C6H4)-m-I -118 ( 41c -2.6 ( 4
a
Percentage inhibition ((SEM, n ) 3). Compared to the
reference full agonist CPA (10 µM) in A1 CHO cells. All compounds
b
were tested at (100Ki. Percentage inhibition ((SEM, n ) 3).
Compared to the reference full agonist Cl-IBMECA (10 µM). All
compounds were tested at (100Ki. c Seemed to behave as inverse
(partial) agonists in this assay (in the [35S]GTPγS assay they
behaved as “normal” (partial) agonists, Table 3).
human adenosine A3 receptors expressed in HEK 293
cells was assessed (Table 2). Note that both the adenos-
ine A1 and A3 receptors are coupled to Gi and inhibit
the production of cAMP. Second, the modulation of [35S]-
GTPγS binding to cell membranes was determined
when compounds 18-35 were bound to the receptors.
Membrane preparations from rat brain cortex (A1 recep-
tors) and from CHO cells stably transfected with human
adenosine A3 receptors were used (no significant stimu-
lation occurred with membranes prepared from the
HEK 293 cells used in the binding studies).
Compounds within the three N-substituted series
showed similar trends in inhibition of the forskolin-
induced cAMP production via the adenosine A1 receptor
(Table 2). Compounds with an intact amino group at
the 6-position and the N-cyclopentyl-substituted deriva-
tives were able to inhibit the cAMP production to a
similar extent, almost as well as the reference full
agonist CPA, whereas the N-(3-iodobenzyl)-substituted
derivatives inhibited the forskolin-induced cAMP pro-
duction somewhat less. Compound 20 had a lower
intrinsic activity than compounds 18 and 19, while only
the N-substituent is different, implicating that a 3-iodo-
benzyl group at N can induce partial agonism for the
adenosine A1 receptor. Furthermore, within the three
N-substituted series the 5′-O-methyl-substituted deriva-
tives had the highest intrinsic activities compared to
the 5′-O-ethyl- and 5′-O-cyclopropyl-substituted ones.
The latter compounds displayed the lowest intrinsic
All compounds were also tested in functional assays.
First, the ability of the compounds 18-35 to inhibit the
forskolin (10 µM) induced cAMP production via human
adenosine A1 receptors expressed in CHO cells or by

2970 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 18
van Tilburg et al.
Ta ble 3. Agonistic Activity of N,5′-Di- and N,2,5′-Trisubstituted Adenosine Analogues Expressed as the Amount of [35S]GTPγS
Bound
adenosine A1 receptor
adenosine A3 receptor
[
35S]GTPγS bound +
[
35S]GTPγS bound +
[
35S]GTPγS bound
1 µM CPA
[
35S]GTPγS bound
1 µM NECA
compd
R1
R2
R3
(% CPA)a
(% CPA)b
(% NECA)c
(% NECA)d
basal
CPA
NECA
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
0
100
0
100
0
0
100
100
CH3
CH3
CH3
CH3
CH3
CH3
C2H5
C2H5
C2H5
C2H5
C2H5
C2H5
c-C3H5
c-C3H5
c-C3H5
c-C3H5
c-C3H5
c-C3H5
H
H
H
Cl
Cl
Cl
H
H
H
Cl
Cl
Cl
H
H
H
Cl
Cl
Cl
H
73 ( 2
68 ( 1
20 ( 0
64 ( 2
53 ( 1
14 ( 1
14 ( 2
33 ( 3
1 ( 2
91 ( 2
74 ( 1
54 ( 1
83 ( 1
72 ( 1
75 ( 2
101 ( 2
53 ( 1
70 ( 11
62 ( 0
47 ( 1
55 ( 7
70 ( 1
53 ( 3
44 ( 12
55 ( 2
39 ( 7
94 ( 3
79 ( 3
63 ( 1
27 ( 2
67 ( 4
51 ( 5
25 ( 7
54 ( 4
38 ( 5
10 ( 4
34 ( 6
30 ( 6
8 ( 4
52 ( 6
38 ( 7
9 ( 5
32 ( 5
31 ( 6
15 ( 7
85 ( 2
72 ( 1
45 ( 3
69 ( 2
74 ( 8
43 ( 6
70 ( 3
61 ( 5
49 ( 8
51 ( 5
47 ( 7
34 ( 10
70 ( 4
54 ( 6
38 ( 6
46 ( 5
57 ( 5
39 ( 7
c- C5H9
CH2(C6H4)-m-I
H
c- C5H9
CH2(C6H4)-m-I
H
c-C5H9
CH2(C6H4)-m-I
H
c-C5H9
CH2(C6H4)-m-I
H
c-C5H9
CH2(C6H4)-m-I
H
31 ( 0
24 ( 1
-3 ( 1
22 ( 2
20 ( 1
-7 ( 3
10 ( 1
8 ( 2
33
34
35
c-C5H9
CH2(C6H4)-m-I
-4 ( 2
a
Agonistic activity of N,2,5′-trisubstituted adenosine analogues expressed as the amount of [35S]GTPγS bound (as a percentage of the
amount of [35S]GTPγS bound by 1 µM CPA) via the rat adenosine A1 receptor ((SEM, n ) 3). Antagonistic activity of N,2,5′-trisubstituted
b
adenosine analogues expressed as the amount of [35S]GTPγS bound in the presence of 1 µM CPA (as a percentage of the amount of
[
35S]GTPγS bound by 1 µM CPA) via the rat adenosine A1 receptor ((SEM, n ) 3). c Agonistic activity of N,5′-di and N,2,5′-trisubstituted
adenosine analogues expressed as the amount of [35S]GTPγS bound (as a percentage of the amount of [35S]GTPγS bound by 1 µM NECA)
d
via the human adenosine A3 receptor ((SEM, n ) 3). Antagonistic activity of N,2,5′-trisubstituted adenosine analogues expressed as
the amount of [35S]GTPγS bound in the presence of 1 µM NECA (as a percentage of the amount of [35S]GTPγS bound by 1 µM NECA) via
the human adenosine A3 receptor. ((SEM, n ) 3).
activities and behaved as partial agonists for the ad-
enosine A1 receptor in this assay. Here, the effect of the
chlorine at the 2-position on the intrinsic activity is
ambiguous.
the large 5′-O-cyclopropyl substituent induced high
affinity, similar to the 5′-O-methyl substituent, but
reduced the intrinsic activity for the adenosine A3
receptor more than the 5′-O-methyl substituent. The
compounds (concentration at 100Ki) were also evaluated
for their antagonistic effects, an important feature of
partial agonists. This evaluation was done by testing
the partial agonists in combination with a reference full
agonist. All compounds showed partially antagonistic
properties for either the adenosine A1 receptor or the
adenosine A3 receptor in the [35S]GTPγS binding assay
in the presence of 1 µM CPA or 1 µM NECA, respec-
tively, by reducing the amount of [35S]GTPγS binding
caused by CPA/NECA to a varying extent (Table 3). In
almost all cases the amount of [35S]GTPγS bound was
less than with CPA/NECA alone, but always higher
than that tested alone with the partial agonists. This
indicates that partial agonists and full agonists have
the same binding site at the receptor.
Similar effects at the adenosine A1 receptor were
observed in a [35S]GTPγS assay (Table 3), in the
presence of these partial agonists, on the level of the G
protein activity. The effect on the intrinsic activity was
more pronounced than in the cAMP assay, with in
general much lower intrinsic activities of the compounds
for the adenosine A1 receptor (Table 3). It must be noted
that in the cAMP assay human adenosine A1 receptors
were used, whereas the studies with [35S]GTPγS were
performed with rat A1 receptors. All compounds behaved
as partial agonists for the adenosine A1 receptor, with
the N-(3-iodobenzyl)-substituted derivatives having the
lowest intrinsic activities. Furthermore, the larger the
5′-substituent, the lower the intrinsic activity, and the
introduction of a 2-chloro substituent seemed to reduce
the intrinsic activity of the compounds for the adenosine
A1 receptor even further. As for the adenosine A1
receptor, the intrinsic activities of the compounds for
the adenosine A3 receptor were quite similar in both
functional assays (Tables 2 and 3). Here too, the
intrinsic activities were lower in the [35S]GTPγS binding
assay (Table 3), and all compounds behaved as partial
agonists for the adenosine A3 receptor as well. Again,
compounds with an unsubstituted 6-amino group had
the highest intrinsic activities, whereas the N-(3-iodo-
benzyl)-substituted derivatives had the lowest. The
large 5′-O-cyclopropyl substituent and the introduction
of chlorine at the 2-position induced low intrinsic
activities for the adenosine A3 receptor, similar to those
for the A1 receptor. Thus, for the adenosine A3 receptor,
Con clu sion s
1,2,3-Tri-O-acetyl-5-O-alkyl-â-D-ribofuranoses were
coupled to an appropriate heterocyclic base, e.g., 6-chloro-
purine or 2,6-dichloropurine, to yield the corresponding
protected nucleosides. Subsequent amination at the
6-position and deprotection of these intermediates gave
the novel 5′-O-alkylated nucleosides 18-35 in good
overall yields. The compounds with an unsubstituted
6-amino group displayed higher adenosine A3 receptor
affinities than those for the A1 receptor. The N-cyclo-
pentyl-substituted derivatives had the highest adeno-
sine A1 receptor affinity (compound 22 with a Ki value
of 16 nM), whereas the N-(3-iodobenzyl)-substituted
derivatives had the highest affinity for the adenosine

5′-O-Alkyl Ethers of Adenosine Derivatives
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 18 2971
Syn th esis of Meth yl 2,3-O-Isopr opylid en e-â-D-r ibofu r -
a n osid e (2). A suspension of 88 g (0.59 mol) of dry D-ribose
in acetone (400 mL), dimethoxypropane (176 mL), and MeOH
(380 mL) was stirred and cooled in an ice bath. Over a period
of 4 h, every 30 min HCl(g) was led through the solution, and
then the mixture was stirred overnight at room temperature.
The mixture was neutralized with pyridine and concentrated
in vacuo. The residue was extracted with H2O and ether. The
organic layers were combined, dried (MgSO4), and concen-
trated: yield 115 g (0.56 mol, 95%); Rf 0.38 (PE40/60-EtOAc,
A3 receptor (compound 26 with a Ki value of 3 nM). The
adenosine A1 receptor affinity was either not affected
or slightly increased with the introduction of chlorine
at the 2-position, while the A3 receptor affinity was
affected by the 2-chlorine, depending on the N-substitu-
ent. In general, chlorine at the 2-position slightly
decreased the A1:A3 selectivity ratio. Furthermore, a 5′-
O-methyl substituent induced the highest adenosine A1
receptor affinity, whereas a 5′-O-ethyl substituent did
so for the A3 receptor.
All compounds were partial agonists in both the cAMP
and [35S]GTPγS assays, although more pronounced in
the latter. The N-(3-iodobenzyl)-substituted derivatives
displayed the lowest intrinsic activities for both the
adenosine A1 and A3 receptors. The introduction of
2-chlorine reduced the intrinsic activity for both receptor
subtypes as well. Hence, although the N-(3-iodobenzyl)
substituent induced the highest affinity for the adeno-
sine A3 receptor, it also reduced the intrinsic activity.
Furthermore, contrary to NECA and MECA, the A3
receptor appeared to accommodate the larger 5′-O-ethyl
and even 5′-O-cyclopropyl better than the smaller 5′-
O-methyl group. This is unlike data for the 5′-carbox-
amido derivatives, in which MECA is more active than
NECA. Increasing the size of the 5′-substituent also
reduced the intrinsic activity for both the adenosine A1
and A3 receptors. Thus, these newly synthesized partial
agonists for both the adenosine A1 and A3 receptors may
be useful tools in further research on adenosine recep-
tors. Compounds 20 and 26 are particularly interesting
in view of their higher affinity for the adenosine A3
receptor than the reference compounds IBMECA and
2-Cl-IBMECA.
1
1:1); H NMR (CDCl3) δ 4.97 (s, 1H, H-1), 4.82-4.57 (2 × d,
2H, J ) 6.18 Hz, H-2,3), 4.42-4.39 (m, 1H, H-4), 3.71-3.62
(m, 2H, H-5), 3.41 (s, 3H, OCH3), 1.48, 1.31 (2 × s, 6H, 2 ×
CCH3) ppm; 13C NMR (CDCl3) δ 112.02 (C(CH3)2), 109.8 (C-
1), 88.13, 85.64, 81.38 (C-2,3,4), 63.80 (C-5), 55.24 (OCH3),
26.24, 24.61 (2 × CH3) ppm.
Gen er a l P r oced u r e for th e Alk yla tion of Com p ou n d
2 to th e Meth yl 5-O-Alk yl-2,3-O-isop r op ylid en e-â-D-r ibo-
fu r a n osid e Der iva tives 3-5. Methyl 2,3-O-isopropylidene-
â-D-ribofuranoside (2; 53.0 g, 0.26 mol) was dissolved in dry
dimethylformamide (DMF; 300 mL). This was cooled (0 °C),
and NaH (60% in mineral oil, 11.5 g, 0.29 mol) was slowly
added. The mixture was allowed to warm to room temperature
and cooled again, and the appropriate alkyl halide (0.31 mol)
was added very slowly. The mixture was stirred at room
temperature overnight. The mixture was treated with MeOH
(100 mL) and concentrated in vacuo. It was coevaporated with
toluene (2×). The (black) mixture was extracted with water
and EtOAc (250 mL each). The water layer was subsequently
extracted with CH2Cl2. The organic layers were combined,
dried (MgSO4), and concentrated. The residue was purified by
column chromatography.
Met h yl 5-O-Met h yl-2,3-O-isop r op ylid en e-â-D-r ib ofu r -
a n osid e (3). The reaction was carried out with 2 (53.0 g, 0.26
mol) and iodomethane (CH3I; 0.31 mol, 19.4 mL). The mixture
was purified by column chromatography (eluent gradient
PE40/60, PE40/60-EtOAc, 2:1): yield 55.1 g (0.25 mol, 97%);
Rf 0.56 (PE40/60-EtOAc, 1:1); 13C NMR (CDCl3) δ 111.71
(C(CH3)2), 108.79 (C-1), 84.76, 84.47, 81.70 (C-2,3,4), 73.18 (C-
5), 58.38 (OCH3), 54.03 (CH2OCH3), 25.98, 24.49 (2 × C(CH3)2)
ppm.
Exp er im en ta l Section
Met h yl 5-O-E t h yl-2,3-O-isop r op ylid en e-â-D-r ib ofu r -
a n osid e (4). The reaction was carried out with 2 (33.4 g, 0.16
mol) and iodoethane (0.20 mol, 15.9 mL). The mixture was
purified by column chromatography (eluent gradient PE40/
60, PE40/60-EtOAc, 1:1): yield 35.2 g (0.15 mol, 92%); Rf 0.62
(PE40/60-EtOAc, 1:1); 13C NMR (CDCl3) δ 111.41 (C(CH3)2),
108.55 (C-1), 84.50, 81.55 (C-2,3,4), 70.78 (CH2CH3), 65.91 (C-
5), 53.85 (OCH3), 25.77, 24.31 (C(CH3)2), 14.48 (CH2CH3) ppm.
Meth yl 5-O-Cyclop r op yl-2,3-O-isop r op ylid en e-â-D-r i-
bofu r a n osid e (5). The reaction was carried out with 2 (43.6
g, 0.21 mol) and cyclopropyl bromide (0.26 mol, 20.6 mL). Prior
to treatment with MeOH, the mixture was stirred at room
temperature. The mixture was purified by column chroma-
tography (eluent gradient PE40/60, PE40/60-EtOAc, 1:1):
yield 19.5 g (79.8 mmol, 38%); Rf 0.70 (PE40/60-EtOAc, 1:1);
13C NMR (CDCl3) δ 111.94 (C(CH3)2), 108.90 (C-1), 84.88, 84.65,
81.85 (C-2,3,4), 71.19 (C-5), 54.38 (OCH), 53.21 (OCH3), 26.15,
24.72 (C(CH3)2), 5.37 (CH2CH2) ppm.
Ch em ica ls a n d Solven ts. Guanosine was obtained from
Aldrich (Aldrich Chemie, Sigma-Aldrich Chemie BV, Zwijn-
drecht, The Netherlands). All other reagents were from
standard commercial sources and of analytical grade. [3H]-
DPCPX, [3H]ZM 241385, and [125I]ABMECA were purchased
from NEN (Hoofddorp, The Netherlands).
Ch r om a togr a p h y. Thin-layer chromatography (TLC) was
carried out using aluminum sheets (20 × 20 cm) with silica
gel F254 from Merck. Spots were visualized under UV light (254
nm). Preparative column chromatography was performed on
silica gel (230-400 mesh, ASTM).
In str u m en ts a n d An a lyses. Elemental analyses were
performed for C, H, and N (Department of Analytical Chem-
istry, Leiden University, The Netherlands). 13C NMR spectra
were measured at 50.1 MHz with a J EOL J NM-FX 200
spectrometer equipped with a PG 200 computer operating in
the Fourier transform mode. 1H NMR spectra were measured
at 200 MHz, using the above-mentioned spectrometer, or at
300 MHz, using a Bruker WM-300 spectrometer equipped with
an ASPECT-2000 computer operating in the Fourier transform
mode. Chemical shifts for 1H and 13C NMR are given in parts
per million (δ) relative to tetramethylsilane (TMS) as internal
standard.
Gen er a l P r oced u r e for th e Dep r otection of Com -
p ou n d s 3-5 to th e 5-O-Alk yl-â-D-r ibofu r a n ose Der iva -
tives 6-8. The appropriate methyl 5-O-alkyl-2,3-O-isopro-
pylidene-â-D-ribofuranose (4.48 mmol) was dissolved in 15 mL
of HCl (0.04 M) and was refluxed for 2 h. The solution was
neutralized with BaCO3, filtered, and concentrated. The
mixture was purified by column chromatography.
All high-resolution mass spectra were measured on
a
Finnigan MAT900 mass spectrometer equipped with a direct
insertion probe for EI experiments (70 eV with resolution 1000)
or on a Finnigan MAT TSQ-70 spectrometer equipped with
an electrospray interface for ESI experiments. Spectra were
collected by constant infusion of the analyte dissolved in 80:
20 methanol/H2O. ESI is a soft ionization technique resulting
in protonated, sodiated species in positive ionization mode and
deprotonated species in negative ionization mode.
5-O-Meth yl-r,â-D-r ibofu r a n ose (6). The reaction was
carried out with methyl 5-O-methyl-2,3-O-isopropylidene-â-
D-ribofuranoside (3; 1 g, 4.58 mmol). The mixture was purified
by column chromatography (eluent 10% MeOH in EtOAc):
yield 0.44 g (2.68 mmol, 59%); Rf 0.34 (10% MeOH in EtOAc);
13C NMR (MeOD) δ 103.12 (C-1, â), 97.78 (C-1, R), 83.94, 82.36,
77.37, 76.88, 72.61, 72.12 (C-2,3,4, R + â), 73.93 (C-5), 59.51,
59.39 (OCH3, R + â) ppm.
Melting points (not corrected) were determined in a Bu¨chi
capillary melting point apparatus.
5-O-Eth yl-r,â-D-r ibofu r a n ose (7). The reaction was car-
ried out with methyl 5-O-ethyl-2,3-O-isopropylidene-â-D-ribo-

2972 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 18
van Tilburg et al.
1
furanoside (4; 1.0 g, 4.85 mmol). The mixture was purified by
column chromatography (10% MeOH in EtOAc): yield 0.56 g
(3.14 mmol, 65%); Rf 0.40 (10% MeOH in EtOAc); 13C NMR
(MeOD) δ 102.88 (C-1, â), 97.51 (C-1, R), 84.09, 82.42, 76.60,
72.53, 72.03 (C-2,3,4, R + â), 73.40 (C-5), 67.83 (CH2CH3), 15.34
(CH3) ppm.
5-O-Cyclop r op yl-r,â-D-r ibofu r a n ose (8). The reaction
was carried out with methyl 5-O-cyclopropyl-2,3-O-isopro-
pylidene-â-D-ribofuranoside (5; 4.4 g, 18 mmol). The mixture
was purified by column chromatography (10% MeOH in
EtOAc): yield 2.09 g (11 mmol, 61%); Rf 0.43 (10% MeOH in
EtOAc); 13C NMR (MeOD) δ 101.60 (C-1, â), 96.41 (C-1, R),
80.77, 79.42, 74.40, 73.53, 72.03 (C-2,3,4, R + â), 69.40 (C-5),
56.97 (CH) 7.83 (CH2CH2) ppm.
Gen er a l P r oced u r e for th e Acyla tion of Com p ou n d s
6-8 to th e 1,2,3-Tr i-O-a cetyl-5-O-a lk yl-r,â-D-r ibofu r a n ose
Der iva tives 9-11. The appropriate 5-O-alkyl-R,â-D-ribofura-
nose (2.68 mmol) was dissolved in 25 mL of pyridine. A
catalytic amount of (dimethylamino)pyridine (DMAP) and
acetic anhydride (8.84 mmol, 843 µL) were added. The mixture
was stirred for 2 h at room temperature, concentrated in vacuo
and coevaporated with toluene. The oil was extracted with
water and EtOAc (25 mL each). The organic layer was dried
(MgSO4), concentrated, and purified by column chromatogra-
phy.
1,2,3-Tr i-O-Acetyl-5-O-m eth yl-r,â-D-r ibofu r a n ose (9).
The reaction was carried out with 5-O-methyl-R,â-D-ribofura-
nose (6; 0.44 g, 2.68 mmol). The mixture was purified by
column chromatography (eluent 10% MeOH in EtOAc): yield
0.70 g (2.41 mmol, 90%); Rf 0.76 (10% MeOH in EtOAc); 13C
NMR (CDCl3) δ 168.98, 168.68, 168.39 (3 × CdO, R + â), 97.64
(C-1, â), 93.64 (C-1, R), 82.81, 80.30, 73.99, 73.67, 70.20, 69.67
(C-2,3,4, R + â), 71.75 (C-5), 58.67 (OCH3), 20.26, 19.70 (3 ×
COCH3) ppm.
Cl2); H NMR (CDCl3) δ 8.75 (s, 1H, H-8), 8.56 (s, 1H, H-2),
6.41 (d, 1H, J ) 6.18 Hz, H-1′), 5.80 (t, 1H, J ) 5.49 Hz, H-2′),
5.60-5.56 (m, 1H, H-3′), 4.39 (q, 1H, J ) 2.41 Hz, H-4′), 3.69
(dq, 2H, J ) 7.21 Hz, J ) 2.40 Hz, H-5′), 3.48 (s, 3H, OCH3),
2.16, 2.03 (2 × s, 6H, 2 × COCH3) ppm.
2,6-Dich lor o-9-(2,3-d i-O-a cetyl-5-O-m eth yl-â-D-r ibofu r -
a n osyl)p u r in e (13). The reaction was carried out with
silylated 2,6-dichloropurine (10.3 mmol) and 9 (2.39 g, 8.24
mmol). The mixture was purified by column chromatography
(eluent gradient 4-6% acetone in CH2Cl2): yield 2.91 g (6.94
mmol, 84%); 1H NMR (CDCl3) δ 8.41 (s, 1H, H-8), 6.18 (d, 1H,
J ) 6.52 Hz, H-1′), 5.60 (dd, 1H, J ) 6.18 Hz, J ) 5.49 Hz,
H-2′), 5.39 (dd, 1H, J ) 5.15 Hz, J ) 2.06 Hz, H-3′), 4.25-
4.23 (m, 1H, H-4′), 3.58-3.48 (m, 2H, H-5′), 3.33 (s, 3H, OCH3),
2.00, 1.88 (2 × s, 6H, 2 × COCH3) ppm.
6-Ch lor o-9-(2,3-di-O-acetyl-5-O-eth yl-â-D-r ibofu r an osyl)-
p u r in e (14). The reaction was carried out with silylated
6-chloropurine (11.6 mmol) and 1,2,3-tri-O-acetyl-5-O-ethyl-
R,â-D-ribofuranose (10; 2.82 g, 9.28 mmol). The mixture was
purified by column chromatography (eluent gradient 4-6%
acetone in CH2Cl2): yield 3.03 g (7.60 mmol, 82%); Rf 0.17 (3%
acetone in CH2Cl2); 1H NMR (CDCl3) δ 8.51 (s, 1H, H-8), 8.45
(s, 1H, H-2), 6.20 (d, 1H, J ) 6.18 Hz, H-1′), 5.64 (t, 1H, J )
5.15 Hz, H-2′), 5.40-5.36 (m, 1H, H-3′), 4.24-4.18 (m, 1H,
H-4′), 3.53 (dq, 2H, J ) 13.73 Hz, J ) 2.06 Hz, H-5′), 3.42 (q,
2H, J ) 6.86 Hz, CH2CH3), 1.94, 1.81 (2 × s, 6H, 2 × COCH3),
1.07 (t, 3H, J ) 6.86 Hz, CH2CH3); 13C NMR (CDCl3) δ 169.33,
168.89 (2 × CdO), 151.70, 151.32, 150.56, 143.26, 131.41 (C-
2,4,5,6,8), 85.46, 82.78, 74.08, 74.86 (C-1′,2′,3′,4′), 69.35, 66.90
(OCH2, C-5′), 20.28, 19.19 (2 × COCH3), 14.71 (CH3) ppm.
2,6-Dich lor o-9-(2,3-d i-O-a cet yl-5-O-et h yl-â-D-r ib ofu r -
a n osyl)p u r in e (15). The reaction was carried out with
silylated 2,6-dichloropurine (10.3 mmol) and 10 (2.51 g, 8.24
mmol). The mixture was purified by column chromatography
(eluent gradient 4-6% acetone in CH2Cl2): yield 2.93 g (6.76
mmol, 82%); 1H NMR (CDCl3) δ 8.66 (s, 1H, H-8), 6.39 (d, 1H,
J ) 6.53 Hz, H-1′), 5.81 (dd, 1H, J ) 6.86 Hz, J ) 5.49 Hz,
H-2′), 5.58 (dd, 1H, J ) 7.21 Hz, J ) 2.06 Hz, H-3′), 4.43-
4.41 (m, 1H, H-4′), 3.75 (dq, 2H, J ) 10.64 Hz, J ) 2.41 Hz,
H-5′), 3.64 (t, 2H, J ) 6.87 Hz, CH2CH3), 2.19, 2.06 (2 × s,
6H, 2 × COCH3), 1.32 (t, 3H, J ) 6.86 Hz, CH2CH3) ppm.
6-Ch lor o-9-(2,3-d i-O-a cet yl-5-O-cyclop r op yl-â-D-r ib o-
fu r a n osyl)p u r in e (16). The reaction was carried out with
silylated 6-chloropurine (9.79 mmol) and 1,2,3-tri-O-acetyl-5-
O-cyclopropyl-R,â-D-ribofuranose (11; 2.48 g, 7.83 mmol). The
mixture was purified by column chromatography (eluent
gradient 4-6% acetone in CH2Cl2): yield 2.09 g (5.10 mmol,
1,2,3-Tr i-O-Acetyl-5-O-eth yl-r,â-D-r ibofu r an ose (10). The
reaction was carried out with 5-O-ethyl-R,â-D-ribofuranose (7;
0.56 g, 3.14 mmol). The mixture was purified by column
chromatography (eluent 10% MeOH in EtOAc): yield 0.85 g
(2.79 mmol, 89%); Rf 0.75 (10% MeOH in EtOAc); 13C NMR
(CDCl3) δ 168.68, 168.42, 168.02 (3 × CdO), 97.43 (C-1, â),
93.40 (C-1, R), 82.69, 80.24, 73.50, 70.08, 69.53 (C-2,3,4, R +
â), 69.38, 68.94 (C-5, R + â), 66.05 (CH2CH3), 19.96, 19.41 (3
× COCH3), 14.18 (CH2CH3) ppm.
1,2,3-T r i-O -Ac e t y l-5-O -c y c lo p r o p y l-r,â-D -r ib o fu r -
a n ose (11). The reaction was carried out with 5-O-cyclopropyl-
R,â-D-ribofuranose (8; 3.42 g, 18.0 mmol). The mixture was
purified by column chromatography (eluent 10% MeOH in
EtOAc): yield 4.74 g (15.0 mmol, 83%); Rf 0.72 (10% MeOH
in EtOAc); 13C NMR (CDCl3) δ 169.27, 168.98, 168.74 (3 × Cd
O), 97.81 (C-1, â), 93.78 (C-1, R), 82.84, 80.36, 73.91, 70.43,
69.88 (C-2,3,4, R + â), 69.79, 69.38 (C-5, R + â), 53.50 (CH),
20.55, 20.02 (3 × COCH3), 5.22 (CH2CH2) ppm.
1
65%); H NMR (CDCl3) δ 8.76 (s, 1H, H-8), 8.59 (s, 1H, H-2),
6.42 (d, 1H, J ) 6.18 Hz, H-1′), 5.83-5.77 (m, 1H, H-2′), 5.57-
5.53 (m, 1H, H-3′), 4.45-4.42 (m, 1H, H-4′), 3.87-3.77 (m, 2H,
H-5′), 3.45-3.42 (m, 1H, CH), 2.19, 2.06 (2 × s, 6H, 2 ×
COCH3), 0.67-0.55 (m, 4H, CH2CH2) ppm.
2,6-Dich lor o-9-(2,3-d i-O-a cetyl-5-O-cyclop r op yl-â-D-r i-
bofu r a n osyl)p u r in e (17). The reaction was carried out with
silylated 2,6-dichloropurine (8.95 mmol) and 11 (2.26 g, 7.16
mmol). The mixture was purified by column chromatography
(eluent gradient 4-6% acetone in CH2Cl2): yield 2.55 g (5.73
mmol, 80%); 1H NMR (CDCl3) δ 8.32 (s, 1H, H-8), 6.10 (d, 1H,
J ) 6.18 Hz, H-1′), 5.52-5.46 (m, 1H, H-2′), 5.28-5.25 (m,
1H, H-3′), 4.20-4.17 (m, 1H, H-4′), 3.60 (q, 2H, J ) 7.56 Hz,
H-5′), 3.24-3.19 (m, 1H, OCH), 1.91, 1.82 (2 × s, 6H, 2 ×
COCH3), 0.43-0.32 (m, 4H, CH2CH2) ppm.
Gen er a l P r oced u r e for th e Am in a tion of Com p ou n d s
12-17 to th e Su bstitu ted Ad en osin e Der iva tives 18-35.
Meth od A (Com p ou n d s 18, 21, 24, 27, 30, a n d 33). The
appropriate 6-chloro-9-(2,3-di-O-acetyl-5-O-alkyl-â-D-ribofura-
nosyl)purine or the appropriate 2,6-dichloro-9-(2,3-di-O-acetyl-
5-O-alkyl-â-D-ribofuranosyl)purine (1.53 mmol) was dissolved
in EtOH/NH3 (30 mL), and the mixture was stirred overnight
at room temperature. The mixture was concentrated and
purified by column chromatography.
Gen er a l P r oced u r e for th e Cou p lin g of Com p ou n d s
9-11 to 6-Ch lor op u r in e or 2,6-Dich lor op u r in e To Give
Com p ou n d s 12-17.
Silyla tion of th e Ba se. The appropriate base (193.8 mg,
1.27 mmol) was treated with 1,1,1,3,3,3-hexamethyldisilazane
(HMDS; 5 mL, 23.7 mmol) and 12.5 µL of chlorotrimethylsilane
(TMSCl; 0.1 mmol) at 130 °C for 20 h. The silylated compound
was concentrated and used without further purification.
Vor br u1 ggen Cou p lin g. To the appropriate silylated base
(12.9 mmol) was added the appropriate ribose (10.3 mmol) in
15 mL of dry 1,2-dichloroethane. The residue was coevaporated
twice with dry 1,2-dichloroethane and subsequently dissolved
in 75 mL of dry 1,2-dichloroethane. The solution was gently
refluxed, and after 5 min TMS triflate (997 µL, 5.16 mmol)
was added. The mixture was refluxed for 2 h, cooled to room
temperature, and diluted with CH2Cl2. It was extracted with
5% NaHCO3 and water. The organic layer was dried (MgSO4),
concentrated, and purified by column chromatography.
6-Ch lor o-9-(2,3-d i-O-a cetyl-5-O-m eth yl-â-D-r ibofu r a n o-
syl)p u r in e (12). The reaction was carried out with silylated
6-chloropurine (12.9 mmol) and 1,2,3-tri-O-acetyl-5-O-methyl-
R,â-D-ribofuranose (9; 3.0 g, 10.3 mmol). The mixture was
purified by column chromatography (eluent 3% acetone in CH2-
Cl2): yield 1.9 g (4.94 mmol, 48%); Rf 0.11 (3% acetone in CH2-
Meth od B (Com p ou n d s 19, 20, 22, 23, 25, 26, 28, 29, 31,
32, 34, a n d 35). The appropriate 6-chloro-9-(2,3-di-O-acetyl-
5-O-alkyl-â-D-ribofuranosyl)purine or the appropriate 2,6-
dichloro-9-(2,3-di-O-acetyl-5-O-alkyl-â-D-ribofuranosyl)pur-
ine (1.53 mmol) was dissolved in EtOH absolute (10 mL). The

5′-O-Alkyl Ethers of Adenosine Derivatives
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 18 2973
appropriate amine (2.3 mmol) and Et3N (1.91 mmol) were
added, and the mixture was refluxed overnight. The mixture
was concentrated, dissolved in EtOH/NH3 (30 mL), and stirred
overnight at room temperature. The mixture was concentrated
again and purified by column chromatography.
8.89 (br s, 1H, NH), 8.35 (s, 1H, H-8), 7.73 (s, 1H, CCHCI),
7.60 (d, 1H, J ) 5.83 Hz, CCHCHCH), 7.33 (d, 1H, J ) 6.12
Hz, CCHCH), 7.12 (t, 1H, J ) 7.55 Hz, CCHCH), 5.82 (d, 1H,
J ) 5.15 Hz, H-1′), 5.54 (d, 1H, J ) 5.84 Hz, OH-2′), 5.30 (d,
1H, J ) 5.49 Hz, OH-3′), 4.61-4.51 (m, 3H, NHCH2, H-2′),
4.09-4.00 (m, 2H, H-3′,4′), 3.56-3.52 (m, 2H, H-5′), 3.28 (s,
3H, OCH3); MS m/ z 532 (M + H)+. Anal. (C18H19ClIN5O4‚
0.3CH3COCH3) C, H, N.
5′-O-Eth yla d en osin e (24). Method A. The reaction was
carried out with 6-chloro-9-(2,3-di-O-acetyl-5-O-ethyl-â-D-ri-
bofuranosyl)purine (14; 663 mg, 1.70 mmol). The mixture was
purified by column chromatography (eluent 5% MeOH in CH2-
Cl2): yield 407 mg (1.38 mmol, 81%); mp 110-112 °C; Rf 0.48
(10% MeOH in CH2Cl2); 1H NMR (DMSO-d6) δ 8.32 (s, 1H,
H-8), 8.14 (s, 1H, H-2), 7.27 (br s, 2H, NH2), 5.89 (d, 1H, J )
4.80 Hz, H-1′), 5.49 (d, 1H, J ) 5.84 Hz, OH-2′), 5.24 (d, 1H,
J ) 5.15 Hz, OH-3′), 4.54 (q, 1H, J ) 5.14 Hz, H-2′), 4.15 (q,
1H, J ) 4.81 Hz, H-3′), 4.00 (q, 1H, J ) 4.46 Hz, H-4′), 3.58
(dq, 2H, J ) 9.27 Hz, J ) 3.78 Hz, H-5′), 3.47 q, 2H, J ) 7.21
Hz, CH2), 1.12 (t, 3H, J ) 7.20 Hz, CH3); MS m/ z 296 (M +
H)+. Anal. (C12H17N5O4‚0.2CH2Cl2) C, H, N.
N-Cyclop en tyl-5′-O-eth yla d en osin e (25). Method B. The
reaction was carried out with 14 (502 mg, 1.26 mmol) and
cyclopentylamine (1.89 mmol, 187 µL). The mixture was
purified by column chromatography (eluent 5% MeOH in CH2-
Cl2): yield 316 mg (0.87 mmol, 69%); mp 134-136 °C; Rf 0.49
(eluent 10% MeOH in CH2Cl2). The product was recrystallized
from CH3CN: 1H NMR (DMSO-d6) δ 8.31 (s, 1H, H-8), 8.19
(s, 1H, H-2), 7.72 (d, 1H, J ) 7.55 Hz, NH), 5.89 (d, 1H, J )
5.15 Hz, H-1′), 5.55-5.25 (m, 2H, OH-2′,3′), 4.53 (t, 1H, J )
4.81 Hz, H-2′), 4.15 (t, 1H, J ) 2.40 Hz, H-3′), 4.01-3.96 (m,
1H, H-4′), 3.62-3.48 (m, 2H, H-5′), 3.46 (q, 2H, J ) 7.21 Hz,
CH2CH3), 1.99-1.82 (m, 2H, cyclopentyl), 1.78-1.52 (m, 4H,
cyclopentyl), 1.11 (t, 3H, J ) 7.21 Hz, CH3); MS m/ z 364 (M
+ H)+. Anal. (C17H25N5O4·0.6CH3CN) C, H, N.
5′-O-Meth yla d en osin e (18). Method A. The reaction was
carried out with 6-chloro-9-(2,3-di-O-acetyl-5-O-methyl-â-d -
ribofuranosyl)purine (12; 682 mg, 1.77 mmol). The mixture
was purified by column chromatography (eluent 5% MeOH in
CH2Cl2): yield 393 mg (1.40 mmol, 79%); mp 112-114 °C; Rf
1
0.43 (10% MeOH in CH2Cl2); H NMR (DMSO-d6) δ 8.28 (s,
1H, H-8), 8.14 (s, 1H, H-2), 7.26 (br s, 2H, NH2), 5.87 (d, 1H,
J ) 5.49 Hz, OH-2′), 5.25 (d, 1H, J ) 5.15 Hz, OH-3′), 4.57 (q,
1H, J ) 6.18 Hz, H-2′), 4.13 (q, 1H, J ) 4.47 Hz, H-3′), 3.99-
3.98 (m, 1H, H-4′), 3.57-3.48 (m, 2H, H-5′), 3.28 (s, 3H, OCH3)
ppm; MS m/ z 282 (M + H)+. Anal. (C11H15N5O4‚0.5H2O) C,
H, N.
N-Cyclop en tyl-5′-O-m eth yla d en osin e (19). Method B.
The reaction was carried out with 12 (589 mg, 1.53 mmol) and
cyclopentylamine (2.3 mmol, 227 µL). The mixture was purified
by column chromatography (eluent 5% MeOH in CH2Cl2):
yield 476 mg (1.36 mmol, 89%); mp 164-166 °C; Rf 0.51 (eluent
10% MeOH in CH2Cl2). The product was recrystallized from
CH3CN: 1H NMR (DMSO-d6) δ 8.28 (s, 1H, H-8), 8.18 (s, 1H,
H-2), 7.70 (d, 1H, J ) 7.55 Hz, NH), 5.89 (d, 1H, J ) 4.80 Hz,
H-1′), 5.50 (d, 1H, J ) 5.15 Hz, OH-2′), 5.26 (d, 1H, J ) 5.15
Hz, OH-3′), 4.56 (q, 1H, J ) 4.12 Hz, H-2′), 4.13 (q, 1H, J )
4.46 Hz, H-3′), 3.99 (q, 1H, J ) 4.12 Hz, H-4′), 3.57-3.42 (m,
2H, H-5′), 3.28 (s, 3H, OCH3), 2.01-1.83 (m, 2H, cyclopentyl),
1.73-1.52 (m, 4H, cyclopentyl); MS m/ z 350 (M + H)+. Anal.
(C16H23N5O4‚0.7CH3CN) C, H, N.
N-(3-Iod oben zyl)-5′-O-m eth yla d en osin e (20). Method B.
The reaction was carried out with 12 (363 mg, 0.94 mmol) and
(3-iodobenzyl)amine hydrochloride (1.41 mmol, 380 mg). The
mixture was purified by column chromatography (eluent 5%
MeOH in CH2Cl2): yield 397 mg (0.80 mmol, 85%); mp 155-
157 °C; Rf 0.54 (eluent 10% MeOH in CH2Cl2); 1H NMR
(DMSO-d6) δ 8.40 (br s, 1H, NH), 8.33 (s, 1H, H-8), 8.21 (s,
1H, H-2), 7.70 (s, 1H, CCHCI), 7.57 (d, 1H, J ) 6.52 Hz,
CCHCHCH), 7.34 (d, 1H, J ) 5.15 Hz, CCHCH), 7.08 (t, 1H,
J ) 8.93 Hz, CCHCH), 5.89 (d, 1H, J ) 4.12 Hz, H-1′), 5.50-
5.47 (m, 1H, OH-2′), 5.27-5.25 (m, 1H, OH-3′), 4.68-4.56 (m,
3H, H-2′, NHCH2), 4.14-4.13 (m, 1H, H-3′), 4.01-3.98 (m, 1H,
H-4′), 3.57-3.53 (m, 2H, H-5′), 3.31 (s, 3H, OCH3); MS m/ z
498 (M + H)+. Anal. (C18H20N5O4) C, H, N.
2-Ch lor o-5′-O-m eth yla d en osin e (21). Method A. The
reaction was carried out with 2,6-dichloro-9-(2,3-di-O-acetyl-
5-O-methyl-â-D-ribofuranosyl)purine (13; 667 mg, 1.59 mmol).
The mixture was purified by column chromatography (eluent
5% MeOH in CH2Cl2): yield 382 mg (1.21 mmol, 76%); mp
200-202 °C; Rf 0.49 (10% MeOH in CH2Cl2); 1H NMR (DMSO-
d6) δ 8.30 (s, 1H, H-8), 7.83 (br s, 2H, NH2), 5.80 (d, 1H, J )
5.15 Hz, H-1′), 5.53 (d, 1H, J ) 6.18 Hz, OH-2′), 5.29 (d, 1H,
J ) 5.49 Hz, OH-3′), 4.52 (q, 1H, J ) 5.84 Hz, H-2′), 4.11-
3.99 (m, 2H, H-3′,4′), 3.55-3.46 (m, 2H, H-5′), 3.28 (s, 3H,
OCH3); MS m/ z 316 (M + H)+. Anal. (C11H14ClN5O4‚1.0CH2-
Cl2) C, H, N.
2-Ch lor o-N-cyclop en t yl-5′-O-m et h yla d en osin e (22).
Method B. The reaction was carried out with 13 (505 mg, 1.2
mmol) and cyclopentylamine (1.8 mmol, 178 µL). The mixture
was purified by column chromatography (eluent 2% MeOH in
CH2Cl2): yield 364 mg (0.95 mmol, 79%); mp 124-126 °C; Rf
0.15 (eluent 2% MeOH in CH2Cl2); 1H NMR (DMSO-d6) δ 8.33
(br s, 1H, NH), 8.30 (s, 1H, H-8), 5.81 (d, 1H, J ) 4.46 Hz,
H-1′), 5.51 (d, 1H, J ) 3.36 Hz, OH-2′), 5.29 (d, 1H, J ) 3.32
Hz, OH-3′), 4.59-4.29 (m, 2H, CH, H-2′), 4.41-3.99 (m, 2H,
H-3′,4′), 3.56-3.52 (m, 2H, H-5′), 1.94-1.92 (m, 2H, cyclopen-
tyl), 1.71-1.50 (m, 4H, cyclopentyl); MS m/ z 384 (M + H)+.
Anal. (C16H22ClN5O4‚0.1CH2Cl2) C, H, N.
5′-O-Eth yl-N-(3-iod oben zyl)a d en osin e (26). Method B.
The reaction was carried out with 14 (367 mg, 0.92 mmol) and
(3-iodobenzyl)amine hydrochloride (1.38 mmol, 372 mg). The
mixture was purified by column chromatography (eluent 5%
MeOH in CH2Cl2): yield 339 mg (0.66 mmol, 72%); mp 164-
166 °C; Rf 0.40 (10% MeOH in CH2Cl2). The product was
recrystallized from CH3CN: 1H NMR (DMSO-d6) δ 8.45 (br s,
1H, NH), 8.37 (s, 1H, H-8), 8.21 (s, 1H, H-2), 7.71 (s, 1H,
CCHCI), 7.57 (d, 1H, J ) 7.90 Hz, CCHCHCH), 7.34 (d, 1H, J
) 6.52 Hz, CCHCH), 7.09 (t, 1H, J ) 6.87 Hz, CCHCH), 5.91
(d, 1H, J ) 4.46 Hz, H-1′), 5.53 (d, 1H, J ) 5.49 Hz, OH-2′),
5.27 (d, 1H, J ) 4.80 Hz, OH-3′), 4.67-4.64 (m, 2H, NHCH2),
4.56 (q, 1H, J ) 4.81 Hz, H-2′), 4.17 (q, 1H, J ) 4.81 Hz, H-3′),
4.00 (q, 1H, J ) 4.12 Hz, H-4′), 3.60 (dq, 2H, J ) 9.61 Hz, J )
3.43 Hz, H-5′), 3.47 (q, 2H, J ) 6.52 Hz, OCH2CH3), 1.12 (t,
3H, J ) 6.52 Hz, CH3); MS m/ z 512 (M + H)+. Anal. (C19H22
IN5O4‚0.3CH3CN) C, H, N.
-
2-Ch lor o-5′-O-eth yla d en osin e (27). Method A. The reac-
tion was carried out with 2,6-dichloro-9-(2,3-di-O-acetyl-5-O-
ethyl-â-D-ribofuranosyl)purine (15; 656 mg, 1.51 mmol). The
mixture was purified by column chromatography (eluent 5%
MeOH in CH2Cl2): yield 388 mg (1.18 mmol, 78%); mp 117-
1
119 °C; Rf 0.50 (10% MeOH in CH2Cl2); H NMR (DMSO-d6)
δ 8.34 (s, 1H, H-8), 7.83 (br s, 2H, NH2), 5.81 (d, 1H, J ) 5.49
Hz, H-1′), 5.53 (d, 1H, J ) 5.83 Hz, OH-2′), 5.28 (d, 1H, J )
5.14 Hz, OH-3′), 4.49 (q, 1H, J ) 4.81 Hz, H-2′), 4.13-4.01
(m, 1H, H-3′), 4.01-3.99 (m, 1H, H-4′), 3.62-3.42 (m, 2H, H-5′),
3.47 (q, 2H, J ) 6.87 Hz, OCH2), 1.12 (t, 3H, J ) 6.87 Hz,
CH3); MS m/ z 330 (M + H)+. Anal. (C12H16ClN5O4‚0.4CH2Cl2)
C, H, N.
2-Ch lor o-N-cyclopen tyl-5′-O-eth yladen osin e (28). Method
B. The reaction was carried out with 15 (505 mg, 1.16 mmol)
and cyclopentylamine (1.74 mmol, 172 µL). The mixture was
purified by column chromatography (eluent 5% MeOH in CH2-
Cl2): yield 323 mg (0.81 mmol, 70%); mp 114-116 °C; Rf 0.55
(10% MeOH in CH2Cl2); 1H NMR (DMSO-d6) δ 8.34 (s, 1H,
H-8), 8.32 (br s, 1H, NH), 5.82 (d, 1H, J ) 5.15 Hz, H-1′), 5.53
(d, 1H, J ) 5.18 Hz, OH-2′), 5.29 (d, 1H, J ) 5.15 Hz, OH-3′),
4.47-4.37 (m, 2H, CH, H-2′), 4.11-4.10 (m, 1H, H-3′), 4.00
(q, 1H, J ) 4.47 Hz, H-4′), 3.62-3.58 (m, 2H, H-5′), 3.47 (q,
2H, J ) 7.21 Hz, CH2CH3), 2.00-1.83 (m, 2H, cyclopentyl),
2-Ch lor o-N-(3-iod oben zyl)-5′-O-m eth yla d en osin e (23).
Method B. The reaction was carried out with 13 (372 mg, 0.89
mmol) and 3-iodobenzylamine hydrochloride (1.34 mmol, 360
mg). The mixture was purified by column chromatography
(eluent 2% MeOH in CH2Cl2): yield 383 mg (0.72 mmol, 81%);
mp 84-86 °C; Rf 0.59 (10% MeOH in CH2Cl2). The product
was recrystallized from CH3COCH3: 1H NMR (DMSO-d6) δ

2974 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 18
van Tilburg et al.
1.71-1.51 (m, 4H, cyclopentyl), 1.12 (t, 3H, J ) 7.21 Hz, CH3);
MS m/ z 399 (M + H)+. Anal. (C17H24ClN5O4) C, H, N.
2-Ch lor o-5′-O-et h yl-N-(3-iod ob en zyl)a d en osin e (29).
Method B. The reaction was carried out with 15 (375 mg, 0.87
mmol) and (3-iodobenzyl)amine hydrochloride (1.31 mmol, 352
mg). The mixture was purified by column chromatography
(eluent 5% MeOH in CH2Cl2): yield 366 mg (0.67 mmol, 77%);
mp 80-82 °C; Rf 0.53 (10% MeOH in CH2Cl2); 1H NMR
(DMSO-d6) δ 8.97 (br s, 1H, NH),8.40 (s, 1H, H-8), 7.74 (s, 1H,
CCHCI), 7.60 (d, 1H, J ) 6.87 Hz, CCHCHCH), 7.35 (d, 1H, J
) 8.24 Hz, CCHCH), 7.12 (t, 1H, J ) 7.21 Hz, CCHCH), 5.84
(d, 1H, J ) 4.46 Hz, H-1′), 5.56 (d, 1H, J ) 5.49 Hz, OH-2′),
5.30 (d, 1H, J ) 5.15 Hz, OH-3′), 4.60 (d, 2H, J ) 4.46 Hz,
NHCH2), 4.51 (d, 1H, J ) 5.15 Hz, H-2′), 4.14-4.11 (m, 1H,
H-3′), 4.02 (q, 1H, J ) 3.43 Hz, H-4′), 3.63-3.58 (m, 2H, H-5′),
3.49 (q, 2H, J ) 6.52 Hz, CH2CH3), 1.13 (t, 3H, J ) 6.52 Hz,
CH3); MS m/ z 546 (M + H)+. Anal. (C19H21ClIN5O4‚0.8HCON-
(CH3)2) C, H, N.
0.51 (10% MeOH in CH2Cl2). The product was recrystallized
from (C2H5)2O: 1H NMR (DMSO-d6) δ 8.34 (br s, 1H, NH), 8.28
(s, 1H, H-8), 5.81 (d, 1H, J ) 5.15 Hz, H-1′), 5.53 (d, 1H, J )
5.84 Hz, OH-2′), 5.29 (d, 1H, J ) 5.15 Hz, OH-3′), 4.50 (q, 1H,
J ) 5.15 Hz, H-2′), 4.50-4.47 (m, 1H, CH), 4.09-3.99 (m, 2H,
H-3′,4′), 3.68-3.63 (m, 2H, H-5′), 1.94-1.89 (m, 2H, cyclopen-
tyl), 1.71-1.50 (m, 4H, cyclopentyl), 0.46-0.41 (m, 4H,
OCHCH2CH2); MS m/ z 410 (M + H)+. Anal. (C18H24ClN5O4‚
0.2(C2H5)2O) C, H, N.
2-Ch lor o-5′-O-cyclop r op yl-N -(3-iod ob e n zyl)a d e n o-
sin e (35). Method B. The reaction was carried out with 17
(483 mg, 1.08 mmol) and (3-iodobenzyl)amine hydrochloride
(1.63 mmol, 439 mg). The mixture was purified by column
chromatography (eluent 5% MeOH in CH2Cl2): yield 435 mg
(0.78 mmol, 72%); mp 94-96 °C; Rf 0.49 (10% MeOH in CH2-
Cl2); 1H NMR (DMSO-d6) δ 8.93 (t, 1H, J ) 6.18 Hz, NH), 8.33
(s, 1H, H-8), 7.73 (s, 1H, CCHCI), 7.59 (d, 1H, J ) 7.90 Hz,
CCHCHCH), 7.34 (d, 1H, J ) 7.55 Hz, CCHCH), 7.11 (t, 1H,
J ) 7.55 Hz, CCHCH), 5.82 (d, 1H, J ) 5.14 Hz, H-1′), 5.54
(d, 1H, J ) 5.84 Hz, OH-2′), 5.30 (d, 1H, J ) 5.15 Hz, OH-3′),
4.61-4.48 (m, 3H, H-2′, NHCH2), 4.09-3.98 (m, 2H, H-3′,4′),
3.68-3.58 (m, 2H, H-5′), 0.45-0.41 (m, 4H, cyclopropyl); MS
m/ z 559 (M + H)+. Anal. (C20H21ClIN5O4‚0.5H2O) C, H, N.
Ra d ioliga n d Bin d in g Stu d ies. Measurements with [3H]-
DPCPX in the absence of GTP were performed according to a
protocol published previously (rat A1).32 Adenosine A2A receptor
(rat) affinities were determined according to Gao et al.33
Adenosine A3 receptor affinities were determined essentially
as described.34,35 Briefly, assays were performed in 50:10:1
buffer (50 mM Tris/10 mM MgCl2/1 mM ethylenediaminetetra-
acetic acid (EDTA) and 0.01% 3-([3-cholamidopropyl]dimethyl-
ammonio)-1-propanesulfonate (CHAPS)) in glass tubes and
contained 50 µL of a HEK 293 cell membrane suspension (10-
30 µg), 25 µL of [125I]ABMECA (final concentration 0.15 nM),
and 25 µL of ligand. Incubations were carried out for 1 h at
37 °C and were terminated by rapid filtration over Whatman
GF/B filters, using a Brandell cell harvester (Brandell, Gaith-
ersburg, MD). Tubes were washed three times with 3 mL of
buffer. Radioactivity was determined in a Beckman 5500B
γ-counter. Nonspecific binding was determined in the presence
of 10-5 M (R)-PIA.
cAMP Assa y for A1 a n d A3. CHO cells expressing either
the human adenosine A1 or the human adenosine A3 receptor
were grown overnight as a monolayer in 24-well tissue culture
plates (400 µL/well, 2 × 105 cells/well). cAMP generation was
performed in Dulbecco’s modified Eagle’s medium (DMEM)/
N-(2-hydroxyethyl)piperazine-N′-2-ethanesulfonic acid (HEPES)
buffer (0.60 g of HEPES/50 mL of DMEM, pH 7.4). To each
well, washed two times with DMEM/HEPES buffer (250 µL)
were added 100 µL of adenosine deaminase (final concentration
5 IU/mL), 100 µL of a mixture of rolipram and cilostamide
(final concentration each 50 µM), and 100 µL of agonist (final
concentration (100 times the Ki value). After incubation for
40 min at 37 °C, 100 µL of forskolin (final concentration 10
µM) was added. After 15 min at 37 °C, the reaction was
terminated by removing the medium and adding 200 µL of
0.1 M HCl. The wells were stored at -20 °C until assay.
The amounts of cAMP were determined after a protocol with
cAMP binding protein36 with the following minor modifications.
As a buffer was used 150 mM K2HPO4/10 mM EDTA/0.2%
bovine serum albumin (BSA) at pH 7.5. Samples (20 µL + 30
µL of 0.1 M HCl) were incubated for at least 2.5 h at 0 °C
before filtration over Whatman GF/B filters. The filters were
additionally rinsed with 2 × 2 mL of Tris·HCl buffer (pH 7.4,
4 °C). The filters were counted in Packard emulsifier safe
scintillation fluid (3.5 mL) after 24 h of extraction.
5′-O-Cyclop r op yla d en osin e (30). Method A. The reaction
was carried out with 6-chloro-9-(2,3-di-O-acetyl-5-O-cyclopro-
pyl-â-D-ribofuranosyl)purine (16; 334 mg, 0.81 mmol). The
mixture was purified by column chromatography (eluent 5%
MeOH in CH2Cl2): yield 172 mg (0.56 mmol, 69%); mp 130-
1
132 °C; Rf 0.49 (10% MeOH in CH2Cl2); H NMR (DMSO-d6)
δ 8.27 (s, 1H, H-8), 8.14 (s, 1H, H-2), 7.28 (br s, 2H, NH2),
5.87 (d, 1H, J ) 4.80 Hz, H-1′), 5.50 (d, 1H, J ) 5.84 Hz, OH-
2′), 5.26 (d, 1H, J ) 5.49 Hz, OH-3′), 4.56 (q, 1H, J ) 4.81 Hz,
H-2′), 4.10 (q, 1H, J ) 4.47 Hz, H-3′), 3.99 (q, 1H, J ) 3.78
Hz, H-4′), 3.67-3.59 (m, 2H, H-5′), 0.44-0.40 (m, 4H, OCH2-
CH2); MS m/ z 308 (M + H)+. Anal. (C13H17N5O4‚0.2CH2Cl2)
C, H, N.
N-Cyclop en tyl-5′-O-cyclop r op yla d en osin e (31). Method
B. The reaction was carried out with 16 (273 mg, 0.66 mmol)
and cyclopentylamine (1.00 mmol, 98 µL). The mixture was
purified by column chromatography (eluent 5% MeOH in CH2-
Cl2): yield 173 mg (0.46 mmol, 70%); mp 126-128 °C; Rf 0.53
(10% MeOH in CH2Cl2). The product was recrystallized from
(C2H5)2O: 1H NMR (MeOD) δ 8.17 (s, 1H, H-8), 8.13 (s, 1H,
H-2), 5.94 (d, 1H, J ) 4.46 Hz, H-1′), 4.44 (t, 1H, J ) 4.80 Hz,
H-2′), 4.18 (t, 1H, J ) 5.15 Hz, H-3′), 4.11-4.06 (m, 1H, H-4′),
3.69 (q, 2H, J ) 9.95 Hz, H-5′), 3.24-3.19 (m, 1H, CH), 2.11-
1.82 (m, 2H, cyclopentyl), 1.72-1.50 (m, 6H, cyclopentyl),
0.48-0.39 (m, 4H, CH2CH2); MS m/ z 376 (M + H)+. Anal.
(C18H25N5O4‚0.5(C2H5)2O) C, H, N.
5′-O-Cyclop r op yl-N-(3-iod oben zyl)a d en osin e
(32).
Method B. The reaction was carried out with 16 (196 mg, 0.48
mmol) and (3-iodobenzyl)amine hydrochloride (0.72 mmol, 193
mg). The mixture was purified by column chromatography
(eluent 5% MeOH in CH2Cl2): yield 186 mg (0.36 mmol, 74%);
mp 110-112 °C; Rf 0.49 (10% MeOH in CH2Cl2); 1H NMR
(DMSO-d6) δ 8.43 (br s, 1H, NH), 8.31 (s, 1H, H-8), 8.21 (s,
1H, H-2), 7.71 (s, 1H, CCHCI), 7.57 (d, 1H, J ) 7.55 Hz,
CCHCHCH), 7.34 (d, 1H, J ) 7.89 Hz, CCHCH), 7.09 (t, 1H,
J ) 7.90 Hz, CCHCH), 5.89 (d, 1H, J ) 5.15 Hz, H-1′), 5.49
(d, 1H, J ) 5.83 Hz, OH-2′), 5.27 (d, 1H, J ) 5.15 Hz, OH-3′),
4.68-4.55 (m, 2H, NHCH2), 4.57 (q, 1H, J ) 5.49 Hz, H-2′),
4.11 (q, 1H, J ) 4.46 Hz, H-3′), 4.00 (q, 1H, J ) 4.11 Hz, H-4′),
3.67-3.62 (m, 2H, H-5′), 0.45-0.40 (m, 4H, CH2CH2); MS m/ z
524 (M + H)+. Anal. (C20H22IN5O4‚0.7CH3OH) C, H, N.
2-Ch lor o-5′-O-cyclop r op yla d en osin e (33). Method A.
The reaction was carried out with 2,6-dichloro-9-(2,3-di-O-
acetyl-5-O-cyclopropyl-â-D-ribofuranosyl)purine (17; 649 mg,
1.46 mmol). The mixture was purified by column chromatog-
raphy (eluent 5% MeOH in CH2Cl2): yield 379 mg (1.11 mmol,
76%); mp 122-124 °C; Rf 0.32 (10% MeOH in CH2Cl2); 1H
NMR (DMSO-d6) δ 8.29 (s, 1H, H-8), 7.83 (br s, 2H, NH2), 5.79
(d, 1H, J ) 5.49 Hz, H-1′), 5.52 (d, 1H, J ) 6.18 Hz, OH-2′),
5.29 (d, 1H, J ) 5.29 Hz, OH-3′), 4.51 (q, 1H, J ) 5.84 Hz,
H-2′), 4.08-3.99 (m, 2H, H-3′,4′), 3.68-3.63 (m, 2H, H-5′),
0.45-0.41 (m, 4H, CH2CH2); MS m/ z 342 (M + H)+. Anal.
(C13H16ClIN5O4‚0.8HCON(CH3)2) C, H, N.
GTP γS Bin d in g. The extent of stimulation of [35S]GTPγS
binding was determined according to the method of Lorenzen
et al.6,7 with some minor modifications. The final incubation
conditions, which allowed the characterization of the synthe-
sized compounds as potential partial agonists of either the
adenosine A1 or A3 receptor, were as follows.
A1. The GDP and NaCl concentrations were 10 µM and 100
mM, respectively. Incubations were performed at 25 °C for 90
min with 2 µg of membrane protein (rat adenosine A1 receptor).
CPA was used as a full agonist in this assay, and its
2-Ch lor o-N-cyclopen tyl-5′-O-cyclopr opyladen osin e (34).
Method B. The reaction was carried out with 17 (543 mg, 1.22
mmol) and cyclopentylamine (1.83 mmol, 180 µL). The mixture
was purified by column chromatography (eluent 5% MeOH in
CH2Cl2): yield 400 mg (0.98 mmol, 80%); mp 104-106 °C; Rf

5′-O-Alkyl Ethers of Adenosine Derivatives
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 18 2975
stimulation of [35S]GTPγ[S] binding to the G protein was set
to 100% (Table 3).
(15) Knutsen, L. J . S.; Sheardown, M. J .; Roberts, S. M.; Mogensen,
J . P.; Olsen, U. B.; Thomsen, C.; Bowler, A. N. Adenosine A1
and A3 selective N-alkoxypurines as novel cytokine modulators
and neuroprotectants. Drug Dev. Res. 1998, 45, 214-221.
(16) Gin, J . B.; Dekker, C. A. The preparation and properties of
O-methylated adenosine derivatives. Biochemistry 1968, 7,
1413-1420.
A3. The GDP concentration was 1 µM, and NaCl was
omitted. Incubations were performed at 25 °C for 90 min with
1-1.5 µg of membrane protein (human adenosine A3 receptor
expressed in CHO cells). NECA was used as a full agonist in
this assay, and its stimulation of [35S]GTPγ[S] binding to the
G protein was set to 100% (Table 3).
(17) Vorbru¨ggen, H.; Ho¨fle, G. On the mechanism of nucleoside
synthesis. Chem. Ber. 1981, 114, 1256-1268.
(18) Vorbru¨ggen, H.; Krolikiewicz, K.; Bennua, B. Nucleoside syn-
thesis with trimethylsilyl triflate and perchlorate as catalysts.
Chem. Ber. 1981, 114, 1234-1255.
(19) Lorenzen, A.; Fuss, M.; Vogt, H.; Schwabe, U. Measurement of
guanine nucleotide-binding protein activation by A1 adenosine
receptor agonists in bovine brain membranes: stimulation of
guanosine-5′-O-(3-[35S]thio)triphosphate binding. Mol. Pharma-
col. 1993, 44, 115-123.
(20) Bessodes, M.; Shamsazar, J .; Antonakis, K. Crown ether cata-
lyzed O-alkylation of carbohydrates and nucleosides. Synthesis
1988, 560-562.
(21) McCall, M. J .; Taylor, M. R. Metal binding to nucleic acid
constituents. The crystal structure of catenadichloro-µ-(9-me-
thyladenine)zinc(II)*. Acta Crystallogr. 1976, B32, 1687-1691.
(22) Leonard, N. J .; Carraway, K. L. 5-Amino-5-deoxyribose deriva-
tives. Synthesis and use in the preparation of ‘reversed’ nucleo-
sides (1a). J . Heterocycl. Chem. 1966, 3, 485-487.
Da ta An a lysis. Apparent Ki and EC50 values were com-
puted from the displacement curves by nonlinear regression
of the competition curves with the software package Prism
(Graph Pad, San Diego, CA).
Ack n ow led gm en t. We thank Karl-Norbert Klotz
(University of Wu¨rzburg, Germany) for providing HEK
293 and CHO cells expressing the human adenosine A3
receptor and Steve Rees (GlaxoWellcome, Stevenage,
U.K.) for the CHO cells expressing the human adenosine
A3 receptor and the CHO cells expressing the human
adenosine A1 receptor.
Refer en ces
(23) Levene, P. A.; Stiller, E. T. Acetone derivatives of D-ribose. J .
Biol. Chem. 1934, 187-201.
(24) Visser, G. M. Synthesis of some modified nucleotide derivatives
and the possible role of mirror-image nucleotides in molecular
evolution. Ph.D. Thesis, Leiden University, The Netherlands,
1986, p 52.
(1) Londos, C.; Honner, R. C.; Dhillon, G. S. cAMP dependent protein
kinase and lipolysis in rat adipocytes. III. Multiple modes of
insulin regulation of lipolysis and regulation of insulin responses
by adenylate cyclase. J . Biol. Chem. 1980, 260, 15139-15145.
(2) Saloranta, C.; Fransilla-Kallunki, A.; Ekstrand, A.; Taskinen,
M.-R.; Groop, L. Modulation of hepatic glucose production by
non-esterified fatty acids in type 2 (non-insulin dependent)
diabetes mellitus. Diabetologia 1991, 34, 409-415.
(3) J acobson, K. A.; Lubitz, D. K. J . E. v.; Daly, J . W.; Fredholm, B.
B. Adenosine receptor ligands: differences with acute versus
chronic treatment. Trends Pharmacol. Sci. 1996, 17, 108-113.
(4) Liang, B. T.; J acobson, K. A. A physiological role of the adenosine
A3 receptor: Sustained cardioprotection. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 6995-6999.
(5) IJ zerman, A. P.; Van der Wenden, E. M.; Von Frijtag Drabbe
Ku¨nzel, J . K.; Mathoˆt, R. A. A.; Danhof, M.; Borea, P. A.; Varani,
K. Partial agonism of theophylline-7-riboside on adenosine
receptors. Naunyn-Schmiedeberg’s Arch. Pharmacol. 1994, 350,
638-645.
(6) Lorenzen, A.; Fuss, M.; Vogt, H.; Schwabe, U. Measurement of
guanine nucleotide-binding protein activation by A1 adenosine
receptor agonists in bovine brain membranes: stimulation of
guanosine-5′-O-(3-[35S]thio)triphosphate binding. Mol. Pharma-
col. 1993, 44, 115-123.
(7) Lorenzen, A.; Guerra, L.; Vogt, H.; Schwabe, U. Interaction of
full and partial agonists of the A1 adenosine receptor with
receptor-G protein complexes in rat brain membranes. Mol.
Pharmacol. 1996, 49, 915-926.
(8) Furchgott, R. F. In Advances in drug research; Harper, N. J .,
Simmonds, A. B., Eds.; Academic Press: 1966; Vol. 3, pp 21-
55.
(9) Clarke, W. P.; Bond, R. A. The exclusive nature of intrinsic
efficacy. Trends Pharmacol. Sci. 1998, 19, 270-276.
(10) Ligt, R. A. F.; Kourounakis, A. P.; IJ zerman, A. P. Inverse
agonism at G protein-coupled receptors: (patho)physiological
relevance and implications for drug delivery. Br. J . Pharmacol.
2000, 130, 1-12.
(11) Lorenzen, A.; Sebastiao, A. M.; Sellink, A.; Vogt, H.; Schwabe,
U.; Ribeiro, J . A.; IJ zerman, A. P. Biological activities of N6,-
C8-disubstituted adenosine derivatives as partial agonists at rat
brain adenosine A1 receptors. Eur. J . Pharmacol. 1997, 334,
299-307.
(25) Vorbru¨ggen, H.; Bennua, B. A new simplified nucleoside syn-
thesis. Chem. Ber. 1981, 114, 1279-1286.
(26) Saneyoshi, M.; Satoh, E. Synthetic nucleosides and nucleotides.
XIII. Stannic chloride catalyzed ribosylation of several 6-sub-
stituted purines. Chem. Pharm. Bull. 1979, 27, 2518-2521.
(27) Roelen, H.; Veldman, N.; Spek, A. L.; von Frijtag Drabbe Ku¨nzel,
J .; Mathot, R. A.; IJ zerman, A. P. N6,C8-Disubstituted adenosine
derivatives as partial agonists for adenosine A1 receptors. J .
Med. Chem. 1996, 39, 1463-71.
(28) Camaioni, E.; DiFrancesco, E.; Vittori, S.; Volpini, R.; Cristalli,
G. Adenosine receptor agonists: Synthesis and biological evalu-
ation of the diastereoisomers of 2-(3-hydroxy-3-phenyl-1-propyn-
1-yl)NECA. Bioorg. Med. Chem. 1997, 5, 2267-2275.
(29) Cristalli, G.; Camaioni, E.; Costanzi, S.; Vittori, S.; Volpini, R.;
Klotz, K. N. Characterization of potent ligands at human
recombinant adenosine receptors. Drug Dev. Res. 1998, 45, 176-
181.
(30) Lohse, M. J .; Klotz, K.-N.; Schwabe, U.; Cristalli, G.; Vittori, S.;
Grifantini, M. 2-Chloro-N6-cyclopentyladenosine: a highly selec-
tive agonist at A1 adenosine receptors. Naunyn-Schmiedeberg’s
Arch. Pharmacol. 1988, 337, 687-689.
(31) Kim, H. O.; Siddiqi, S. M.; Olah, M. E.; Stiles, G. L.; K. A. J .
2-Substitution of N6-benzyladenosine-5′-uronamides enhances
selectivity for A3 adenosine receptors. J . Med. Chem. 1994, 37,
3614-3621.
(32) Pirovano, I. M.; IJ zerman, A. P.; van Galen, P. J . M.; Soudijn,
W. The influence of molecular structure of N6-(ω-aminoalkyl)-
adenosines on adenosine receptor affinity and intrinsic activity.
Eur. J . Pharmacol. 1989, 172, 185-193.
(33) Gao, Z.-G.; IJ zerman, A. P. Allosteric modulation of A2A adenos-
ine receptors by amiloride analogues and sodium ions. Biochem.
Pharmacol. 2000, 60, 669-676.
(34) Olah, M. E.; Gallo-Rodriquez, C.; J acobson, K. A.; Stiles, G. L.
125I-4-Aminobenzyl-5′-N-methylcarboxamidoadenosine, a high
affinity radioligand for the rat A3 adenosine receptor. Mol.
Pharmacol. 1994, 45, 978-982.
(35) van Galen, P. J . M.; Van Bergen, A. H.; Gallo-Rodriquez, C.;
Melman, N.; Olah, M. E.; IJ zerman, A. P.; Stiles, G. L.; J acobson,
K. A. A binding site model and structure-acitivity relationships
for the rat A3 adenosine receptor. Mol. Pharmacol. 1994, 45,
1101-1111.
(12) Van der Graaf, P. H.; Van Schaick, E. A.; Visser, S. A. G.; De
Greef, H. J . M. M.; IJ zerman, A. P.; Danhof, M. Mechanism-
based pharmacokinetic-pharmacodynamic modeling of antili-
polytic effects of adenosine A1 receptor agonists in rats: Pre-
diction of tissue-dependent efficacy in vivo. J . Pharmacol. Exp.
Ther. 1999, 290, 702-709.
(36) Van der Wenden, E. M.; Hartog-Witte, H. R.; Roelen, H. C. P.
F.; Von Frijtag Drabbe Ku¨nzel, J . K.; Pirovano, I. M.; Mathoˆt,
R. A. A.; Danhof, M.; Van Aerschot, A.; Lidaks, M. J .; IJ zerman,
A. P.; Soudijn, W. 8-substituted adenosine and theophylline-7-
riboside analogues as potential partial agonists for the adenosine
A1 receptor. Eur. J . Pharmacol.-Mol. Pharmacol. Sect. 1995, 290,
189-199.
(13) Van Tilburg, E. W.; von Frijtag Drabbe Ku¨nzel, J .; Groote, M.;
Vollinga, R. C.; Lorenzen, A.; IJ zerman, A. P. N6,5′-Disubstituted
adenosine derivatives as partial agonists for the human adenos-
ine A3 receptor. J . Med. Chem. 1999, 42, 1393-1400.
(14) Van der Wenden, E. M.; Carnielli, M.; Roelen, H. C. P. F.;
Lorenzen, A.; Von Frijtag Drabbe Ku¨nzel, J . K.; IJ zerman, A.
P. 5_-Substituted adenosine analogs as new high-affinity partial
agonists for the adenosine A1 receptor. J . Med. Chem. 1998, 41,
102-108.
J M001114O