Design, synthesis, and evaluation of novel A2A adenosine receptor agonists

Article abstract of DOI:10.1021/jm0003642

We have been interested in the design, synthesis, and evaluation of novel adenosine A₂A agonists. Through the use of comparative molecular field analysis (CoMFA) we have generated a training model that includes 78 structurally diverse A_2A<

Full text of DOI:10.1021/jm0003642

J . Med. Chem. 2001, 44, 531-539
531
Design , Syn th esis, a n d Eva lu a tion of Novel A2A Ad en osin e Recep tor Agon ists
J ayson M. Rieger,† Milton L. Brown,† Gail W. Sullivan,‡ J oel Linden,‡ and Timothy L. Macdonald*,†
Departments of Chemistry and Medicine, University of Virginia, McCormick Road, P.O. Box 400319,
Charlottesville, Virginia 22904-4319
Received August 21, 2000
We have been interested in the design, synthesis, and evaluation of novel adenosine A2A
agonists. Through the use of comparative molecular field analysis (CoMFA) we have generated
a training model that includes 78 structurally diverse A2A agonists and correlated their affinity
for isolated rat brain receptors with differences in their structural and electrostatic properties.
We validated this model by predicting the activity of a test set that included 24 additional A2A
agonists. Our CoMFA model, which incorporates the physiochemical property of dipole and
selects against A1 receptor activity, generated a correlated final model (r2 ) 0.891) that provides
for enhanced A2A selectivity and predictability. Synthesis, pharmacological evaluation, and
modeling of four novel ligands further validate the utility and predictive power (r2 ) 0.626) of
the CoMFA model.
Ch a r t 1. General Adenosine Agonist Pharmacophore
In tr od u ction
Adenosine is an endogenous molecule that has many
important biological functions and is a common sub-
structure found in ATP, RNA, and several coenzymes.1
It is the native ligand for adenosine receptors, also
referred to as P1 purinergic receptors, which are distinct
from P2 purinergic nucleotide receptors. The adenosine
receptor class can be further divided into four receptor
subtypes which have been identified and cloned from
several mammalian tissue types: A1, A2A, A2B, and A3.2
These receptors are known to elicit their biological
responses through prototypical seven-transmembrane
spanning G-protein coupled receptors.3 Agonists and
antagonists of these receptors are being explored as
potential therapeutic agents for: cardiovascular media-
tion,4 ischemia-reperfusion injury,5 inflammation,6 Par-
kinson’s disease,7 and schizophrenia.8 We are particu-
larly interested in the impact of adenosine A2A receptor
agonists on inflammation and have recently published
several reports5,9 describing the reduction of these
responses utilizing the novel ligands described here.
Because of the many potential biological implications
of stimulating the various subclasses of adenosine
receptors, much effort has been spent trying to elucidate
SARs (structure-activity relationships) for adenosine
analogues. This has led to the synthesis and evaluation
of a great number of adenosine analogues.10-17 From
this work, a general pharmacophore for A2A agonists can
be produced and is summarized in Chart 1. In agree-
ment with this general pharmacophore, several ligands
have evolved as standards for their specific receptor
subclasses and include NECA (78), CGS21680 (113),
CPA, and Cl-IB-MECA (Chart 2).
Ch a r t 2. Several Common Selective Adenosine
Receptor Agonists
explain the observed experimental data for the ligands
that have been synthesized and screened against the
adenosine receptor subtypes. Recently adenosine A1
agonists,18 adenosine A2A antagonists,19 and adenosine
A3 receptor antagonists20-23 and agonists24 have been
the subject of modeling efforts. Several19,21,22,24 of these
efforts have successfully employed the use of CoMFA
(comparative molecular field analysis), which develops
a predictive three-dimensional model based on differ-
ences in steric and electrostatic fields and binding
Because of the large amount of research that has been
done on adenosine and its receptors, the need for models
to summarize and correlate the available data is im-
mense. Various SAR techniques have been employed to
* To whom correspondence should be addressed. Phone: 804-924-
7718. Fax: 804-982-2302. E-mail: tlm@virginia.edu.
† Department of Chemistry.
‡ Department of Medicine.
10.1021/jm0003642 CCC: $20.00 © 2001 American Chemical Society
Published on Web 01/16/2001

532 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4
Rieger et al.
Sch em e 1. Coupling Scheme for Synthesis of Novel
2-Alkynyl-NECAs
Sch em e 2. First Synthetic Route Toward Ligand
Synthesisa
a
Reaction conditions: (A) (i) NaH, THF, (ii) TosCl; (B) lithium
acetylide ethylenediamine complex, DMSO.
Sch em e 3. Final Synthetic Route Toward Ligand
Synthesisa
affinities. However, predictive models for adenosine A2A
agonists are presently lacking. Therefore, in this study
we have used CoMFA to correlate a diverse group of
adenosine structural modifications in order to predict
a new, highly potent, and selective class of adenosine
A2A agonists.
Resu lts
Syn th esis. A plethora of substitutions have been
made at the C2 position of adenosine (Chart 1) to
enhance the activity and selectivity of compounds for
the various receptor subclasses. Many of these substitu-
tions are represented by compounds 1-78 which are
found in Table 1. We chose to explore novel 2-alkynyl-
substituted derivatives of NECA to validate our model.
The novel 2-alkynyladenosines (2-AAs) designed and
evaluated in this report required the coupling of a series
of terminal alkynes to 2-iodo-NECA (114)17 utilizing
TEA and CuI in DMF/acetonitrile with catalytic Pd-
(PPh3)4 and PPh3 in a classical Sonogoshira cross-
coupling fashion as shown in Scheme 1. Many different
attempts, including changing temperature, palladium
source, and component ratio, ultimately only afforded
the desired products in low to moderate yields.
The proposed compounds, 109-112, required the
synthesis of a series of novel alkynes. We envisioned
the formation of these alkynes from commercially avail-
able trans-1,4-cyclohexanedimethanol (79). We first
proposed formation of the desired alkynes by mono-
tosylation, followed by insertion of the alkyne, and
modification of the remaining free hydroxyl to the
various proposed functionalities (carboxylic acid, methyl
ester, or acetate). Unfortunately, the addition of lithium
acetylide-ethylenediamine complex gave an intractable
mixture of products as seen by MS and NMR believed
to contain the ether 82 as shown in Scheme 2.
a
Reaction conditions: (A) TBSCl, imidazole, DMF; (B) TosCl,
pyridine; (C) lithium acetylide, DMSO; (D) TBAF, THF; (E) acetic
anhydride, pyridine, THF; (F) J ones oxidation; (G) TMS-di-
azomethane, MeOH/CHCl3.
Introduction of the terminal acetylene was accom-
plished through the in situ formation of lithium acetyl-
ide27 by bubbling acetylene gas into a solution of lithium
wire in liquid ammonia. After addition of DMSO, the
ammonia was removed and tosylate 84 was added
ultimately affording compound 85. Interestingly, some
deprotected alcohol 86 (approximately 10-20%) was
also observed upon working up the reaction. In an effort
to increase the overall yield, the crude mixture was
treated with TBAF hydrate, generating 86 in an overall
yield of 93% through both steps. This alkyne was either
coupled to 2-iodo-NECA (114) directly or converted to
a carboxylic acid, acetate, or methyl ester. Treating 86
with acetic anhydride and pyridine gave the acetate 87.
J ones oxidation of 86 gave the carboxylic acid 88, and
the methyl ester 89 was formed by treating 88 with
TMS-diazomethane. These compounds, 86-89, were
coupled with 2-iodo-NECA to form the novel A2A ago-
nists ATL2037 (112), ATL193 (111), ATL146a (109), and
ATL146e (110), respectively.
To circumvent this problem we utilized the synthetic
protocol outlined in Scheme 3. The reaction of an excess
of 79 with TBSCl (tert-butyl-dimethylsilyl chloride) and
imidazole in DMF gave the mono-TBS alcohol. Attempts
to enhance the monosilation of the diol through the use
of TBSCl and NaH in THF25 were unsuccessful. We then
introduced the tosyl group to 83 under typical condi-
tions,26 preparing the molecule for introduction of the
acetylene moiety and formation of 84.
P h a r m a cology. The four novel compounds synthe-
sized in this report were evaluated by competitive
binding assays employing the A2A receptors in rat
striatal membranes ([125I]APE)28 or A1 receptors of rat
cortex ([125I]ABA).29 Compounds 109-112 were potent
agonists, ranging in binding affinities from Ki ) 20 nM
to 0.5 nM at the A2A receptor (see Table 2). They also

Novel A2A Adenosine Receptor Agonists
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4 533
Ta ble 1. Structures and Biological Data for Compounds in the Training Set
compd
R
NECA* (R′) A1 (nM) A2A (nM) log(norm A1) -log(norm A2A
)
dipole energy (kcal)
1
2
3
4
5
6
7
8
2-ethyl-1-butoxy
3-ethyl-1-pentoxy
3-pentyl-1-oxy
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
N
N
2700a
1490a
460a
4600a
530a
150a
1000a
3200a
120a
22a
-5.276
-5.534
-6.045
-4.561
-5.230
-5.503
-5.503
-5.260
-5.666
-5.163
-5.929
-6.594
-7.881
-5.594
-5.151
-5.327
-5.725
-5.346
-5.503
-5.992
-6.310
-6.758
-6.850
-6.531
-6.009
-6.054
-6.561
-7.017
-6.429
-6.026
-6.128
-5.628
-5.009
-5.310
-5.758
-5.293
-6.240
-5.631
-5.497
-5.142
-5.180
-5.860
-6.063
ND
-5.367
-6.697
-6.874
-6.386
-6.526
-6.241
-7.086
-6.767
-7.060
-6.173
-6.321
-6.903
-6.089
-4.829
-5.690
-4.681
-5.075
-5.664
-7.067
-6.872
-6.516
-6.166
5.324
6.262
6.811
5.987
5.482
6.908
7.644
7.644
7.731
5.231
5.644
7.756
7.106
5.811
5.908
6.756
7.262
6.353
5.756
7.239
7.106
7.644
6.148
7.223
7
7.396
7.201
7.945
7.430
7.482
7.385
6.430
6.468
6.555
6.783
6.142
6.839
6.394
6.981
5.716
5.828
6.647
7.560
6.021
6.204
8.813
6.523
6.953
6.778
7.602
7.276
7.675
7.454
6.813
7.578
8.550
8.081
6.015
8.182
7.217
7.415
8.114
7.732
7.811
7.396
6.774
1.66
1.71
1.73
3.45
1.89
1.65
2.02
1.85
1.90
1.57
1.64
1.74
2.46
1.83
1.58
1.64
1.59
1.95
1.78
1.24
0.84
2.13
1.98
0.85
2.16
2.01
2.04
1.85
1.80
1.47
2.07
1.78
1.71
1.79
1.78
1.68
2.55
3.37
2.55
2.18
0.90
2.46
6.34
2.95
2.33
2.93
3.00
2.40
2.27
3.15
3.63
4.86
3.74
2.50
2.74
2.50
2.50
2.24
3.42
2.88
2.78
3.62
2.39
2.41
2.25
2.45
46.235
47.865
45.431
52.200
42.335
53.581
43.426
43.700
44.163
43.570
42.908
42.488
40.942
41.882
42.636
42.445
43.000
42.237
42.627
42.181
42.033
42.060
42.311
42.194
42.202
43.492
42.741
42.758
44.000
44.191
45.561
42.040
46.848
42.223
42.594
41.894
37.774
37.792
37.665
40.192
37.584
37.736
39.991
38.331
38.696
38.479
37.971
37.930
46.536
40.539
43.980
61.210
43.955
37.953
38.717
38.031
39.019
39.005
38.732
39.558
40.618
40.593
51.132
41.041
42.299
41.240
cyclopentyloxy
14000a
3000a
1600a
1600a
2800a
1100a
3500a
600a
cyclohexylmethoxy
2-cyclopentylethoxy
2-cyclohexylethoxy
3-cyclohexyl-1-propoxy
4-cyclohexyl-1-butoxy
phenoxy
benzyloxy
2-phenylethoxy
Cl
22a
9
18a
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
5700a
2200a
17a
130a
6.7a
76a
ethoxy
1300a
3600a
2400a
960a
2300a
1600a
520a
250a
89a
1500a
1200a
170a
53a
3-phenyl-1-propoxy
4-phenyl-1-butoxy
5-phenyl-1-pentoxy
(2S)-phenyl-1-butoxy
(3S)-phenyl-1-butoxy
2-(2-chlorophenyl)ethoxy
2-(3-chlorophenyl)ethoxy
2-(4-chlorophenyl)ethoxy
2-hydroxyethoxy
2-(3-fluorophenyl)ethoxy
2-(4-fluorophenyl)ethoxy
2-(2-methylphenyl)ethoxy
2-(3-methylphenyl)ethoxy
2-(4-methylphenyl)ethoxy
2-(2-methoxyphenyl)ethoxy
2-(4-methoxyphenyl)ethoxy
2-(3,4-dimethoxyphenyl)ethoxy
1-propoxy
2-(3,4,5-dimethoxyphenyl)ethoxy
1-butoxy
1-hexyloxy
2-methyl-1-propoxy
p-PhNH2
p-PhCF3
p-PhF
p-PhCONH2
o-PhCHO
m-PhCHO
p-PhNO2
p-Ph(CH2)COOtBu
1-naphthyl
(CH2)3Ph
3-pyridyl
4-pyridyl
2-furyl
2-thienyl
2-thiazolyl
CH2-N-imidazolyl
CH2-N-morpholinyl
Ph
p-PhOH
(CH2)3CH3
430a
1700a
56a
76a
22a
72a
690a
58a
150a
500a
450a
140a
49a
97a
39a
61a
11a
190a
480a
380a
1200a
5000a
2500a
890a
2600a
599b
2432b
3312b
7498b
6864b
1435b
900b
ND
36a
32a
40a
360a
330a
270a
160a
700a
113b
315b
81.4b
1500b
1158b
176b
21.5b
743b
488b
4463b
209b
139b
428b
310b
597b
1.2b
234b
87b
130b
19.5b
85.4b
178b
41.3b
16.5b
27.4b
90.5b
698b
497b
120b
20.6b
2.2b
9.7c
130b
NH(CH2)2Ph
NHCH2Ph
473c
8604c
1184c
12100c
4880c
1257c
88d
1130c
NH(CH2)2(4-fluorophenyl)
N(CH3)((CH2)2Ph)
NH(CH2)2(4-(CH2COOH)phenyl)
NH(CH2)2(4-(CH2COOMe)phenyl)
c-C5H9
c-C6H11
c-C6H11(CH2)2
Ph
7.7c
71c
45c
9c
12d
10d
26d
109d
138d
313d
701d

534 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4
Ta ble 1. (Continued)
Rieger et al.
compd
67
68
69
70
71
72
73
74
75
76
77
78
(NECA)
R
NECA* (R′)
A1 (nM)
A2A (nM)
log(norm A1)
-log(norm A2A
)
dipole
energy (kcal)
Ph(CH2)2
N
N
N
N
N
N
N
N
N
N
N
400d
186e
154e
126e
202e
10.1e
81.4e
38.1e
18.5e
128e
244e
13.3e
37d
-6.409
-6.854
-6.936
-7.023
-6.819
-8.120
-7.213
-7.543
-7.857
-7.017
-6.736
-8.000
7.243
7.767
8.218
8.646
8.011
7.782
7.490
7.148
8.068
8.085
7.986
8.000
2.22
2.35
2.29
2.29
2.22
3.41
2.64
2.85
2.72
2.22
2.65
2.61
40.377
40.601
40.698
40.869
41.205
40.923
40.963
41.213
41.741
41.420
41.466
40.318
CH2CH3
21.2e
7.5e
(CH2)2CH3
(CH2)3CH3
(CH2)5CH3
CH2OH
(CH2)2OH
CH2OCH3
CH2O(CH2)2CH3
(CH2)2O(CH2)2CH3
(CH2)3OCH2CH3
H
2.8e
12.1e
20.5e
40.1e
88.1e
10.6e
10.2e
12.8e
12.4e
a
*Y refers to R′ ) NECA modification (N-ethylcarboxamide); N refers to R′ ) CH2OH. Reported in ref 11. b Reported in ref 10. c Reported
d
in ref 15. Reported in ref 12. e Reported in ref 16. Normalized data found by multiplying or dividing the respective paper’s IC50 value for
NECA by a factor that converts NECA to a standard value of 10 nM for both A1 and A2A assays. ND ) no data available.
Ta ble 2. Structures and Biological Data for Compounds in the Test Set
compd
R
NECA* (R′)
A1 (nM)
A2A (nM) log(norm A1) -log(norm A2A
)
dipole energy (kcal)
90
91
92
93
94
95
96
97
98
2-(4-methyl-1-pentoxy)
Cl
2-[2-(3-fluorophenyl)ethoxy]
2-[2-(3-methoxyphenyl)ethoxy]
2-(2-propoxy)
CH2Ph
2-pyridyl
2-thiazolyl
CH2-N-imadazolyl
p-PhCH3
p-PhCH2CN
p-PhOCH3
p-ClC6H4(CH2)2NH
p-(HOOCCH2O)C6H4(CH2)2NH
c-C5H9CH2
c-C5H9(CH2)2
c-C6H11CH2
(CH2)4CH3
H
N
N
N
N
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
N
N
N
Y
Y
Y
Y
NA
2100a
6.7a
130a
-5.39
-7.88
-6.45
-5.83
-4.41
-7.58
-6.96
-7.09
-6.78
-6.32
-5.81
-6.40
-5.59
-5.14
-6.80
-6.88
-6.69
-6.89
-6.69
-5.50
-6.54
-6.82
-8.04
-5.45
6.87
7.11
6.95
6.48
5.12
8.67
6.94
7.28
7.68
7.33
7.15
7.17
8.19
7.45
8.45
8.28
8.00
8.36
7.52
7.23
7.88
7.74
8.85
7.71
1.90
2.46
1.28
1.23
1.19
2.63
2.63
1.10
5.03
5.03
3.11
2.43
3.38
4.87
2.61
2.25
2.48
2.25
2.23
1.02
1.90
2.80
2.33
4.02
43.200
40.942
42.121
44.106
41.924
38.655
38.229
44.029
61.243
38.203
37.629
39.567
38.586
41.875
50.941
51.167
41.102
41.019
40.701
40.559
44.449
42.512
38.740
39.973
76a
180a
760a
20000a
27.4b
114b
85.4b
178b
500b
1603b
414b
1505c
4250c
162d
136d
208d
170e
272e
292
110a
320a
7400a
1.6b
89.8b
41.3b
16.5b
36.3b
54.8b
52.4b
7.5c
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
42c
2.3d
3.4d
6.5d
5.4e
37.6e
20.1
4.5
4-(COOH)C6H10CH2
4-(COOCH3)C6H10CH2
4-(CH2OAc)C6H10CH2
4-(CH2OH)C6H10CH2
NA
28.1
14.0
6.2
0.84
0.48
1800a
19a
a
*Y refers to R′ ) NECA modification (N-ethylcarboxamide); N refers to R′ ) CH2OH. Reported in ref 11. b Reported in ref 10. c Reported
d
in ref 15. Reported in ref 12. e Reported in ref 16. Normalized data found by multiplying or dividing the respective paper’s IC50 value for
NECA by a factor that converts NECA to a standard value of 10 nM for both A1 and A2A assays. NA ) not available.
demonstrated moderate selectivity against the A1 recep-
tor, with nearly 2-15-fold differences in binding affinity
favoring the A2A receptor. These compounds all incor-
porate key SAR features known to contribute to high
A2A affinity and selectivity: namely an aliphatic alkyne
with a polar terminal group at C2 and an ethylcarboxa-
mide group at the 5′ position.
Molecu la r Mod elin g. CoMFA was employed as a
tool for elucidating the structural and electrostatic
requirements for high-affinity A2A receptor binding.
Using previously reported A2A agonists that also had
available A1 receptor data, we compiled the representa-
tive list of compounds found in Tables 1 and 2. We
designed our CoMFA model both to be selective for
potent A2A activity and to minimize activity at the A1
receptor subtype. We hypothesized that this would
provide us with a more useful and global model that
was predictive for both potency at A2A and selectivity
over A1. Furthermore, traditional QSAR observations
indicated that several of the more potent compounds
contained terminal polar functionalities. This led us to
explore the use of molecular dipole moment as an
additional physical descriptor.
To generate the models we first obtained the X-ray
crystal structure for adenosine30 (Cambridge ref code
) ADENOS01) and NECA31 (Cambridge ref code )
KEMYEG). Through the use of the AGGREGATE func-
tion within SYBYL, we fixed these low-energy confor-

Novel A2A Adenosine Receptor Agonists
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4 535
F igu r e 1. Plot of final non-cross-validated observed versus
predicted for A2A model with no independent variables for the
training set.
F igu r e 2. Plot of final non-cross-validated observed versus
predicted for A2A model with A1 and dipole as independent
variables for the training set.
mations and subsequently added the C2 substitutions
described in Tables 1 and 2. All compounds were aligned
through the root-mean-square (rms) fit of N3, C6, N7,
and N9 of their respective purine rings. Before perform-
ing partial-least-squares (PLS) analysis on the com-
pounds, it was necessary to normalize the Ki values
because they were obtained from multiple sources. For
consistency, NECA was assigned a Ki of 10 nM and all
compounds were adjusted appropriately. To allow for
selectivity, A2A values were incorporated into the mo-
lecular spreadsheet as -log(Ki), whereas A1 data was
incorporated as log(Ki). This approach represents a
mathematically valid and useful method for generating
receptor subtype-selective CoMFA models. In this method
the displayed CoMFA fields represent potential modi-
fications/substitutions requisite to enhance A2A potency
and selectivity over A1 receptors simultaneously.
After molecular alignment, PLS analysis with cross-
validation led to the optimum number of components
(4, 4, 5, 6) and r2 (0.622, 0.541, 0.653, 0.646) for the
cv
A1, A2A, A1/A2A, and A1/A2A/dipole models, respectively.
For the latter two models that included additional
descriptors (A1 data and molecular dipole), those values
were incorporated as additional independent variables
along with the CoMFA value, while A2A was maintained
as the dependent variable. Using the optimum number
of components, non-cross-validated PLS analysis was
performed and resulted in r2 values of 0.877, 0.817,
0.863, and 0.890, respectively. Through the addition of
A1 receptor binding and molecular dipole moment, the
r2 for the A2A model increases by 0.073 and the standard
error of estimate decreased by nearly 0.1 log unit. This
data is graphically represented in Figures 1 and 2.
The most important part of validating a model is the
evaluation of that model for accuracy in predicting the
F igu r e 3. Plot of final non-cross-validated observed versus
predicted for A2A model with no independent variables for the
test set.
The difference in predictive powers of the model that
includes A1 values as well as molecular dipole moment
is also apparent when viewing the CoMFA contour maps
that are generated, as seen in Figures 5 and 6. There is
a noticeable increase in the amount of sterically re-
stricted space as indicated by the larger yellow fields
as well as the new presence of sterically and electro-
statically favorable fields near the 5′ position of the
adenosine backbone.
test set. The test set clearly indicates that the A1/A2A
/
dipole model has far greater predictive powers (r2
)
Discu ssion a n d Con clu sion s
0.626) in comparison to the model of only A2A data (r2
) 0.438). Furthermore, the model predicts high binding
affinity for the four novel compounds synthesized, 109-
112, which is confirmed by the experimental binding
values. This data is summarized in Figures 3 and 4.
We have successfully generated a CoMFA model (r2
) 0.890) based on 78 structurally diverse C2-modified
adenosine derivatives. Our model, which is predictive
for A2A activity, also allows for selectivity against the

536 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4
Rieger et al.
F igu r e 6. Steric and electrostatic CoMFA contour maps for
A2A model with A1 and dipole as independent variables with
compounds 10 (pink) and 70 (gray).
carbon of the purine ring indicating decreased activity
with compounds that explore that area of space. This
supports the hypothesis that there is a narrow pocket
that opens up to a much larger binding region with
which many of the C2-substituted compounds interact.
Thus, while compounds with short linkers have poorer
A2A activity, they are generally more A1-active. The
model also appears to strongly suggest that the NECA
modification does indeed make a contribution to the
potency/selectivity of A2A agonists. This is evident from
the new CoMFA contour fields generated near the
adenosine 5′ region in the A1/A2A/dipole model.
F igu r e 4. Plot of final non-cross-validated observed versus
predicted for A2A model with A1 and dipole as independent
variables for the test set.
To further validate our model we also embarked on
the synthesis and evaluation of four compounds, 109-
112, from a novel structural class. Our experimentally
determined normalized A2A binding values (59, 13, 18,
and 1.4 nM, respectively) correlate quite well with our
predicted values (10, 1.5, 1.2, and 0.34 nM, respectively)
from our CoMFA model. Thus, our model will continue
to be quite useful in our quest to design additional
potent and selective A2A agonists. We plan to continue
this work through the examination of the functional
utility of these compounds as mediators of ischemia-
reperfusion injury and inflammation. We are also cur-
rently examining these compounds in cloned human
receptors to examine species differences and will report
on these results shortly.
F igu r e 5. Steric and electrostatic CoMFA contour maps for
A2A model with no independent variables with compounds 10
(pink) and 70 (gray).
A1 receptor subtype through incorporation of A1 Ki
values as an additional independent variable. This
represents the first A2A adenosine receptor CoMFA
model that also incorporates a selectivity factor against
A1 binding as well as the first CoMFA model for
adenosine A2A agonists. Furthermore, we found that the
molecular dipole moment is a useful physical descriptor
that enhances the models predictive ability. We have
used this model to predict the A2A rat striatum Ki values
for an additional 24 compounds not found in the training
set. While the model that includes A1 Ki values and
molecular dipole moment led to only a modest difference
in r2 for the training set, a pronounced enhancement
was apparent for the test set (r2 ) 0.626 vs r2 ) 0.438).
Our model also very accurately distinguishes between
ortho-, meta-, and para-substituted phenyl derivatives
20-22, 24-31, and 37-43 as well as a large diversity
of functionalized and nonfunctionalized hydrocarbons.
In comparing the CoMFA contour plots for the two
models, the impact of the A1 data is readily apparent
in the larger yellow field directly adjacent to the C2
Exp er im en ta l Section
All melting points were determined with a Thomas-Hoover
capillary melting point apparatus and are uncorrected. Nuclear
magnetic resonance spectra for proton (1H NMR) were recorded
on a 300-MHz GE spectrophotometer. The chemical shift
values are expressed in ppm (parts per million) relative to
tetramethylsilane. For data reporting: s ) singlet, d ) doublet,
t ) triplet, q ) quartet, and m ) multiplet. Mass spectra were
measured on a Finnigan LcQ Classic. High-resolution mass
spectrometry (HRMS) data was provided by the Nebraska
Center for Mass Spectrometry. Analytical HPLC was done on
a Waters 2690 separation module with a Waters symmetry
C8 (2.1 × 150 mm) column operated at room temperature.
Compounds were eluted at 200 µL/min with 70:30 acetonitrile/
water, containing 0.5% acetic acid, with UV detection at 214
nm using a Waters 486 tunable detector. Preparative HPLC
was performed on a Shimadzu Discovery HPLC with a Shim-
pack VP-ODS C18 (20 × 100 mm) column operated at room

Novel A2A Adenosine Receptor Agonists
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4 537
temperature. Compounds were eluted at 30 mL/min with a
gradient 20-80% of water (containing 0.1% TFA) to methanol
over 15 min with UV detection at 214 nm using a SPD10A
VP tunable detector. All final compounds presented here were
determined to be greater than 98% pure by HPLC. Flash
chromatography was performed on silica 60A gel (230-400
mesh) using the method of Still.32 Analytical thin-layer chro-
matography was done on Merck Kieselgel 60 F254 aluminum
sheets. All reactions were done under a nitrogen atmosphere
in flame-dried glassware unless otherwise stated.
δ 3.45 (d, J ) 6.2, 2 H), 2.10 (d, J ) 6.2, 2 H), 1.9 (s, 1 H),
1.94-1.69 (m, 4 H), 1.52-1.34 (m, 2 H), 1.16-0.83 (m, 4 H);
13C NMR (CDCl3) δ 83.8, 69.5, 69.0, 40.8, 37.7, 32.3, 29.7, 26.5.
(4-P r op -2-yn ylcycloh exyl)m eth yl Aceta te (87). To a
solution of 960 mg (6.31 mmol) of 86 in 6 mL DMF were added
0.62 mL (7.57 mmol) pyridine and 0.78 mL (8.27 mmol) acetic
anhydride. The reaction was allowed to stir overnight at room
temperature. After 16 h some starting material still remained.
The reaction was then heated at 75 °C for 3 h. The solvent
was then removed under reduced pressure to yield a yellow
oil which was purified by flash chromatography on silica gel
eluting with 1:3 ether/petroleum ether to yield 1.12 g (91%) of
87 as an oil: 1H NMR (CDCl3) δ3.87 (d, J ) 6.2 Hz, 2 H), 2.06
(d, J ) 4.3 Hz, 2 H), 2.03 (s, 3 H), 1.98-1.93 (m, 1 H), 1.92-
1.83 (m, 2 H), 1.83-1.74 (m, 2 H), 1.63-1.36 (m, 2 H), 1.12-
0.90 (m, 4 H); 13C NMR (CDCl3) δ 171.7, 83.7, 69.9, 69.6, 37.4,
37.3, 32.1, 29.7, 26.5, 21.4; APCI m/z (rel intensity) 195 (M+,
30), 153 (M+, 70), 135 (M+, 100).
4-P r op -2-yn ylcycloh exa n eca r boxylic Acid (88). A solu-
tion of chromium trioxide (600 mg, 6.0 mmol) in 1.5 M H2SO4
(2.6 mL, 150 mmol) was cooled to 5 °C and added to a solution
of 86 (280 mg, 1.84 mmol) in acetone (15 mL). The mixture
was allowed to warm to room temperature and allowed to stir
overnight. 2-Propanol (4 mL) was added to the green/black
solution, which turned light blue after 1 h. After adding water
(15 mL), the solution was extracted with CHCl3 (6 × 25 mL).
The organics were pooled and concentrated in vacuo to yield
a white solid. The solid was dissolved in ether (50 mL) and
extracted with 1 M NaOH (2 × 30 mL). The basic extracts
were pooled, acidified with 10% HCl, and re-extracted with
ether (3 × 30 mL). The ether layers were then dried with
sodium sulfate and concentrated in vacuo to yield a white solid
which was recrystallized from acetone/water to yield 88 (222
mg, 73%) as white needles: mp 84-85 °C; 1H NMR (CDCl3) δ
2.30-2.23 (m, 1 H), 2.17-2.11 (m, 2 H), 2.07-2.03 (m, 2 H),
1.97-1.91 (m, 3H), 1.51-1.39 (m, 3 H), 1.13-1.01 (m, 2 H);
13C NMR (CDCl3) δ 182.5, 83.8, 69.6, 40.7, 37.7, 32.3, 29.6,
26.5; APCI m/z (rel intensity) 165 (M-, 100).
Meth yl 4-P r op -2-yn ylcycloh exa n eca r boxyla te (89). To
a solution of 88 (240 mg, 1.45 mmol) in 7:3 CH2Cl2/MeOH (10
mL) was added TMS-diazomethane (2.0 M in hexanes) (0.9
mL, 1.8 mmol) in 0.2-mL aliquots until the color remained
yellow. The reaction was allowed to stir for an additional 0.25
h at room temperature after which time glacial acetic acid was
added dropwise until the solution became colorless. The
reaction was then concentrated in vacuo to an oil which was
purified by flash chromatography on silica gel using ether/
petroleum ether (1:9) to yield 89 (210 mg, 80%) as a clear oil:
1H NMR (CDCl3) δ 3.60 (s, 3H), 2.25-2.13 (m, 1 H), 2.08-
1.94 (m, 3 H), 1.95-1.90 (m, 2 H), 1.49-1.31 (m, 3 H), 1.10-
0.93 (m, 2 H); 13C NMR (CDCl3) δ 176.7, 83.3, 69.8, 51.9, 43.4,
36.7, 31.9, 29.2, 26.3; APCI m/z (rel intensity) 181 (MH+, 100).
Gen er a l Meth od for th e P r ep a r a tion of 2-AAs 109-112.
To a solution of 2-iodo-NECA (0.025 mmol) and terminal
alkyne (0.045 mmol) in DMF/acetonitrile (0.5 mL/1 mL) were
added TEA (0.76 mmol), Pd(PPh3)4 (0.0038 mmol), CuI (0.0046
mmol), and PPh3 (0.0038 mmol). The clear yellow solution was
heated to 60 °C for 12 h under a nitrogen atmosphere to yield
[4-(ter t-Bu t yld im et h ylsila n yloxym et h yl)cycloh exyl]-
m eth a n ol (83). To a 100-mL flask containing 79 (4.0 g, 27.8
mmol) in DMF (40 mL) were added TBDMSCl (3.56 g, 23.6
mmol) and imidazole (3.79 g, 55.6 mmol). The reaction was
allowed to stir at 25 °C for 16 h after which time saturated
aqueous LiBr (50 mL) was added and the reaction extracted
with ether (2 × 50 mL). The ether layers were pooled and
extracted again with LiBr (2 × 35 mL) resulting in the ether
layer becoming a clear solution. The ether layer was then
concentrated in vacuo and flash chromatographed on a silica
gel column eluting with 1:2 ether/petroleum ether to yield 83
(3.80 g, 62%) as a homogeneous oil: 1H NMR (CDCl3) δ 3.46
(d, J ) 6.2 Hz, 2 H), 3.39 (d, J ) 6.2 Hz, 2 H), 1.95-1.72 (m,
4 H), 1.65 (m, 1 H), 1.40 (m, 1 H), 1.03-0.89 (m, 4 H), 0.88 (s,
9 H), 0.04 (s, 6 H); 13C NMR (CDCl3) δ 69.2, 69.1, 41.2, 41.1,
29.5, 26.5, 18.9, -4.8; APCI m/z (rel intensity) 259 (MH+, 100).
Tolu en e-4-su lfon ic Acid 4-(ter t-Bu tyld im eth ylsila n yl-
oxym eth yl)cycloh exylm eth yl Ester (84). To a 100-mL flask
containing 83 (3.4 g, 13.2 mmol) in CHCl3 (30 mL) were added
tosyl chloride (3.26 g, 17.1 mmol) and pyidine (3.2 mL, 39.6
mmol). The reaction was allowed to stir at 25 °C for 14 h after
which time the reaction was concentrated in vacuo to yield a
wet white solid. To this solid was added ether (50 mL) and
the solid was filtered and subsequently washed with additional
ether (2 × 50 mL). The ether layers were pooled, concentrated
in vacuo to yield a clear oil which was flash chromatographed
on a silica gel column eluting with 1:4 ether/petroleum ether
to yield 84 (4.5 g, 83%) as a white solid: 1H NMR (CDCl3) δ
7.78 (d, J ) 7.7, 2 H), 7.33 (d, J ) 7.7 Hz, 2 H), 3,81 (d, J )
6.2 Hz, 2H), 3.37 (d, J ) 6.2, 2 H), 2.44 (s, 3 H), 1.95-1.72 (m,
4 H), 1.65 (m, 1 H), 1.40 (m, 1 H), 1.03-0.89 (m, 4 H), 0.88 (s,
9 H), 0.04 (s, 6 H); 13C NMR (CDCl3) δ 145.1, 133.7, 130.3,
128.4, 75.8, 68.9, 40.7, 38.0, 29.1, 26.5, 22.1, 18.9, -4.9; APCI
m/z (rel intensity) 413 (MH+, 100).
(4-P r op -2-yn ylcycloh exyl)m eth a n ol (86). A three-neck
250-mL flask equipped with a gas inlet tube and dry ice
condenser was cooled to -78 °C and charged with liquid
ammonia (40 mL). To the reaction mixture was added lithium
wire (600 mg, 86.4 mmol) generating a deep blue solution.
Subsequently, the mixture was allowed to stir for 1 h.
Acetylene, that was passed through a charcoal drying tube,
was added to the ammonia until all the lithium had reacted
and the solution turned colorless, at which time the flow of
acetylene was stopped, the acetylene-inlet tube and condenser
removed and the flask outfitted with a thermometer. DMSO
(20 mL) was added and the ammonia evaporated with a warm
water bath until the mixture reached a temperature of 30 °C
and stirred at this temperature for 2 h until the solution
stopped bubbling. The mixture was then cooled to 5 °C and
84 (11.25 g, 27.3 mmol) was added dissolved in DMSO (10 mL)
while maintaining the temperature at 5 °C. The mixture was
allowed to stir at 5 °C for 0.5 h and then gradually warmed to
room temperature and stirred for an additional 18 h. The
brown/black reaction mixture was poured slowly over ice (300
g) and extracted with ether (4 × 100 mL), dried with
anhydrous sodium sulfate, and concentrated in vacuo to yield
a yellow oil. The oil was subsequently dissolved in THF (200
mL) and changed to a brownish color upon addition of TBAF
hydrate (11.20 g, 35.5 mmol). The solution was allowed to stir
for 24 h under N2 atmosphre, after which time it was quenched
with water (200 mL) and extracted with ether (3 × 100 mL).
The ether was then concentrated in vacuo and chromato-
graphed on a silica gel column eluting with 1:1 ether/petroleum
ether to yield 86 (3.91 g, 93%) as a yellow oil: 1H NMR (CDCl3)
a
brown/black solution. After the starting material was
completely consumed, as judged by TLC (CHCl3/MeOH 4:1),
the solvent was removed in vacuo to yield a syrup that was
purified via silica gel column chromatography using 6% MeOH/
CHCl3 to yield pure products.
4-{3-[6-Am in o-9-(5-et h ylca r b a m oyl-3,4-d ih yd r oxyt et -
r a h y d r o fu r a n -2 -y l )-9 H -p u r i n -2 -y l ]p r o p -2 -y n y l }-
cycloh exa n eca r boxylic Acid (109). The reaction of 110 with
5 equiv of LiOH in THF/water for 6 h gave 109 (7 mg, 72%) as
a white solid which was crystallized from MeOH/H2O (0.1%
TFA) after purification by reverse phase HPLC: 1H NMR
(DMSO-d6) δ 8.70 (s, 1 H), 8.41 (s, 1 H), 7.62 (s, 2 H), 5.89 (d,
J ) 7.25 Hz, 1 H), 4.53 (m, 1 H), 4.27 (s, 1 H), 4.08 (d, J ) 3.6
Hz, 1 H), 2.29 (d, J ) 6.4 Hz, 2 H), 2.15-1.99 (m, 1 H), 1.92-
1.76 (m, 4 H), 1.52-1.38 (m, 1 H), 1.38-1.19 (m, 2 H), 1.02 (t,
J ) 6.3 Hz 3 H); 13C NMR (DMSO-d6) 176.7, 169.2, 155.6,

538 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4
Rieger et al.
148.9, 145.2, 141.6, 119.0, 87.7, 85.0, 84.6, 81.6, 73.1, 71.9, 43.2,
35.9, 33.3, 31.2, 28.3, 25.6, 15.0; HRMS (FAB) m/z 474.2196
[(M + H)+ cacld for C22H29N6O6 474.2182].
fashion, adenosine was used as the core for construction of
derivatives 1-36, 63-77, 90-94, and 104-108. The corre-
sponding X-ray portion of each analogue was held constant
using AGGREGATE and the added functional groups were
energy-minimized with the Tripos force field,33 without solvent,
using default bond angles and distances while neglecting
electrostatics. In this manner, we obtained the lowest-energy
conformation for each compound in Tables 1 and 2. In cases
where long alkyl chains with multiple conformations were
possible, the extended chain conformation was chosen.
Ch a r ge Ca lcu la tion s. All applied Coulson charges were
calculated using AM1 (MOPAC) within SYBYL. The atomic
charges were calculated in the singlet state with net charge
equal to zero, keyword NOINTER, and normal convergence
without geometry optimization and applied to the correspond-
ing minimized molecular structure.
4-{3-[6-Am in o-9-(5-et h ylca r b a m oyl-3,4-d ih yd r oxyt et -
r a h y d r o fu r a n -2 -y l )-9 H -p u r i n -2 -y l ]p r o p -2 -y n y l }-
cycloh exa n eca r boxylic Acid Meth yl Ester (110). The
reaction of 89 with 2-iodo-NECA under the general conditions
described above gave 110 (74 mg, 60%) as a white solid: 1H
NMR (CD3OD) δ 8.23 (s, 1 H), 5.92 (d, J ) 7.7 Hz, 1 H), 4.69-
4.65 (dd, J ) 7.7 Hz, 4.6 Hz, 1 H), 4.40 (s, 1 H), 4.24 (d, J )
4.6 Hz, 1 H), 3.59 (s, 3 H), 3.49-3.31 (m, 2 H), 2.31 (d, J ) 6.6
Hz, 2 H), 2.10-2.09 (m, 1 H), 2.01-1.89 (m, 4 H), 1.61-1.32
(m, 5 H), 1.13 (t, J ) 7.3 Hz, 3 H); 13C NMR (CD3OD) δ 177.1,
171.1, 156.3, 149.3, 146.7, 142.4, 119.7 89.6, 86.0, 85.5, 81.6,
74.0, 72.2, 51.2, 43.2, 36.8, 34.2, 31.8, 28.9, 26.2, 14.4; HRMS
(FAB) m/z 487.2325 [(M + H)+ cacld for C23H31N6O6 487.2305].
Acetic Acid 4-{3-[6-Am in o-9-(5-eth ylca r ba m oyl-3,4-d i-
h ydr oxytetr ah ydr ofu r an -2-yl)-9H-pu r in -2-yl]pr op-2-yn yl}-
cycloh exylm eth yl Ester (111). The reaction of 87 with
2-iodo-NECA under the general conditions described above
gave 111 (78 mg, 62%) as a white solid: 1H NMR (CD3OD) δ
8.22 (s, 1 H), 5.92 (d, J ) 8.1 Hz, 1 H), 4.70-4.66 (dd, J ) 8.1
Hz, 4.6 Hz, 1 H), 4.40 (d, J ) 1.2 Hz, 1 H), 4.25-4.23 (dd, J )
4.6 Hz, 1.2 Hz, 1 H), 3.83 (d, J ) 6.5, 2 H), 3.53-3.31 (m, 2
H), 2.29 (d, J ) 6.5 Hz, 2 H), 1.97 (s, 3 H), 1.93-1.89 (m, 2 H),
1.79-1.75 (m, 2 H), 1.64-1.42 (m, 2 H), 1.12 (t, J ) 7.3 Hz, 3
H), 1.09-0.91 (m, 4 H); 13C NMR (CD3OD) δ 172.0, 171.2,
156.2, 149.3, 146.7, 142.5, 119.7, 89.6, 86.3, 85.5, 81.5, 74.0,
72.2, 69.6, 37.4, 37.2, 34.2, 32.1, 29.4, 26.4, 19.9, 14.5; HRMS
(FAB) m/z 501.2469 [(M + H)+ cacld for C24H33N6O6 501.2462].
5-{6-Am in o-2-[3-(4-h yd r oxym eth ylcycloh exyl)p r op -1-
yn yl]p u r in -9-yl}-3,4-d ih yd r oxyt et r a h yd r ofu r a n -2-ca r -
boxylic Acid Eth yla m id e (112). The reaction of 86 (30 mg,
0.2 mmol) with 2-iodo-NECA (28 mg, 0.07 mmol) under the
general conditions described above gave 112 (7 mg, 24%) as a
white solid: 1H NMR (CD3OD) δ 8.22 (s, 1 H), 5.92 (d, J ) 7.7
Hz, 1 H), 4.70-4.66 (dd, J ) 7.7 Hz, 4.8 Hz, 1 H), 4.40 (d, J )
1.2 Hz, 1 H), 4.25-4.23 (dd, J ) 4.8 Hz, 1.2 Hz, 1 H), 3.51-
3.37 (m, 2 H), 3.31 (d, J ) 6 Hz, 2 H), 2.30 (d, J ) 6.8 Hz, 2
H), 1.94-1.89 (m, 2 H), 1.83-1.78 (m, 2 H), 1.64-1.42 (m, 2
H), 1.12 (t, J ) 7.3 Hz, 3 H), 1.09-0.91 (m, 4 H); 13C NMR
(CD3OD) δ 170.3, 155.4, 148.5, 146.0, 141.6, 118.8, 88.7, 85.5,
84.6, 80.6, 73.1, 71.3, 66.8, 39.6, 36.9, 33.3, 31.5, 28.6, 25.6,
13.5; HRMS (FAB) m/z 459.2373 [(M + H)+ cacld for C22H31N6O5
459.2356].
Molecu la r Mod elin g a n d CoMF A Ca lcu la tion s. CoMFA
values, using default parameters except where noted, were
calculated through the QSAR option of SYBYL 6.5 on a silicon
graphics Octane computer. The CoMFA grid spacing was 2.0
Å in the x, y, and z directions, and the grid region was
automatically generated by the CoMFA routine to encompass
all molecules with an extension of 4.0 Å in each direction. A
sp3 carbon (for sterics) and a charge of +1.0 (for electrostatics)
were used as probes to generate the interaction energies at
each lattice point. The default value of 30 kcal/mol was used
as the maximum electrostatic and steric energy cutoff.
Nor m a liza tion of Biologica l Da ta . NECA (78) was
normalized to a value of 10 nM, and all others compounds were
adjusted appropriately by dividing or multiplying the ana-
logue’s experimental IC50 by the appropriate scaling factor. The
same normalization protocol was done for both A1 and A2A
receptor data.
Molecu la r Align m en t. All conformations for each com-
pound in the training set (Table 1) were aligned with the FIT
routine within SYBYL without changing the low-energy
conformation previously obtained. All compounds were rms fit
by aligning each atom of the adenine ring of the analogue to
that of adenosine.
P a r tia l Lea st Squ a r es (P LS) Regr ession An a lysis.
Cross-validated and non-cross-validated PLS analyses for A1
and A2A receptors were performed within the SYBYL/QSAR
routine. Separate cross-validations of the dependent columns
log(adenosine A1 IC50) and -log(adenosine A2A IC50) within the
CoMFA columns were performed without column filtering. In
an additional experiment, we also constructed a model that
contained multiple independent variables (dipole moment and
log(A1 IC50) while maintaining -log(A2A IC50) as the dependent
variable). As an example of our method for obtaining the non-
cross-validated results, the analysis for adenosine A2A was
scaled by the CoMFA standard deviation and generated an
optimum number of components equal to 4 and r2 ) 0.541.
PLS analysis with non-cross-validation, performed with 4
components and 2.0 kcal/mol column filtering, resulted in
dropping 2886 of 3146 columns and gave a standard error of
estimate of 0.364, with a probability of r2 ) 0 (n1 ) 4, n2 ) 70)
equal to 0.000, an F value of 78.242, and a final r2 ) 0.817.
The relative steric 0.520 and electrostatic 0.480 contributions
to the adenosine A2A model were contoured as the standard
deviation multiplied by the coefficient at 80% for favored steric
(contoured in green) and favored positive electrostatic (con-
toured in blue) effects and at 20% for disfavored steric
contoured in yellow and favored negative electrostatic (con-
toured in red) effects, as shown in Figure 5. In a similar
manner, we derived a CoMFA model using adenosine A1 IC50
as well as adenosine A1 IC50/dipole as descriptors. We ulti-
mately used the model generated from the adenosine A1 IC50
/
dipole descriptors, and the resulting CoMFA electrostatic and
steric maps are shown in Figure 6. Using the appropriate
CoMFA model and PREDICT PROPERTY in the QSAR
module of SYBYL, we predicted adenosine A1 and A2A IC50
s
using the appropriate model for compounds 90-113.
Recep tor Bin d in g Da ta . Cortex and striatum were dis-
sected from rats brain. They were rinsed with phosphate-
buffered saline, pH 7.4. There were then homogenized in HE
buffer (10 mM Na-HEPES, pH 7.4, 1 mM EDTA), plus a
mixture of protease inhibitors (100 µM PMSF, 100 µM ben-
zamidine, 2 µM leupeptin, 2 µM pepstatin A, 2 µM aprotinin)
and centrifuged at 20000g for 20 min. Pellets were resus-
pended and washed twice in the HE buffer plus protease
inhibitors. Finally, the pellet was resuspended in the same
buffer plus 10% sucrose and frozen until use.
Da ta An a lysis. Statistical correlations were performed
using J andel Scientific Software. All equations and correlations
were calculated through the use of linear regressions based
on the normalized data using SigmaStat for Windows Version
1.0. The plots shown in Figures 1-4 were constructed using
SigmaPlot version 1.02a to which the best-fit linear regression
line was added.
Con for m a tion a l An a lysis. We obtained X-ray structures
from the Cambridge Database for adenosine and NECA and
used these as representative low-energy conformations. The
X-ray structure of NECA was used as a core to construct all
NECA derivatives 37-62, 95-103, and 109-112. In a similar
For radioligand binding, striatum and cortex membranes
were diluted to 25 µg/50 µL of striatum or 40 µg/50 µL of cortex
and aliquoted in triplicate. Membranes were incubated with
5 mM Mg, 1 unit/mL adenosine deaminase and either 0.5 nM
[
125I]APE (striatum) or 125I]ABA (cortex) and increasing
[
concentrations of competing agonist or antagonist for 3 h at
room temperature. Tubes were filtered using a Brandel
harvester and counted for 1 min in a gamma counter.
Abbr evia tion s: CGS 21680, 2-[[4-(2-carboxyethyl)phenyl]-

Novel A2A Adenosine Receptor Agonists
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4 539
ethylamino]-5′-N-ethylcarbamoyladenosine; CPA, N6-cyclopen-
tyladenosine; DMSO, dimethyl sulfoxide; [125I]ABA, 2-[2-(4-
amino-3-[125I]iodophenyl)ethylamino]adenosine; [125I]APE, [125I]-
aminobenzyladenosine; NECA (5′-N-ethylcarboxamidoadeno-
sine); PPh3, triphenylphosphine; SAR, structure-activity re-
lationship: TBAF, tetrabutylammonium fluoride.
(14) Camaioni, E.; Di Francesco, E.; Vittori, S.; Volpini, R.; Cristalli,
G. Adenosine Receptor Agonists: Synthesis and Biological
Evaluation of the Diastereomers of 2-(3-Hydroxy-3-pheynyl-1-
propyn-1-yl)NECA. Bioorg. Med. Chem. 1997, 5, 2267-2275.
(15) Hutchinson, A. J .; Williams, M.; de J esus, R.; Yokoyama, R.; Oei,
H. H.; Ghai, G. R.; Webb, R. L.; Zoganas, H. C.; Stone, G. A.;
J arvis, M. F. 2-(Arylalkylamino)adenosin-5′-uronamides: A new
Class of Highly Selective Adenosine A2 Receptor Ligands. J . Med.
Chem. 1990, 33, 1919-1924.
(16) Matsuda, A.; Shinozaki, M.; Yamaguchi, T.; Homma, H.; Nomoto,
R.; Miyasaka, T.; Watanabe, Y.; Abiru, T. Nucleosides and
Nucleotides. 103. 2-Alkynyladenosines: A Novel Class of Selec-
tive Adenosine A2 Receptor Agonists with Potent Antihyperten-
sive Effects. J . Med. Chem. 1992, 35, 241-252.
(17) Cristalli, G.; Eleuteri, A.; Vittori, S.; Volpini, R.; Lohse, M. J .;
Klotz, K. 2-Alkynyl Derivatives of Adenosine and Adenosine-5′-
N-ethyluronamide as Selective Agonists at A2 Adenosine Recep-
tors. J . Med. Chem. 1992, 35, 2363-2368.
(18) Graaf, P. H. V.; Nilsson, J .; Schaick, E. A. V.; Danhof, M.
Multivariate Quantitative Structure-Pharmacokinetic Relation-
ships (QSPKR) Analysis of Adenosine A1 Receptor Agonists in
Rat. J . Pharm. Sci. 1999, 88, 306-312.
Ack n ow led gm en t. This research was made possible
by the University of Virginia Department of Chemistry
as well as through funding provided by RO1 HL37942.
M.L.B. thanks the UNCF/Merck Science Institute for
funding support. We also thank The University of
Nebraska Center for Mass Spectrometry for performing
the HRMS analyses. Further appreciation is extended
to Christine Dieckhaus for LC/MS work and Melissa
Marshall for expert technical assistance with the recep-
tor binding assays.
Su p p or tin g In for m a tion Ava ila ble: Tables summariz-
ing the predicted, actual, and residual values for the training
and test sets as well as details on the PLS analyses. This
material is available free of charge via the Internet at http://
pubs.acs.org.
(19) Baraldi, P. G.; Borea, P. A.; Bergonzoni, M.; Cacciari, B.; Ongini,
E.; Recanatini, M.; Spaluto, G. Comparative Molecular Field
Analysis (CoMFA) of a Series of Selective Adenosine Receptor
A2A Antagonists. Drug Dev. Res. 1999, 46, 126-133.
(20) Moro, S.; Li, A.; J acobsen, K. Molecular Modeling Studies of
Human A3 Adenosine Antagonists: Structural H omolgy and
Receptor Docking. J . Chem. Inf. Comput. Sci. 1998, 38, 1239-
1248.
Refer en ces
(21) Li, A.; Moro, S. Forsyth, N.; Melman, N.; J i, X.; J acobson, K. A.
Synthesis, CoMFA Analysis, and Receptor Docking of 3,5-Diacyl-
2,4-Dialkylpyridine Derivatives as Selective A3 Adenosine Re-
ceptor Antagonists. J . Med. Chem. 1999, 42, 706-721.
(22) Moro, S.; van Rhee, A. M.; Sanders, L. H.; J acobson, K. A.
(1) Muller, C. E.; Scior, T. Adenosine Receptors and Their Modula-
tors. Pharm. Acta Helv. 1993, 68, 77-111.
(2) Ralevic, V.; Burnstock, G. Receptors for Purines and Pyrim-
idines. Pharmacol. Rev. 1998, 50, 413-492.
(3) DeNinno, M. P. Adenosine. Annu. Rep. Med. Chem. 1998, 33,
111-121.
(4) Mubagwa, K.; Mullane, K.; Flameng, W. Role of Adenosine in
the Heart and Circulation. Cardiovasc. Res. 1996, 32, 797-813.
(5) Okusa, M. D.; Linden, J .; Macdonald, T.; Huang, L. Selective
Flavanoid Derivatives as Adenosine Receptor Antagonists:
A
Comparasion of the Hypothetical Receptor Binding Site Based
on a Comparative Molecular Field Analysis Model. J . Med.
Chem. 1998, 41, 46-52.
A2A adenosine receptor activation reduces ischemia-reperfusion
(23) Li, A.; Moro, S.; Melman, N.; J i, X.; J acobson, K. A. Structure-
Activity Relationships and Molecular Modeling of 3,5-Diacyl-
2,4-dialkylpyridine Derivatives as Selective A3 Adenosine Re-
ceptor Antagonists. J . Med. Chem. 1998, 41, 3186-3201.
(24) Siddiqi, S. M.; Pearlstein, R. A. Sanders, L. H.; J acobson, K. A.
Comparative Molecular Field Analysis of Selective A3 Adenosine
Receptor Agonists. Bioorg. Med. Chem. 1995, 3, 1331-1343.
(25) McDougal, P. G.; Rico, J . G.; Oh, Y.; Condon, B. D. A Convenient
Procedure for the Monosilylation of Symmetric 1,n-Diols. J . Org.
Chem. 1986, 51, 3388-3390.
(26) Kabalka, G. W.; Varma, M.; Varma, R. S. Tosylation fo Alcohols.
J . Org. Chem. 1986, 52, 2486-2388.
(27) Manchand, P. S.; Schwartz, A.; Wolff, S.; Belica, P. S.; Madan,
P.; Patel, P.; Saposnik, S. J . Synthesis of an Orally Active PAF
Antagonist of The N-[4-(3-Pyridinyl)butyl]pentadeienamide Class.
Heterocycles 1993, 35, 1350-1369.
injury in rat kidney. AJ P: Renal 1999, 277, F404-F412.
(6) Cronstein, B. N. Adenosine, an Endogenous antiinflammatory
agent. J . Appl. Physiol. 1994, 76, 5-13.
(7) Grondin, R.; Bedard, P. J .; Tahar, A. H.; Bregoire, L.; Mori, A.;
Kase, H. Antiparkinsonian effect of a new selective adenosine
A2A receptor antagonist in MPTP-treated monkeys. Neurology
1999, 52, 1673-1677.
(8) Dixon, D. A.; Fenix, L. A.; Kim, D. M.; Raffa, R. B. Indirect
Modulation of Dopamine D2 Receptors as Potential Pharmaco-
therapy for Schizophrenia: I. Adenosine Agonists. Ann. Phar-
macol. 1999, 33, 480-488.
(9) McPherson, J . A.; Bishop, G G.; Sanders, J . M.; Rieger, J . M.;
Heselbacher, S. E.; Gimple, L. W.; Powers, E. R.; Macdonald,
T.; Sullivan, G.; Linden, J .; Sarembock, I. J . Adenosine A2A
Receptor Stimulation Reduces Inflammation and Neointimal
Growht in
submitted.
a Murine Carotid Ligation Model. Circulation,
(28) Luthin, D. R.; Olsson, R. A.; Thompson, R. D.; Sawmiller, D. R.;
Linden, J . Characterization of Two Affinity States of Adenosine
A2A Receptors with a New Radioligand, 2-[2-(4-amino-3-[125I]-
iodophenyl)ethylamino]adenosine. Mol. Pharmacol. 1995, 47,
307.
(10) Cristalli, G.; Camaioni, E.; Vittori, S.; Volpini, R.; Borea, P. A.;
Conti, A.; Dionisotti, S.; Ongini, E.; Monopoli, A. 2-Aralkynyl
and 2-Heteroalkynyl derivatives of Adenosine-5′N-ethylurona-
mide as Selective A2A Adenosine Receptor Agonists. J . Med.
Chem. 1995, 38, 1462-1472.
(11) Daly, J . W.; Padgett, W. L.; Secunda, S. I.; Thompson, R. D.;
Olsson, R. A. Structure-Activity relationships for 2-substituted
Adenosines at A1 and A2 Adenosine Receptors. Pharmacology.
1993, 46, 91-100.
(12) Abiru, T.; Miyashita, T.; Watanabe, Y.; Yamaguchi, T.; Machida,
H.; Matsuda, A. Nucleosides and Nucleotides. 107. 2-(Cycloal-
kylalkynyl)adenosines: Adenosine A2 Receptor Agonists with
Potent Antihypertensive Effects. J . Med. Chem. 1992, 35, 2253-
2260.
(13) Vittori, Z.; Camaioni, E.; Di Francesco, E.; Volpini, R.; Monopoli,
A.; Dionisotti, S.; Ongini, E.; Cristalli, G. 2-Alkenyl and 2-Alyky
Derivatives of Adenosine and Adenosine-5′-N-Ethyluronamide:
Different Afinity and Selectivity of E- and Z-Diastereomers at
(29) Linden, J .; Patel, A.; Sadek, S. [125I]aminobenyzladenosine, a
New Radioligand with Improved Specific Binding to Adenosine
Receptors in Heart. Circ. Res. 1985, 56, 279.
(30) Klooster, W. T.; Ruble, J . R.; Craven, B. M.; McMullan, R. K.
Acta Crystallogr., Sect. B 1991, 47, 376.
(31) Moos, W. H.; Hamilton, H. W.; Ortwine, D. F.; Taylor, M. D.;
McPhail, A. T. Nuclesides Nucleotides 1989, 8, 449.
(32) Still, W. C.; Kahn, M.; Mitra, A. Rapid Chromatographic
Technique for Preparative Separations With Moderate Resolu-
tion. J . Org. Chem. 1978, 43, 2923-2925.
(33) Clark, M. D.; Cramer, R. D., III; Opdenbosch, N. V. J . Comput.
Chem. 1989, 10.
A2A Adenosine Receptors. J . Med. Chem. 1996, 39, 4211-4217.
J M0003642