Modeling the Adenosine Receptors: Comparison of the Binding Domains of A2A Agonists and Antagonists

Article abstract of DOI:10.1021/jm0300431

A three-dimensional model of the human A_2A adenosine receptor (AR) and its docked ligands was built by homology to rhodopsin and validated with site-directed mutagenesis and the synthesis of chemically complementary agonists. Different binding

Full text of DOI:10.1021/jm0300431

J . Med. Chem. 2003, 46, 4847-4859
4847
Mod elin g th e Ad en osin e Recep tor s: Com p a r ison of th e Bin d in g Dom a in s of A2A
Agon ists a n d An ta gon ists
Soo-Kyung Kim,† Zhan-Guo Gao,† Philippe Van Rompaey,§ Ariel S. Gross,† Aishe Chen,†
Serge Van Calenbergh,§ and Kenneth A. J acobson*,†
Molecular Recognition Section, Laboratory of Bioorganic Chemistry, National Institute of Diabetes and Digestive and Kidney
Diseases (NIDDK), National Institutes of Health (NIH), Bethesda, Maryland 20892 and Laboratorium voor Medicinale Chemie
(FFW), Harelbekestraat 72 B-9000 Gent, Belgium
Received J anuary 27, 2003
A three-dimensional model of the human A2A adenosine receptor (AR) and its docked ligands
was built by homology to rhodopsin and validated with site-directed mutagenesis and the
synthesis of chemically complementary agonists. Different binding modes of A2AAR antagonists
and agonists were compared by using the FlexiDock automated docking procedure, with manual
adjustment. Putative binding regions for the 9H-purine ring in agonist NECA 3 and the 1H-
[1,2,4]triazolo[1,5-c]quinazoline ring in antagonist CGS15943 1 overlapped, and the exocyclic
amino groups of each were H-bonded to the side chain of N6.55. For bound agonist, H-bonds
formed between the ribose 3′- and 5′-substituents and the hydrophilic amino acids T3.36, S7.42
,
and H7.43, and the terminal methyl group of the 5′-uronamide interacted with the hydrophobic
side chain of F6.44. Formation of the agonist complex destabilized the ground-state structure of
the A2AAR, which was stabilized through a network of H-bonding and hydrophobic interactions
in the transmembrane helical domain (TM) regions, facilitating a conformational change upon
activation. Both flexibility of the ribose moiety, required for the movement of TM6, and its
H-bonding to the receptor were important for agonism. Two sets of interhelical H-bonds involved
residues conserved among ARs but not in rhodopsin: (1) E131.39 and H2787.43 and (2) D522.50
,
with the highly conserved amino acids N2807.45 and S2817.46, and N2847.49 with S913.39. Most of
the amino acid residues lining the putative binding site(s) were conserved among the four AR
subtypes. The A2AAR/3 complex showed a preference for an intermediate conformation about
the glycosidic bond, unlike in the A3AR/3 complex, which featured an anti-conformation.
Hydrophilic amino acids of TMs 3 and 7 (ribose-binding region) were replaced with anionic
residues for enhanced binding to amine-derivatized agonists. We identified new neoceptor
(T88D)-neoligand pairs that were consistent with the model.
In tr od u ction
The adenosine receptors (ARs), classified as A1, A2A
2B, and A3 subtypes, represent a physiologically and
antagonists have been proposed as novel therapeutics
for Parkinson’s disease and may also be active as
cognition enhancers, neuroprotective and antiallergic
agents, analgesics, and positive inotropics.8-10 Although
the physiological effects mediated through the A2AAR
have been extensively investigated, only a few molecular
modeling studies, which used a low-resolution rhodopsin
template,4,11 have explored the binding properties of
agonists and antagonists at the A2AAR. The develop-
ment of more potent and/or selective A2AAR antagonists
and agonists is still being pursued intensively.
,
A
pharmacologically important family of G protein-coupled
receptors (GPCRs).1 ARs are important pharmacological
targets in the treatment of a variety of diseases because
of their key roles in controlling numerous physiological
processes. For example, many therapeutic agents under
development for treatment of central nervous system
disorders, inflammatory diseases, asthma, kidney fail-
ure, and ischemic injuries exert their effects via interac-
tions with ARs.
We are particularly interested in the ligand interac-
tions of the A2AAR because of current interest in its
relation to inflammatory conditions and movement
disorders2,3 and because the results of point mutagenesis
are known for this subtype.4-6 A2AAR agonists are also
potentially useful for the treatment of cardiovascular
diseases, such as hypertension, ischemic cardiomy-
opathy, inflammation, and atherosclerosis,7 and A2AAR
Extensive mutagenesis was carried out for both the
A1 and A2A ARs and to a lesser extent for the A2B and
A3 ARs.1,12 The retinal binding site of rhodopsin,13 a G
protein-coupled photoreceptor, and the putative ligand-
binding sites on ARs, as deduced by using mutational
analysis, overlap extensively. Most of the essential
residues required for recognition of AR agonists and/or
antagonists, which bind within the transmembrane
helical domains (TMs) 3, 5, 6, and 7,1 coincide largely
with the corresponding amino acids of the binding site
of cis-retinal in rhodopsin, although there are additional
interaction sites within TMs 6 and 7 of the ARs in
comparison with the binding site of rhodopsin (Table
1). Recently, the ground-state X-ray structure of rhodop-
sin with 2.8-Å resolution13 has advanced our under-
* Corresponding author: Dr. K. A. J acobson, Chief, Molecular
Recognition Section, Bldg. 8A, Rm. B1A-19, NIH, NIDDK, LBC,
Bethesda, MD 20892-0810, Tel.: 301-496-9024, Fax: 301-480-8422,
E-mail: kajacobs@helix.nih.gov.
† National Institutes of Health.
§ Laboratorium voor Medicinale Chemie.
10.1021/jm0300431 CCC: $25.00 © 2003 American Chemical Society
Published on Web 10/10/2003

4848 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23
Kim et al.
Ta ble 1. Mutational Analysis for Selected Residues of the ARs with Respect to Ligand Binding
a
b
rhodopsin A2A
mutational results
A1
mutational results
T: increased Ag affinity55
TM 1 A421.37
T11
E13
G14
E16
M441.39
Q: slight reduction of Ag but not Ant affinity24
A/Q: Ag affinity reduced 4- to 40-fold;
little change in Ant affinity25
P531.48
L591.54
L22
C28
Y43
S47
D52
P25
I31
C46
S50
D55
L: modest reduction of Ag affinity55
C: NC in radioligand binding55
A/S: NC in ligand binding56
A: NC in ligand binding25
TM2 Y742.41
N782.45
D832.50
A: increase in Ag affinity with NC in Ant affinity;
disrupted regulation of Ag binding by sodium ions25
S932.60
TM3
F62
C77
F79
L65
C80
M82
V83
A84
C85
F: NC in radioligand binding55
A/S: no detectable radioligand binding56
F: NC in radioligand binding55
L1123.27
E1133.28 I80
G1143.29 A81
F1153.30
C82
A: NC in radioligand binding, S: Ag
affinity reduced 4- to 13-fold; NC in Ant affinity56
F1163.31
F83
P 86
V87
F: substantial reduction of Ag binding55
A: NC in ligand affinity55
A1173.32 V84
A/D: loss of Ag & Ant radioligand binding,
L: slight v in Ag (N6-substituent)
& V in Ant affinity57
T1183.33
L85
L88
T91
Q92
A: substantial V of Ag & Ant binding55
A: substantial V of Ag & Ant binding55
G1213.36 T88
A/S/R/D/E: substantial V in Ag but not Ant activity6
E1223.37 Q89
A: slight v in Ag and Ant activity,
D: slight v in Ag but not Ant affinity,
N/S/L: marginal changes in ligand binding,
H/R: Ant binding affected6
A: marginal changes in ligand binding6
A: marginal changes in ligand binding6
A: substantial V of Ag & Ant binding55
H95A in A3: substantial V of Ag & Ant binding12
I1233.38
A1243.39
S90
S91
S93
S94
A: NC in radioligand binding25
A: no detectable Ag & Ant binding,
T: minor changes25
EL2
E151 A/Q/D: loss of Ag & Ant binding,
∼1000-fold V in Ag potency5
K152A in A3: substantial v in Ant binding.
NC in Ag binding12
Ser 186
Cys187
Ile189
A165
C166
F168
K168
C169 A/S: no detectable radioligand binding56
F171
E172
E169 A: loss of Ag & Ant binding, ∼1000-fold V
in Ag potency, Q: gain in N6-substituted Ag affinity5
Tyr 191
D170 K: NC in ligand binding5
K173
S176
F184
N185
F186
P173 R: NC in ligand binding5
TM5 Y2065.41
F180 A: minor changes in ligand binding4
M2075.42 N181 S: modest reduction of Ag binding4
F 2085.43
F 182 A: loss of Ag & Ant binding,
Y, W: modest reduction of Ag binding4
F 2125.47
V186
F242
W2656.48 W246
V190
F243
TM6 F 2616.44
W247 W243A,F in A3: substantial V in Ant,
NC in Ag binding, 400-fold V in Ag potency12
Y2686.51
L249
L250
A2696.52 H250 A: loss of Ag & Ant binding, no Ag activity
in functional assays, F, Y: modest reduction
of Ag binding; no effect on Ant binding,
N: slight v in Ag affinity,
H251 L: Ant affinity reduced 4-fold;
NC in Ag affinity (Bovine A1 AR)58
minor changes in Ant affinity4,57
A2726.55
F2736.56
F2766.59
N253 A: loss of Ag & Ant radioligand binding4
C254 A: minor changes in ligand binding4
F 257 A: loss of Ag & Ant radioligand binding4
N254 N250A in A3: loss of Ag & Ant binding12
C255 A/S: NC in binding56
L258
H2786.61 C259
C262 G: NC in radioligand binding5
TM7 M2887.35 M270
A2927.39 I274 A: loss of Ag & Ant binding,
30-fold V in Ag potency4
C260 A/S: NC in binding56
C263 A/S: NC in binding56
I270
I274
M: canine/bovine A1 AR binding selectivity59
F 2937.40
V275
F265
A2957.42 S277 A: substantial V in only Ag activity and potency,
T277 A: substantial V in Ag affinity; NC in Ant affinity,
T/C/N/E: marginal changes in binding4,6
S: modest V in Ag affinity; NC in Ant affinity59,60,61
K2967.43 H278 A: loss of Ag & Ant binding; 300-fold V in Ag potency,
Y: modest reduction of Ag binding; NC on Ant binding,
D/E: marginal changes in binding4,47
H278 L: loss of Ag & Ant binding (Bovine A1 AR)58
H272E in A3: substantial V of Ag & Ant binding
except N6-substituted Ag act.17
A2997.46
S281 A: loss of Ag and Ant radioligand binding;
no Ag activity in functional assay, T: enhanced
activity for Ag, N: marginal changes in ligand binding4,47
S281
R309
C309 A/S: NC in radioligand binding56
Bold entries in the second column indicate residues found to be within the retinal binding site, 5.0 Å. The residues with bold face in
ARs were demonstrated important for ligand binding in point-mutation experiments.1 Ag)agonist; Ant)antagonist; NC) no change.
a
b
The mutational results of the A2AAR in the present study are included. The mutational results of the A3AR12 are included and indicated
as A3. The numbering scheme used here in the superscript is based on the most conserved positions in each helix. Each identifier is
composed of a number from 1 to 7 that identifies the helix and, separated by a period, a number associated with a position in that helix.
The position number is given relative to the most conserved residue in that helix, which takes the number 50 (N in TM1, D in TM2,
Rsfrom DRYsin TM3, W in TM4 and P in TMs 5, 6 and 7).

Modeling the Adenosine Receptors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23 4849
Ch a r t 1. Structures of the Known A2AAR Agonists and
Ch a r t 2. Structures of the A2AAR Agonists Containing
Amino Groups for Electrostatic Interaction with
Negatively Charged Side Chains of Mutant A2AARsa
Antagonists Used in Initial Receptor Dockinga
a
Compounds 5, 6, and 8 were used in the previous neoceptor
study of the A3AR.17
a
Sch em e 1. Synthetic Route Used To Prepare
Compound 7a
Affinities (Ki: nM) and selectivity ratios are 1, CGS15943,
nonselective antagonists; 2, A2A-selective antagonists (hA2A: 0.22,
hA1: 2160, hA1/hA2A
: 9820); 3, nonselective agonist; 4, A2A-
selective agonist (rA2A: 3.5, rA1: 1017, A1/A2A: 291).18,24
standing of the structure and activation of GPCRs.
Although the sequence homology between rhodopsin
and GPCRs is very low, different amino acids or
alternate microdomains can support similar deviations
from the regular R-helical structure, thereby resulting
in a similar tertiary structure.14 Thus, it may be possible
to extend the binding site and activation models of ARs
by using indirect structural data.
In the present study, we constructed a three-dimen-
sional model of the TMs, composed of seven R-helical
segments, including portions of the extracellular and
cytosolic loops, for the A2AAR, with the high-resolution
structure of rhodopsin as a template.11 In this paper,
we describe a comparison of the binding characteristics
of A2A agonist and antagonist ligands (Chart 1). In
addition, using an A2A-selective antagonist 2 and agonist
4, we investigated the bound conformations and the
basis for selectivity for the human A2AAR. We also made
comparisons to our previously published A3AR model,15,16
which was derived with similar methods. Finally, the
model was tested experimentally by making comple-
mentary changes in the structures of agonist ligands
(e.g., introduction of amino groups into known ligands,
resulting in adenosine derivatives 5-8; Chart 2) and
the A2AAR, to form neoceptor-neoligand pairs.17
a
Reagents: (a) (i) TBAF, THF, rt; (ii) toluoyl chloride, pyridine,
rt; (b) (i) 50% CH3COOH, 50 °C (ii) (CH3CO)2O, pyridine, rt; (c)
silylated 6-chloropurine, TMSOTf, dry 1,2-dichloroethane, reflux;
(d) (i) 3-iodobenzylamine HCl, Et3N, EtOH, reflux (ii) 7 N NH3 in
methanol, rt; (e) Ph3P, NH4OH, pyridine, rt.
lylated 6-chloropurine22 gave 12. Displacement of the
chloro atom with 3-iodobenzylamine and deprotection
in 7 N NH3 in MeOH produced the 3′-C-azidomethyl
nucleoside 13 in an overall yield of 51% (from 11),
improving the N6-(3-iodobenzyl)adenine coupling previ-
ously described.17 Reduction of the azido moiety with
triphenylphosphine gave access to the 3′-C-aminomethyl
nucleoside 7.
Con str u ction of th e A2AAR Molecu la r Mod el. As
has been done with recent GPCR modeling,15,23 we have
included loop regions, which serve as topological con-
straints to aid in determining the packing of the
attached R-helices within the TM bundle. To decrease
the large deviation at the end of the TM bundle, all
three-dimensional structures of the A2AAR except the
C-terminal region were constructed with the use of
homology modeling from the X-ray structure of rhodop-
sin.13 To validate the reliability of the model, stereo-
chemical accuracy, packing quality, and folding reli-
ability were checked. A Ramachandran plot was used
to compare the φ and æ angles with the crystal structure
of rhodopsin. All helical amino acids were located in the
region of the Ramachandran plot favoring a right-
handed R-helix. Only 3.3% of the residues of the loops
were in a sterically disallowed region. Calculated ω
angles for the A2A model, as in the experimental
structure of rhodopsin, indicated the absence of cis-
peptide bonds. The chirality of all CR atoms except Cys
showed an S-configuration. Indicative of the packing
quality, there were no bump regions in the calculated
Resu lts a n d Discu ssion
Ch em ica l Syn th esis. The synthesis of 9-(3-C-amino-
methyl-3-deoxy-â-D-ribofuranosyl)-N6-(3-iodobenzyl)-
adenine (7) is depicted in Scheme 1. The 3-C-azido-
methyl sugar 9 was prepared in six steps from 1,2-O-
isopropylidene-D-xylofuranose by reported proce-
dures.18-20 Removal of the 5-O-(tert-butyldimethyl)silyl
group of 9 with tetrabutylammonium fluoride (TBAF)
and protection of the 5-OH group as a toluoyl ester gave
10. Acidic deblocking of the 1,2-O-isopropyl moiety of
10, followed by acetylation furnished the peracylated
3-C-azidomethyl sugar 11. It should be pointed out that
the protecting group exchange was necessary to avoid
pyranose ring formation that, in our hands, occurred
upon acidic deprotection of 9 (contrary to what was
reported19). Vorbru¨ggen-type coupling21 of 11 with si-

4850 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23
Kim et al.
F igu r e 1. Stereoview showing major interhelical H-bonding networks detected in the unoccupied hA2AAR, between TM1 and
TM7 which do not occur in rhodopsin (B) and those involving in addition TM2 and TM3 as found in rhodopsin (A). The H-bonds
are represented in yellow.
A
2AAR model. Indicative of the folding reliability, the
in rhodopsin (Figure 1B). As previously proposed,11
there was H-bonding between the side chains of E131.39
and H2787.43. Mutational results24 indicated that E131.39
rmsd (root-mean-squared deviation) between backbone
atoms in the TMs of the calculated A2AAR and the
template molecule was calculated to be 1.27 Å. Both
structures, especially the parts with well-defined sec-
ondary structures, showed overall similarities.
Con ser ved H-Bon d Netw or k s. The transmem-
brane region of rhodopsin is stabilized by a number of
interhelical H-bonds and hydrophobic interactions, most
of which are mediated by highly conserved residues in
GPCRs.13 For example, in the case of the unoccupied
and the corresponding residue in the A1AR, E161.39
,
facilitated agonist binding. Another residue, D522.50 in
the A2AAR, was highly stabilized by an H-bonding
network among the highly conserved amino acids,
N2807.45, S2817.46, and N2847.49, which also formed
H-bonds with S913.39. Unlike in rhodopsin, there were
additional H-bonds in the ARs for D522.50 interacting
with hydrophilic amino acids; whereas in rhodopsin, the
amino acids that corresponded to those participating in
the H-bond network were all alanine residues except
N2847.49. The corresponding amino acid, D552.50 in the
A1AR, was responsible for sodium binding.25
A
2AAR (Figure 1A), N241.50 interacted with the backbone
carbonyl groups of S2817.46 and D522.50, as was observed
in rhodopsin, which had interhelical H-bonds between
the highly conserved N551.50 and the backbone carbonyl
groups of A2997.46 and D832.50. Another Asn residue,
Hyd r op h obic Region . Among hydrophobic interac-
tions, the conserved W2466.48 was typically surrounded
by hydrophobic residues from TMs 3, 6, and 7, as was
observed for the human A3AR15 and another GPCR, the
thyrotropin-releasing hormone receptor.26 Those hydro-
phobic amino acids near W2466.48 were V84, L85, L87,
F242, L243, and P248. The hydrophilic aromatic resi-
dues H250 and H278 were also in proximity. The indole
N782.45, in rhodopsin formed H-bonds to S1273.42, T1604.49
,
and W1614.50. The corresponding amino acid in the
2AAR, S472.45, showed the same hydrophilic interaction
A
with S943.42 and W1294.50. With respect to the highly
conserved (D/E)R(Y/W) motif in GPCRs, the carboxylate
of E1343.49 in rhodopsin formed a salt bridge with the
guanidium group of the adjacent R1353.50, which was
also associated with E247 and T251 in TM6. The
corresponding amino acids in the A2AAR were D101-
R102-Y103. The analogous interactions occurred in the
ring of W2466.48 also formed an H-bond with N2807.45
,
.
which was stabilized through H-bonding with D522.50
Agonist binding would cause a rotation of the Trp side
chain, disrupting these interhelical interactions. Thus,
the intramolecular contact network might be destabi-
lized, facilitating the conformational change to activate
the A2AAR. The experimental results15 were consistent
with this hypothesis: the W243A6.48 mutant A3AR
displayed normal agonist binding but no activity in a
functional assay.
A
2AAR, i.e., the salt bridge of R1023.50 with D1013.49 and
E228 in TM6. For the NPXXY motif in the TM7 of
GPCRs, the hydroxyl group of Y3067.53 was close to
N732.40 in rhodopsin, which was also highly conserved
among GPCRs. The same result appeared with the
calculated A2AAR structure, i.e., the OH group of
Y2887.53 in the A2AAR was located in proximity to the
side chain of N42.2.40
Bin d in g of a Kn ow n Agon ist a n d An ta gon ist to
Mu ta n t Hu m a n A2AARs. A variety of negatively
charged side chains were substituted in the A2AAR at
Two important interhelical H-bonding interactions for
highly conserved sequences took place in ARs but not

Modeling the Adenosine Receptors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23 4851
nist 1. These findings contrasted with the selective
decrease of agonist affinity in the S277A mutant recep-
tor and the inability of the H278A mutant receptor to
bind either agonist or antagonist radioligand.4 A total
of six mutationssT88D, T88E, Q89D, S277E, H278D,
and H278Esdid not alter the binding affinity of the
antagonist 1.
Liga n d Dock in g. A crystallographic determination
of the human A2A receptor structure would be a better
method by which to analyze the conformational implica-
tions of our mutagenesis experiments; however, pres-
ently no structure is available. While such studies are
underway,27 no structure of close homology to the
receptor is available. Thus, we resorted to the widely
used, however imprecise, method of rhodopsin-based
homology modeling of Family 1 GPCRs. Rhodopsin-
based homology modeling is not an automatic method
for obtaining a realistic structure for a given GPCR, but
rather requires time-consuming custom treatment ac-
cording to known pharmacological data.28 In the present
study, both agonists and antagonists are docked in the
human A2A receptor model. It is to be emphasized that,
in general, docking of agonists to GPCR models is
subject to even greater uncertainty than antagonist
since the template consists of the inactive state of
rhodopsin. Often multiple modes of docking of a given
agonist or antagonist ligand are observed,29 and the
selection of preference of one docking mode in such cases
must be based on diverse pharmacological data, rather
than on computational results alone. Nevertheless, the
increasing level of refinement of rhodopsin-based ho-
mology modeling, for example in the addition of the
extracellular loops, has yielded useful insights and
results.30 In the present study, the introduction of
complementary functional groups on the ligand and
receptor is meant to overcome the problem of ambiguity
of docking modes in GPCR modeling.
F igu r e 2. Effect of the agonist NECA 3 and antagonist
CGS15943 1 on the binding of the radiolabeled antagonist
[3H]ZM241385 to the human A2AAR. Membranes (10-20 µg
of protein) from COS-7 cells transfected with wild type (9) or
mutant receptors of T88D (1), T88E ([), Q89D (b), S277E (0),
H278D (4), or H278E (3) were incubated with 1.0 nM
[3H]ZM241385 in duplicate, together with increasing concen-
trations of the competing compounds, in a final volume of 0.4
mL of Tris‚HCl buffer (50 mM, pH 7.4) at 25 °C for 120 min.
Results were from a representative experiment.
Ta ble 2. Binding Affinity of the Nonselective Agonist 3 and
Antagonist 1 at WT and Mutant Human A2AARsa
Ki (nM) or % displacement
mutant receptor
3
1
We used an automated docking procedure (Flexi-
Dock)31 with manual adjustment to determine the most
energetically favorable binding location and orientation
for several ligands and compared these results with our
previously reported experimental results.4 An energeti-
cally favorable docking mode for each ligand, consistent
with the experimental results, was selected from several
possible models, which were subjected to a test of
consistency with pharmacological results. The most
general structural distinction between agonists and
antagonists is that only an agonist requires a ribose
ring, or more specifically, the 3′- and 5′-ribose substit-
uents that interact directly with the A2AAR.
WT
21.4 ( 8.7
20% at 10 µM
0% at 10 µM
1.5 ( 0.4
29.2 ( 6.3
19.1 ( 3.4
24.6 ( 8.3
0.84 ( 0.22
0.91 ( 0.09
0.67 ( 0.17
0.35 ( 0.12
1.0 ( 0.3
T88D
T88E
Q89D
S277E
H278D
H278E
0.41 ( 0.14
0.52 ( 0.06
a
Membranes from COS-7 cells transfected with WT or mutant
2AAR cDNA were incubated with 1.0 nM [3H]ZM241385 in
duplicate, together with increasing concentrations of the competing
compounds, in a final volume of 0.4 mL of Tris‚HCl buffer (50 mM,
pH 7.4) at 25 °C for 120 min. The Ki values are expressed as mean
( standard error from three independent experiments.
A
hydrophilic positions in TMs 3 and 7 predicted to be in
proximity to the bound ribose moiety. Radioligand
binding studies showed that T88D and T88E mutant
receptors were able to bind the nonselective AR antago-
nist 1 but not the nonselective AR agonist 3 (Figure 2
and Table 2). Similarly, the T88A mutation was already
shown to substantially decrease agonist but not antago-
nist affinity.6 For the Q89D mutant receptor, the
binding affinity of agonist 3 was increased approxi-
mately 14-fold. Its Ala mutation also increased agonist
and antagonist binding affinity, but its N/S/L mutation
produced only marginal changes in ligand affinity, and
its H/R mutation affected only antagonist affinity.6
However, S277E, H278D, and H278E mutations did not
affect the binding of either the agonist 3 or the antago-
To explain the different binding properties of antago-
nists and agonists at the A2AAR, we calculated the
conformations of representative ligands, and these
structures were docked within the human A2AAR model.
The result of FlexiDock showed overlapped binding
modes for antagonists and agonists (Figures 3 and 4).
Although additional mutational experiments were needed
to confirm the reliability of the docked complexes, the
result of FlexiDock correlated well with known muta-
tional results.4 The agonist binding region was similar
to our previous A2AAR docking results,4 which did not
consider the flexibility of the ligand.
An ta gon ist Bin d in g. 5-Amino-9-chloro-2-(2-furyl)-
1,2,4-triazolo[1,5-c]quinazoline (CGS15943) 1 is a potent

4852 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23
Kim et al.
F igu r e 3. The complex of the A2AAR with agonists and antagonists. (A) CGS15943, a nonselective antagonist; (B) an A2A-selective
antagonist; (C) NECA, a nonselective agonist; (D) an A2A-selective agonist. All ligands are represented in yellow. The amino acids
of the A2AAR that are depicted in red and orange participated in the hydrophilic and hydrophobic interactions, respectively.
and nonselective adenosine antagonist that we used as
a modeling template in previous studies.32 The major
mode of interaction of 1 with the A2AAR consisted of
hydrophobic interactions (Figures 3A and 4A). Large
hydrophobic pockets consisting mostly of residues at
TMs 3, 6, and 7 interacted with the ligand. Hydrophobic
amino acids that participated in these interactions with
the ligand were L853.33, I1354.56, L167 (EL2), F168
(EL2), F1825.43, V1865.47, W246,6.48 and L2496.51 near the
quinazoline ring and I803.28, V843.32, and I2747.39 in
proximity to the furan ring. One important hydro-
philic interaction was an H-bond formed between the
produced a highly A2A-selective antagonist 2 (Ki ) 0.22
nM, hA1/hA2A) 9820), which did not significantly in-
teract with either A2B or A3 ARs.34 Its derived structure-
activity relationships indicate that the tricyclic structure
of the pyrazolotriazolopyrimidine, the presence of the
furan ring, the exocyclic 5-amino group, and the aryl-
alkyl substituent on the nitrogen at the 7-position are
probably essential for both affinity and selectivity for
the A2AAR subtype. To investigate the structural re-
quirements that increase high affinity and selectivity
for the A2AAR, we carried out a docking study of this
A2A-selective antagonist 2. Additional hydrophobic in-
teractions between the 4-aminophenylpropyl moiety and
the hydrophobic pocket at TM4 and TM5 and H-bonding
of the 4-amino group with N145 (EL2) both could
contribute to the increase of A2AAR affinity (Figure 3B).
An equilibrium thermodynamic study, which showed a
higher binding equilibrium enthalpy value for the
pyrazolotriazolopyrimidine derivative containing a hy-
droxyl group at the para-position of the phenyl ring than
for its analogue lacking the hydroxyl group, suggested
the presence of an electrostatic interaction (probably an
H-bond) involved in recognition of the hydroxyl moiety
within the binding site.35
exocyclic amino group at the 5-position and N2536.55
.
Additional weak H-bonding between the side chain of
N1815.42 and N6 of the CGS15943 served to increase the
thermal stability of the complex. This docking result was
consistent with our previously reported Ala mutant
receptorssF182A, H250A, N253A, I274A, and H278As
all of which lost the high-affinity binding of both A2AAR
agonists and antagonists. The aromatic residue H250
also appeared to be a required component of this mainly
hydrophobic pocket. H-Bonding to this residue was not
essential, as indicated by retention of function in F and
Y mutant receptors.4
A series of pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]pyrimi-
dine derivatives was developed by Baraldi and co-
workers as potent and selective nonxanthine A2A an-
tagonists.33 The addition of the 4-aminophenylpropyl
group at the N7-position of the triazolopyrimidine ring
The A2AAR sequence alignment indicated that most
of the amino acids in the putative binding site within
5Å of the A2A-selective antagonist 2 were conserved
among ARs. Highly conserved amino acids were L853.33
T883.36, G1364.57, P1394.60, F168 (EL2), M1775.38, F1825.43
,
,

Modeling the Adenosine Receptors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23 4853
group. The size of the side chain at positions 4.56 and
5.47 may be important for an optimal van der Waals
interaction, although the hydrophobic character of these
two side chains is similar: with a longer side chain at
this position, i.e., I5.47 in the A3AR, steric repulsion
occurred, and with a shorter size, i.e., V4.56 in A1 and
A3, van der Waals attractive forces were not optimal.
Thus, the molecular modeling results were supported
by the evidence that N5-, N7-, and N8-substituents of
pyrazolotriazolopyrimidine analogues controlled the
subtype selectivity of binding to ARs.38
Agon ist Bin d in g. The environment surrounding the
purine ring of A2A agonists binding in our current A2AAR
model was very similar to that of the triazolopyrimidine
ring of antagonists, including the H-bonding of the
exocyclic amino group with the highly conserved N2536.55
.
As we expected from the structural differences between
the agonists and antagonists, a characteristic feature
of agonist binding was additional H-bond formation of
the 3′-OH with H2787.43 and of the 5′-amide group of 3
with T883.36 and S2777.42 (Figures 3C, 3D, and 4B).
However, with our former A2AAR model4 we reported
different hydrophilic interactions: (1) H-bonds between
the 5′-OH in adenosine or 5′-NH in 3 and S2777.42 and
H2787.43 and (2) H-bonds between T883.36 and both the
2′-O and 3′-O. In both cases, T883.36 and S2777.42 were
important only for agonist but not antagonist binding.
H2787.43 participated in the binding differently, as its
hydrophilic interaction with agonists and aromatic
interaction with antagonists were the result of the dual
role of the imidazole ring. Our present model is consist-
ent with several experimental results concerning rec-
ognition of ribose-containing ligands (i.e., adenosine
agonists): (1) The hydrophilic interactions at S2777.42
and H2787.43 were required for high-affinity binding of
agonists but not antagonists.6 The toleration of ligand
binding in the S277E and H278D/E mutant receptors
suggested that the acidic residues in the mutant recep-
tors retained the ability to form an H-bond with the
ligand, i.e., at the 3′-OH group. With respect to the 5′-
carbonyl group of 3, the model suggested that a weak
H-bond (2.42 Å and 152.62°), or alternately a water-
mediated H-bond, was possible. (2) Radioligand binding
studies (Figure 2 and Table 2) showed that T88D and
T88E mutant receptors were able to bind the antagonist
1 but not the agonist 3. Similarly, the T88A mutation
was already shown to substantially decrease agonist but
not antagonist affinity.6 Thus the reason that mutation
of T88 was specific for diminishing the affinity of the
ribosides appeared to be the proximity of this residue
to agonists but not antagonists based on the docking
result. (3) Replacement of H278 with other aromatic
residues was not tolerated in ligand binding.4
F igu r e 4. Detailed interactions with 1 (A) and 3 (B) in the
putative A2A binding site. The residues in the double-squared
box are highly conserved among G protein-coupled receptors,
and those in the single-squared box are conserved amino acids
among ARs.
W2466.48, N2536.55, I2747.39, and H2787.43. However, in
the docked complex the amino acids located near the
N,5 N,7 and N8 positions of the triazolo[1,5-c]pyrimidine
varied. The proximity of L167 (E for A1, Q for A3) and
H2506.52 (S for A3) to the ribose 5′-position and Q893.37
(H for A3) to the adenine 8-position could account for
the reported experimental result that the N5-unsubsti-
tuted derivatives showed high affinity for the A2AAR
subtype, whereas the N5-phenylcarbamoyl and N8-
methyl substituents at the pyrazolo[4,3-e]-1,2,4-triazolo-
[1,5-c]pyrimidine as the same core molecule increased
affinity and selectivity for the A3AR subtype (Ki ) 0.16
nM, hA1/hA3 ) 3,710, hA2A/hA3 ) 2381, hA2B/hA3
)
1390):36 On one hand, hydrophilic amino acids such as
Q167 and S247 in the A3AR near the putative binding
region for N5 substituents would be expected to increase
the selectivity for the A3 subtype through additional
H-bonding with carbamoyl groups. On the other hand,
the bulky and aromatic side chains such as L167 and
H250 in the A2AAR made it easy to accommodate an
unsubstituted N5-amino group. H95 in the A3AR near
the 8-position binding region had a more hydrophobic
To better understand the increase of A2A selectivity
on substitution of the C2 of the adenine ring, we
performed the docking of the (E)-2-phenylpentenyl
derivative of NECA 4 (Ki values, nM: rA2A, 3.5; rA1,
1020; A1/A2A) 291).39 Additional hydrophobic interac-
tions between the phenylpentenyl moiety at the 2-posi-
tion of the adenine ring and the hydrophobic pocket
formed from TM4 and TM5 increased the A2AAR affin-
ity, similar to the result shown for the A2A-selective
antagonist binding. The interaction sites of the ad-
ditional phenyl rings were very close to each other. The
character than did the homologous Ser in the A1, A2A
,
and A2B ARs, possibly explaining the fact that the
hydrophobic factor at the N8-position was important for
A3AR binding.37 Residues V1865.47 (I for A3) and I1354.56
(V for A1 and A3) appeared to contribute to the high
affinity at the A2AAR through interaction with N7-
substitutions of the pyrazole ring,37 such as a propyl

4854 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23
Kim et al.
F igu r e 5. Stereoview showing the superimposition of the bound conformations of the A2A-selective antagonist 2 and agonist 4 in
the putative binding sites. The colors represent atom type.
superimposition of the bound conformations of the A2A
-
selective agonist 4 and antagonist 2 in their putative
binding sites demonstrated the partial overlap of the
binding sites of the adenine and triazolopyrimidine rings
and their C2 and N7 substituents (Figure 5). The
energies of A2AAR complex formation with 4 and its
geometric isomer were calculated. The complex with (E)-
2-(phenylpentenyl) derivative 4 was more energetically
favored than the one with the corresponding(Z)-2-(phenyl-
pentenyl) analogue, consistent with experimental re-
sults showing that the E-isomers in this series were
more potent and selective than were the Z-isomers.38
The anti-conformation of the glycosidic bond of 3 and
other agonists was energetically favorable as an active
conformation for A3AR binding;15 this was supported by
both a molecular modeling study and a binding prefer-
ence for the methanocarba-ring system in the (N)-
conformation, which favors the anti-conformation.40,41
However, the A2AAR complex with its agonist showed
a preference for the intermediate conformation about
the glycosydic bond. The ø angles (O4′-C1′-N9-C8) of
3 and 4 docked in the A2AAR were -74.4° and -66.2°,
respectively. There might be subtle differences in bind-
ing requirements, conformational preferences, and local
environments among subtypes of ARs, although most
of the amino acid sequences in the putative binding sites
(TM regions) are conserved among the four types of ARs.
The electrostatic potential maps of the A2A and A3 ARs
showed some differences in the shape and electrostatic
nature of their binding sites near the second extra-
cellular loop (Figure 6).
Con for m a tion a l Hyp oth esis for Activa tion of th e
2AAR. Agonist binding was significantly different than
F igu r e 6. Electrostatic potential map made with the Grasp
program. Putative binding sites of the A2AAR (top) and A3AR
(bottom), with the second extracellular loop removed from the
extracellular view. NECA 3 is represented as a stick model.
Blue and red color potential indicates the electronic positive
and negative, respectively.
A
antagonist binding in the region of the ribose ring, as
expected from the requirement for a ribose ring in the
agonist but not in the antagonist. Thus, the putative
ribose-binding region is probably involved in receptor
activation. For the A3AR,15 we tried to gain insight into
the distinct structural requirements for binding and
activation by using ligand effects on cyclic AMP produc-
tion in intact CHO (Chinese hamster ovary) cells and
docking a few adenosine analogues. An interesting
result concerned the conserved W2436.48 side chain in
the A3AR, which was involved in recognition of the
classical (nonnucleoside) A3AR antagonists but not
adenosine-derived ligands and which displayed a char-
acteristic movementscounterclockwise rotation as viewed
from the exofacial sidesexclusively upon docking of
agonists.15 We concluded similarly for the A2AAR that
a significant distinction between agonists and antago-
nists was whether ligand binding could effect the
movement of W2466.48 side chain (Figure 3). Additional
binding of the ribose 5′-substituents shown in the
agonist complex induced the movement of the side-chain
of W2466.48. That conformational change might disrupt
a network of H-bonding of W2466.48 with N2807.45, which
participated in the H-bonding network of D522.50, as well
as hydrophobic interactions of W2466.48 with hydro-
phobic residues from TMs 3, 6, and 7, thus facilitating
a conformational change upon receptor activation. If the
A2AAR behaves like the A3AR,15 then flexibility of the
ribose moiety and specific recognition elements at the
3′- and 5′-positions, to permit the movement of TM6,
would be important for agonism at the A2AAR. The
modeling result indicated that the overall flexibility and
binding elements of the ribose ring were important for
distinguishing the receptor interactions of agonists and
antagonists. It also correlated with recent studies42
based on electron paramagnetic resonance and fluores-
cence spectroscopy, which suggested an outward move-
ment of the cytoplasmic end of TMs 3 and 6, as well as

Modeling the Adenosine Receptors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23 4855
F igu r e 7. Effect of the amino derivatives of adenosine 8 (A), 5 (B), 6 (C), and 7 (D) on the binding of the radiolabeled antagonist
[3H]ZM241385 to the human A2AAR. Membranes (10-20 µg of protein) from COS-7 cells transfected with wild type or T88D
mutant receptor. Membranes were incubated with 1.0 nM [3H]ZM241385 in duplicate, together with increasing concentrations of
the competing compounds, in a final volume of 0.4 mL of Tris‚HCl buffer (50 mM, pH 7.4) at 25 °C for 120 min. Results were from
a representative experiment. The Ki values were from three independent experiments and are listed in Table 3.
Ta ble 3. Binding Affinity of Amino Derivatives of Adenosine
an anticlockwise rotation of TM6 around its helical axis
as viewed from the extracellular side.
(see Chart 2) at WT and Mutant Human A2AARsa
Ki (µM)
Two of the three most conserved prolines in GPCRs,
P6.50 and P7.50, occur on TM6 and TM7. P6.50 is in
proximity to the binding site, i.e., W2466.48. P7.50 was
near N2847.49, which associated with D522.50, the puta-
tive sodium-binding site, through H-bonding. It was pro-
posed that in rhodopsin P6.50 acts as a flexible hinge,
straightening TM6 upon light-induced activation.43
Thus, these two proline residues that are conserved
among GPCRs would facilitate the agonist-induced
movement of TM6 and subsequently TM7 to rearrange
intracellular loop (IL) 3 and helix 8, which are known
to be important for the receptor-G protein interface.44
Am in o Der iva tives of Ad en osin e a s Neoliga n d s
for Neocep tor s Der ived fr om th e A2AAR. The cur-
rent model, like the A3AR model,15 suggested the ribose
moiety was in proximity to hydrophilic residues of TM3.
This hypothesis was tested experimentally by making
complementary changes in the structures of the agonist
ligand (i.e., introducing a positively charged ammonium
group) and the receptor (i.e., introducing a negatively
charged carboxylate group) to provide a new electro-
static interaction as the basis of neoceptor-neoligand
pairs. Four amino derivatives (Chart 2) were tested in
binding to a variety of Asp and Glu mutant receptors
(Table 3). The 5′-aminoethyluronamide 8 (equivalent to
appending an amino group at the end of the 5′-
substituent of 3) displayed a large selective enhance-
ment in binding to the T88D mutant receptor (Figure
7). Also, compound 7 displayed a significant enhance-
ment of affinity at the same mutant receptor. However,
other possible electrostatic interactions of the ribose 3′-
amino groups of compounds 6 and 7 with negatively
5
6
7
8b
WT
534 ( 51
407 ( 142
492 ( 161
52 ( 9
604 ( 125
-
67 ( 13
18 ( 7
57 ( 8
6.4 ( 0.9
88 ( 36
87 ( 36
64 ( 5
79 ( 16
6.5 ( 1.9
64 ( 28
58 ( 18
67 ( 11
59 ( 19
56 ( 6
46.1 ( 4.2
4.4 ( 1.6
57 ( 19
3.6 ( 1.0
41 ( 8
121 ( 32
72 ( 14
T88D
T88E
Q89D
S277E
H278D
H278E
-
a
Membranes from COS-7 cells transfected with WT or mutant
2AAR cDNA were incubated with 1.0 nM [3H]ZM241385 in
duplicate, together with increasing concentrations of the competing
compounds, in a final volume of 0.4 mL of Tris‚HCl buffer (50 mM,
pH 7.4) at 25 °C for 120 min. The Ki values are expressed as mean
A
b
( standard error from three independent experiments. 8,
MRS3366.
charged mutant receptors, such as H278E, as suggested
in the model, were not supported in the binding assay.
For the docking studies of neoceptor and neoligand,17
three mutant receptorssT88D, S277D, and H278Es
were optimized through a molecular dynamics (MD)
procedure after the mutation of each side chain. T88,
S277, and H278 residues in the inactive state of A2AAR
preferred the gauche+ (g+) rotamer conformation,
which also occurred in the bound state of the A2AAR/3
complex. However, in the T88D mutant receptor, the
side chain of aspartate residue was in the trans (t) form,
because of new H-bonding of the carboxylate group with
N1815.42. Thus, there was a local conformational change
with respect to the wild type in the position and
direction of the aspartate side chain. This conforma-
tional change was consistent with the binding profile
of the neoceptor. Initially, on the basis of the model of
the unoccupied native A2AAR, the 5′-amino derivative

4856 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23
Kim et al.
F igu r e 8. Stereoview showing the complex of the T88D neoceptor with neoligand. (A) Compound 8 with 5′-aminoethyluronamide
group. (B) Compound 7 with a 3′-aminomethyl group (represented in the protonated form). The neoligands are represented as
ball-and-stick models. The intermolecular H-bonding between neoceptor and neoligand is displayed as yellow color. The W2466.48
of the T88D A2AAR in red shows the movement of the indole side chain upon ligand binding.
5 was expected to display an enhanced affinity for the
neoceptor T88D. However, 5 did not display an increase
of binding affinity, whereas compound 8 with the
extended ammonium group was enhanced in affinity.
The docking result of compound 8 suggested the pos-
sibility of the salt bridge (2.82 Å) and H-bonding (2.03
Å) between acidic residues of D88 and the terminal
ammonium ion but the loss of H-bonding at the 3′- and
5′-position, as suggested in 3 binding to the wild-type
A2AAR (Figure 8A). Compound 7 was additionally
substituted at the N6-position and also displayed bind-
ing enhancement at the T88D mutant receptor. In the
docking of 7, a different binding mode that was ener-
getically unfavorable in the A2AAR/3 complex was
suggested, because with side-chain geometry that was
identical to that of the A2AAR/3 complex, it was impos-
sible for the 3′-ammonium group of 7 to interact with
the side chain of D88. In its complex, the conjugation
of H-bonding between D88 and 3′-ammonium ion through
the 2′-OH group was suggested. An additional interac-
tion of the N6-benzyl group of 7 with the second
extracellular loop, especially through hydrophobic in-
teraction with F168, was evident in the model (Figure
8B); however, there was no interaction of the 5′-OH with
this neoceptor. The conformational search of compounds
6 and 7 indicated that the lowest-energy conformer of
compound 7 had more stable intramolecular H-bonding
of the 2′-OH group and the 3′-ammonium ion (1.86 Å)
than was displayed by compound 6 (2.45 Å). It was
consistent with the experimental result: i.e., compound
7 showed a greater increase of binding affinity to its
neoceptor than compound 6 did. Thus, we identified new
neoceptor (T88D)-neoligand pairs consistent with the
theoretical model, even though the binding affinities to
the T88D neoceptor were still low compared to the
endogenous ligand binding to the wild-type A2AAR.
Con clu sion s
The molecular modeling results clearly delineated the
interactions involved in the binding of agonists and
antagonists, which correlated well with known experi-
mental results. Flexibility and binding character at the
3′- and 5′-positions of the ribose ring were required for
a movement of TM6, which was correlated with receptor
activation. The docking complexes provided insight into
the conformational and binding requirements for ago-
nists and antagonists at the A2AAR. Structural differ-
ences between the A2AAR and the A3AR include the
microenvironment surrounding docked agonist and
antagonist ligands and the glycosidic bond angle of
docked agonists. Structural similarities between the
A2AAR and the A3AR include H-bond networks involving
the putative sodium binding site, a hydrophobic region
surrounding W2466.48, and the proposed rotation of TM6
to induce receptor activation. The introduction of a
hydrophilic moiety such as ribose into an otherwise
hydrophobic region destabilizes the inactive ground
state of the receptor and thus would facilitate activation.
Furthermore, this study is intended to facilitate further
design of both adenosine analogues and nonpurines
targeted for improved A2AAR affinity and selectivity,
including neopceptor-neoligand pairs.17 The identifica-
tion of the T88D receptor as a new neoceptor that may
be activated selectively by synthetic ligands, such as 7
and 8, provides a new means to apply the beneficial
effects of A2AAR activation.1-3 Moreover, this putative

Modeling the Adenosine Receptors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23 4857
bonds were 3000 iteration, 3-kcal energy cutoffs, and chirality
checking. In all cases, MMFF force field50 and charge were
applied with the use of distance-dependent dielectric constants
and conjugate gradient method until the gradient reached 0.05
kcal ‚ mol-1‚Å-1. After clustering the low-energy conformers
from the result of the conformational search, the representa-
tive ones from all groups were reoptimized by semiempirical
molecular orbital calculations with the PM3 method in the
MOPAC 6.0 package.51
electrostatic interaction, which enhances affinity by
anchoring the ligand in the agonist binding site, also
serves to validate the model. The positively charged
neoligands were docked in the neoceptor to suggest an
electrostatic explanation for the selectively enhanced
affinity.
Exp er im en ta l Section
A human A2AAR model was built by using homology model-
ing from the recently published X-ray structure of bovine
rhodopsin,13 as we previously described.17 Multiple-sequence
alignment data of selected GPCRs were used for the construc-
tion of human A2AAR TM domains.52 For the model of the
second extracellular loop, EL2, two beta-sheet domains in
rhodopsin were first aligned, including the disulfide bond
between Cys77 and Cys166, and then other parts were added
or deleted. To construct the N-terminal region and the intra-
and extracellular loops, each alignment was manually adjusted
to preserve the overall shape of the loop. Acetyl and N-methyl
groups blocked the N-terminal and C-terminal regions, re-
spectively, to prevent electrostatic interactions. All helices with
backbone constraints and loops were separately minimized.
After being combined, all structures were reminimized initially
with backbone constraints in the secondary structure and then
without any constraints. The Amber all-atom force field53 with
fixed dielectric constant of 4.0 was used for all calculations,
terminating when the conjugate gradient reached 0.05
Ma ter ia ls. Human A2AAR cDNA (expression vector pSVL-
2A) was kindly provided by Dr. Marlene J acobson (Merck
A
Research Labs, West Point, PA). Taq polymerase for the
polymerase chain reaction was purchased from PerkinElmer
(Norwalk, CT). All enzymes used in this study were obtained
from New England Biolabs (Boston, MA). Oligonucleotides
used were synthesized by Bioserve Biotechnologies (Laurel,
MD). [3H]ZM241385 (4-(2-[7-amino-2-(2-furyl)1,2,4]triazolo-
[2,3a][1,3,5]triazin-5-ylamino]ethyl)phenol, 17 Ci/mmol) was
from Tocris Cookson Ltd. (Bristol, United Kingdom). Adenosine
deaminase, CGS15943, and NECA were obtained from Sigma
(St. Louis, MO). All other compounds were obtained from
standard commercial sources and were of analytical grade.
Site-Dir ected Mu ta gen esis. The protocols used were as
described in the QuikChange Site-directed Mutagenesis Kit
(Stratagene, La J olla, CA). Mutations were confirmed by DNA
sequencing.
Tr a n sien t Exp r ession of Wild -Typ e a n d Mu ta n t Re-
cep tor s in COS-7 Cells. COS-7 cells were grown in 100-mm
cell culture dishes containing Dulbecco’s modified Eagle’s
medium supplemented with 10% fetal bovine serum, 100 units
penicillin/mL, 100 µg streptomycin/mL, and 2 µmol glutamine/
mL. Cells were washed with phosphate-buffered saline (with
calcium) and then transfected with plasmid DNA (10 µg/dish)
by the DEAE (diethylaminoethyl)-dextran method45 for 1 h.
The cells were then treated with 100 µM chloroquine for 3 h
in culture medium and cultured for an additional 48 h at 37
°C and 5% CO2.
Mem br a n e P r ep a r a tion . After 48 h of transfection, COS-7
cells were harvested and homogenized with a Polytron ho-
mogenizer. The homogenates were centrifuged at 20 000 g for
20 min, and the resulting pellet was resuspended in the 50
mM Tris‚HCl buffer (pH 7.4) and stored at -80 °C in aliquots.
The protein concentration was determined by using the
method of Bradford.46
Ra d ioliga n d Bin d in g Assa y. The procedures of [3H]-
ZM241385 binding to wild-type and mutant human A2AARs
was as previously described.47 Briefly, membranes (10-20 µg
of protein) were incubated with 1.0 nM [3H]ZM241385 in
duplicate, together with increasing concentrations of the
competing compounds, in a final volume of 0.4 mL Tris‚HCl
buffer (50 mM, pH 7.4) at 25 °C for 120 min. Binding reactions
were terminated by filtration through Whatman GF/B glass-
fiber filters under reduced pressure with an MT-24 cell
harvester (Brandel, Gaithersburg, MD). Filters were washed
three times with ice-cold buffer and placed in scintillation vials
with 5 mL scintillation fluid, and bound radioactivity was
determined by using a liquid scintillation counter.
kcal‚mol-1‚Å-1
.
For the conformational refinement of the A2AAR and their
mutant receptors, the optimized structures were then used as
the starting point for subsequent 50-ps MD, during which the
protein backbone atoms in the secondary structures were
constrained as in the previous step. The options of MD at 300
K with a 0.2-ps coupling constant were a time step of 1 fs and
a nonbonded update every 25 fs. The lengths of bonds with
hydrogen atoms were constrained according to the SHAKE
algorithm.54 The average structure from the last 10-ps trajec-
tory of MD was reminimized with backbone constraints in the
secondary structure and then without all constraints as
described above.
For the accuracy of the three-dimensional A2AAR model, the
distribution of the main chain torsion angles æ and ψ was
examined in a Ramachandran plot. Also, all ω angles for the
peptide planarity were measured. We checked that all CR
atoms of the receptor backbone amino acid residues were of
the L-configuration. The rmsd between backbone atoms in all
helices was compared with the X-ray structure of rhodopsin
as a template. The coordinates of the optimized human A2AAR
model (1upe) are available from the Protein Data Bank at
www.rcsb.org/pdb/.
Flexible docking was facilitated through the FlexiDock
utility in the Biopolymer module of SYBYL 6.7.1 (Tripos, St.
Louis, MO). During flexible docking, only the ligand was
defined with rotatable bonds. After the hydrogen atoms were
added to the receptor, atomic charges were recalculated by
using Kollman All-atom for the protein and Gasteiger-Hu¨ckel
for the ligand. H-bonding sites were marked for all residues
in the active site and for ligands that were able to act as
H-bond donor or acceptor. Ligands were variously preposi-
tioned in the putative binding cavity guided by point mutation
results as a starting point for FlexiDock. Default FlexiDock
parameters were set at 3000-generation for genetic algorithms.
To increase the binding interaction, the torsion angles of the
side chains that directly interacted within 5 Å of the ligands,
according to the results of FlexiDock, were manually adjusted.
Finally, the complex structure was minimized by using an
Amber force field with a fixed dielectric constant (4.0), until
Sta tistica l An a lysis. Binding and functional parameters
were estimated with GraphPad Prism software (GraphPad,
San Diego, CA). IC50 values obtained from competition curves
were converted to Ki values by using the Cheng-Prusoff
equation.48 Data were expressed as mean ( standard error.
Molecu la r Mod elin g. All calculations were performed on
a Silicon Graphics (Mountain View, CA) Octane workstation
(300 MHz MIPS R12000 (IP30) processor). All ligand struc-
tures were constructed with the use of the Sketch Molecule of
SYBYL 6.7.1.49 A conformational search of antagonists 1 and
2 was performed by grid search of flexible bonds, rotating at
60°, 180°, and -60° for the propyl group at the N7 position
and 0° or 180° for the bond between the furan and quinazoline
rings. The low-energy conformers from the grid search were
reoptimized, removing all constraints. A random search for
agonists 3 and 4 was performed to obtain thermally stable
conformations. The options of random search for all rotatable
the conjugate gradient reached 0.1 kcal‚mol-1‚Å-1
.
Ack n ow led gm en t. We thank Dr. Marlene J acobson
(Merck, West Point, PA) for the gift of the plasmid
coding the human A2AAR. We thank Dr. Gary Stiles
(Duke University, Durham, NC) for the gift of CHO cells

4858 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23
Kim et al.
(17) J acobson, K. A.; Gao, Z. G.; Chen, A.; Barak, D.; Kim, S. A.; Lee,
K.; Link, A.; van Rompaey, P.; van Calenbergh, S.; Liang, B. T.
Neoceptor Concept Based on Molecular Complementarity in
GPCRs: A Mutant Adenosine A3 Receptor with Selectively
Enhanced Affinity for Amine-Modified Nucleosides. J . Med.
Chem. 2001, 44, 4125-4136.
(18) Parr, I. B.; Horenstein, B. A. New Electronic Analogues of the
Sialyl Cation: N-Functionalized 4-Acetamido-2,4-dihydroxy-
piperidines. Inhibition of Bacterial Sialidases. J . Org. Chem.
1997, 62, 7489-7494.
(19) Lin, T.-S.; Zhu, J .-L.; Dutschman, G. E.; Cheng, Y.-C.; Prusoff,
W. H. Syntheses and Biological Evaluations of 3′-Deoxy-3′-C-
Branched-Chain-Substituted Nucleosides. J . Med. Chem. 1993,
36, 353-362.
expressing the human A3AR, and Dr. J oel Linden
(University of Virginia, Charlottesville, VA) for the gift
of CHO cells expressing the human A1AR and A2AARs.
S.-K.K. and Z.-G.G. thank Gilead Sciences (Foster City,
CA) for financial support. We thank Dr. Neli Melman
for technical assistance.
Su p p or tin g In for m a tion Ava ila ble: Synthetic methods
for preparation of compound 7. This material is available free
of charge via the Internet at http://pubs.acs.org.
(20) Filichev, V. V.; Brandt, M.; Pedersen, E. B. Synthesis of an Aza
Analogue of the 2-Deoxy-d-ribofuranose and Its Homologues.
Carbohydr. Res. 2001, 333, 115-122.
(21) Vorbru¨ggen, H.; Krolikiewicz, K.; Bennua, B. Nucleoside Syn-
thesis with Trimethylsilyl Triflate and Perchlorate as Catalysts.
Chem. Ber. 1981, 114, 1234-1255.
(22) van Tilburg, E. W.; van der Klein, P. A. M.; von Frijtag Drabbe
Ku¨nzel, J .; de Groote, M.; Stannek, C.; Lorenzen, A.; IJ zerman,
A. P. 5′-O-Alkyl Ethers of N, 2-Substituted Adenosine Deriva-
tives: Partial Agonists for the Adenosine A1 and A3 Receptors.
J . Med. Chem. 2001, 44, 2966-2975.
(23) Kontoyianni, M.; DeWeese, C.; Penzotti, J . E.; Lybrand, T. P.
Three-Dimensional Models for Agonist and Antagonist Com-
plexes with â2-Adrenergic Receptor. J . Med. Chem. 1996, 39,
4406-4420.
(24) IJ zerman, A. P.; Von Frijtag Drabbe Ku¨nzel, J . K.; Kim, J .; J iang,
Q.; J acobson, K. A. Site-Directed Mutagenesis of the Human
Adenosine A2A Receptor. Critical Involvement of Glu13 in Agonist
Recognition. Eur. J . Pharmacol. 1996, 310, 269-272.
(25) Barbhaiya, H.; McClain, R.; IJ zerman, A.; Rivkees, S. A. Site-
Directed Mutagenesis of the Human A1 Adenosine Receptor:
Influences of Acidic and Hydroxy Residues in the First Four
Transmembrane Domains on Ligand Binding. Mol. Pharmacol.
1996, 50, 1635-1642.
(26) Colson, A.-O.; Perlman, J . H.; J insi-Parimoo, A.; Nussenzveig,
D. R.; Osman, R.; Gershengorn, M. C. A hydrophobic cluster
between transmembrane helices 5 and 6 constrains the thyro-
tropin-releasing hormone receptor in an inactive conformation.
Mol. Pharmacol. 1998, 54, 968-978.
(27) Weiss, H. M.; Grisshammer, R. Purification and characterization
of the human adenosine A2A receptor functionally expressed in
Escherichia coli. Eur. J . Biochem. 2002, 269, 82-92.
(28) Archer, E.; Maigret, B.; Escrieut, C.; Pradayrol, L.; Fourmy, D.
Rhodopsin crystal: new template yielding realistic models of
G-protein-coupled receptors? Trends Pharm. Sci. 2003, 24, 36-
40.
(29) Chambers, J . J .; Nichols, D. E. A homology-based model of the
human 5-HT2A receptor derived from an in silico activated
G-protein coupled receptor. J . Comput. Aided Mol. Des. 2002,
16, 511-520.
(30) Filipek, S.; Teller, D. C.; Palczewski, K.; Stenkamp, R. The
crystallographic model of rhodopsin and its use in studies of
other G protein-coupled receptors. Annu. Rev. Biophys. Biomol.
Struct. 2003, in press.
(31) J udson, R. Genetic Algorithms and Their Use in Chemistry. In
Reviews in Computational Chemistry, Vol. 10; Lipkowitz, K. B.,
Boyd, D. B., Eds.; VCH Publishers: New York. 1997; pp 1-73.
(32) Moro, S.; Li, A.-H.; J acobson, K. A. Molecular Modeling Studies
of Human A3 Adenosine Antagonists: Structural Homology and
Receptor Docking. J . Chem. Inform. Comput. Sci. 1998, 38,
1239-1248.
(33) Baraldi, P. G.; Cacciari, B.; Spalluto, G. P.; de las Infantas y
Villatoro, M. J . P.; Zocci, C.; Dionisotti, S.; Ongini, E. Pyrazolo-
[4,3-e]-1,2,4-triazolo[1,5-c]pyrimidine Derivatives: Potent and
Selective A2A Adenosine Antagonists. J . Med. Chem. 1996, 39,
1164-1171.
(34) Baraldi, P. G.; Cacciari, B.; Romagnoli, R.; Spalluto, G.; Mo-
nopoli, A.; Ongini, E.; Varani, K.; Borea, P. A. 7-Substituted
5-Amino-2-(2-furyl)pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]pyrim-
idines as A2A Adenosine Receptor Antagonists: A Study on the
Importance of Modifications at the Side Chain on the Activity
and Solubility. J . Med. Chem. 2002, 45, 115-126.
(35) Baraldi, P. G.; Cacciari, B.; Spalluto, G.; Bergonzoni, M.;
Dionisotti, S.; Ongini, E.; Varani, K.; Borea, P. A. Design,
Synthesis, and Biological Evaluation of a Second Generation of
Pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]pyrimidines as Potent and
Selective A2A Adenosine Receptor Antagonists. J . Med. Chem.
1998, 41, 2126-2133.
(36) Baraldi, P. G.; Cacciari, B.; Moro, S.; Spalluto, G.; Pastorin, G.;
Ros, T. D.; Klotz, K.-N.; Varani, K.; Gessi, S.; Borea, P. A.
Synthesis, Biological Activity, and Molecular Modeling Inves-
tigation of New Pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]pyrimidine
Derivatives as Human A3 Adenosine Receptor Antagonists. J .
Med. Chem. 2002, 45, 770-780.
Refer en ces
(1) Fredholm, B. B.; IJ zerman, A. P.; J acobson, K. A.; Klotz, K.-N.;
Linden, J . International Union of Pharmacology. XXV. Nomen-
clature and Classification of Adenosine Receptors. Pharmacol.
Rev. 2001, 53, 527-552.
(2) Cassada, D. C.; Tribble, C. G.; Long, S. M.; Kaza, A. K.; Linden,
J .; Rieger, J . M.; Rosin, D.; Kron, I. L.; Kern, J . A. Adenosine
A2A agonist reduces Paralysis after Spinal Cord Ischemia:
Correlation with A2A Receptor Expression on Motor Neurons.
Ann. Thorac. Surg. 2002, 74, 846-849.
(3) Ohta, A.; Sitkovsky, M. Role of G-Protein-Coupled Adenosine
Receptors in Downregulation of Inflammation and Protection
from Tissue Damage. Nature 2001, 414, 916-920.
(4) Kim, J .; Wess, J .; van Rhee, A. M.; Scho¨neberg, T.; J acobson, K.
A. Site-directed Mutagenesis Identifies Residues Involved in
Ligand Recognition in the Human A2a Adenosine Receptor. J .
Biol. Chem. 1995, 270, 13987-13997.
(5) Kim, J .; J iang, Q.; Glashofer, M.; Yehle, S.; Wess, J .; J acobson,
K. A. Glutamate Residues in the Second Extracellular Loop of
the Human A2a Adenosine Receptor Are Required for Ligand
Recognition. Mol. Pharmacol. 1996, 49, 683-691.
(6) J iang, Q.; van Rhee, M.; Kim. J .; Yehle, S.; Wess, J .; J acobson,
K. A. Hydrophilic Side Chains in the Third and Seventh
Transmembrane Helical Domains of Human A2A Adenosine
Receptors Are Required for Ligand Recognition. Mol. Pharmacol.
1996, 50, 512-521.
(7) Stone, T. W.; Collis, M. G.; Williams, M.; Miller, L. P.; Karasawa,
A.; Hillaire-Buys, D. Adenosine: Some Therapeutic Applications
and Prospects. In Pharmacological Sciences: Perspectives for
Research and Therapy in the Late 1990s; Cuello, A. C., Collier,
B., Eds.; Birkhauser Verlag: Basel, Switzerland, 1995; pp 303-
309.
(8) Richardson, P. J .; Kase, H.; J enner, P. G. Adenosine A2A Receptor
Antagonists as New Agents for the Treatment of Parkinson’s
Disease. Trends Pharmacol. Sci. 1997, 18, 338-344.
(9) Ledent, C.; Vaugeois, J . M.; Schiffmann, S. N.; Pedrazzini, T.;
El Yacoubi, M.; Vanderhaeghen, J . J .; Costentin, J .; Heath, J .
K.; Vassart, G.; Parmentier, M. Aggressiveness, Hypoalgesia and
High Blood Pressure in Mice Lacking the Adenosine A2A Recep-
tor. Nature (London) 1997, 388, 674-678.
(10) Baraldi, P. G.; Cacciari, B.; Spalluto, G.; Borioni, A.; Viziano,
M.; Dionisotti, S.; Ongini, E. Current Developments of A2A
Adenosine Receptor Antagonists. Curr. Med. Chem. 1995, 2,
707-722.
(11) IJ zerman, A. P.; van der Wenden, E. M.; van Galen, P. J . M.;
J acobson, K. A. Molecular Modeling of Adenosine Receptors. II.
The Ligand Binding Site on the Rat A2a Receptor. Eur. J .
Pharmacol.: Mol. Pharmacol. 1994, 268, 95-104.
(12) Gao, Z.-G.; Chen, A.; Barak, D.; Kim, S.-K.; Mu¨ller, C. E.;
J acobson, K. A. Identification by Site-directed Mutagenesis of
Residues Involved in Ligand Recognition and Activation of the
Human A3 Adenosine Receptor. J . Biol. Chem. 2002, 277,
19056-19063.
(13) Palczewski, K.; Kumasaka, T.; Hori, T.; Behnke, C. A.; Mo-
toshima, H.; Fox, B. A.; Le Trong, I.; Teller, D. C.; Okada, T.;
Stenkamp, T. E.; Yamamoto, M.; Miyano, M. Crystal Structure
of Rhodopsin: A G Protein-Coupled Receptor. Science 2000, 289,
739-745.
(14) Ballesteros, J . A.; Shi, L.; J avitch, J . A. Structural Mimicry in
G Protein-Coupled Receptors: Implications of the High-Resolu-
tion Structure of Rhodopsin for Structure-Function Analysis
of Rhodopsin-Like Receptors. Mol. Pharmacol. 2001, 60, 1-19.
(15) Gao, Z.-G.; Kim, S.-K.; Biadatti, T.; Chen, Q.; Lee, K.; Barak,
D.; Kim, S. G.; J ohnson, C. R.; J acobson K. A. Structural
Determinants of A3 Adenosine Receptor Activation: Nucleoside
Ligands at the Agonist/Antagonist Boundary. J . Med. Chem.
2002, 45, 4471-4484.
(16) Our most recent model of the human A3AR (by Dr. Soo-Kyung
Kim) is now available on the pdb ftp site. Entry in PDB
format (compressed): pdblo74.ent. Z.; Entry in mmCIF
format (compressed): lo74.cif. Z. See http:/ / www.rcsb.org/ pdb/
latest_news.html#models_removal2.

Modeling the Adenosine Receptors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23 4859
(37) Baraldi, P. G.; Cacciari, B.; Romangnoli, R.; Spalluto, G.; Moro,
S.; Klotz, K.-N.; Leung, E.; Varani, K.; Gessi, S.; Merighi, S.;
Borea, P. A. Pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]pyrimidine De-
rivatives as Highly Potent and Selective Human A3 Adenosine
Receptor Antagonists: Influence of the Chain at the N8 Pyrazole
Nitrogen. J . Med. Chem. 2000, 43, 4768-4780.
(38) Baraldi, P. G.; Cacciari, B.; Romangnoli, R.; Spalluto, G.; Klotz,
K.-N.; Leung, E.; Varani, K.; Gessi, S.; Merighi, S.; Borea, P. A.
Pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]-pyrimidine Derivatives as
Highly Potent and Selective Human A3 Adenosine Receptor
Antagonists. J . Med. Chem. 1999, 42, 4473-4478.
(49) Sybyl Molecular Modeling System, version 6.7.1. Tripos Inc., St.
Louis, MO.
(50) Halgren, T. A. MMFF VII. Characterization of MMFF94,
MMFF94s, and Other Widely Available Force Fields for Con-
formational Energies and for Intermolecular-Interaction Ener-
gies and Geometries. J . Comput. Chem. 1999, 20, 730-748.
(51) Stewart, J . J . P. MOPAC: A Semiempirical Molecular Orbital
Program. J . Comput. Aided Mol. Des. 1990, 4, 1-105.
(52) van Rhee, A. M.; J acobson, K. A. Molecular Architecture of G
Protein-Coupled Receptors. Drug Devel. Res. 1996, 37, 1-38.
(53) Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, J r.,
K. M.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell, J .
W.; Kollman, P. A. A Second Generation Force Field for the
Simulation of Proteins, Nucleic Acids and Organic Molecules.
J . Am. Chem. Soc. 1995, 117, 5179-5197.
(54) Ryckaert, J . P.; Ciccotti, G.; Berendsen, H. J . C. Numerical
Integration of the Cartesian Equations of Motion for a System
with Constraints: Molecular Dynamics of n-alkanes. J . Comput.
Phys. 1977, 23, 327-333.
(55) Rivkees, S. A.; Barbhaiya, H.; IJ zerman A. P. Identification of
the adenine binding site of the human A1 adenosine receptor.
J . Biol. Chem. 1999, 274, 3617-3621.
(56) Scholl, D. J .; Wells, J . N. Serine and alanine mutagenesis of the
nine native cysteine residues of the human A1 adenosine
receptor. Biochem. Pharmacol. 2000, 60, 1647-1654.
(57) J iang Z.; Lee, B. X.; Glashofer, M.; van Rhee, A. M.; J acobson,
K. A. Mutagenesis reveals structure-function parallels between
human A2A-adenosine receptors and the biogenic amine family.
J . Med. Chem. 1997, 40, 2588-2595.
(58) Olah, M. E.; Ren H.; Ostrowski, J .; J acobson, K. A.; Stiles, G. L.
Cloning, expression, and characterization of the unique bovine
A1 adenosine receptor. Studies on the ligand binding site by site-
directed mutagenesis. J . Biol. Chem. 1992, 267, 10764-10770.
(59) Tucker, A. L.; Robeva, A. S.; Taylor, H. E.; Holeton, D.; Bockner,
M.; Lynch, K. R.; Linden, J . A1 adenosine receptors. Two amino
acids are responsible for species differences in ligand recognition.
J . Biol. Chem. 1994, 269, 27900-27906.
(39) Vittori, S.; Camaioni, E.; Francesco, E. D.; Volpini, R.; Monopoli,
A.; Dionisotti, S.; Ongini, E.; Cristalli, G. 2-Alkenyl and 2-Alky
Derivatives of Adenosine and Adenosine-5′-N-Ethyluronamide:
Different Affinity and Selectivity of E- and Z-Diastereomers at
A2A Adenosine Receptors. J . Med. Chem. 1996, 39, 4211-4217.
(40) Lee, K.; Ravi, R. G.; J i, X-d.; Marquez, V. E.; J acobson, K. A.
Ring-Constrained (N)Methanocarba-Nucleosides as Adenosine
Receptor Agonists: Independent 5′-Uronamide and 2′-Deoxy
Modifications. Bioorg. Med. Chem. Lett. 2001, 11, 1333-1337.
(41) J acobson, K. A.; J i, Xi-d.; Li, A.-H.; Melman, N.; Siddiqui, M.
A.; Shin, K.-J .; Marquez, V. E.; Ravi, R. G. Methanocarba
Analogues of Purine Nucleosides as Potent and Selective Ad-
enosine Receptor Agonists. J . Med. Chem. 2000, 43, 2196-2203.
(42) Farrens, D. L.; Altenbach, C.; Yang, K.; Hubell, W. L.; Khorana,
H. G. Requirements of rigid-body motion of transmembrane
helices for light activation of rhodopsin. Science 1996, 274, 768-
770.
(43) Shi, L.; Liapakis, G.; Xu, R.; Guarneri, F.; Ballesteros, J . A.;
J avitch, J . A. â-Adrenergic Receptor Activation: Modulation of
the Proline Kink in Transmembrane 6 by a Rotamer Toggle
Switch. J . Biol. Chem. 2002, 277, 40989-40996.
(44) Liu, J .; Conklin, B. R.; Blin, N.; Yun, J .; Wess, J . Identification
of a Receptor/G-Protein Contact Site Critical for Signaling
Specificity and G-Protein Activation. Proc. Natl. Acad. Sci.
U.S.A. 1995, 92, 11642-11646.
(45) Cullen, B. R. Use of Eukaryotic Expression Technology in the
Functional Analysis of Cloned Genes. Methods Enzymol. 1987,
152, 684-704.
(60) Townsend-Nicholson, A.; Schofield, P. R. A threonine residue
in the seventh transmembrane domain of the human A1 adeno-
sine receptor mediates specific agonist binding. J . Biol. Chem.
1994, 269, 2373-2376.
(46) Bradford, M. M. A Rapid and Sensitive Method for the Quan-
titation of Microgram Quantities of Protein Utilizing the Prin-
ciple of Protein-Dye Binding. Anal. Biochem. 1976, 72, 248-
254.
(61) Dalpiaz, A.; Townsend-Nicholson, A.; Beukers, M. W.; Schofield,
P. R.; IJ zerman, A. P. Thermodynamics of full agonist, partial
agonist, and antagonist binding to wild-type and mutant ad-
enosine A1 receptors. Biochem. Pharmacol. 1998, 56, 1437-
1445.
(47) Gao, Z.-G.; J iang, Q.; J acobson, K. A.; IJ zerman, A. P. Site-
Directed Mutagenesis Studies of Human A2A Adenosine Recep-
tors. Involvement of Glu13 and His278 in Ligand Binding and
Sodium Modulation. Biochem. Pharmacol. 2000, 60, 661-668.
(48) Cheng, Y.-C.; Prusoff, W. H. Relationship between the Inhibition
Constant (Ki) and the Concentration of Inhibitor Which Causes
50 Percent Inhibition (I50) of an Enzymatic Reaction. Biochem.
Pharmacol. 1973, 22, 3099-3108.
J M0300431