Structure-kinetics relationships of Capadenoson derivatives as adenosine A1 receptor agonists

Article abstract of DOI:10.1016/j.ejmech.2015.07.023

Abstract We report the synthesis and biological evaluation of new derivatives of Capadenoson, a former drug candidate that was previously advanced to phase IIa clinical trials. 19 of the 20 ligands show an affinity below 100 nM at the human adenosine A₁ receptor (hA₁AR) and display a wide range of residence times at this target (from approx. 5 min (compound 10) up to 132 min (compound 5)). Structure-affinity and structure-kinetics relationships were established, and computational studies of a homology model of the hA₁AR revealed crucial interactions for both the affinity and dissociation kinetics of this family of ligands. These results were also combined with global metrics (Ligand Efficiency, cLogP), showing the importance of binding kinetics as an additional way to better select a drug candidate amongst seemingly similar leads.

Full text of DOI:10.1016/j.ejmech.2015.07.023

European Journal of Medicinal Chemistry 101 (2015) 681e691
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
Structure-kinetics relationships of Capadenoson derivatives as
adenosine A1 receptor agonists
Julien Louvel*, Dong Guo, Marjolein Soethoudt, Tamara A.M. Mocking, Eelke B. Lenselink,
Thea Mulder-Krieger, Laura H. Heitman, Adriaan P. IJzerman
Division of Medicinal Chemistry, Leiden Academic Centre for Drug Research, Leiden University, P.O. Box 9502, 2300 RA Leiden, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history:
We report the synthesis and biological evaluation of new derivatives of Capadenoson, a former drug
candidate that was previously advanced to phase IIa clinical trials. 19 of the 20 ligands show an affinity
below 100 nM at the human adenosine A1 receptor (hA1AR) and display a wide range of residence times
at this target (from approx. 5 min (compound 10) up to 132 min (compound 5)). Structure-affinity and
structure-kinetics relationships were established, and computational studies of a homology model of the
hA1AR revealed crucial interactions for both the affinity and dissociation kinetics of this family of ligands.
These results were also combined with global metrics (Ligand Efficiency, cLogP), showing the importance
of binding kinetics as an additional way to better select a drug candidate amongst seemingly similar
leads.
Received 11 February 2015
Received in revised form
10 July 2015
Accepted 11 July 2015
Available online 15 July 2015
Keywords:
Residence time
Structure-affinity relationships
Structure-kinetics relationships
Adenosine A1 receptor
Extracellular loops
© 2015 Elsevier Masson SAS. All rights reserved.
1. Introduction
(GPCRs). Their physiological and clinical relevance have been
extensively studied [1]. Since the early 1990s, many synthetic li-
The adenosine receptors (A1, A2A, A2B and A3), all of which accept
adenosine as an endogenous ligand, belong to the G protein-
coupled receptor superfamily of cell membrane-bound proteins
gands that selectively bind to a determined number of receptor
subtypes have been developed, along with different functional
profiles: partial or full agonists and antagonists/inverse agonists.
Among these, compounds such as Capadenoson (Bayer) have
emerged as non-ribose agonists for the A1 Adenosine Receptor
(A1AR) and the A2BAR [2,3]. In the meantime, Otsuka Pharmaceu-
Abbreviations: ADA, adenosine deaminase; Ac, acetyl; BCA, bicinchoninic acid;
BEDROC, Boltzmann-enhanced discrimination of receiver-operating characteristic;
BSA, bovine serum albumin; CPA, N6-cyclopentyladenosine; CHO, Chinese hamster
ovary; DMF, N,N-dimethylformamide; DMSO, dimethylsulfoxide; DPCPX, 1,3-
dipropyl-8-cyclopentyl-xanthine; DTT, dithiothreitol; EC50, concentration of unla-
beled ligand which elicits 50% of the maximum response (Emax) in a functional
assay at membranes of CHO cells stably expressing the adenosine A1 receptor;
EDTA, ethylenediaminetetraacetic acid; EF, Enrichment factor; Emax, maximum
response elicited by an unlabeled ligand in a functional assay (relatively to CPA) at
membranes of CHO cells stably expressing the adenosine A1 receptor; FBS, fetal
bovine serum; G418, Geneticin; GDP, guanosine diphosphate; GF, glass filter; GTP,
ticals described
a
series of 4-amino-6-aryl-5-cyano-2-
thiopyrimidines as selective agonists for the A2AAR [4]. Using this
scaffold, our group has recently reported a series of derivatives that
are selective for the A1AR. 5 These compounds are shown in Fig. 1.
In the latter paper we emphasized the relationships between the
structure and binding kinetics of the new compounds. Indeed it has
been shown 23 that dissociation kinetics, embodied by residence
time (RT), may be a more accurate reflection of in vivo behavior
than in vitro affinity and/or potency [6e8]. RT is defined as the
reciprocal of the ligand's dissociation rate constant (RT ¼ 1/koff) and
represents the lifetime of the ligand-receptor complex [9]. Inci-
dentally, Capadenoson was advanced to clinical studies, but failed
to show heart rate reduction in patients with atrial fibrillation in
phase IIa clinical trials [10].
guanosine triphosphate; GTPg g-thio]triphosphate; HPLC, high-
S, guanosine 50-O-[
pressure liquid chromatography; IC50, concentration of unlabeled ligand which
displaces 50% of [3H]DPCPX binding to membranes of CHO cells stably expressing
the adenosine A1 receptor; Ki, affinity of ligand; koff, dissociation rate constant at
the hA1AR; kon, association rate constant at the hA1AR; KRI, kinetic rate index;
NMM, N-methylmorpholine; NMR, nuclear magnetic resonance; PBS, phosphate
buffered saline; RT, residence time; SAR, structure-affinity relationships; SKR,
structure kinetics relationships; THF, tetrahydrofuran; TLC, thin layer chromatog-
raphy; Tris, tris(hydroxymethyl)aminomethane.
In the current study we report the synthesis and evaluation of 6-
amino-4-aryl-3,5-dicyano-2-thiopyridines as (selective) A1AR ag-
onists. These compounds are derivatives of Capadenoson. They
* Corresponding author.
E-mail address: j.a.louvel@lacdr.leidenuniv.nl (J. Louvel).
http://dx.doi.org/10.1016/j.ejmech.2015.07.023
0223-5234/© 2015 Elsevier Masson SAS. All rights reserved.

682
J. Louvel et al. / European Journal of Medicinal Chemistry 101 (2015) 681e691
Fig. 1. Examples of non-ribose agonists for A1AR (top and bottom left), A1AR/A2AAR/A2BAR (top right), A2AAR (bottom right).
were evaluated in radioligand displacement binding assays and in a
competition association assay. The compounds are selective over
the other AR subtypes and show high affinities and a broad range of
kinetic profiles at the hA1AR, and their Structure-Kinetics Re-
lationships (SKR) were established. Their binding mode was
analyzed in a homology model of hA1AR, potentially shedding light
on key structural features of the receptor that are involved in ligand
dissociation. A representative set of ligands was also characterized
in a functional assay, in which they emerged as A1 receptor partial
or full agonists.
compound of interest at two time points and calculated the kinetic
rate index (KRI), as defined by the ratio of bindings at the earlier
and later time points, respectively; a KRI value above 1.0 indicates a
slower dissociation from the target, while a value below 1.0 predicts
a faster dissociation rate, as compared to the dissociation rate of the
radioligand. Subsequently to this screening, the compounds with a
KRI above 1.4 (Capadenoson had a KRI value of 1.42) were then
subjected to a full competition association assay so as to quanti-
tatively determine their kinetic parameters (kon, koff) and thus their
RT.
2.2.1. Structure-affinity relationships and structure-kinetics
relationships
2. Results and discussion
All binding and kinetic data at the A1AR discussed in this section
are gathered in Table 1.
2.1. Design and synthesis
The S-substituted thiopyridines were synthesized in a two-step
procedure. First, thiopyridines 1e3 were obtained in a previously
reported one-pot, two-step reaction between malononitrile, thio-
phenol and various substituted benzaldehydes in the presence of
piperidine and triethylamine at the reflux of ethanol, followed by
displacement of thiophenolate by sodium sulfide at 80 ꢀC in DMF
[3]. Then the alkylation of the obtained thiols 1e3 with various
alkyl chlorides in the presence of NaHCO3 or DBU at r.t. in DMF
afforded compounds 5e23 (Scheme 1) [3a,11]. Capadenoson 4 was
synthesized using a reported procedure [11].
2.2.1.1. Aromatic substituent. Starting from Capadenoson, a brief
screening of the substituent attached to the distal phenyl ring was
carried out.
2.2.1.1.1. Structure-affinity relationships. In our hands, 4 (Capa-
denoson) had a high affinity (1.4 nM) at the A1AR. Truncation of the
terminal hydroxyethoxy fragment to a methoxy group led to a
slightly decreased affinity (5, 5 nM). This result is different from our
previous findings on a series of cyanopyrimidine analogs, for which
a methoxy substituent at this position had led to a drastic drop in
affinity (Ki < 100 nM) [5]. Further shortening of the chain to a hy-
droxy substituent did not affect binding (6, 1.5 nM). Finally, intro-
The alkyl chlorides 29e31 and 33, 34 that were not commer-
cially available were synthesized according to the method
described previously [5], or by chlorination or bromination of 4-
(chloromethyl)-2-(4-chlorophenyl)thiazole 32 (Scheme 1) [12].
ducing
a 3,4-methylenedioxy substituent, which had proven
valuable in the previous study [5], yielded a similar affinity, com-
parable to that of 5 (7, 4.6 nM).
2.2. Biology
All the compounds were tested in a radioligand displacement
assay on the adenosine A1 receptor in the presence of 2.5 nM [3H]
DPCPX [13]. All compounds, except 9, displayed a high affinity for
the human A1AR (Ki < 100 nM). Subsequently, compounds with
high A1AR affinity were screened in a dual-point competition as-
sociation assay in the presence of [3H]DPCPX (RT ¼ 4.8 min), which
we recently applied to similar studies [5,14]. In brief, we deter-
mined specific [3H]DPCPX binding in the presence of an unlabeled
2.2.1.1.2. Structure-kinetics relationships. All four compounds dis-
played a KRI value above 1.4, 4 having the lower (1.42). 6 and 7 had
similar KRI values (2.07 and 2.00, respectively), with associated
residence times of 36 and 63 min respectively. The big difference
between the residence times with identical KRI values can be
explained by a big spread in the KRI value of 6 (values of 1.6 and
2.6). Noteworthy is that these three compounds (4, 6 and 7) have
longer residence times at the A1AR than their cyanopyrimidine

J. Louvel et al. / European Journal of Medicinal Chemistry 101 (2015) 681e691
683
Scheme 1. Synthesis of Compounds 4e22 and alkylating agents 29e31 and 33e34.a
counterparts [5], which indicates a crucial influence of the addi-
tional cyano substituent on dissociation rates. 5 showed the highest
KRI value (2.55), along with the longest residence time of the whole
series of compounds in our study (132 min). Taken all together, the
compounds with a more lipophilic substituent (5 and 7) show
slower dissociation from the A1AR than the compounds bearing a
hydroxyl group (4 and 6). As far as association rates are concerned,
4 displays the highest value (kon ¼ 2.4 ꢁ 107 Mꢂ1 minꢂ1), six to ten
range. When the phenyl ring was exchanged for a morpholine ring,
a higher affinity was measured (12, 3.2 nM). These results seem to
be in contrast with our previous findings on the cyanopyrimidine
analogs of these compounds, for which the affinity decreased in the
morpholino < phenyl < H order [5]. This could be due to the
presence of an additional cyano substituent on the central het-
erocycle, and/or the methoxy group instead of a methylenedioxy
group on the distal phenyl ring, both of which could slightly change
the binding pose of the ligand.
The substitution pattern of the phenyl ring was then studied.
The introduction of a halogen on the 4-position did not lead to a
significant increase in affinity (4-F, 13, 5.0 nM vs H, 11, 8.4 nM). An
increase in size and lipophilicity in the halogen series also led to
similar results (F, 11; Cl, 5; Br, 14; I, 15). In contrast to the analogous
cyanopyrimidines, moving the chloro substituent along the ring did
not change the affinity much (4-Cl, 5, 5.0 nM; 3-Cl, 16, 12 nM; 2-Cl,
17,11 nM), which seems to indicate that the dihedral angle between
the thiazolyl and phenyl rings does not have a small value (which
would be disfavored by the presence of a 2-chloro substituent) [16]
in the binding pose. Also, dichlorination on the 3 and 4 position
gave a similar affinity (18, 11 nM), which is again in contrast with
our previous findings on cyanopyrimidine analogs. Replacing the
chloro substituent with a methoxy group did not change the af-
finity (19, 9.3 nM), while the introduction of a lipophilic methyl
group led to a twofold loss of binding (20, 19 nM).
times higher than
5
(kon
¼
2.6
ꢁ
106 Mꢂ1 minꢂ1),
6
(kon ¼ 4.3 ꢁ 106
M
ꢂ1 minꢂ1) and 7 (kon ¼ 2.4 ꢁ 106 Mꢂ1 minꢂ1).
2.2.1.2. Thiazole substituent. With the 4-methoxy substituent on
the phenyl ring adjacent to the pyridine core the influence of the
nature of the substituents on the thiazole ring was then studied.
2.2.1.2.1. 5-Position. Two compounds bearing a halogen atom (8, Cl
and 9, Br) on the 5-position of the thiazole ring were synthesized to
study the potential for halogen bonding at this position [15].
Introduction of a chloro substituent led to an 18-fold decrease in
affinity (8, 88 nM vs 5, 5 nM), while a bromo substituent further
halved the affinity (9, 162 nM). The latter compound was not
evaluated in the competition association assay because it displayed
an affinity above the 100 nM threshold. 8, on the other hand, had a
KRI of 1.79, with an associated residence time of 36 min. Both af-
finity and RT show that no halogen bonding seems to be taking
place, because both the affinity and the residence time decrease
from 5 to 8.
Finally, the phenyl ring was changed into a 2-pyridyl ring: the
introduction of a nitrogen atom at this position with a halogen
substituent at the para position (relative to the thiazole) leads to an
increase of the sigma-hole on the halogen atom, which could
reinforce halogen bonding with residues in the vicinity of the
ligand [15]. The effect of the pyridyl ring itself was negligible in
terms of affinity (21, 9.5 nM vs 11, 8.4 nM). Addition of a chlorine or
a bromine atom led to a 5- and 14-fold decrease when compared to
2.2.1.2.2. 2-Position
2.2.1.2.2.1. Structure-affinity relationships. The unsubstituted thia-
zole (R3 ¼ H) gave a high affinity (10, 1.3 nM), the highest of this
subseries of compounds. Addition of a phenyl ring led to a signifi-
cant decrease in affinity (11, 8.4 nM), albeit still in the nanomolar

684
J. Louvel et al. / European Journal of Medicinal Chemistry 101 (2015) 681e691
Table 1
Structuredaffinity relationships and Structure-Kinetics relationships of 4e23.
d
nr.
R1
R3
X
Ki (nM) SEMa
1.4 0.15
KRIb
kon (Mꢂ1 minꢂ1
)
koff (minꢂ1 e (RT (min))f
)
4
4-O(CH2)2OH
4-Cl-Ph
H
1.42
(1.29; 1.55)
2.55
(1.56; 3.54)
2.07
(1.58; 2.56)
2.00
2.4 0.90 ꢁ 107
2.6 0.6 ꢁ 106
4.3 1.6 ꢁ 106
2.4 0.25 ꢁ 106
1.4 0.26 ꢁ 106
0.036 0.013
(27.8)
0.0076 0.0003
(132)
0.028 0.01
(35.7)
0.016 0.001
(62.5)
5
6
7
8
4-OMe
4-OH
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
H
H
H
Cl
5.0 0.57
1.5 0.47
4.6 1.1
3,4-OCH2O
4-OMe
(1.90; 2.20)
1.79
88
8
0.028 0.002
(35.7)
(1.72; 1.86)
9
10
4-OMe
4-OMe
4-Cl-Ph
H
Br
H
162 14
1.3 0.38
ec
ec
ec
ec
ec
1.00
(0.97; 1.03)
1.23
(1.16; 1.30)
1.12
(0.91; 1.33)
1.56
(1.49; 1.63)
1.55
(1.51; 1.59)
2.07
(1.83; 2.32)
1.60
(1.32; 1.88)
1.53
(1.32; 1.74)
2.35
(2.23; 2.46)
1.22
(1.19; 1.25)
1.25
(1.25; 1.26)
1.08
(1.11; 1.05)
1.64
11
12
13
14
15
16
17
18
19
20
21
22
23
4-OMe
4-OMe
4-OMe
4-OMe
4-OMe
4-OMe
4-OMe
4-OMe
4-OMe
4-OMe
4-OMe
4-OMe
4-OMe
Ph
H
H
H
H
H
H
H
H
H
H
H
H
H
8.4 0.80
3.2 1.2
2.3 0.21
3.8 0.34
3.9 0.41
12 0.19
11 0.68
11 0.78
9.3 0.95
19 2.7
ec
ec
ec
morpholine
4-F-Ph
ec
4.6 0.90 ꢁ 106
2.4 0.50 ꢁ 106
1.6 0.40 ꢁ 106
7.7 2.2 ꢁ 105
9.4 2.5 ꢁ 105
1.1 0.20 ꢁ 106
ec
0.033 0.007
(30.3)
0.020 0.005
(50.0)
0.011 0.004
(90.9)
0.025 0.007
(40.0)
4-Br-Ph
4-I-Ph
3-Cl-Ph
2-Cl-Ph
3,4-diCl-Ph
4-OMe-Ph
4-Me-Ph
2-Pyr
0.024 0.007
(41.7)
0.019 0.005
(52.6)
ec
ec
ec
ec
9.5 0.90
24 4.3
ec
5-Cl-2-Pyr
5-Br-2-Pyr
3.4 0.28 ꢁ 106
1.4 0.20 ꢁ 106
0.035 0.008
(28.6)
0.030 0.002
(33.3)
(1.64; 1.64)
1.71
(1.66; 1.75)
53 8.6
a
Ki SEM (n ¼ 3), obtained from radioligand binding assays with [3H]DPCPX on CHO cell membranes stably expressing human adenosine A1 receptors.
KRI (n1, n2) (n ¼ 2), obtained from dual-point competition association assays with [3H]DPCPX on CHO cell membranes stably expressing human adenosine A1 receptors.
Not determined.
b
c
d
e
f
kon SEM (n ꢃ 3), obtained from competition association assays with [3H]DPCPX on CHO cell membranes stably expressing human adenosine A1 receptors.
koff SEM (n ꢃ 3), obtained from competition association assays with [3H]DPCPX on CHO cell membranes stably expressing human adenosine A1 receptors.
RT (residence time) ¼ 1/koff
.
the phenyl-ring analogs, respectively (22, 24 nM and 23, 53 nM).
One explanation for the loss of affinity could be that the halogen
substituent interacts with the binding pocket not through halogen
bonding but lipophilic or dipolar interactions. As a result an in-
crease of the size of the sigma-hole would diminish these in-
teractions through local depletion of electron density on the
chlorine atom.
the A1AR. Indeed, para-fluoro (13)-, chloro (5)-, bromo (14)- and
iodo (15)-substituted compounds all displayed KRI values above
1.50, with associated residence times ranging from 30 min (13) to
132 min (5). Surprisingly, there does not seem to be any progression
in the halogen series in terms of size, lipophilicity or electronega-
tivity that would correlate with the evolution of residence time,
since RT follows the F < Br < I < Cl order. When looking at associ-
ation rates, however, there is a somewhat clearer trend following
the increase of electronegativity or
s parameter. Indeed, the fluoro-
2.2.1.2.2.2. Structure-kinetics relationships. The ligand with the
shortest residence time was the smaller one (R3 ¼ H, 10), with a KRI
of 1.00. Adding a phenyl group slightly decreased the dissociation
rate (11, KRI ¼ 1.23), while introducing a morpholino substituent
had virtually no effect (12, KRI ¼ 1.12). Introduction of a halogen
atom on the phenyl ring dramatically increased residence time at
substituted compound associates almost twice as fast as its chloro
and bromo analogs, which in turn associate 1.5 times faster than
the iodo compound. Moving the chloro substituent to the meta or
ortho position has deleterious effects both on residence time and
association rate constant; indeed, 16 and 17 both dissociate more

J. Louvel et al. / European Journal of Medicinal Chemistry 101 (2015) 681e691
685
than 3 times faster and associate 30 times slower than 5.
Dichlorination at the meta and para positions led to a more than
twofold decrease of residence time when compared to the para-
chlorinated derivative (18, 53 min vs 5, 132 min), but slightly longer
than the meta-chlorinated compound (16, 40 min). The same trend
was observed in terms of evolution of the association rate constant.
Introduction of a methoxy or methyl group at the 4-position led to a
greatly increased dissociation rate (19, KRI ¼ 1.22; 20, KRI ¼ 1.25).
Changing the phenyl ring into a 2-pyridyl ring led to a shorter
residence time in the case of the unsubstituted derivative (21,
KRI ¼ 1.08 vs 11, KRI ¼ 1.23), as well as for the compound with a
chloro substituent para to the thiazole (22, RT ¼ 29 min vs 5,
RT ¼ 132 min). A slight decrease was observed in the case of the
bromo derivative (23, RT ¼ 33 min vs 14, RT ¼ 48 min).
[20], the reference radioligand for A2AAR, at a concentration of
M. At A2BAR, compounds 11 and 10 showed modest to good
affinity (63% displacement of the A2BAR radioligand [3H]PSB603 at
M and Ki ¼ 23 nM [21], respectively). This trend was the same as
with the cyanopyrimidine analogs [5]. At the A3AR, the displace-
ment of the radioligand [3H]PSB11 at 1
M of the selected com-
pounds stayed below 50% [22], indicating a selectivity of at least
100-fold for the A1AR. The most selective compound across all
subtypes was 4 (Table 2).
1
m
1
m
m
2.2.4. Functional assays
The same six compounds were tested in a [35S]GTP
gS functional
assay. In our hands, 4 gave the response (Emax ¼ 97%) of a full
agonist (CPA [23], 100%), and so did 11, 12 and 21. Compounds 5 and
10 showed a lower Emax of 83% and 55%, respectively, which are
characteristic of partial agonists (Table 2). All tested compounds
showed EC50 values in the nanomolar range, between 1.1 nM for 4
and 7.6 nM for 21 (Table 2), indicating that these ligands are potent
agonists at the hA1AR.
2.2.2. Ligand efficiency and Lipophilicity [17]
Ligand Efficiency (LE) is one of the many metrics used in drug
discovery for the selection of lead compounds; it represents the
binding energy per (non-hydrogen) atom of a given compound to a
target (LE ¼ DG=N, where DG is the Gibbs energy of binding and N
the number of non-hydrogen atoms) [18]. The higher LE, the more
efficient the binding, and the more likely it is to be a good drug
candidate. Reference compound 4 has an LE of 0.35, making it a
good drug candidate for this metric. From the other compounds in
this study, most (16) were in the 0.32 ꢄ LE ꢄ 0.38 range. Out of the
remaining four, three had a ligand efficiency below 0.30 (8, 9 and
23), and one above 0.4 (10, LE ¼ 0.47). The compound with the
longest residence time had the same ligand efficiency (5, LE ¼ 0.35)
as Capadenoson (4). Only its affinity was slightly lower than that of
4 (5.0 nM vs 1.4 nM). The calculated clogP of 4 is 5.04, whereas for 5
it is 5.73. In the end, this deviation from the ‘rule of 5’ [19] could be
compensated by the significant gain in residence time. Also note-
worthy is that the least lipophilic compound 10 (clogP ¼ 3.10) is
also the one with the shortest residence time; the opposite does not
hold true, however, because the most lipophilic compound that was
also screened for binding kinetics does not have the longest resi-
dence time (8, clogP ¼ 6.50, RT ¼ 36 min). For compounds with the
same ligand efficiency as 5, it appears that there are compounds
which are more lipophilic, yet dissociate faster from the A1AR (14,
clogP ¼ 5.89, RT ¼ 50 min; 15, clogP ¼ 6.05, RT ¼ 91 min). Hence,
there is no obvious correlation between overall compound lip-
ophilicity and residence time.
2.3. Computational studies
Finally, we decided to further investigate the ligand-receptor
interactions through docking studies. To this end, a homology
model of the hA1AR receptor was constructed (see Experimental
details).
When looking at the docking pose of 5 (Fig. 2), it appears that this
compound (and the others in this study) binds with the distal six-
membered aromatic ring pointing inside the binding cavity, which
is comparable to the binding mode of LUF5834 in the A2A receptor
[24]. The pyridine ringof 5 interacts with Phe171 (located on EL2) via
a pep interaction, which corresponds to the interaction of LUF5834
and Phe168 observed at the A2A receptor. The interaction of Asn254
with the amino group and the vicinal nitrile (Asn253 in A2AAR) via
hydrogen bonding is similar too in the two receptor models.
For the distal nitrile group, however, an interaction occurs with
Asn70, while in the A2A receptor a bridging water molecule be-
tween Ala63 and Tyr9 plays this role [24]. Another residue, Lys265,
located in EL3, also seems to entertain a crucial interaction with the
ligands; it is involved in two hydrogen bonds, one with the nitrogen
atom of the central pyridine ring of the ligand, and one with the
nitrogen atom of the thiazole ring. The substituent borne by the
thiazole ring (i.e. the substituted aromatic ring) does not appear to
be involved in any specific interaction as far as the model can show,
and the 4-chloro substituent appears to nest into the surface
defined by the backbone of TM7 in the vicinity of Lys265. The rest of
the aromatic system seems to be exposed to the solvent, since it is
in the vicinity of Tyr271 which is forming a hydrogen bond with
Asn70.
2.2.3. Selectivity
Six compounds representative of the structural variations
introduced in the study, as well as showing a good affinity at the
hA1AR, were selected to be tested for their selectivity against the
other adenosine receptor subtypes. All compounds were selective
over the A2AAR, showing negligible displacement of [3H]ZM241385
Table 2
Selectivity data and functional profile of selected compounds.
Displ. at A2AAR at 1
m
M (%)a
Displ. at A2BAR at 1
m
M (%)b
Displ. at A3AR at 1
m
M (%)c
EC50 (nM)d
Emax (%)e
4
5
10
11
12
21
0.0 (ꢂ6.0, 3.0)
5.8 (7.6, 4.0)
2.5 (5.0, 0.0)
20 (31, 8.5)
1.2 (2.4, 0.0)
19 (21, 18)
50 (46, 54)
27 (22, 33)
51 (48, 54)
22 (15, 30)
1.1 0.80
2.9 0.7
1.5 0.58
1.9 0.41
1.9 0.51
7.6 0.72
97 6.7
83 3.7
55 4.1
105 9.8
104 5.4
99 2.6
0.0 (ꢂ10, 1.0)
0.0 (ꢂ7.0, ꢂ11)
0.0 (ꢂ17, ꢂ2.0)
0.0 (ꢂ13, 1.0)
23
2
63 (70, 55)
0.0 (ꢂ15, 1.0)
0.0 (ꢂ30, ꢂ38)
a
% displacement at 1
m
m
m
M concentrations of specific [3H]ZM241385 binding on HEK 293 cell membranes stably expressing human adenosine A2A receptors.
b
c
% displacement at 1
% displacement at 1
M concentrations of specific [3H]PSB603 binding on CHO cell membranes stably expressing human adenosine A2B receptors or affinity (cpd 10).
M concentrations of specific [3H]PSB11 binding on CHO cell membranes stably expressing human adenosine A3 receptors.
d
e
Concentration causing half-maximal stimulation of specific binding of [35S]GTP
g
S in CHO cell membranes expressing human adenosine A1 receptors (n ¼ 3).
S in CHO cell membranes expressing human adenosine A1 receptors, with 0% being the binding in the absence of any ligand (buffer)
M (n ¼ 3).
Maximum specific binding of [35S]GTP
g
and 100% being the reference agonist CPA at a concentration of 100
m

686
J. Louvel et al. / European Journal of Medicinal Chemistry 101 (2015) 681e691
the hA1AR. In addition to structure-affinity relationships, structure-
kinetics relationships were established, and it was shown that the
residence time of Capadenoson (4) can be increased from 28 min up
to 132 min by replacing its 4-(2-hydroxyethoxy) substituent with a
4-methoxy substituent (5). The crucial importance of the substit-
uent borne by the thiazole moiety was also highlighted, showing
that a 4-halogen-bearing phenyl ring allowed for slower dissocia-
tion rates, whereas the use of more flexible and polar groups (12)
led to shorter residence times. Furthermore, the docking pose of 5
in a homology model of hA1AR revealed the potential importance of
EL3 in receptoreligand complex dissociation, suggesting the
halogen substituent to act as a lipophilic cap. A selection of com-
pounds was also characterized in a functional assay, showing par-
tial to full (55e105%) activation of the hA1AR. These results, taken
together with the similar drug-like properties of 4 (LE ¼ 0.35,
cLogP ¼ 5.04) and 5 (LE ¼ 0.35, cLogP ¼ 5.73), seem to indicate the
importance of another hit selection criterion, namely residence
time, in the early stages of drug discovery. Lack of efficacy in clinical
studies, as was the case for 4 [10], might be linked to a shorter
residence time for the target in case, whereas on-target toxicity
may be associated with longer residence times [27]. Whether 5
would be a better lead candidate than 4 is an open question, but
addressing the kinetics of the drugereceptor interaction provides
room for a more detailed and educated hit and lead optimization
process.
4. Experimental section
4.1. Chemistry
All solvents and reagents were purchased from commercial
sources and were of analytical grade. Demineralised water is simply
referred to as H2O, as was used in all cases unless stated otherwise
(i.e. brine). 1H and 13C NMR spectra were recorded on a Bruker AV
400 liquid spectrometer (1H NMR, 400 MHz; 13C NMR, 100 MHz) at
ambient temperature. Chemical shifts are reported in parts per
Fig. 2. Docking pose of 5 (LUF6941) in the homology model of hA1AR, showing
hydrogen bonding interactions (yellow dashed lines) with the residues of the binding
site. A: side view. B: top view, with molecular surfaces of both the ligand and the
receptor; receptor surface color code: grey ¼ hydrophobic, red ¼ negatively charged,
blue ¼ positively charged; ligand atoms color code: white ¼ carbon, blue ¼ nitrogen,
green ¼ chlorine, yellow ¼ sulfur, red ¼ oxygen, white ¼ hydrogen. (For interpretation
of the references to color in this figure legend, the reader is referred to the web version
of this article.)
million (ppm), are designated by d and are downfield to the internal
standard tetramethylsilane (TMS) in CDCl3. Coupling-constants are
reported in Hz and are designated as J. Analytical purity of the final
compounds was determined by high pressure liquid chromatog-
Although homology modeling is speculative in nature, particu-
larly for the extracellular loops of the receptor, a few hypotheses
can be made using the abovementioned observations. The thiazole
moiety seems to play a crucial role in binding to EL3 via Lys265,
thereby anchoring EL3 in a specific conformation. Out of this
interaction the formation of a cavity arises, in which the chloro
substituent seems to fit best. The chlorine atom could then act as a
“lipophilic cap”, preventing the access of water molecules to the
binding site and the hydrogen bonds [25], as illustrated by the
slower dissociation rate displayed by 5. This seems to be in agree-
ment with the increased dissociation rate observed when the size
raphy (HPLC) with a Phenomenex Gemini 3
m C18 110A column
(50 ꢁ 4.6 mm, 3 m), measuring UV absorbance at 254 nm. Sample
m
preparation and HPLC method was e unless stated otherwise e as
follows: 0.3e0.8 mg of compound was dissolved in 1 mL of a 1:1:1
mixture of CH3CN/H2O/tBuOH and eluted from the column within
15 min, with a three component system of H2O/CH3CN/1% TFA in
H2O, decreasing polarity of the solvent mixture in time from 80/10/
10 to 0/90/10. All compounds showed a single peak at the desig-
nated retention time and are at least 95% pure. Liquid chromatog-
raphyemass spectrometry (LCeMS) analyses were performed
using Thermo Finnigan Surveyor e LCQ Advantage Max LC-MS
system and a Gemini C18 Phenomenex column (50 ꢁ 4.6 mm,
of the s-hole of the chlorine atom is increased, thereby allowing it
to interact with water through halogen-bonding more readily [26].
Increasing the size of the halogen atom (to Br or I, compounds 14 or
15) leads to a decrease in residence time, which could be due to the
looser or less compact association between the ligand and the re-
ceptor in this region. When replacing the phenyl ring with a mor-
pholino group (12), the exposed substituent may not be lipophilic
enough to slow down dissociation, and the same would go in the
absence of a substituent (10).
3 mm). The sample preparation was the same as for HPLC analysis.
The elution method was set up as follows: 1e4 min isocratic system
of H2O/CH3CN/1% TFA in H2O, 80:10:10, from the 4th min, a
gradient was applied from 80:10:10 to 0:90:10 within 9 min, fol-
lowed by 1 min of equilibration at 0:90:10 and 1 min at 80:10:10.
Thin-layer chromatography (TLC) was routinely performed to
monitor the progress of reactions, using aluminum coated Merck
silica gel F254 plates. Purification by column chromatography was
achieved by use of Grace Davison Davisil silica column material
(LC60A 30e200 micron). Solutions were concentrated using a
Heidolph laborota W8 2000 efficient rotary evaporation apparatus
and by a high vacuum on a Binder APT line Vacuum Drying Oven.
The procedure for a series of similar compounds is given as a
3. Conclusion
A series of derivatives of Capadenoson (including Capadenoson
itself) were synthesized that display high affinity and selectivity at

J. Louvel et al. / European Journal of Medicinal Chemistry 101 (2015) 681e691
687
general procedure for all within that series, annotated by the
4.10. 3-Bromo-4-(chloromethyl)-2-(4-chlorophenyl)thiazole 34
numbers of the compounds. The procedure for the synthesis of
similar compounds is given as a general procedure; deviations
therefrom are given below for each compound.
To a solution of 4-(chloromethyl)-2-(4-chlorophenyl)thiazole 32
(244 mg, 1 mmol) in Ac2O (35 mL) was added bromine (0.10 mL,
3 mmol). After 8 h at room temperature, H2O (200 mL) was added,
the precipitate was filtered and washed with H2O, and dried under
vacuum to afford 34 as a solid (282 mg, 87%). 1H NMR (400 MHz,
4.2. General procedure to thioamides 27e28
CDCl3)
d 7.82e7.79 (m, 2H), 7.44e7.40 (m, 2H), 4.72 (s, 2H).
Benzonitriles 24e25 (1 eq.) were dissolved in pyridine (1.5 M),
triethylamine (1.1 eq.) and ammoniumsulfide (1.1 eq.) were added
and the mixture was heated to 50 ꢀC. After completion of the re-
action, the mixture was extracted three times with EtOAc, dried
with MgSO4, filtered and concentrated in vacuo.
4.11. General procedure to final aminopyridines 5e23
Central aminopyridines 1e3 (1 eq.), alkylating agents (1 eq.),
and NaHCO3 (1 eq.) were dissolved in DMF (1 mL) and stirred at rt.
The work-up procedure after completion of the reaction is
described for each compound independently.
4.3. 5-Chloropyridine-2-carbothioamide 27
Prepared from 5-chloropicolinonitrile 24, yield 57%. 1H NMR
4.12. 2-Amino-6-{[2-(4-chlorophenyl)thiazol-4-yl]methylthio}-4-
(4-methoxyphenyl)pyridine-3,5-dicarbonitrile 5
(400 MHz, MeOD)
d
8.63 (d, J ¼ 8.4 Hz, 1H), 8.55 (s, 1H), 7.95 (d,
J ¼ 8.8 Hz, 1H).
Prepared from central aminopyridine 2 and 4-(chloromethyl)-2-
(4-chlorophenyl)thiazole. H2O was added to the reaction mixture
and the precipitate was filtered. The product was recrystallized
from CHCl3 and washed with EtOAc and MeOH over filter. Light
beige solid (392 mg, 0.8 mmol, 53%). 1H NMR (400 MHz, DMSO)
4.4. 5-Bromopyridine-2-carbothioamide 28
Prepared from 5-bromopicolinonitrile 25, yield 48%. 1H NMR
(400 MHz, MeOD)
J ¼ 8.8 Hz, 1H).
d
8.63 (d, J ¼ 8.4 Hz, 1H), 8.55 (s, 1H), 7.95 (d,
d
7.95e7.92 (m, 3H), 7.57e7.55 (m, 2H), 7.50e7.47 (m, 2H), 7.10 (d,
J ¼ 7.6 Hz, 2H), 4.63 (s, 2H), 3.83 (s, 3H). 13C NMR (101 MHz, DMSO)
4.5. General procedure to alkylating agents 29e31
d
165.9, 165.7, 160.8, 159.7, 158.1, 152.5, 134.9, 131.7, 130.2, 129.4,
127.8, 125.8, 125.8, 118.9, 115.5, 114.1, 93.4, 85.9, 55.4, 29.3. HPLC tR:
12.3 min.
Pyridinecarbothioamides 26e28 (1 eq.) and dichloroacetone
(1 eq.) were dissolved in toluene (1 M) and heated to reflux (110 ꢀC).
After completion of the reaction, the mixture was cooled to room
temperature and extracted with EtOAc (3 times) and the organic
layer was washed (brine), dried (MgSO4), filtered and concentrated
in vacuo. The product was purified by column chromatography.
4.13. 2-Amino-6-({[2-(4-chlorophenyl)thiazol-4-yl]methyl}thio)-4-
(4-hydroxyphenyl)pyridine-3,5-dicarbonitrile 6
Prepared from central aminopyridine 3 and 4-(chloromethyl)-2-
(4-chlorophenyl)thiazole. H2O was added to the reaction mixture,
extracted three times with EtOAc, dried with MgSO4, filtered and
concentrated in vacuo. The product was purified by Preparative TLC
(50%e100% EtOAc in PE). Light yellow solid (83 mg, 0.17 mmol,
4.6. 4-(Chloromethyl)-2-(pyridin-2-yl)thiazole 29
Prepared from benzothioamide 26. Eluent for column chroma-
tography: (1% MeOH in CH2Cl2), yield 47%. 1H NMR (400 MHz,
34%). 1H NMR (400 MHz, DMSO)
d
10.07 (s, 1H), 7.95e7.91 (m, 3H),
7.56 (d, J ¼ 8.4 Hz, 2H), 7.36 (d, J ¼ 8.4 Hz, 2H), 6.90 (d, J ¼ 8.4 Hz,
2H), 4.63 (s, 2H). 13C NMR (101 MHz, DMSO)
165.9, 165.7, 159.8,
CDCl3)
d
8.61 (d, J ¼ 4.4 Hz, 1H), 8.20 (d, J ¼ 8.0 Hz, 1H), 7.80 (td,
J ¼ 7.8, 1.5 Hz, 1H), 7.42 (s, 1H), 7.35e7.32 (m, 1H), 4.76 (s, 2H).
d
159.5, 158.4,152.5,134.9,131.7, 130.3, 129.3,127.8, 124.1, 118.8, 115.6,
115.4, 93.3, 85.8, 29.3. HPLC tR: 11.3 min.
4.7. 4-(Chloromethyl)-2-(5-chloropyridin-2-yl)thiazole 30
Prepared from benzothioamide 27. Eluent for column chroma-
4.14. 2-Amino-6-{[2-(3,4-methylenedioxyphenyl)thiazol-4-yl]
methylthio}-4-(4-methoxyphenyl)pyridine-3,5-dicarbonitrile 7
tography: (10% EtOAc in Pet Ether), yield 70%. 1H NMR (400 MHz,
CDCl3)
d
8.55 (d, J ¼ 2.0 Hz, 1H), 8.15 (d, J ¼ 8.4 Hz, 1H), 7.78 (dd,
J ¼ 8.4, 2.4 Hz, 1H), 7.43 (s, 1H), 4.77 (s, 2H).
Prepared from central aminopyridine 1 and 4-(chloromethyl)-2-
(4-chlorophenyl)thiazole. H2O was added to the reaction mixture
and the precipitate was filtered. Beige/yellow powder (50%). 1H
4.8. 4-(Chloromethyl)-2-(5-bromopyridin-2-yl)thiazole 31
NMR (400 MHz, DMSO)
7.13 (s, 1H), 7.08 (d, J ¼ 8.0 Hz, 1H), 7.02 (d, J ¼ 8.0 Hz, 1H), 6.13 (s,
2H), 4.62 (s, 2H). 13C NMR (101 MHz, DMSO)
165.8, 165.7, 159.6,
158.0,152.5,149.0,147.3,134.9,131.7,129.4,127.8,127.3,123.0,118.9,
115.4, 109.0, 108.6, 101.8, 93.5, 86.1, 29.3. HPLC tR: 11.2 min.
d
8.50e7.80 (m, 5H), 7.55 (d, J ¼ 8.4 Hz, 2H),
Prepared from benzothioamide 28. Eluent for column chroma-
tography: (10% EtOAc in Pet Ether), yield 40%. 1H NMR (400 MHz,
d
CDCl3)
d
8.65 (d, J ¼ 1.2 Hz, 1H), 8.09 (d, J ¼ 8.4 Hz, 1H), 7.93 (dd,
J ¼ 8.4, 2.4 Hz, 1H), 7.43 (s, 1H), 4.75 (s, 2H).
4.9. 3-Chloro-4-(chloromethyl)-2-(4-chlorophenyl)thiazole 33
4.15. 2-Amino-6-({[5-chloro-2-(4-chlorophenyl)thiazol-4-yl]
methyl}thio)-4-(4-methoxyphenyl)pyridine-3,5-dicarbonitrile 8
To a solution of 4-(chloromethyl)-2-(4-chlorophenyl)thiazole 32
(244 mg, 1 mmol) in Ac2O (35 mL) was added a mixture of
concentrated HCl (0.48 mL) and HNO3 (0.36 mL). After 1 h at room
temperature, H2O (200 mL) was added, the precipitate was filtered
and washed with H2O, and dried under vacuum to afford 33 as a
Prepared from central aminopyridine 2 and 33. H2O was added
to the reaction mixture and the precipitate was filtered. The
product was washed with EtOAc and methyl tert-butyl ether. Yield
46%. 1H NMR (400 MHz, DMSO)
d
8.00 (br s, 2H), 7.89 (d, J ¼ 7.6 Hz,
2H), 7.57 (d, J ¼ 7.6 Hz, 2H), 7.50 (d, J ¼ 7.8 Hz, 2H), 7.10 (d, J ¼ 7.8 Hz,
2H), 4.70 (s, 2H), 3.83 (s, 3H). 13C NMR (101 MHz, DMSO)
165.5,
solid (233 mg, 84%). 1H NMR (400 MHz, CDCl3)
d 7.81e7.78 (m, 2H),
7.43e7.41 (m, 2H), 4.71 (s, 2H).
d

688
J. Louvel et al. / European Journal of Medicinal Chemistry 101 (2015) 681e691
163.2, 160.9, 159.7, 158.1, 148.4, 135.6, 130.8, 130.2, 129.4, 127.6,
125.7, 122.9, 115.4, 114.1, 93.4, 86.2, 55.4, 26.8. HPLC tR: 11.8 min. MS
(ESIþ) m/z ¼ 523.93.
29.3. HPLC tR: 11.7 min.
4.21. 2-Amino-6-{[2-(4-bromophenyl)thiazol-4-yl]methylthio}-4-
(4-methoxyphenyl)pyridine-3,5-dicarbonitrile 14
4.16. 2-Amino-6-({[5-bromo-2-(4-chlorophenyl)thiazol-4-yl]
methyl}thio)-4-(4-methoxyphenyl)pyridine-3,5-dicarbonitrile 9
Prepared from central aminopyridine 2 and 4-(chloromethyl)-2-
(4-bromophenyl)thiazole. H2O was added to the reaction mixture
and the precipitate was filtered. The product was washed with
EtOAc over filter. Light beige solid (182 mg, 0.34 mmol, 68%). 1H
Prepared from central aminopyridine 2 and 34. H2O was added
to the reaction mixture and the precipitate was filtered. The
product was washed with EtOAc. Yield 91%. 1H NMR (400 MHz,
NMR (400 MHz, DMSO)
7.48 (d, J ¼ 7.2 Hz, 2H), 7.10 (d, J ¼ 7.2 Hz, 2H), 4.63 (s, 2H), 3.83 (s,
3H). 13C NMR (101 MHz, DMSO)
165.9, 165.7, 160.8, 159.7, 158.1,
d
7.91e7.86 (m, 3H), 7.69 (d, J ¼ 6.8 Hz, 2H),
DMSO)
2H), 7.50 (d, J ¼ 7.5 Hz, 2H), 7.10 (d, J ¼ 7.5 Hz, 2H), 4.69 (s, 2H), 3.83
(s, 3H). 13C NMR (101 MHz, DMSO)
166.0, 165.6, 160.9, 159.7, 158.1,
d
8.00 (br s, 2H), 7.89 (d, J ¼ 7.2 Hz, 2H), 7.57 (d, J ¼ 7.2 Hz,
d
d
152.5, 132.2, 132.0, 128.0, 125.8, 123.6, 118.8, 115.5, 114.1, 93.4, 85.9,
55.3, 29.3. HPLC tR: 12.6 min.
150.8, 135.5, 130.8, 130.3, 129.4, 127.6, 125.7, 115.5, 115.4, 114.1, 107.7,
93.3, 86.2, 55.4, 28.0. HPLC tR: 8.4 min (50e100% gradient of
acetonitrile). MS (ESIþ) m/z ¼ 569.87.
4.22. 2-Amino-6-{[2-(4-iodophenyl)thiazol-4-yl]methylthio}-4-(4-
methoxyphenyl)pyridine-3,5-dicarbonitrile 15
4.17. 2-Amino-4-(4-methoxyphenyl)-6-(thiazol-4-ylmethylthio)
pyridine-3,5-dicarbonitrile 10
Prepared from central aminopyridine 2 and 4-(chloromethyl)-2-
(4-iodophenyl)thiazole. H2O was added to the reaction mixture and
the precipitate was filtered. The precipitate was washed with EtOAc
over filter, filtrate was concentrated in vacuo, yielding the pure
product as a light brown solid (177 mg, 0.30 mmol, 60%). 1H NMR
Prepared from central aminopyridine 2 and 4-(chloromethyl)-
thiazole hydrochloride, using 2 equivalents of NaHCO3. Beige solid
(154 mg, 0.41 mmol, 82%). 1H NMR (400 MHz, DMSO)
d
9.07 (s, 1H),
7.86 (s, 1H), 7.48 (d, J ¼ 8.8 Hz, 2H), 7.10 (d, J ¼ 8.8 Hz, 2H), 4.63 (s,
2H), 3.84 (s, 3H). 13C NMR (101 MHz, DMSO)
166.0, 160.8, 159.7,
(400 MHz, DMSO)
J ¼ 8.0 Hz, 2H), 7.48 (d, J ¼ 8.8 Hz, 2H), 7.10 (d, J ¼ 8.4 Hz, 2H), 4.63
(s, 2H), 3.83 (s, 3H). 13C NMR (101 MHz, DMSO)
166.0, 165.9, 160.8,
d
7.90 (s, 1H), 7.87 (d, J ¼ 8.4 Hz, 2H), 7.71 (d,
d
158.1, 154.3, 151.9, 130.2, 125.8, 117.7, 115.4, 114.1, 93.3, 85.9, 55.3,
29.3. HPLC tR: 9.8 min.
d
159.7, 158.1, 152.5, 138.0, 132.3, 130.2, 127.9, 125.8, 20.0, 118.7, 115.5,
114.1, 97.1, 93.4, 85.9, 55.3, 29.3. HPLC tR: 12.8 min.
4.18. 2-Amino-4-(4-methoxyphenyl)-6-[(2-phenylthiazol-4-yl)
methylthio]pyridine-3,5-dicarbonitrile 11
4.23. 2-Amino-6-{[2-(3-chlorophenyl)thiazol-4-yl]methylthio}-4-
(4-methoxyphenyl)pyridine-3,5-dicarbonitrile 16
Prepared from central aminopyridine 2 and 4-(chloromethyl)-2-
phenylthiazole. H2O was added to the reaction mixture and the
precipitate was filtered. Off-white/light brown solid (157 mg,
Prepared from central aminopyridine 2 and 4-(chloromethyl)-2-
(3-chlorophenyl)thiazole. H2O was added to the reaction mixture
and the precipitate was filtered. Beige/yellow powder (388 mg,
0.34 mmol, 68%). 1H NMR (400 MHz, DMSO)
d
7.93e7.92 (m, 2H),
7.88 (s, 1H), 7.49 (t, J ¼ 3.2 Hz, 5H), 7.10 (d, J ¼ 8.8 Hz, 2H), 4.64 (s,
2H), 3.83 (s, 3H). 13C NMR (101 MHz, DMSO)
167.0, 166.0, 160.8,
0.79 mmol, 100%). 1H NMR (400 MHz, DMSO)
d
7.89 (s, 3H),
7.55e7.48 (m, 4H), 7.10 (d, J ¼ 8.4 Hz, 2H), 4.64 (s, 2H), 3.83 (s, 3H).
13C NMR (101 MHz, DMSO)
165.9, 160.9, 159.7, 152.6, 134.7, 134.0,
d
159.7, 158.1, 152.3, 132.8, 130.5, 130.3, 130.2, 129.3, 126.1, 125.8,
119.6, 118.3, 115.5, 114.1, 93.4, 85.9, 55.3, 29.3. HPLC tR: 11.8 min.
d
131.3, 130.2, 130.1, 125.8, 125.4, 124.9, 119.3, 115.5, 114.1, 93.4, 55.4,
29.2. HPLC tR: 12.3 min.
4.19. 2-Amino-4-(4-methoxyphenyl)-6-[(2-morpholinothiazol-4-
yl)methylthio]pyridine-3,5-dicarbonitrile 12
4.24. 2-Amino-6-{[2-(2-chlorophenyl)thiazol-4-yl]methylthio}-4-
(4-methoxyphenyl)pyridine-3,5-dicarbonitrile 17
Central aminopyridine
2
(1 eq.), 4-(chloromethyl)-2-
morpholinothiazole (1 eq.), and DBU (2 eq.) were dissolved in
DMF (1 mL) and stirred at rt. After completion of the reaction, H2O
was added to the reaction mixture and the precipitate was filtered.
Beige solid (181 mg, 0.39 mmol, 78%). 1H NMR (400 MHz, DMSO)
Prepared from central aminopyridine 2 and 4-(chloromethyl)-2-
(2-chlorophenyl)thiazole. H2O was added to the reaction mixture
and the precipitate was filtered. The product was purified by Pre-
parative TLC (50% EtOAc in PE). Yellow solid (93 mg, 0.19 mmol,
d
7.48 (d, J ¼ 8.8 Hz, 2H), 7.10 (d, J ¼ 8.4 Hz, 2H), 6.97 (s, 1H), 4.36 (s,
2H), 3.83 (s, 3H), 3.69 (m, 4H), 3.35-3.34 (m, 4H). 13C NMR
(101 MHz, DMSO) 171.1, 166.3, 162.4, 160.9, 159.7, 158.1, 130.3,
38%). 1H NMR (400 MHz, DMSO)
d
8.19e8.17 (m, 1H), 8.06 (s, 1H),
7.65e7.63 (m, 1H), 7.51e7.48 (m, 4H), 7.10 (d, J ¼ 8.8 Hz, 2H), 4.67 (s,
2H), 3.83 (s, 3H). 13C NMR (101 MHz, DMSO)
166.0, 162.2, 160.9,
d
d
125.9, 115.6, 114.1, 106.1, 93.4, 85.8, 65.4, 55.4, 48.1, 47.9, 30.0. HPLC
tR: 8.8 min.
159.8,158.2,151.3,131.2, 131.0,130.8,130.7, 130.7, 130.3,127.8, 125.8,
120.3, 115.5, 114.1, 93.4, 85.9, 55.4, 29.2. HPLC tR: 12.1 min.
4.20. 2-Amino-6-{[2-(4-fluorophenyl)thiazol-4-yl]methylthio}-4-
(4-methoxyphenyl)pyridine-3,5-dicarbonitrile 13
4.25. 2-Amino-6-{[2-(3,4-dichlorophenyl)thiazol-4-yl]methylthio}-
4-(4-methoxyphenyl)pyridine-3,5-dicarbonitrile 18
Prepared from central aminopyridine 2 and 4-(chloromethyl)-2-
(4-fluorophenyl)thiazole. H2O was added to the reaction mixture
and the precipitate was filtered. The product was washed with
acetone over filter. Off-white powder (93 mg, 0.20 mmol, 40%).1H
Prepared from central aminopyridine 2 and 4-(chloromethyl)-2-
(3,4-dichlorophenyl)thiazole. H2O was added to the reaction
mixture and the precipitate was filtered. Light beige solid (224 mg,
0.43 mmol, 86%). 1H NMR (400 MHz, DMSO)
d 8.11 (s, 1H), 7.97 (s,
NMR (400 MHz, DMSO)
J ¼ 8.4 Hz, 2H), 7.36e7.32 (m, 2H), 7.10 (d, J ¼ 8.4 Hz, 2H), 4.63 (s,
2H), 3.83 (s, 3H). 13C NMR (101 MHz, DMSO)
165.9, 160.8, 158.1,
152.3, 130.2, 128.5, 128.4, 125.8, 118.4, 116.4, 116.2, 115.5, 114.1, 55.4,
d
7.99e7.96 (m, 2H), 7.87 (s, 1H), 7.48 (d,
1H), 7.89 (d, J ¼ 8.4 Hz, 1H), 7.75 (d, J ¼ 8.4 Hz, 1H), 7.48 (d,
J ¼ 8.4 Hz, 2H), 7.10 (d, J ¼ 8.8 Hz, 2H), 4.64 (s, 2H), 3.83 (s, 3H). 13
C
d
NMR (101 MHz, DMSO) d 165.8, 164.1, 160.8, 159.7, 158.1, 152.7,
133.2, 132.7, 132.1, 131.5, 130.2, 127.4, 126.2, 125.8, 119.7, 115.5, 114.1,

J. Louvel et al. / European Journal of Medicinal Chemistry 101 (2015) 681e691
689
93.4, 85.9, 55.3, 29.2. HPLC tR: 13.1 min.
121.2, 120.8, 115.5, 114.1, 93.4, 85.9, 55.3, 29.3. HPLC tR: 11.1 min. MS
(ESIþ) m/z ¼ 536.93.
4.26. 2-Amino-4-(4-methoxyphenyl)-6-{[2-(4-methoxyphenyl)
thiazol-4-yl]methylthio}pyridine-3,5-dicarbonitrile 19
4.31. Biology
Prepared from central aminopyridine 2 and 4-(chloromethyl)-2-
(4-methoxyphenyl)thiazole. H2O was added to the reaction
mixture and the precipitate was filtered. Beige solid (100 mg,
4.31.1. Chemicals and reagents
[3H]1,3-dipropyl-8-cyclopentyl-xanthine was purchased from
ARC Inc. (St. Louis, USA). [3H]ZM241385 was obtained from Tocris
Cookson, Ltd. (U.K.). [3H]PSB603 and [3H]PSB11 were obtained from
Prof. C. Müller (University of Bonn, Germany). Adenosine deami-
nase (ADA) was purchased from Boehringer Mannheim (Man-
nheim, Germany). DPCPX was from Sigma (St. Louis, MO, U.S.A).
Bicinchoninic acid (BCA) and BCA protein assay reagent were ob-
tained from Pierce Chemical Company (Rockford, IL, U.S.A.). Chinese
hamster ovary cells stably expressing the hA1R were obtained from
Prof. Steve Hill (University of Nottingham, UK). HEK 293 cells stably
expressing the human adenosine A2A, A2B and A3 receptor were
gifts from Dr. J. Wang (Biogen, U.S.A.) and Dr. K.-N. Klotz (University
of Würzburg, Germany), respectively. All other chemicals were of
analytical grade and obtained from standard commercial sources.
0.39 mmol, 78%). 1H NMR (400 MHz, DMSO)
d
7.86 (d, J ¼ 8.4 Hz,
2H), 7.77 (s, 1H), 7.48 (d, J ¼ 8.4 Hz, 2H), 7.10 (d, J ¼ 8.4 Hz, 2H), 7.04
(d, J ¼ 8.8 Hz, 2H), 4.61 (s, 2H), 3.83 (s, 3H), 3.82 (s, 3H). 13C NMR
(101 MHz, DMSO)
d 166.9, 166.0, 160.9, 160.8, 159.7, 158.1, 151.9,
130.2, 127.7, 125.8, 125.7, 117.2, 115.5, 114.6, 114.1, 93.4, 85.9, 55.4,
29.4. HPLC tR: 11.7 min.
4.27. 2-Amino-4-(4-methoxyphenyl)-6-[(2-p-tolylthiazol-4-yl)
methylthio]pyridine-3,5-dicarbonitrile 20
Prepared from central aminopyridine 2 and 4-(chloromethyl)-2-
(4-methylphenyl)thiazole. H2O was added to the reaction mixture
and the precipitate was filtered. The product was washed with
EtOAc over filter. Off-white/light beige solid (98 mg, 0.21 mmol,
4.32. Binding assays
42%). 1H NMR (400 MHz, DMSO)
d
7.83e7.80 (m, 3H), 7.48 (d,
J ¼ 8.0 Hz, 2H), 7.29 (d, J ¼ 7.2 Hz, 2H), 7.10 (d, J ¼ 8.0 Hz, 2H), 4.63 (s,
2H), 3.83 (s, 3H), 2.34 (s, 3H). 13C NMR (101 MHz, DMSO)
167.1,
Binding assays were carried out as reported previously [5].
Displacements of radioligands in the presence of 1 mM of unla-
d
166.0, 160.8, 159.7, 158.1, 152.1, 140.2, 130.3, 130.2, 129.8, 126.0,
125.8, 119.0, 117.8, 115.5, 114.1, 93.4, 85.9, 55.3, 29.4, 20.9. HPLC tR:
12.2 min.
beled compounds are an average of at least two independent ex-
periments, each consisting of two replicates. Affinity values are an
average of at least three independent experiments, each consisting
of two replicates [28].
4.28. 2-Amino-6-[(2-pyridinethiazol-4-yl)methylthio]-4-(4-
methoxyphenyl)pyridine-3,5-dicarbonitrile 21
4.33. Competition association assays
Prepared from central aminopyridine 2 and 29. H2O was added
to the reaction mixture and the precipitate was filtered. The
product was triturated in CH2Cl2. Yield 26%. 1H NMR (400 MHz,
Competition association assays were carried out as reported
previously [5].
DMSO)
7.97 (s, 2H), 7.50 (d, J ¼ 3.2 Hz, 3H), 7.11 (d, J ¼ 7.2 Hz, 2H), 4.66 (s,
2H), 3.84 (s, 3H). 13C NMR (101 MHz, DMSO)
168.3, 165.9, 160.8,
159.7, 158.1, 152.7, 150.2,149.7, 137.8, 130.2, 125.8, 125.2,120.7, 119.2,
115.5, 114.1, 93.4, 85.9, 55.3, 29.4. HPLC tR: 9.9 min. MS (ESIþ) m/
z ¼ 457.13.
d
8.63 (s, 1H), 8.50e7.80 (br. s, 2H), 8.11 (d, J ¼ 6.8 Hz, 2H),
4.34. Functional assays
d
[
35S]GTP
g, saponin, 3
buffer (50 mM Tris, 100 mM NaCl, 5 mM MgCl2, pH 7.4, 1 mM EDTA,
1 mM DTT, and 0.05% BSA) containing 3 M GDP and a range of
concentrations of ligand at 25 ꢀC for 30 min. After this 20 l of [35S]
GTP S (final concentration 0.3 nM) was added and incubation
g
S binding assay: Membrane homogenates (CHO-hA1,
3
m
mg) were equilibrated in 80 l total volume of assay
m
m
m
4.29. 2-Amino-6-{[2-(4-chloropyridine)thiazol-4-yl]methylthio}-4-
g
(4-methoxyphenyl)pyridine-3,5-dicarbonitrile 22
continued for 90 min at 25 ꢀC. The incubation was terminated by
filtration over Whatman GF/B filters under reduced pressure with a
Brandell harvester or through 96-well GF/B filter plates using a
PerkinElmer Filtermate-harvester. Filters were washed three times
with ice-cold buffer and placed in scintillation vials. Emulsifier Safe
(3.5 mL) was added, and after 2 h radioactivity was counted in an
Prepared from central aminopyridine 2 and 30. H2O was added
to the reaction mixture and the precipitate was filtered. The
product was triturated in CH2Cl2. Yield 44%. 1H NMR (400 MHz,
DMSO)
d
8.69 (s, 1H), 8.60e7.75 (br s, 2H), 8.12 (d, J ¼ 8.4 Hz, 1H),
8.08 (dd, J ¼ 8.5, 1.3 Hz, 1H), 8.01 (s, 1H) 7.48 (d, J ¼ 8.5 Hz, 2H), 7.10
LKB rack b scintillation counter.
(d, J ¼ 8.5 Hz, 2H), 4.65 (s, 2H), 3.83 (s, 3H). 13C NMR (101 MHz,
DMSO)
d
166.8, 165.9, 160.8, 159.7, 158.1, 153.0, 148.7, 148.3, 137.5,
4.35. Data analysis
132.1, 130.2, 125.8, 121.2, 120.4, 115.5, 114.1, 93.4, 85.9, 55.3, 29.3.
HPLC tR: 10.8 min. MS (ESIþ) m/z ¼ 491.07.
Ki values were calculated by use of a nonlinear regression curve-
fitting program (GraphPad Prism 5, GraphPad Software Inc., San
Diego, CA). KD values of the radioligands were 1.6 nM, 1.0 nM,
0.4 nM and 4.9 nM for [3H]DPCPX, [3H]ZM241385, [3H]PSB603 and
[3H]PSB11 on the adenosine A1R, A2AR, A2BR and A3R, respectively.
The data from the functional assays were generated and normal-
4.30. 2-Amino-6-{[2-(4-bromopyridine)thiazol-4-yl]methylthio}-
4-(4-methoxyphenyl)pyridine-3,5-dicarbonitrile 23
Prepared from central aminopyridine 2 and 31. H2O was added
to the reaction mixture and the precipitate was filtered. The
product was triturated in CH2Cl2. Yield 63%. 1H NMR (400 MHz,
ized to the value obtained with 100 mM CPA (set at 100%).
DMSO)
8.03 (t, J ¼ 9.6 Hz, 2H), 7.48 (d, J ¼ 8.2 Hz, 2H), 7.10 (d, J ¼ 8.3 Hz, 2H),
4.65 (s, 2H), 3.83 (s, 3H). 13C NMR (101 MHz, DMSO)
166.9, 165.9,
160.8, 159.7, 158.1, 153.0, 150.4, 148.9, 140.3, 130.2, 125.8, 121.3,
d
8.77 (s, 1H), 8.60e7.75 (br. s, 2H), 8.20 (d, J ¼ 8.4 Hz, 1H),
4.36. Computational modeling
d
A homology model of the adenosine A1 receptor was con-
structed using the homology modeling tool within the software

690
J. Louvel et al. / European Journal of Medicinal Chemistry 101 (2015) 681e691
selectivity profile compared to the reference agonist N-ethyl-
carboxamidoadenosine, J. Med. Chem. 47 (2004) 3707e3709;
b) L.C.W. Chang, J.K. von Frijtag Drabbe Künzel, T. Mulder-Krieger,
R.F. Spanjersberg, S.F. Roerink, G. van den Hout, M.W. Beukers, J. Brussee,
package Maestro (version 3.7). [29e31]. The model was based on
the active structure of the Adenosine A2A receptor co-crystalized
with UK-432097 (PDB: 3QAK) [32]. The alignment between the
receptor and templates was performed using ClustalW and GPCRDB
for the TM-regions [33]. Based on this alignment a homology model
was constructed using the energy-based scoring function. Next,
two rounds of ligand-guided optimization were performed [34]. For
this we used a set of 76 in house LUF ligands [35]. These ligands
were matched with 50 decoys each using the DUD-E web service
[36]. In addition to the decoys we added 18 LUF compounds that
had been found to be inactive [35]. To ensure proper docking of the
LUF ligands and to force a bidentate interaction with Asn2546.55, H-
bond constraints were used and the VdW radii were lowered to
0.50 Å. After docking 65 models were generated using the top three
scoring ligands (LUF7052 [35], 4, 5). These ligands were refined
using 100 steps of Minimization Monte Carlo present in Prime.
[31,32]. Next the best scoring model was selected based on early
enrichment (BEDROC-160.9) [37]. We further optimized this model
using extended loop sampling on EL3 [31,32] returning a total of 50
models with different conformations. From these models the best
performing model showed excellent enrichment (BEDROC-
160.9 ¼ 0.844/EF 1% ¼ 48%).
A.P. IJzerman,
A series of ligands displaying a remarkable agonistic-
antagonistic profile at the adenosine A1 receptor, J. Med. Chem. 48 (2005)
2045e2053.
[4] a) Otsuka Pharmaceuticals WO2005105778 A2;
b) Otsuka Pharmaceuticals US20120022077 A1.
[5] J. Louvel, D. Guo, M. Agliardi, T.A.M. Mocking, R. Kars, T.P. Pham, L. Xia, H. de
Vries, J. Brussee, L.H. Heitman, A.P. IJzerman, Agonists for the adenosine A1
receptor with tunable residence time. A case for non-ribose 4-amino-6-aryl-5-
cyano-2-thiopyrimidines, J. Med. Chem. 57 (2014) 3213e3222.
[6] R.A. Prentis, Y. Lis, S.R. Walker, Pharmaceutical innovation by the seven UK-
owned pharmaceutical companies (1964e1985), Br. J. Clin. Pharmacol. 25
(1988) 387e396.
[7] a) D.A. Smith, B.C. Jones, D.K. Walker, Design of drugs involving the concepts
and theories of drug metabolism and pharmacokinetics, Med. Res. Rev. 16
(1996) 243e266;
b) R. Zhang, F. Monsma, The importance of drug-target residence time, Curr.
Opin. Drug. Disc. Dev. 12 (2009) 488e496;
c) R. Zhang, F. Monsma, Binding kinetics and mechanism of action: toward the
discovery and development of better and best in class drugs, Exp. Opin. Drug
Disc. 5 (2010) 1023e1029;
d) R.A. Copeland, D.L. Pompliano, T.D. Meek, Drug-target residence time and
its implications for lead optimization, Nat. Rev. Drug Discov.
5 (2006)
730e739.
[8] a) M.R. Dowling, S.J. Charlton, Quantifying the association and dissociation
rates of unlabelled antagonists at the muscarinic M3 receptor, Br. J. Pharmacol.
148 (2006) 927e937;
b) I. Van Liefde, G. Vauquelin, SartaneAT1 receptor interactions: In vitro evi-
dence for insurmountable antagonism and inverse agonism, Mol. Cell. Endo-
crinol. 302 (2008) 237e243.
Author information
Notes
[9] R.A. Copeland, The dynamics of drug-target interactions: drug-target resi-
dence time and its impact on efficacy and safety, Exp. Opin. Drug Disc. 5
(2010) 305e310.
[10] Bayer Healthcare Aktiengesellschaft clinical trial report BAY 68-4986/12679.
[11] Bayer Healthcare Aktiengesellschaft WO2007073855 A1.
[12] O. Oniga, I. Grosu, S. Mager, I. Simiti, Heterocycles LXXVIII. Electrophilic
substitution of 20-phenyl-4R-2,40-bisthiazoles, Monatsh. Chem. 129 (1998)
661e669.
The authors declare no competing financial interests.
Author contributions
J.L. conceived the study, A.P.IJ. and L.H.H. supervised the project.
The chemical synthesis was designed and supervised by J.L. and
performed by M.S. and J.L. The bioassays were supervised by D.G.
and J.L. and performed by J.L., T.M.-K. and T.A.M.M. The computa-
tional study was performed by E.B.L. The manuscript was written
by J.L., E.B.L. and A.P.IJ. All authors have given approval to the final
version of the manuscript.
[13] M.J. Lohse, K.N. Klotz, J. Lindenborn-Fotinos, M. Reddington, U. Schwabe,
R.A. Olsson, 8-Cyclopentyl-1,3-dipropylxanthine (DPCPX) e a selective high
affinity antagonist radioligand for A1 adenosine receptors, Naunyn-Schmie-
deberg's Arch. Pharmacol. 336 (1987) 204e210.
[14] D. Guo, E.J. van Dorp, T. Mulder-Krieger, J.P. van Veldhoven, J. Brussee,
A.P. IJzerman, L.H. Heitman, Dual-point competition association assay: a fast
and high-throughput kinetic screening method for assessing ligand-receptor
binding kinetics, J. Biomol. Screen. 18 (2013) 309e320.
[15] R. Wilcken, M.O. Zimmermann, A. Lange, A.C. Joerger, F.M. Boeckler, Principles
and applications of halogen bonding in medicinal chemistry and chemical
biology, J. Med. Chem. 56 (2013) 1363e1388.
[16] J.J. Dynes, F.L. Baudais, R.K. Boyd, Inter-ring dihedral angles in polychlorinated
biphenyls from photoelectron spectroscopy, Can. J. Chem. 63 (1985)
1292e1299.
[17] Ligand efficiencies were calculated using the equation LE (kcal/mol heavy
atom) ¼ (1.374/N) pKi (at 293 K) clogP values were calculated using Chem-
axon Marvin: http://www.chemaxon.com/marvin/sketch/index.jsp.
[18] A.L. Hopkins, C.R. Groom, A. Alex, Ligand efficiency: a useful metric for lead
selection, Drug Discov. Today 9 (2004) 430e431.
[19] a) C.A. Lipinski, F. Lombardo, B.W. Dominy, P.J. Freeney, Experimental and
computational approaches to estimate solubility and permeability in drug
discovery and development settings, Adv. Drug. Deliv. Rev. 23 (1997) 3e25;
b) ibid Adv. Drug. Deliv. Rev. 46 (2001) 3e26.
Funding sources
NWO TOP subsidy (714.011.001) to A. P. IJzerman.
Acknowledgment
The authors would like to thank J. P. D. van Veldhoven for
carefully reviewing the manuscript.
Appendix A. Supplementary data
[20] S.P.H. Alexander, P.J. Millns, [3H]ZM241385ean antagonist for adenosine A2A
receptors in rat brain, Eur. J. Pharmacol. 411 (2001) 205e210.
[21] T. Borrmann, S. Hinz, D.C.G. Bertarelli, W. Li, N.C. Florin, A.B. Scheiff,
C.E. Müller, 1-alkyl-8-(piperazine-1-sulfonyl)phenylxanthines: development
and characterization of adenosine A2B receptor antagonists and a new radi-
oligand with subnanomolar affinity and subtype specificity, J. Med. Chem. 52
(2009) 3994e4006.
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.ejmech.2015.07.023.
References
[22] C.E. Müller, M. Diekmann, M. Thorand, V. Ozola, [3H]8-Ethyl-4-methyl-2-
phenyl-(8R)-4,5,7,8-tetrahydro-1H-imidazo[2,1-i]-purin-5-one ([3H]PSB-11),
a novel high-affinity antagonist radioligand for human A3 adenosine re-
ceptors, Bioorg. Med. Chem. Lett. 12 (2002) 501e503.
[1] a) B.B. Fredholm, A.P. IJzerman, K.A. Jacobson, J. Linden, C.E. Müller, Interna-
tional Union of basic and clinical pharmacology. LXXXI. Nomenclature and
classification of adenosine receptorsdan update, Pharmacol. Rev. 63 (2011)
1e34;
[23] W.H. Moos, D.S. Szotek, R.F. Bruns, N6-cycloaklyladenosines. Potent, A1 se-
lective adenosine agonists, J. Med. Chem. 28 (1985) 1383e1384.
[24] J.R. Lane, C. Klein Herenbrink, G.J.P. van Westen, J.A. Spoorendonk,
C. Hoffmann, A.P. IJzerman, A novel nonribose agonist, LUF5834, engages
residues that are distinct from those of adenosine-like ligands to activate the
adenosine A2A receptor, Mol. Pharmacol. 81 (2012) 475e487.
b) K.A. Jacobson, Z.G. Gao, Adenosine receptors as therapeutic targets, Nat.
Rev. Drug Disc. 5 (2006) 247e264;
c) B.B. Fredholm, J.F. Chen, S.A. Masino, J.M. Vaugeois, Actions of adenosine at
its receptors in the CNS: insights from knockouts and drugs, Annu. Rev.
Pharmacol. Toxicol. 45 (2005) 385e412.
[25] P. Schmidtke, F. Javier Luque, J.B. Murray, X. Barril, Shielded hydrogen bonds
as structural determinants of binding kinetics: application in drug design,
J. Am. Chem. Soc. 133 (2011) 18903e18910.
[2] Bayer Aktiengesellschaft WO2001025210.
[3] a) M.W. Beukers, L.C.W. Chang, J.K. von Frijtag Drabbe Künzel, T. Mulder-
Krieger, R.F. Spanjersberg, J. Brussee, A.P. IJzerman, New, non-adenosine, high-
potency agonists for the human adenosine A2B receptor with an improved
[26] P. Auffinger, F.A. Hays, E. Westhoff, P. Shing Ho, Halogen bonds in biological

J. Louvel et al. / European Journal of Medicinal Chemistry 101 (2015) 681e691
691
molecules, Proc. Nat. Acad. Sci. U. S. A. 101 (2004) 16789e16794.
[27] D. Guo, J.M. Hillger, A.P. IJzerman, L.H. Heitman, Drug-target residence time e
a case for G protein-coupled receptors, Med. Res. Rev. 34 (2014) 856e892.
[28] T. Borrmann, S. Hinz, D.C.G. Bertarelli, W. Li, N.C. Florin, A.B. Scheiff,
C.E. Müller, 1-alkyl-8-(piperazine-1-sulfonyl)phenylxanthines: development
and characterization of adenosine A2B receptor antagonists and a new radi-
oligand with subnanomolar affinity and subtype specificity, J. Med. Chem. 52
(2009) 3994e4006.
[29] M.P. Jacobson, D.L. Pincus, C.S. Rapp, T.J. Day, B. Honig, D.E. Shaw, R.A. Friesner,
A hierarchical approach to all-atom protein loop prediction, Proteins 55
(2004) 351e367.
[30] M.P. Jacobson, R.A. Friesner, Z. Xiang, B. Honig, On the role of the crystal
environment in determining protein side-chain conformations, J. Mol. Biol.
320 (2002) 597e608.
[34] a) V. Katrich, M. Rueda, R. Abagyan, Ligand-guided receptor optimization,
Methods Mol. Biol. 857 (2012) 189e205;
b) V. Katritch, V.-P. Jaakola, J.R. Lane, J. Lin, A.P. IJzerman, M. Yeager,
I. Kufareva, R.C. Stevens, R. Abagyan, Structure-based discovery of novel
chemotypes for adenosine A2A receptor antagonists, J. Med. Chem. 53 (2010)
1799e1809;
c) V. Katritch, I. Kufareva, R. Abagyan, Structure based prediction of subtype-
selectivity for adenosine receptor antagonists, Neuropharmacology 60 (2011)
108e115;
d) P. Kolb, K. Phan, Z.-G. Gao, A.C. Marko, A. Sali, K.A. Jacobson, Limits of ligand
selectivity from docking to models: in silico screening for A1 adenosine re-
ceptor antagonists, PLoS ONE 7 (2012) 1e9;
e) I. Kufareva, V. Katritch, Participants of Gpcr dock 2013, S. Grudinin,
R.C. Stevens, R. Abagyan, Advances in GPCR modeling evaluated by the GPCR
dock 2013 assessment: meeting new challenges, Structure 22 (2014)
1120e1139.
€
€
[31] Schrodinger, Release 2014-3: Prime, Version 3.7, Schrodinger, LLC, New York,
NY, 2014.
[32] F. Xu, H. Wu, V. Katritch, G.W. Han, K. a. Jacobson, Z.-G. Gao, V. Cherezov,
R.C. Stevens, Structure of an agonist-bound human A2A adenosine receptor,
Sci. (N.Y.) 332 (2011) 322e327.
[33] V. Isberg, B. Vroling, R. van der Kant, K. Li, G. Vriend, D. Gloriam, GPCRDB: an
information system for G protein-coupled receptors, Nucleic Acids Res. 42
(2014) D422eD425.
[35] See supplementary information for more details.
[36] M.M. Mysinger, M. Carchia, J.J. Irwin, B.K. Shoichet, Directory of useful decoys,
enhanced (DUD-E): better ligands and decoys for better benchmarking,
J. Med. Chem. 55 (2012) 6582e6594.
€
[37] Small-Molecule Drug Discovery Suite 2014-3, Glide, Version 6.4, Schrodinger,
LLC, New York, NY, 2014.