Binding kinetics of ZM241385 derivatives at the human adenosine A 2A receptor

Article abstract of DOI:10.1002/cmdc.201300474

Classical drug design and development rely mostly on affinity- or potency-driven structure-activity relationships (SAR). Thus far, a given compound's binding kinetics have been largely ignored, the importance of which is now being increasingly recognized. In the present study, we performed an extensive structure-kinetics relationship (SKR) study in addition to a traditional SAR analysis at the adenosine A_2A receptor (A _2AR). The ensemble of 24 A_2AR compounds, all triazolotriazine derivatives resembling the prototypic antagonist ZM241385 (4-(2-((7-amino-2-(furan-2-yl)-[1,2,4]triazolo[1,5-a][1,3,5]triazin-5-yl)amino) ethyl)phenol), displayed only minor differences in affinity, although they varied substantially in their dissociation rates from the receptor. We believe that such a combination of SKR and SAR analyses, as we have done with the A _2AR, will have general importance for the superfamily of G protein-coupled receptors, as it can serve as a new strategy to tailor the interaction between ligand and receptor. Above and beyond: Insight into the binding kinetics of ZM241385 derivatives at the human adenosine A_2A receptor has provided additional information beyond a traditional structure-activity relationship (SAR) analysis. The strategy, combining a structure-kinetics relationship investigation and SAR, can serve as an important tool for more directed medicinal chemistry efforts in the future.

Full text of DOI:10.1002/cmdc.201300474

CHEMMEDCHEM
FULL PAPERS
DOI: 10.1002/cmdc.201300474
Binding Kinetics of ZM241385 Derivatives at the Human
Adenosine A2A Receptor
Dong Guo, Lizi Xia, Jacobus P. D. van Veldhoven, Marc Hazeu, Tamara Mocking,
Johannes Brussee, Adriaan P. IJzerman, and Laura H. Heitman*[a]
Classical drug design and development rely mostly on affinity-
or potency-driven structure–activity relationships (SAR). Thus
far, a given compound’s binding kinetics have been largely ig-
nored, the importance of which is now being increasingly rec-
ognized. In the present study, we performed an extensive
structure–kinetics relationship (SKR) study in addition to a tradi-
tional SAR analysis at the adenosine A2A receptor (A2AR). The
ensemble of 24 A2AR compounds, all triazolotriazine derivatives
resembling the prototypic antagonist ZM241385 (4-(2-((7-
amino-2-(furan-2-yl)-[1,2,4]triazolo[1,5-a][1,3,5]triazin-5-yl)ami-
no)ethyl)phenol), displayed only minor differences in affinity,
although they varied substantially in their dissociation rates
from the receptor. We believe that such a combination of SKR
and SAR analyses, as we have done with the A2AR, will have
general importance for the superfamily of G protein-coupled
receptors, as it can serve as a new strategy to tailor the inter-
action between ligand and receptor.
Introduction
G protein-coupled receptors (GPCRs) are among the largest
and most heavily investigated drug targets in the drug re-
search community. Traditional early phase drug design and dis-
covery campaigns of GPCRs largely depend on equilibrium af-
finity- or potency-based structure–activity relationships (SAR).
This approach of lead optimization allows a quick synthesis–
evaluation feedback loop to pool abundant candidate com-
pound assemblies for further drug evaluation. Nevertheless,
this classical SAR approach does not seem to predict clinical ef-
ficacy very well, which is evidenced by the high levels of attri-
tion during the translation of lead compound in vitro activity
into in vivo and clinical evaluation. To address this issue, sever-
al recent reviews have emphasized the importance of binding
kinetics, and in particular, the lifetime of a drug–target binary
complex (i.e., drug–target residence time (RT)), as a critical dif-
ferentiator and predictor for drug efficacy and safety.[1] In addi-
tion to the RT, the association rate of a ligand–receptor interac-
tion, which reflects the “target engagement time” (ET), should
also be taken into consideration in the early phases of drug re-
search. This is especially important for designing drugs that re-
quire a fast onset of action and potentially for drugs that act
on temporarily existing targets, such as protein–protein inter-
actions.[2] Therefore, an extensive structure–kinetics relation-
ship (SKR) investigation, in addition to the traditional SAR anal-
ysis, can be of great use in the early phases of candidate drug
optimization.
The human adenosine A2A receptor (A2AR) is a subtype of ad-
enosine receptors (other subtypes are A1, A2B, and A3) belong-
ing to the superfamily of GPCRs.[3] Antagonists for this receptor
have been reported as potential treatment for Parkinson’s dis-
ease.[4] As such, many compounds with high A2AR affinities
have been developed,[5] including the reference antagonist
ZM241385
(4-(2-((7-amino-2-(furan-2-yl)-[1,2,4]triazolo[1,5-a]-
[1,3,5]triazin-5-yl)amino)ethyl)phenol), a triazolotriazine deriva-
tive.[6] These compounds were well characterized and opti-
mized in terms of their binding affinity and thus benchmarked
for later medicinal chemistry attempts targeting the A2AR. For
example, Vu and colleagues synthesized ZM241385 derivatives
with increased bioavailability.[7] However, the success rate of
the developed A2AR antagonists in clinical trials is disappoint-
ingly low. On one hand, this indicates that the results of cur-
rently used preclinical animal models do not translate well into
clinical studies. On the other hand, traditional, affinity-directed
SAR alone may not be sufficient enough to select candidates
for preclinical tests, especially when comparing compounds
that are otherwise biologically or chemically similar. Although
A2AR antagonists have been previously investigated extensively
in terms of their affinity or potency, little is known about their
binding kinetics thus far. It is of great importance to be able to
optimize the kinetic profiles of such compounds, in addition to
[a] D. Guo, L. Xia, J. P. D. van Veldhoven, M. Hazeu, T. Mocking, Dr. J. Brussee,
Prof. A. P. IJzerman, Dr. L. H. Heitman
Division of Medicinal Chemistry
Leiden Academic Centre for Drug Research (LACDR)
P.O. Box 9502, 2300 RA Leiden (The Netherlands)
E-mail: l.h.heitman@lacdr.leidenuniv.nl
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201300474.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 752 – 761 752

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
their affinity, by medicinal
chemistry efforts. Thus, we de-
cided to further extend the
series of ZM241385 derivatives
by progressively modifying the
C2 position to generate an in-
sight into both SKR and SAR. We
believe that the present study
adds knowledge to our current
understanding of drug design
and development for A2AR an-
tagonists. Hopefully, this meth-
odology of combining both SKR
and SAR can be generally ap-
plied to other drug targets as
well in the future.
Results and Discussion
Chemical synthesis
Synthesis routes are depicted in
Schemes 1 and 2. In total, an en-
semble of 24 triazolotriazine de-
rivatives (12a–x) was obtained.
Notably, compounds 12a, 12b,
and 12x were previously report-
ed by Vu et al.[7c] and were re-
Scheme 2. Preparation of key intermediates. Reagents and conditions: a) piperazine, N,N-dimethylacetamide,
1658C; b) bis(2-chloroethyl)amine hydrochloride, Na2CO3, 1308C, butanol; c) tBuONa, BINAP, Pd2(dba)3, toluene, N2,
110 8C; d) K2CO3, NaI, butanone, reflux; e) H2NNH2·H2O, EtOH; f) 3-bromoalkylphthamide, K2CO3, DMF, 708C;
g) H2NNH2·H2O, EtOH, 708C.
synthesized in the present study,
although using a different syn-
were carried out via N-alkylation of the commercially available
piperazine derivatives (9a–d and 9i–u) or the in-house synthe-
sized phenylpiperazines (9v–x), which were derived from 6–
8,[10] to obtain the appropriate N-phthalimide-protected alkyl
piperazines (10a–d and 10i–x). This was followed by deprotec-
tion of the phthalimide to afford the free amines (11 a–d and
11 i–x).
thetic approach. All compounds (12a–x) were synthesized be-
ginning from furan-2-carbohydrazide (1) to generate 7-amino-
2-(furyl)-5-methylthio[1,2,4]triazolo[1,5-a][1,3,5]triazine (4), fol-
lowing the synthetic approach reported by Dolzhenko et al.
and Jçrg et al.[8] Subsequently, 4 was oxidized with 3-chloro-
perbenzoic acid (mCPBA) to afford the corresponding sulfox-
ide/sulfone mixture 5,[9] which was substituted with a variety
of commercially available amines (11 f–h) to generate 12 f–h,
or with in-house prepared amines (11 a–d and 11 i–x) to gener-
ate 12a–d and 12i–x. Compound 12e was obtained by the N-
Boc deprotection of 12d.
SAR and SKR of triazolotriazine derivatives
The SAR and SKR analyses were initiated by testing two com-
pounds, 12a and 12b, then chemical modifications were grad-
ually introduced to these two compounds (Table 1). Several ob-
servations were made: 1) the molecule with a two-
carbon spacer was superior to the compound with
a three-carbon spacer. The former (12a, Ki =0.30ꢀ
For the preparation of intermediate amines 11 a–d and 11 i–
x, synthetic routes are depicted in Scheme 2. In brief, reactions
0.08 nm; RT=164ꢀ32 min) displayed
a fourfold
higher affinity and 41-fold longer RT than the latter
(12b, Ki =1.3ꢀ0.1 nm; RT=4ꢀ1 min). Such variation
in linker length also resulted in different association
rates (12a, kon =0.051ꢀ0.005 nmꢁ1·minꢁ1; 12b, kon
=
0.16ꢀ0.06 nmꢁ1·minꢁ1). 2) Upon different degrees of
C2-phenylpiperazine modification, the ligand affinities
were moderately to largely affected, while their re-
ceptor RT values were drastically shortened (12d–h),
except for the Boc-protected intermediate 12d. This
compound, in fact, had a 30- and 20-fold improved
Scheme 1. General synthesis route to 24 triazolotriazine derivatives. Reagents and condi-
tions: a) S-methylisothiourea sulfate (2:1), 4% NaOH(aq) room temp.; b) H2O, room temp.;
c) 1. (MeS)2C=NCN, heat, 1808C, 2. CH2Cl2/CH3OH (2:1), reflux; d) mCPBA (70% strength),
CH2Cl2, 08C!room temp.; e) Et3N, CH3CN.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 752 – 761 753

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Table 1. Binding affinities and kinetics of ZM241385 and 12a–h.
hA2A
R
hA1R
[b]
[b]
Compd
R
Ki [nm][a]
kon [nmꢁ1 minꢁ1
]
koff [minꢁ1
]
RT [min][c]
Ki [nm]
ZM241385
0.40ꢀ0.03
0.13ꢀ0.06
0.014ꢀ0.003
71ꢀ21
255[23]
12a
12b
12c
0.30ꢀ0.08
1.3ꢀ0.1
3.8ꢀ0.8
0.051ꢀ0.005
0.16ꢀ0.06
0.20ꢀ0.1
0.0061ꢀ0.0020
0.25ꢀ0.01
164ꢀ32
4ꢀ1
8%[d]
30%[d]
22%[d]
0.35ꢀ0.03
3ꢀ1
12d
1.5ꢀ0.1
0.030ꢀ0.003
0.012ꢀ0.004
83ꢀ17
2%[d]
12e
12 f
12g
12h
45ꢀ0.1
31ꢀ6
0.0063ꢀ0.0030
0.0057ꢀ0.0020
0.046ꢀ0.020
0.020ꢀ0.002
0.24ꢀ0.10
0.25ꢀ0.10
0.24ꢀ0.10
0.62ꢀ0.08
4ꢀ1
4ꢀ1
4ꢀ1
2ꢀ0
10%[d]
11%[d]
21%[d]
32%[d]
8.1ꢀ0.5
64ꢀ1
[a] Displacement of specific [3H]ZM241385 binding from the hA2AR at 48C. [b] kon and koff values were determined in a competition association assay at
48C. [c] RT (residence time)=1/koff. [d] Percent displacement of specific [3H]DPCPX binding from the hA1R at 1 mm at 258C. Data are the mean ꢀSEM of
three separate experiments, each performed in duplicate.
affinity and RT, respectively, relative to its truncated analogue
(12e). Moreover, ET values were also significantly influenced
by chemical modifications on the phenylpiperazine moiety.
Specifically, 12g displayed the fastest association rate (0.046ꢀ
0.020 nmꢁ1·minꢁ1). (3) Receptor RT values of 12a or 12d high-
light the preference for an electron-withdrawing effect at the
piperazine amine moiety. The insertion of an additional carbon
into 12a (between the piperazine and the phenyl group; 12g)
reversed the electron-withdrawing effect to a donating effect,
which resulted in decreased A2AR affinity (8.1ꢀ0.5 nm) and RT
(4ꢀ1 min). Replacement of the nitrogen by a carbon atom on
the “right side” of the piperazine (compare 12a and 12c) re-
sulted in strongly decreased RT values that further confirmed
the importance of the nitrogen in maintaining A2AR affinity and
RT (Table 1, Figure 1 for 12c). Notably, the ET and RT of 12c
were the shortest for this series of compounds (except for the
nearly 20-fold lower-affinity compound 12h) without a large
compromise on affinity. Taken together, these results highlight
the importance of the C2-phenylpiperazine-ethyl group, and
more specifically show that the electron-deficient nitrogen (on
the right side of the piperazine) has a role in preserving a tight
ligand–receptor interaction.
The SAR and SKR were further analyzed with 16 phenyl-sub-
stituted 12a analogues (Table 2, 12i–x). Upon para substitu-
tion at the phenyl ring (12i–n), no significant change in ligand
affinity (Ki values <1 nm) was observed, except for 12n, which
had a 4.7-fold decrease in affinity (1.4ꢀ0.2 nm). This decrease
was probably caused by steric hindrance induced by the bulky
phenyl substituent, which presumably also limited its RT to
29ꢀ2 min and decreased the association rate to 0.018ꢀ
0.002 nmꢁ1·minꢁ1. In contrast, the other para-substituted com-
pounds (12i–m) displayed a similar duration of in vitro recep-
tor occupancy as ZM241385 (RT=71ꢀ21 min). In comparison
with the convergent results upon para-position modifications,
ortho-substituted analogues (12o–r) displayed divergent affini-
ties and binding kinetics. Specifically, an ortho-methoxy sub-
stituent (12p) displayed decreased A2AR affinity and RT relative
to its para-substituted analogue (12m), while methyl- (12o) or
halogen-substituted (12q, 12r) analogues displayed increased
A2AR affinities and RT values. For most compounds, disubstitu-
tion of the phenylpiperazine did not dramatically change their
affinities or binding kinetics (12t–w). Interestingly for 12x,
which has ortho- and para-fluoro substituents, an exceptionally
long receptor RT of 323ꢀ25 min was found (Table 2) that was
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 752 – 761 754

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Table 2. Binding affinities and kinetics of compounds 12i–12x.
hA2AR
hA1R
[b]
[b]
Compd
R
Ki [nm][a]
kon [nmꢁ1·minꢁ1
]
koff [minꢁ1
]
RT [min][c]
Ki [nm]
ZM241385
12a
12i
12j
12k
12l
12m
12n
12o
12p
12q
12r
12s
12t
12u
12v
12w
12x
Table 1
4-H
4-CH3
4-Cl
4-CF3
4-F
4-OCH3
4-Ph
2-CH3
2-OCH3
2-Cl
0.40ꢀ0.03
0.30ꢀ0.08
0.79ꢀ0.06
0.29ꢀ0.10
0.38ꢀ0.10
0.54ꢀ0.05
0.51ꢀ0.10
1.4ꢀ0.2
0.13ꢀ0.06
0.014ꢀ0.003
0.0061ꢀ0.0020
0.016ꢀ0.006
0.018ꢀ0.006
0.020ꢀ0.005
0.020ꢀ0.005
0.0079ꢀ0.0020
0.034ꢀ0.010
0.0075ꢀ0.0020
0.070ꢀ0.070
0.0065ꢀ0.0010
0.011ꢀ0.002
0.012ꢀ0.003
0.012ꢀ0.001
0.015ꢀ0.004
0.014ꢀ0.001
0.0083ꢀ0.0010
0.0031ꢀ0.0002
71ꢀ21
164ꢀ32
63ꢀ18
56ꢀ11
50ꢀ37
50ꢀ8
255[23]
8%[d]
0.051ꢀ0.005
0.062ꢀ0.020
0.090ꢀ0.010
0.072ꢀ0.009
0.10ꢀ0.02
14%[d]
35 ꢀ13
64%[d]
57%[d]
11%[d]
32%[d]
29%[d]
19%[d]
30%[d]
28%[d]
62%[d]
29%[d]
21 ꢀ4
80ꢀ24
27%[d]
22%[d]
0.064ꢀ0.005
0.018ꢀ0.002
0.062ꢀ0.002
0.032ꢀ0.003
0.068ꢀ0.016
0.052ꢀ0.012
0.055ꢀ0.009
0.11ꢀ0.02
127ꢀ19
29ꢀ2
0.13ꢀ0.04
3.5ꢀ0.7
133ꢀ21
14ꢀ11
154ꢀ25
91ꢀ15
83ꢀ14
78ꢀ9
0.13ꢀ0.03
0.12ꢀ0.05
0.29ꢀ0.03
0.16ꢀ0.01
0.31ꢀ0.10
0.15ꢀ0.02
0.24ꢀ0.05
0.33ꢀ0.04
2-F
3-F
2,4-diCH3
3,4-diCl
2,4-diCl
2-F, 4-OCH3
2,4-diF
0.10ꢀ0.01
67ꢀ11
70ꢀ5
0.11ꢀ0.01
0.054ꢀ0.005
0.034ꢀ0.004
120ꢀ65
323ꢀ25
[a] Displacement of specific [3H]ZM241385 binding from the hA2AR at 48C. [b] kon and koff values were determined in a competition association assay at
48C. [c] RT (residence time)=1/koff. [d] Percent displacement of specific [3H]DPCPX binding from the hA1R at 1 mm at 258C. Data are the mean ꢀSEM of
three separate experiments, each performed in duplicate.
much longer than the simple sum of the RT values of the mon-
ofluorinated analogues (12l, 12r, and 12s) and almost fivefold
longer than the RT of ZM241385. From Figure 1B, it also fol-
lows that 12x had a much longer RT than ZM241385 (the
radioligand), as a typical “overshoot” in specific radioligand
binding was observed.[11] By contrast, if a competitor dissoci-
ates faster from its target than the radioligand, the specific
binding of the radioligand will slowly and monotonically ap-
proach equilibrium over time, as observed for 12c (Fig-
ure 1B).[11]
in the para position, yet this substituent in other positions
(e.g., 12q, ortho substituent) did not exhibit lower selectivity
for the A2AR over the A1R.
Functional characterization of 12x and 12c in a cAMP assay
Subsequently, the compounds with the longest and shortest
RT with high affinity (i.e., 12x and 12c) were functionally char-
acterized in an A2AR agonist-induced cAMP assay, which re-
vealed their antagonistic behavior. Firstly, it follows from Fig-
ure 1C that both 12x and 12c induced a concentration-depen-
dent decrease of intracellular cAMP levels with 16-fold differ-
ence in their IC50 values, which were 1.4ꢀ0.1 nm and 21.8ꢀ
0.5 nm, respectively (Table 3). Secondly, pre-treatment of
HEK293 hA2AR cells with different concentrations of 12x before
stimulation with an AR agonist (i.e., 5’-N-ethylcarboxamidoade-
nosine, NECA) induced insurmountable antagonism (Table 3).
In other words, the NECA concentration–effect curve was shift-
ed to the right, with a concomitant decrease in the maximal
Almost all prepared phenylpiperazine triazolotriazine deriva-
tives displayed high selectivity over the human adenosine A1
receptor (A1R), i.e., <50% of [3H]DPCPX displacement on A1R
at 1 mm (Tables 1 and 2). Notably, the compound with the lon-
gest RT toward A2AR, 12x, had very good A2AR selectivity over
the A1R (Table 2). In contrast, 12j, 12u, and 12v, with mono-
or dichloro substitution, lost some selectivity over the A1R (Ki
values at the A1R ranged from 20 nm to 80 nm, Table 2). Inter-
estingly, all of these compounds contain a chloro substituent
Table 3. Functional characterization of 12x and 12c in a cAMP assay.
Compd
IC50 [nm][a]
Pre-incubation[a]
Co-incubation[a]
Mode of antagonism[b]
pA2
Schild slope
pA2
Schild slope
12c
12x
1.4ꢀ0.1
8.62ꢀ0.29
0.93ꢀ0.11
8.57ꢀ0.06
9.69ꢀ0.03
0.93ꢀ0.02
1.13ꢀ0.01
Competitive surmountable
Competitive insurmountable
21.8ꢀ0.5
NA
NA
[a] Antagonist potency values were determined from concentration–response curves for 12x and 12c in the presence of 100 nm NECA with a 30 min co-in-
cubation; antagonists were pre-incubated for 30 min or co-incubated with NECA at concentrations ranging from 100 mm to 0.1 nm. Data are the mean ꢀ
SEM of three separate experiments, each performed in duplicate at 22–258C. NA: not available. [b] See text for further explanation.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 752 – 761 755

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
steric mode of inhibition, which would be proven by suppres-
sion of the maximal response in the co-incubation experi-
ment.[12] Notably, the generated pA2 values of 12x and 12c in
this experimental setup were similar to their pKi values, and
the derived Schild slopes were close to unity (Table 3). Togeth-
er, this confirmed that 12x or 12c bound fully competitively
with NECA, and the insurmountable A2AR antagonism of 12x
was a result caused by so-called hemi-equilibrium during the
functional assay, due to its long A2AR RT profile.[12b]
It needs to be pointed out that functional characterization
and determination of antagonist binding kinetics were per-
formed at 48C and thus do not reflect in vivo RT values, that
is, at body temperature. However, it is reasonable to expect
that the ranking of the compound RT values at 48C will agree
with those at higher, more physiological, temperatures. One
example at the A2AR is the agonist UK432097, which has previ-
ously been shown to have a fivefold longer RT than agonist
CGS21680.[15b] This difference at 48C apparently translates into
a distinct duration of action in vivo reported by Mantell et al.,
that is, 8 h for UK432097 and less than 1 h for CGS21680.[14]
Generation of a kinetics map and physicochemical
correlation plots
Next, we plotted an on-/off-rate graph, or “kinetics map”, in-
cluding the data for all A2AR ligands obtained in this study
(Figure 3).[13] This kinetics map depicted the ligand–receptor
binding affinity (KD, represented by parallel diagonal lines) as
detailed kinetic rates that reflect the process of target ET (kon,
y axis) and the target RT (koff, x axis), respectively. We observed
that the compounds can be divided into three groups: firstly,
both the kon and koff values could vary by one order of magni-
tude (Figure 3, Group A), while the KD remained within
a narrow range (0.1–0.3 nm), as mentioned above. This indicat-
ed that compounds with the same affinity may have many dif-
ferent combinations of on- and off-rates, even within the same
scaffold and target system. Such information, often ignored or
unavailable in traditional SAR studies, can in fact be highly de-
cisive in translating the in vitro profile in of a lead to in vivo
pharmacokinetics (PK) and/or pharmacodynamics (PD) behav-
ior.[14] Secondly, compounds sharing the same off-rate may
bear divergent KD values, due to different on-rates (Group B). It
has been shown in several cases that a slow compound disso-
ciation rate is pivotal for high in vivo efficacy.[15] Thus, in retro-
spect, one could imagine that many compounds with promis-
ing koff values were overlooked simply due to their low scores
in classical affinity- or potency-dominated evaluations. Thirdly,
the same holds for a compound’s on-rate (Group C), i.e.,
merely focusing on the KD value of a ligand can result in com-
pounds without the desired on-rate, as exemplified by candi-
date drugs aimed at acute diseases where a rapid onset of
action is desired (e.g., acute respiratory distress syndrome).[2a,f]
Taken together, the kinetics map provides a detailed interpre-
tation of a ligand–receptor binding process with a full invento-
ry of kon, koff, and KD values of a series of compounds.
Figure 1. a) Displacement of specific [3H]ZM241385 binding from the hA2A
receptor by two representative compounds, namely 12x and 12c.
b) [3H]ZM241385 competition association binding in the absence of ligand
(control) and in the presence of 10ꢁKi of unlabeled 12x or 12c. Data were
fitted to the equation described in the Experimental section to calculate the
kon and koff values for unlabeled ligands. Representative graphs are from one
experiment performed in duplicate (see Tables 1 and 2 for affinity and kinet-
ic values); c) Concentration–effect curves for 12x and 12c in a cAMP assay
(percentage relative to 100 mm NECA). Data were obtained by adding
HEK293 hA2AR cells to the mixture of the antagonist (12x or 12c) and
100 nm NECA for a 30 min incubation. Data are expressed as mean ꢀSEM
from at least three independent experiments (see Table 3 for potency
values).
response (Figure 2A). Conversely, 12c displayed surmountable
A2AR antagonism (Table 3), shifting the NECA curve to the right
without affecting its maximal response (Figure 2B). In addition,
the pA2 value for 12c generated from a Schild plot was 8.62ꢀ
0.29, which was similar to its pKi value (8.40ꢀ0.10), and the
Schild slope was close to unity (0.93ꢀ0.11), suggesting that
12c competed with NECA for the same receptor binding site.
To further examine whether 12x and 12c both bound to the
same site as the agonist, we also performed a co-incubation
experiment with 12x or 12c in the presence of NECA. It fol-
lows from Figure 2C,D that in this experimental setup, both
compounds produced a shift to the right in the NECA dose–re-
sponse curve without suppression of the maximal response, in-
dicative of a competitive interaction. Hence, these findings
oppose that insurmountable antagonism resulted from an allo-
Molecular and physicochemical properties of the synthesized
phenylpiperazine analogues (12i–x and 12a) and their puta-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 752 – 761 756

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
is the same, we specifically fo-
cused on the properties of C2-
phenylpiperazine
fragments.
Several descriptors were selected
to be further examined in corre-
lation plots. These included the
size (molecular weight, Mr [Da]),
surface (molecular surface area,
MSA), lipophilicity (logP), and
charge (ionization constant, pKa)
of the fragment. There was no
obvious linear correlation be-
tween the association/dissocia-
tion rates and Mr, logP, MSA, or
pKa values. We also performed
a multiple linear regression anal-
ysis to check whether com-
pound binding kinetics were di-
rected by a combination of two
or more of the physicochemical
descriptors. However, no signifi-
cant correlation was found (sig-
nificance F>0.05 in all cases). Al-
together, this indicated that the
binding kinetics were com-
pound-specific and that there is
no general trend in the correla-
tion of their molecular and phys-
icochemical properties.
Figure 2. cAMP experiments were performed on HEK293 human embryonic kidney cells stably expressing the
hA2AR at room temperature (22–258C). a) 12x or b) 12c were incubated for 30 min prior to challenge of the ade-
nosine receptor agonist NECA at concentrations ranging from 100 mm to 0.1 nm for another 30 min. c) 12x or
d) 12c were co-incubated with NECA at concentrations ranging from 100 mm to 0.1 nm for 30 min. The agonist
curves were generated in the presence of increasing concentrations of antagonist, namely 0.3-, 1-, 3- and 10-fold
their respective Ki values. Data were normalized according to the maximal response produced by 100 mm NECA.
The shift in agonist EC50 was determined to perform Schild analyses. Data are expressed as mean ꢀSEM from at
least three independent experiments performed in duplicate.
Importance of the C2-phenylpiperazine fragment and its
location
In this study, we observed that, upon minor chemical modifica-
tions of the phenylpiperazine side chain on the triazolotriazine
scaffold (Tables 1 and 2), binding affinity of the derivatives un-
derwent only subtle changes, while their binding kinetics were
very sensitive to such structural variations. For instance, upon
substitution of 12a, 12k (para-trifluoromethyl substituted) dis-
played a similar Ki value as 12a, while its off-rate was increased
3.4-fold. In another case, the on- and off-rates of 12u (meta-,
para-chloro disubstituted) were increased and decreased, re-
spectively, by a similar magnitude (approximately twofold),
leading to an unchanged affinity value relative to 12a. Table 2
as a whole exemplifies the difficulty of selecting a next-stage
candidate based on SAR alone. Most compounds have sub-
nanomolar affinity, and compound 12x does not stand out in
any particular way.
Figure 3. Kinetics map (y axis: kon, nmꢁ1·minꢁ1; x axis: koff, minꢁ1) of all A2AR li-
gands tested in this study. The kinetically derived affinity (KD =koff/kon) is rep-
resented by parallel diagonal lines. Group A: compounds that had varied kon
and koff values across several orders of magnitude, while the KD remained
within a similar range (0.1–0.3 nm). Group B: compounds that had the same
koff values, but divergent KD values. Group C: compounds that had the same
kon values, but divergent KD values.
The lack of large changes in binding affinities might be ex-
pected, given the absence of direct interactions between the
phenylhydroxy group and the receptor in a recently deter-
mined high-resolution crystal structure of ZM241385-bound
A2AR.[17] In this structure, the phenylhydroxy group points away
from the binding pocket toward the extracellular space. Like-
wise, in another crystal structure of UK432097-bound A2AR, the
bulky tail of agonist UK432097 at the adenine C2 position ex-
tends out of the ligand-binding cavity.[18] Notably, it was re-
tive relationship with the on- and off-rates were examined in
efforts to identify key factor(s) affecting their binding kinetics
(Figure 4). As the main scaffold of these series of compounds
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 752 – 761 757

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
study of the b1- and b2-adrener-
gic receptors by Dror et al.[19]
They found that several beta
blockers and one beta agonist
all traverse the same well-de-
fined, dominant pathway as they
bind to the b1- and b2-adrenergic
receptors, initially making con-
tact with a so-called “vestibule”
on the receptor’s extracellular
surface. Interestingly, this holds
true for the ligand binding dy-
namics of the M3 muscarinic ace-
tylcholine receptor too. Simula-
tion results indicated that as tio-
tropium binds to or dissociates
from the receptor, it pauses at
an alternative binding site in the
extracellular vestibule.[20] Taken
together, such an extracellular
vestibule appears to play an im-
portant role in the on- or off-tra-
jectory to and from the binding
pocket of a GPCR.[19] We there-
fore assume that the derivatives
of ZM241385 also transiently in-
teract with a similar extracellular
region of the A2AR. This is in ac-
cordance with our observation
that a change in the C2-phenyl-
piperazine group significantly af-
fected the ligand association
and dissociation rates at the
A2AR, while their Ki values were
minimally changed.
Conclusions
We have exemplified an exten-
sive SKR in addition to a tradi-
tional SAR analysis at the A2AR.
Compound 12x, the high-affinity
A2AR ligand previously reported
by Vu et al.,[7c] was revealed to
have an exceptionally long RT
(323 min). Compared with tradi-
tional SAR analysis, such a kinetic
insight provided a further ration-
Figure 4. Molecular descriptors of a) size (Mr), b) lipophilicity (logP), c) molecular surface area (MSA, 3D), and
d) charge (pKa) of the substituted C2-phenylpiperazine fragments and their correlation with the log values of on-
(left side) and off-rates (right side). No clear linear correlation was observed between the association/dissociation
rates and Mr (for association: R2 =0.0089, P=0.7180; for dissociation: R2 =0.1059, P=0.2024), logP (for association:
R2 =0.0023, P=0.8563; for dissociation: R2 =0.0731, P=0.2938), MSA (for association: R2 =0.1629, P=0.1082; for
dissociation: R2 =0.2107, P=0.0638), or pKa (for association: R2 =0.0018, P=0.8727; for dissociation: R2 =0.0187,
P=0.6010).
cently published that this agonist also has a slow association
rate and a long RT at the A2AR.[15b] Based on these findings, we
postulated that the C2-phenylpiperazine group protrudes out-
ward without forming direct interactions with residues in the
binding pocket of the triazolotriazine core. Instead, it may in-
teract with residues that are located in the extracellular loops
or the adjacent regions of the binding cavity along its trajecto-
ry of associating to or dissociating from the receptor. Such rea-
soning is supported by a recent molecular dynamics simulation
ale to support the selection of 12x from otherwise chemically
and biologically similar compounds for further testing. Kinetics
mapping of all tested A2AR ligands also provided a detailed in-
terpretation of the ligand–receptor binding process. Next,
a functional comparison between 12c and 12x in different
assay formats (co-application versus pre-incubation) further re-
vealed competitive insurmountable antagonism of 12x at the
hA2AR—a phenomenon distinct from that of 12c. In addition,
investigation of the molecular properties indicated that the
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 752 – 761 758

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
ganic layer was washed with H2O and brine (2ꢁ100 mL each) and
dried over anhydrous Na2SO4. After removing the solvent, 4 was
obtained as a white solid (6.51 g, two-step yield: 54%): 1H NMR
(400 MHz, [D6]DMSO): d=12.08 (brs, 1H, NH), 7.68 (s, 1H), 6.67 (s,
1H), 6.54 (s, 1H), 6.08 ppm (s, 2H).
ligand–receptor binding kinetics were most likely driven by
specific interactions between the ligand and the receptor. As
an extension of the current study, it would be of great interest
to subject compounds having similar affinity yet different bind-
ing kinetics to (pre)clinical tests. This would show how relevant
the variations in RT and on-/off-rates are in terms of in vivo effi-
cacy and duration of action. We believe that SKR, in combina-
tion with traditional SAR, can serve as an important tool for
more directed medicinal chemistry efforts in the future.
2-(Furan-2-yl)-5-(methylthio)-[1,2,4]triazolo[1,5-a][1,3,5]triazin-7-
amine (4): A mixture of amine 3 (43.4 mmol, 6.5 g) and dimethyl
N-cyanodithio(imino)carbonate (47.8 mmol, 7.0 g) was heated at
1808C in a stream of nitrogen for 4 h, then cooled to room tem-
perature to add 30 mL of a CH2Cl2/CH3OH (1:1) solution. The mix-
ture was stirred at reflux for another 1.5 h, followed by filtration.
The solids were washed by the solution, and the solvent was re-
moved under reduced pressure. The residue was purified by
column chromatography, eluting with CH2Cl2 containing increasing
amounts of EtOAc (0–50%). This gave compound 4 as a pale-
Experimental Section
Chemical synthesis
1
yellow solid (3.2 g, yield: 30%): H NMR (400 MHz, CD3OD): d=7.91
General: All solvents and reagents were purchased from commer-
cial sources and were of analytical grade. Demineralized water is
simply referred to as H2O and 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,
101 MHz) at room temperature. Chemical shifts are reported in
parts per million (ppm), are designated by d, and are downfield of
the internal standard tetramethylsilane (TMS). Coupling constants
are reported in Hz and are designated as J. High-resolution mass
spectrometry was performed by the Leiden Institute of Chemistry
and recorded by direct injection (2 mL of a 2 mm solution in H2O/
CH3CN; 50:50; v/v and 0.1% formic acid) on a mass spectrometer
(Thermo Finnigan LTQ Orbitrap) equipped with an electrospray ion
source in positive mode (source voltage 3.5 kV, sheath gas flow 10,
capillary temperature 2758C), with resolution (R)=60000 at m/z
400 (mass range m/z=150–2000) and calibrated for dioctylphtha-
late (m/z=391.28428). Analytical purity of the final compounds
was determined by high performance liquid chromatography
(HPLC) with a Phenominex Gemini 3u C18 110A column (50ꢁ
4.6 mm, 3 mm), measuring UV absorbance at 254 nm. Sample prep-
aration and HPLC method were as follows, unless stated otherwise:
0.3–0.8 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, de-
creasing polarity of the solvent mixture over time from 80:10:10 to
90:0:10. All compounds showed a single peak at the designated
RT and are at least 95% pure. Thin-layer chromatography (TLC) was
routinely consulted to monitor the progress of reactions, using alu-
minum-coated Merck silica gel F254 plates. Purification by column
chromatography was achieved by use of Grace Davison Davisil
silica column material (LC60A, 30–200 mm). Solutions were concen-
trated using a Heidolph Laborota W8 2000 evaporation apparatus
and by high vacuum on a Binder APT line vacuum drying oven.
The procedure for a series of similar compounds is given as a gen-
eral procedure for all within that series, annotated by the numbers
of the compounds.
(dd, J=1.7, 0.7 Hz, 1H), 7.12 (dd, J=3.4, 0.7 Hz, 1H), 6.70 (dd, J=
3.4, 1.7 Hz, 1H), 2.55 ppm (s, 3H).
2-(Furan-2-yl)-5-(methylsulfonyl)-[1,2,4]triazolo[1,5-a]-
[1,3,5]triazin-7-amine (5): A solution of mCPBA (70%, 32.5 mmol,
8.0 g) in CH2Cl2 (20 mL) was added to a stirred, ice-cooled suspen-
sion of the sulfide (R=MeSO) (13.0 mmol, 3.2 g) in CH2Cl2 (50 mL).
The resulting solution was stirred overnight (08C!room tempera-
ture). The solvent was removed, and EtOH (70 mL) was added to
the residue. The solid was collected by filtration, washed with
EtOH, and dried in a vacuum oven to give a white solid (R=
1
MeSO2) (yield: 3.2 g, 88%): H NMR (400 MHz, [D6]DMSO): d=9.81
and 9.48 (2ꢁs due to dimer formation,[8c] 2H, NH2), 7.98 (dd, J=
1.2, 0.8 Hz, 1H), 7.34 (dd, J=2.4, 0.8 Hz, 1H), 6.73 (dd, J=2.4,
1.2 Hz, 1H), 3.35 ppm (s, 3H).
1-(2,4-Dichlophenyl)piperazine (9v): A mixture of 6 (16.7 mmol,
2.0 mL) and piperazine (83.6 mmol, 7.2 g) in 10 mL of N,N-dimethy-
lacetamide was heated in the microwave at 1658C for 6.5 h, after
which 6 was consumed, as shown by TLC. H2O and CH2Cl2 were
added, and the pH value was adjusted to 1 with 1m HCl(aq). The
aqueous layer was washed three times with CH2Cl2 and subse-
quently brought to pH 12 with 5m NaOH(aq). After extraction of the
basified aqueous layer with CH2Cl2, the combined organic layers
were washed four times with H2O, dried over MgSO4, and concen-
trated in vacuo to yield 9v as a yellow oil (yield: 2.4 g, 61%):
1H NMR (400 MHz, CDCl3): d=7.36 (d, J=2.4 Hz, 1H), 7.19 (dd, J=
8.8, 2.4 Hz, 1H), 6.95 (d, J=8.4 Hz, 1H), 3.06–3.02 (m, 5H, 2ꢁCH2
and NH), 2.99–2.96 ppm (m, 4H).
1-(2-Fluoro-4-methoxyphenyl)piperazine hydrochloride (9w): A
mixture of bis(2-chloroethyl)amine hydrochloride (9.59 mmol, 1.7 g)
(7) and 1-BuOH (20 mL) was treated slowly with 2-fluoro-4-methox-
ybenzenamine (9.14 mmol, 1.3 g) at room temperature. After the
addition, the mixture was stirred at reflux for 48 h and then
cooled. The solid was filtered and rinsed with CH3OH and Et2O to
1
give 9w as a white solid (yield: 660 mg, 29%): H NMR (400 MHz,
CDCl3): d=9.21 (brs, 2H, NH, and HCl), 7.07–7.02 (m, 1H), 6.86–
6.83 (m, 1H), 6.74–6.71 (m, 1H), 3.72 (s, 3H), 3.28–3.20 (m, 4H),
3.14–3.10 ppm (m, 4H).
2-(Furan-2-carboxamido) guanidine (2): A mixture of hydrazide
1
(0.1 mol, 12.6 g) and S-methylisothiourea sulfate (0.05 mol,
13.9 g) in an 1% aqueous NaOH solution (400 mL) was stirred at
room temperature for 72 h. The precipitated solid (2), was filtered,
washed with ice water, and used in next step without further pu-
rification: 1H NMR (400 MHz, [D6]DMSO): d=10.77 (brs, 1H, NH),
7.56 (s, 1H), 6.88 and 6.76 (2ꢁ s due to dimer formation, 2H, NH2),
6.64 (d, J=2.8 Hz, 1H), 6.45 ppm (dd, J=2.0, 0.8 Hz, 1H).
1-(2,4-Difluorophenyl)piperazine) (9x): A mixture of piperazine
(24.9 mmol, 2.14 g), 1-bromo-2,4-difluorobenzene (8) (4.1 mmol,
0.8 g), tBuONa (5.8 mmol, 0.56 g), BINAP (0.25 mmol, 0.16 g), and
Pd2(dba)3 (0.083 mmol, 0.048 g) in dry toluene was heated at
110 8C under a nitrogen atmosphere for 24 h. The mixture was fil-
tered over Celite and rinsed with CH2Cl2. The solution was washed
with H2O and brine (2ꢁ10 mL each), dried over Na2SO4, and the
solvent was evaporated in vacuo. The residue was purified by silica
5-(Furan-2-yl)-2H-1,2,4-triazol-3-amine
(3):
Guanidine
2
(53.6 mmol, 9.0 g) was stirred in a 1:1 mixture of EtOAc/H2O
(400 mL) for 3 h. After extraction with EtOAc (2ꢁ150 mL), the or-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 752 – 761 759

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
gel via CH2Cl2/CH3OH (10:1) to give compound 9x as a pale-yellow
oil (yield: 337 mg, 42%): 1H NMR (400 MHz, CDCl3): d=6.91–6.87
(m, 1H), 6.83–6.78 (m, 2H), 3.06 (t, J=3.6 Hz, 4H), 3.00–2.97 (m,
4H), 1.80 ppm (s, 1H, NH).
expressing the hA2AR (HEK293 hA2AR) were kindly provided by Dr. J.
Wang (Biogen/IDEC, Cambridge, MA). Chinese hamster ovary cells
stably expressing the hA1R (CHOhA1R) were kindly provided by
Prof. Steve Hill (University of Nottingham, UK). All other chemicals
were of analytical grade and obtained from standard commercial
sources.
General procedure for the preparation of compounds 10a–d
and 10i–x: A mixture of the appropriate phthalimide-protected
alkyl bromide (7.5 mmol), piperazine derivative (5 mmol) and K2CO3
(10 mmol) in DMF (5 mL) was stirred at 708C overnight. The reac-
tion mixture was cooled, washed with H2O (5 mL), and extracted
with EtOAc (3ꢁ10 mL each). The organic phase was then com-
bined, dried over Na2SO4, and evaporated in vacuo to give the
crude product, which was recrystallized from EtOH and/or CH3OH
or purified by chromatography (petroleum ether/EtOAc).
Cell culture and membrane preparation: Cell culture and mem-
brane preparation were performed as reported previously.[15b,21]
Radioligand displacement assay: Radioligand displacement from
the hA1R and hA2AR was determined using the displacement assay
as described previously.[15b,21]
Radioligand competition association assay: The binding kinetics
of unlabeled A2AR ligands were determined at 48C using the com-
petition association assay as described previously.[15b]
2-(2-(4-Phenylpiperazin-1-yl)ethyl)isoindoline-1,3-dione
(10a):
Compound 10a was obtained as a pale-yellow solid after column
chromatography with petroleum ether/EtOAc (5:1–1:1) (yield:
1.5 g, 43%): H NMR (400 MHz, CDCl3): d=7.86–7.80 (m, 4H), 7.25–
7.24 (m, 2H), 6.91–6.88 (m, 2H), 6.80 (t, J=8.1 Hz, 1H), 3.87 (t, J=
6.3 Hz, 2H), 3.15–3.12 (m, 4H), 2.72–2.68 ppm (m, 6H).
cAMP assay: HEK293 hA2AR cells were cultured as a monolayer on
10 cm ø culture plates to 80%-90% confluency. Cells were harvest-
ed and centrifuged two times at 200ꢁg for 5 min. The amount of
cAMP produced was determined with the LANCE ultra cAMP 384
kit (PerkinElmer, Groningen, Netherlands). In general, 1000 cells per
well were seeded on 384-well plates and incubated at room tem-
perature (22–258C). cAMP was generated in the stimulation buffer
[N-2-hydroxyethylpiperazine-N’-ethanesulfonic acid (HEPES), 5 mm;
0.1% (w/v) BSA; cilostamide, 50 mm; rolipram, 50 mm; adenosine
deaminase (ADA), 0.8 IUmLꢁ1] Concentration–effect curves for 12x
and 12c were obtained by adding HEK293 hA2AR cells to a mixture
of antagonist (12x or 12c) and 100 nm NECA (prepared in the
stimulation buffer) for a 30 min co-incubation. For assessment of
either surmountable or insurmountable behavior, the antagonists
(12x and 12c) were pre-incubated for 30 min or co-incubated with
the agonist NECA at concentrations ranging from 100 mm to
0.1 nm for a duration of 30 min, with antagonist concentrations of
0.3-, one-, two- and tenfold their respective Ki values. The incuba-
tion was stopped by adding detection mix and antibody solution,
according to the instructions of the manufacturer. The generated
fluorescence intensity was quantified on an EnVision Multilabel
Reader (PerkinElmer, Groningen, Netherlands). The resulting data
were normalized according to the maximal response produced by
100 mm NECA. The shift in agonist EC50 was determined to perform
Schild analyses.
1
General procedure for preparation of compounds 11a–d and
11i–x: An excess of hydrazine hydrate was added (1–5 mL) to a so-
lution of isoindoline-1,3-dione (3 mmol) in EtOH (25 mL), and the
mixture was stirred at 708C overnight. The solvent was removed in
vacuo, and EtOAc was added to the residue. The solids were fil-
tered, dried over Na2SO4, and the solvent was evaporated in vacuo
to give the crude product as a pale-yellow oil or solid. The crude
product was used in the next step without further purification.
General procedure for preparation of compounds 12a–x: A mix-
ture of the respective amine (0.75 mmol), the sulfone/sulfoxide
mixture (5) (0.50 mmol), and Et3N (1.0 mmol) was dissolved in
CH3CN and stirred at reflux overnight. After removing the solvent
in vacuo, EtOAc was added, and the organic phase was washed
with H2O and brine (2ꢁ10 mL each), dried over Na2SO4, and evapo-
rated. The residue was purified by column chromatography using
silica gel and EtOAc or EtOAc/CH3OH to afford a white or off-white
solid.
2-(Furan-2-yl)-N5-(2-(4-phenylpiperazin-1-yl)ethyl)-[1,2,4]triazolo-
[1,5-a][1,3,5]triazine-5,7-diamine (12a): Eluting with EtOAc/
CH3OH (10:1) afforded the title compound as a white powder
(yield: 55 mg, 14%): 1H NMR (400 MHz, MeOD): d=7.68 (s, 1H),
7.22 (t, J=8.4 Hz, 2H), 7.12 (d, J=3.2 Hz, 1H), 6.96 (d, J=8.4 Hz,
2H), 6.84–6.81 (m, 1H), 6.60 (d, J=1.6 Hz, 1H), 3.61–3.60 (m, 2H),
3.20 (t, J=4.8 Hz, 4H), 2.72 (t, J=4.8 Hz, 4H), 2.68–2.67 ppm (m,
2H). 13C NMR (101 MHz, [D6]DMSO): d=161.6, 159.7, 156.3, 151.5,
150.5, 146.7, 145.1, 129.4, 119.2, 115.8, 112.4, 112.1, 57.1, 53.2, 48.7,
38.5 ppm. HRMS (ESI) m/z [M+H]+ calcd for C20H24N9O+: 406.2026,
found: 406.2092; HPLC: tR =11.96 min, purity=96.7%.
Data analysis: All experimental data was analyzed using GraphPad
Prism 5.0 (GraphPad Software Inc., San Diego, CA). KD and Bmax
values of [3H]ZM241385 at hA2AR membranes were obtained from
Guo et al.[15b] IC50 values obtained from competition displacement
binding data were converted into Ki values using the Cheng–Prus-
off equation.[22] Association and dissociation rates for unlabeled li-
gands were calculated by fitting the data in the competition asso-
ciation model using kinetics of competitive binding:[11]
KA ¼ k1½Lꢂþk2
KB ¼ k3½Lꢂþk4
ð1Þ
ð2Þ
ð3Þ
ð4Þ
ð5Þ
ð6Þ
Pharmacological characterization
Materials: [3H]ZM241385 (specific activity 50 Cimmolꢁ1
) and
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
[3H]1,3-dipropyl-8-cyclopentyl-xanthine ([3H]DPCPX, specific activity
116.7 Cimmolꢁ1) were purchased from ARC Inc. (St. Louis, USA). Un-
labeled ZM241385 was a gift from Dr. S. M. Poucher (AstraZeneca,
Macclesfield, UK). CGS21680 was a gift from Dr. R. A. Lovell (Ciba-
Geigy, Summit, NJ). NECA (5’-N-ethylcarboxamidoadenosine) and
DPCPX were purchased from Sigma–Aldrich (Steinheim, Germany).
Adenosine deaminase (ADA) was purchased from Boehringer
Mannheim (Mannheim, Germany). Bicinchoninic acid (BCA) and
BCA protein assay reagent were obtained from Pierce Chemical
Company (Rockford, IL, USA). Human embryonic kidney cells stably
2
S ¼ ½ðKAꢁKBÞ þ 4 ꢃ k1 ꢃ k3 ꢃ L ꢃ Iꢂ
KF ¼ 0:5 ðKAþKBþSÞ
KS ¼ 0:5 ðKAþKBꢁSÞ
Q ¼ Bmax ꢃ k1 ꢃ L ꢃ ðKFꢁKSÞꢁ1
ꢁ1
ꢁ1
ꢁ1
Y ¼ Q ꢃ ½k4 ꢃ ðKFꢁKSÞ ꢃ KF ꢃ KS þ ðk4ꢁKFÞ ꢃ KF
ð7Þ
FꢃX
ꢁ1
SꢃX
ꢃ eꢁK ꢁðK4ꢁKSÞ ꢃ KS ꢃ eꢁK
ꢂ
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 752 – 761 760

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Where X is the time (min), Y is the specific [3H]ZM241385 binding
(DPM), k1 and k2 are the kon (nmꢁ1·minꢁ1) and koff (minꢁ1) values of
[3H]ZM241385 obtained from Guo et al.,[15b] L is the concentration
of [3H]ZM241385 used (nm), Bmax is the total binding (DPM), and I is
the concentration of unlabeled ligand (nm). Fixing these parame-
ters allows the following parameters to be calculated: k3, which is
the kon value (nmꢁ1·minꢁ1) of the unlabeled ligand, and k4, which is
the koff value (minꢁ1) of the unlabeled ligand. Molecular property
descriptors (Mr, logP, MSA, pKa) of the substituted C2-phenylpipera-
zine were calculated using MarinSketch 5.11 (ChemAxon, Hungary).
(Multiple) Linear regression analysis was done using Microsoft
Excel 2003.
Smits, X. Jin, D. Phadke, T. Engber, C. Huang, J. Reilly, S. Tam, D. Grant,
G. Hetu, L. Chen, J. Zhang, R. C. Petter, J. Med. Chem. 2004, 47, 4291–
4299; c) C. B. Vu, P. Shields, B. Peng, G. Kumaravel, X. Jin, D. Phadke, J.
Wang, T. Engber, E. Ayyub, R. C. Petter, Bioorg. Med. Chem. Lett. 2004,
14, 4835–4838.
[8] a) A. V. Dolzhenko, A. V. Dolzhenko, W. K. Chui, Tetrahedron 2007, 63,
12888–12895; b) A. V. Dolzhenko, G. Pastorin, A. V. Dolzhenko, W. K.
Chui, Tetrahedron Lett. 2009, 50, 2124–2128; c) M. Jçrg, M. Agostino, E.
Yuriev, F. S. Mak, N. D. Miller, J. M. White, P. J. Scammells, B. Capuano,
Struct. Chem. 2013, 24, 1241–1251; d) M. Jçrg, J. Shonberg, F. S. Mak,
N. D. Miller, E. Yuriev, P. J. Scammells, B. Capuano, Bioorg. Med. Chem.
Lett. 2013, 23, 3427–3433.
[9] P. W. R. Caulkett, G. Jones, M. McPartlin, N. D. Renshaw, S. K. Stewart, B.
Wright, J. Chem. Soc. Perkin Trans. 1 1995, 801–808.
[10] a) R. E. Tenbrink (Pharmacia & Upjohn Company, MI, USA), US Patent
No. US5912246, 1999; b) G. E. Martin, R. J. Elgin, Jr., J. R. Mathiasen, C. B.
Davis, J. M. Kesslick, W. J. Baldy, R. P. Shank, D. L. DiStefano, C. L. Fedde,
M. K. Scott, J. Med. Chem. 1989, 32, 1052–1056; c) S. Morita, K. Kitano,
J. Matsubara, T. Ohtani, Y. Kawano, K. Otsubo, M. Uchida, Tetrahedron
1998, 54, 4811–4818.
[11] H. J. Motulsky, L. C. Mahan, Mol. Pharmacol. 1984, 25, 1–9.
[12] a) G. Vauquelin, I. Van Liefde, P. Vanderheyden, Trends Pharmacol. Sci.
2002, 23, 514–518; b) T. Kenakin, S. Jenkinson, C. Watson, J. Pharmacol.
Exp. Ther. 2006, 319, 710–723; c) G. Vauquelin, A. Szczuka, Neurochem.
Int. 2007, 51, 254–260.
Acknowledgements
The authors thank Dr. Ron Dror and Dr. Albert Pan (D. E. Shaw
Research, USA) for fruitful discussions, and Dr. Julien Louvel and
Mr. Maris Vilums for helpful comments on the manuscript. This
project was supported financially by the Innovational Research
Incentive Scheme of the Netherlands Research Organization
(NWO; VENI-Grant 11188 to L.H.).
[13] a) D. A. Sykes, M. R. Dowling, S. J. Charlton, Mol. Pharmacol. 2009, 76,
543–551; b) D. Guo, T. Mulder-Krieger, A. P. IJzerman, L. H. Heitman, Br.
J. Pharmacol. 2012, 166, 1846–1859.
Keywords: binding kinetics · G protein-coupled receptors ·
[14] a) S. J. Mantell, P. T. Stephenson, S. M. Monaghan, G. N. Maw, M. A. Tre-
vethick, M. Yeadon, R. F. Keir, D. K. Walker, R. M. Jones, M. D. Selby, D. V.
Batchelor, S. Rozze, H. Chavaroche, T. J. Hobson, P. G. Dodd, A. Lemaitre,
K. N. Wright, E. F. Stuart, Bioorg. Med. Chem. Lett. 2008, 18, 1284–1287;
b) S. J. Mantell, P. T. Stephenson, S. M. Monaghan, G. N. Maw, M. A. Tre-
vethick, M. Yeadon, D. K. Walker, M. D. Selby, D. V. Batchelor, S. Rozze, H.
Chavaroche, A. Lemaitre, K. N. Wright, L. Whitlock, E. F. Stuart, P. A.
Wright, F. Macintyre, Bioorg. Med. Chem. Lett. 2009, 19, 4471–4475.
[15] K. Andersson, R. Karlsson, S. Lofas, G. Franklin, M. D. Hamalainen, Expert
Opin. Drug Discovery 2006, 1, 439–446.
[16] J. Gabrielsson, H. Dolgos, P. G. Gillberg, U. Bredberg, B. Benthem, G.
Duker, Drug Discovery Today 2009, 14, 358–372.
[17] W. Liu, E. Chun, A. A. Thompson, P. Chubukov, F. Xu, V. Katritch, G. W.
Han, C. B. Roth, L. H. Heitman, A. P. IJzerman, V. Cherezov, R. C. Stevens,
Science 2012, 337, 232–236.
antagonism
· structure–activity relationships · structure–
kinetics relationships
[1] a) R. Zhang, F. Monsma, Expert Opin. Drug Discovery 2010, 5, 1023–
1029; b) R. A. Copeland, D. L. Pompliano, T. D. Meek, Nat. Rev. Drug Dis-
covery 2006, 5, 730–739; c) D. C. Swinney, Nat. Rev. Drug Discovery
2004, 3, 801–808; d) D. C. Swinney, Lett. Drug Des. Discovery 2006, 3,
569–574; e) S. Nfflnez, J. Venhorst, C. G. Kruse, Drug Discovery Today
2012, 17, 10–22.
[2] a) E. M. Rosethorne, R. J. Turner, R. A. Fairhurst, S. J. Charlton, Naunyn-
Schmiedeberg’s Arch. Pharmacol. 2010, 382, 255–263; b) C. C. Govern,
M. K. Paczosa, A. K. Chakraborty, E. S. Huseby, Proc. Natl. Acad. Sci. USA
2010, 107, 8724–8729; c) J. Foote, H. N. Eisen, Proc. Natl. Acad. Sci. USA
1995, 92, 1254–1256; d) G. Schreiber, Curr. Opin. Struct. Biol. 2002, 12,
41–47; e) T. Selzer, S. Albeck, G. Schreiber, Nat. Struct. Biol. 2000, 7,
537–541; f) D. A. Sykes, M. R. Dowling, J. Leighton-Davies, T. C. Kent, E.
Renard, A. Trifilieff, S. J. Charlton, J. Pharmacol. Exp. Ther. 2012, 343,
520–528.
[18] F. Xu, H. Wu, V. Katritch, G. W. Han, K. A. Jacobson, Z. G. Gao, V. Chere-
zov, R. C. Stevens, Science 2011, 332, 322–327.
[19] R. O. Dror, A. C. Pan, D. H. Arlow, D. W. Borhani, P. Maragakis, Y. Shan, H.
Xu, D. E. Shaw, Proc. Natl. Acad. Sci. USA 2011, 108, 13118–13123.
[20] A. C. Kruse, J. Hu, A. C. Pan, D. H. Arlow, D. M. Rosenbaum, E. Rose-
mond, H. F. Green, T. Liu, P. S. Chae, R. O. Dror, D. E. Shaw, W. I. Weis, J.
Wess, B. K. Kobilka, Nature 2012, 482, 552–556.
[21] D. Guo, E. J. van Dorp, T. Mulder-Krieger, J. P. van Veldhoven, J. Brussee,
A. P. IJzerman, L. H. Heitman, J. Biomol. Screening 2013, 18, 309–320.
[22] Y. Cheng, W. H. Prusoff, Biochem. Pharmacol. 1973, 22, 3099–3108.
[23] E. Ongini, S. Dionisotti, S. Gessi, E. Irenius, B. B. Fredholm, Naunyn-
Schmiedeberg’s Arch. Pharmacol. 1999, 359, 7–10.
[3] B. B. Fredholm, A. P. IJzerman, K. A. Jacobson, J. Linden, C. E. Müller,
Pharmacol. Rev. 2011, 63, 1–34.
[4] K. A. Jacobson, Z. G. Gao, Nat. Rev. Drug Discovery 2006, 5, 247–264.
[5] a) B. C. Shook, P. F. Jackson, ACS Chem. Neurosci. 2011, 2, 555–567;
b) C. E. Müller, K. A. Jacobson, Biochim. Biophys. Acta Biomembr. 2011,
1808, 1290–1308; c) U. Shah, R. Hodgson, Curr. Opin. Drug Discov. Devel.
2010, 13, 466–480.
[6] a) S. M. Poucher, J. R. Keddie, P. Singh, S. M. Stoggall, P. W. Caulkett, G.
Jones, M. G. Coll, Br. J. Pharmacol. 1995, 115, 1096–1102; b) P. W. R.
Caulkett, G. Jones, M. G. Collis, S. M. Poucher, (Imperial Chemical Indus-
try PLC, London, UK), European Patent No. EP459702, 1991.
[7] a) C. B. Vu, D. Pan, B. Peng, G. Kumaravel, G. Smits, X. Jin, D. Phadke, T.
Engber, C. Huang, J. Reilly, S. Tam, D. Grant, G. Hetu, R. C. Petter, J. Med.
Chem. 2005, 48, 2009–2018; b) C. B. Vu, B. Peng, G. Kumaravel, G.
Received: November 21, 2013
Revised: January 27, 2014
Published online on March 3, 2014
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 752 – 761 761