Development of novel adenosine receptor ligands based on the 3-amidocoumarin scaffold

Article abstract of DOI:10.1016/j.bioorg.2015.05.008

With the aim of finding new adenosine receptor (AR) ligands presenting the 3-amidocoumarin scaffold, a study focusing on the discovery of new chemical entities was carried out. The synthesized compounds 1-8 were evaluated in radioligand binding (A₁

Full text of DOI:10.1016/j.bioorg.2015.05.008

Bioorganic Chemistry 61 (2015) 1–6
Contents lists available at ScienceDirect
Bioorganic Chemistry
journal homepage: www.elsevier.com/locate/bioorg
Development of novel adenosine receptor ligands based
on the 3-amidocoumarin scaffold
Maria J. Matos a, , Santiago Vilar b,c, Sonja Kachler d, Maria Celeiro b, Saleta Vazquez-Rodriguez a,
⇑
Lourdes Santana b, Eugenio Uriarte b, George Hripcsak c, Fernanda Borges a, , Karl-Norbert Klotz d
⇑
a CIQUP, Departamento de Química e Bioquímica, Faculdade de Ciências, Universidade do Porto, 4169-007 Porto, Portugal
b Departamento de Química Orgánica, Facultad de Farmacia, Universidad de Santiago de Compostela, 15782 Santiago de Compostela, Spain
c Department of Biomedical Informatics, Columbia University Medical Center, New York, NY 10032, USA
d Institut für Pharmakologie und Toxikologie, Universität Würzburg, 97078 Würzburg, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 12 February 2015
Available online 22 May 2015
With the aim of finding new adenosine receptor (AR) ligands presenting the 3-amidocoumarin scaffold, a
study focusing on the discovery of new chemical entities was carried out. The synthesized compounds
1–8 were evaluated in radioligand binding (A1, A2A and A3) and adenylyl cyclase activity (A2B) assays
in order to determine their affinity for human AR subtypes. The 3-benzamide derivative 4 showed the
Keywords:
3-Amidocoumarins
Adenosine ligands
Theoretical ADME properties
Molecular modeling calculations
highest affinity of the whole series and was more than 30-fold selective for the A3 AR (Ki = 3.24 lM).
The current study supported that small structural changes in this scaffold allowed modulating the affinity
resulting in novel promising classes of A1, A2A, and/or A3 AR ligands. We also performed docking
calculations in hA2A and hA3 to identify the hypothetical binding mode for the most active compounds.
In addition, some ADME properties were calculated in order to better understand the potential of these
compounds as drug candidates.
Ó 2015 Elsevier Inc. All rights reserved.
1. Introduction
treatment of cancer [8]. Finally, A3 AR ligands have been linked
to inflammatory diseases, such as rheumatoid arthritis and psoria-
Adenosine
receptors
(ARs)
are
members
of
the
sis, liver cancer, neurodegeneration, hepatitis, asthma and to the
protective effects in cardiac ischemia and liver regeneration [4].
In the last years, the search for selective and potent ligands toward
individual AR subtypes has been intensified, as the role of these AR
in many therapeutic areas is continuously expanding [4–6].
Coumarins are an important family of compounds extensively
studied and described due to their significant role in medicinal
chemistry [9,10]. The simplicity of the chemical framework, the
synthetic accessibility and substitution variability of these hetero-
cyclic compounds make them relevant molecules with different
properties such as anti-cancer, antiviral, anti-inflammatory,
antimicrobial, enzyme-inhibition and antioxidant [11]. Our
research group has described different coumarins as potential AR
ligands [12–14]. In particular, some amide and carbamate deriva-
tives were studied [14]. The activity profile depicted by some pre-
viously described aliphatic amidocoumarins was the basis for this
new study [14]. To further explore the importance of this frame-
work as the basis for AR ligands, a selected series of coumarin
G-protein-coupled receptor family and represent important phar-
macological targets in the treatment of a variety of diseases in
which different stressful processes, such as inflammation, hypoxia,
ischemia or trauma, are involved [1–3]. Their presence on almost
every cell in different organs makes them an interesting target
for the pharmacological intervention in many pathophysiological
situations [4]. Adenosine plays its many roles through binding to
one or more of the four AR subtypes A1, A2A, A2B and A3 [5]. All
the AR subtypes are closely related to specific biochemical pro-
cesses, therefore they are involved in different pathologies. A1
ligands are being investigated as promising agents in the therapy
of cardiovascular diseases and pain [6]. A2A AR ligands are impor-
tant compounds for the treatment of neurodegenerative diseases,
specifically Alzheimer’s and Parkinson’s diseases [7]. A2B antago-
nists and dual A2B/A3 antagonists are under development due to
the role they play in asthma and diabetes and selective adenosine
A3 agonists and antagonists are under consideration for the
derivatives bearing an
a-b unsaturated bond (aliphatic, aromatic
or heteroaromatic) attached to the amide group was synthesized,
purified and characterized. Pharmacological evaluation, theoretical
⇑
Corresponding authors.
E-mail addresses: mariacmatos@gmail.com (M.J. Matos), mfernandamborges@
gmail.com (F. Borges).
http://dx.doi.org/10.1016/j.bioorg.2015.05.008
0045-2068/Ó 2015 Elsevier Inc. All rights reserved.

2
M.J. Matos et al. / Bioorganic Chemistry 61 (2015) 1–6
evaluation of ADME properties, docking calculations and SAR stud-
ies of the new series of the amidocoumarins 1–8 were also carried
out. To deeply study this family of compounds, and to complete the
previous results [14], with the compounds described in this manu-
script we also performed docking calculations to better understand
the activity and selectivity against the four AR.
receptors. The results were expressed as Ki values (dissociation
constants), which were calculated with the program SCTFIT [19].
Due to the lack of a suitable radioligand for hA2B receptors, the
potency of antagonists at the hA2B receptor (expressed on CHO
cells) was determined by inhibition of NECA-stimulated adenylyl
cyclase activity. The 50% inhibitory concentration (IC50) for inhibi-
tion of cAMP (cyclic adenosine monophosphate) production was
determined and converted to a Ki value using the Cheng and
Prusoff equation. Ki values (Table 1) are reported as geometric
means of three independent experiments with each tested concen-
tration of compound measured in duplicate. As an interval esti-
mate for the dissociation constants, 95% confidence intervals are
given in parentheses. Details for pharmacological experiments
are described in previous work [13].
2. Materials and methods
2.1. Chemistry
Starting materials and reagents were obtained from commercial
suppliers and were used without further purification (Sigma–
Aldrich). Melting points (mp) are uncorrected and were deter-
mined with a Reichert Kofler thermopan or in capillary tubes in a
Büchi 510 apparatus. 1H NMR (300 MHz) and 13C NMR
(75.4 MHz) spectra were recorded with a Bruker AMX spectrome-
ter using CDCl3 as solvent. Chemical shifts (d) are expressed in ppm
using TMS as an internal standard. Coupling constants (J) are
expressed in Hz. Spin multiplicities are given as s (singlet), dd
(doublet of doublets), td (triplet of doublets) and m (multiplet).
2.3. Docking calculations
We used the Molecular Operating Environment (MOE) software
[20] and the Schrödinger package [21] to carry out homology mod-
eling of the hA3 AR and molecular docking simulations,
respectively.
Mass
spectrometry
was
carried
out
with
a
Glide SP [22] molecular docking simulations were carried out
for the most active compounds for the hA2A and hA3 receptors to
rationalize the selectivity shown by the compounds. To run the cal-
culations we used the hA2A crystal structure (PDB: 3EML) [23] and
a homology model for the hA3. The protein pocket structure and
the hypothetical binding modes for the compounds were opti-
mized through MM–GBSA in Prime [24]. Using this protocol
RMSD values of 0.69 and 1.90 between the calculated and the
co-crystallographic poses of the ligands in the 3EML [23] and
3UZC [25] hA2A receptors were obtained.
Hewlett-Packard-5972-MSD spectrometer. Elemental analyses
were performed with a Perkin Elmer 240B microanalyzer and are
within 0.4% of calculated values in all cases. Flash chromatography
(FC) was performed on silica gel (Merck 60, 230–400 mesh); ana-
lytical TLC was performed on pre-coated silica gel plates (Merck
60 F254). Organic solutions were dried over anhydrous Na2SO4.
Concentration and evaporation of the solvent after reaction or
extraction was carried out on
Rotavapor) operating at reduced pressure. The analytical results
showed >95% purity for all compounds.
a rotary evaporator (Büchi
2.3.1. Homology modeling
2.1.1. Preparation of the precursor 3-amino-4-hydroxycoumarin
The commercially available 4-hydroxy-3-nitrocoumarin
(Sigma–Aldrich) (2.5 mmol) was dissolved in ethanol and a cat-
alytic amount of Pd/C was added to the mixture. The solution
was stirred, at room temperature, under H2 atmosphere, for 5 h.
After the completion of the reaction, the mixture was filtered to
eliminate the catalyst. The obtained crude product was then puri-
fied by FC (hexane/ethyl acetate, 9:1) to give the desired coumarin,
in a yield of 90%.
We used the hA2A crystallized structure (PDB: 3EML) [23] as a
template to generate the models. We followed the same protein
alignment as described by Katritch et al. [26], considering the most
conserved residues in the TM helices. We assessed the protein
geometry taking into account Phi–Psi angles and Ramachandran
plots, bond lengths, bond angles, dihedrals, side chain rotamers,
and non-bonded contacts. We docked high affinity ligands to the
hA3 homology model through the Induce Fit Docking Workflow
[21] to optimize the protein pocket. Different protein pocket con-
formations were evaluated for their ability to discriminate: (1) true
2.1.2. General procedure for the preparation of 3-amidocoumarins 1–8
The
3-aminocoumarin
(Sigma–Aldrich)
or
Table 1
3-amino-4-hydroxycoumarin (1 mmol) was dissolved in dichloro-
methane (9 mL), then pyridine (1.1 mmol) was added and the mix-
ture was cooled to 0 °C. Differently substituted acid chloride
(Sigma–Aldrich) (1.1 mmol) was added drop-wise at this tempera-
ture, and the mixture was stirred overnight at room temperature.
The batch was evaporated and purified by column chromatography
(hexane/ethyl acetate, 9:1) to give the desired compounds 1–8
[15–18].
Binding affinity of coumarins 1–8 for human A1, A2A and A3 ARs expressed in CHO
cells.
Comp.
hA1
hA2A
Ki (
hA3
Selectivity
Ki (l
M)
l
M)
Ki (l
M)
hA1/hA3 hA2A/hA3
1 [14]
53.9
(35.9–81.1)
8.95
>100
7.16
(5.70–9.00)
>100
7.5
>14
2
12.2
<0.09
<0.12
(5.60–14.3) (8.98–16.7)
3
4
>100
>100
>100
>100
>100
3.24
–
>31
–
>31
2.2. Pharmacology
(2.85–3.69)
>100
>100
5
6
7
>100
>100
16.2
(8.68–30.2)
5.18
>100
>100
>100
–
–
–
–
–
The affinity of compounds 1–8 for the human AR subtypes hA1,
hA2A, hA3, was determined with radioligand competition experi-
ments in Chinese hamster ovary (CHO) cells that were stably
transfected with the individual receptor subtypes. The radioligands
used were 1 nM (2R,3R,4S,5R)-2-(2-chloro-6-cyclopentylamino-p
>60
<0.27
8
>100
9.79
(7.66–12.5)
86.40
0.53
0.08
>10
(3.44–7.79)
urin-9-yl)-5-hydroxymethyl-tetrahydro-3,4-diol
for hA1, 10 nM (1-(6-amino-9H-purin-9-yl)-1-deoxy-N-ethyl-b
-D-ribofuronamide) for hA2A and 1 nM
([3H]NECA)
2-(1-hexynyl)-N6-methyladenosine [3H] ([3H]HEMADO) for hA3
([3H]CCPA)
Theophylline 6.77
[13]
>1.71
>1.2
(4.07–11.30) (1.02–2.90) (73.60–101.30)
Values are geometric means of three experiments and given with 95% confidence
intervals in parentheses.
,

M.J. Matos et al. / Bioorganic Chemistry 61 (2015) 1–6
3
ligands from decoys, and (2) different sets of sub-type selective
high affinity compounds. We used ROC curves to assess perfor-
mance in the tests. The best homology model showed an area
under the ROC curve (AUROC) for test 1 of 0.92 (22 hA3 true posi-
tive ligands collected in Katritch et al. [26] and 200 random
decoys) and for test 2 of 0.82 (22 hA3 true positives and 22
hA2A + 22 hA1 compounds as false positives) [26]. The best hA3
models were retained for further docking studies.
prepared by
a
reduction of the commercially available
3-nitro-4-hydroxycoumarin, in ethanol, with Pd/C as catalyst, in
H2 atmosphere, with a yield of 90% [13]. An acylation reaction of
the 3-aminocoumarins with the conveniently substituted acid
chloride, using pyridine in dichloromethane, from 0 °C to
room temperature, afforded the differently substituted
3-amidocoumarins (1–8) in yields between 80% and 90% [14–18].
The reaction conditions and chemical characterization of the new
compounds are detailed in methods.
N-(Coumarin-3-yl)acrylamide (2) Yield: 84%. mp: 168–169 °C.
1H NMR (300 MHz, CDCl3): 5.90 (dd, 1H, CH, J = 9.9, J = 1.5 Hz), 6.35
(dd, 1H, CH2, J = 16.9, J = 9.9 Hz), 6.51 (dd, 1H, CH2, J = 16.9,
J = 1.5 Hz), 7.30–7.38 (m, 2H, H-6, H-8), 7.46–7.57 (m, 2H, H-5,
H-7), 8.28 (s, 1H, NH), 8.82 (s, 1H, H-4). 13C NMR (75 MHz,
DMSO-d6) d (ppm): 111.6, 115.1, 119.1, 119.2, 120.5, 123.2,
124.3, 125.1, 125.7, 145.3, 154.1, 159.6. MS m/z (%): 216 (12),
215 (M+, 80), 161 (97), 133 (27), 104 (11), 77 (22), 55 (100).
Anal. Elem. Calc. for C12H9NO3: C, 66.97; H, 4.22. Found: C, 66.98;
H, 4.25.
N-(4-Hydroxycoumarin-3-yl)acrylamide (3) Yield: 80%. mp:
159–160 °C 1H NMR (300 MHz, CDCl3): 6.01 (dd, 1H, CH, J = 9.9,
J = 1.3 Hz), 6.45 (dd, 1H, CH2, J = 16.7, J = 9.9 Hz), 6.60 (dd, 1H,
CH2, J = 16.7, J = 1.3 Hz), 7.36–7.43 (m, 2H, H-6, H-8), 7.58 (td,
1H, H-7, J = 7.5, J = 1.6 Hz), 8.04 (dd, 1H, H-5, J = 7.9, J = 1.6 Hz),
8.31 (s, 1H, NH), 13.95 (s, 1H, OH). 13C NMR (75 MHz, DMSO-d6):
102.5, 115.5, 117.0, 119.8, 124.9, 125.1, 126.1, 130.4, 151.3,
160.0, 164.3, 167.5. MS m/z (%): 232 (9), 231 (M+, 44), 177 (71),
148 (11), 121 (40), 65 (16). Anal. Elem. Calc. for C12H9NO4: C,
62.34; H, 3.92. Found: C, 62.36; H, 3.95.
2.4. Theoretical evaluation of absorption, distribution, metabolism and
excretion properties
The absorption, distribution, metabolism and excretion (ADME)
properties of the studied compounds were calculated using the
Molinspiration property programme. LogP was calculated using
the methodology developed by Molinspiration [27] as a sum of
fragment-based contributions and correction factors. Topological
polar surface area (TPSA) was calculated based on the methodology
published by Ertl et al. as a sum of fragment contributions [28].
Oxygenand nitrogen-centered polar fragments were considered.
Polar surface area (PSA) has been shown to be a very good descrip-
tor characterizing drug absorption, including intestinal absorption,
bioavailability, Caco-2 permeability and blood–brain barrier pene-
tration. The method for calculation of molecule volume developed
at Molinspiration is based on group contributions. These have been
obtained by fitting the sum of fragment contributions to ‘real’
threedimensional (3D) volume for a training set of about 12,000,
mostly drug-like molecules. 3D molecular geometries for a training
set were fully optimized by the semi-empirical AM1 method.
N-(Coumarin-3-yl)furan-2-carboxamide (8) Yield: 90%. mp:
1
183–184 °C. H NMR (300 MHz, DMSO-d6): 6.74 (dd, 1H, H-40,
3. Results and discussion
J = 3.6, J = 1.8 Hz), 7.34–7.58 (m, 4H, H-5, H-6, H-8, H-50), 7.77
(td, 1H, H-7, J = 8.0, J = 1.4 Hz), 8.00 (dd, 1H, H-30, J = 1.8,
J = 0.8 Hz), 8.58 (s, 1H, H-4), 9.26 (s, 1H, NH). 13C NMR
(75 MHz, DMSO-d6): 112.1, 113.2, 115.7, 116.8, 122.9, 124.3,
126.2, 126.8, 127.7, 145.8, 146.5, 149.5, 155.3, 159.8. MS m/z
(%): 256 (16), 255 (M+, 79), 227 (7), 132 (6), 95 (100), 77 (10).
Anal. Elem. Calc. for C14H9NO4: C, 65.88; H, 3.55. Found: C,
65.86; H, 3.53.
3.1. Chemistry
The described derivatives were efficiently synthesized
according to the protocol outlined in Scheme 1. Coumarins 1–8
were prepared starting from 3-aminocoumarin or from
3-amino-4-hydroxycoumarin, which were synthesized as previ-
ously described [13,15]. The 3-amino-4-hydroxycoumarin was
Scheme 1. Reagents and conditions: (a) H2, EtOH, Pd/C, r.t., 5 h; (b) R1COCl, pyridine, dichloromethane, 0 °C to r.t., overnight.

4
M.J. Matos et al. / Bioorganic Chemistry 61 (2015) 1–6
3.2. Pharmacological study
in the hA3 showed an energy contribution in the interaction with
different residues distinct from the mode of interaction with the
hA2A AR (Fig. 1b). This result is in accordance with the lack of affin-
ity of compound 2 for the hA3 AR. The interaction scores are calcu-
lated as the sum of Coulomb, van der Waals and hydrogen bonding
energies. Fig. 1b also shows the importance of the residue Phe168
in ligand recognition.
The affinity of the synthesized 3-amidocoumarins for A1, A2A
,
and A3 ARs was tested in radioligand binding assays. The affinity
for the A2B AR was determined in a functional assay (inhibition
of agonist-stimulated adenylyl cyclase activity) [29,30]. The
detailed methodology is described in methods. The binding data
for A1, A2A and A3 ARs are shown in Table 1. None of the derivatives
Compound 4, the most potent A3 AR ligand in the series, showed
a hypothetical binding mode similar to compound 2. The com-
pound docked to the hA3 AR orientated the benzamide substituent
toward the upper region of the cavity and the coumarin ring is
located in the bottom of the pocket (Fig. 2a). The phenyl group of
the compound is located close to the hydrophobic residues
Val169 and Leu264 (residues not present in the hA2A).
Lipophilicity and steric size of the substituent seem to be impor-
tant factors for hA3AR affinity. Compound 4 also establishes hydro-
gen bond interactions with the residue Asn250. Previous results
performed with other scaffolds at the hA3 AR [32,33] showed
similar ligand poses as reported in this study. Although a similar
binding mode was found for compound 4 in the hA2A AR, the com-
pound is placed slightly shifted toward the upper region causing a
disruption in the interaction energy with the residue Asn250
(Fig. 2b and c). Other possible binding modes of compound 4
extracted from docking studies for the hA2A did not show relevant
interactions with the cited residue. Additional information on the
selectivity between hA2A and hA3 was also acquired by calculating
the residue energy contributions to the interaction with compound
4 (Fig. 2c).
showed measurable affinity for the A2B AR (Ki > 20 lM).
3.3. Theoretical ADME properties
The prediction of the ADME properties plays an important role
in the drug design process because these properties account for the
failure of about 60% of all drugs in the clinical phases. Therefore,
preliminary data for theoretical ADME profiles of the newly
synthesized 3-amidocoumarins, were determined (Table 2). The
lipophilicity, expressed as the octanol/water partition coefficient
(represented as logP), was also calculated using the
Molinspiration property calculation program [27,28,31].
From the data obtained, one can notice that all the coumarin
derivatives possess logP values compatible with those required
to cross membranes. From the data obtained from the prediction
of ADME properties it can be observed that no violations of
Lipinski’s rule (molecular weight, logP, number of hydrogen
donors and acceptors) were found making the coumarin deriva-
tives promising agents. Topological polar surface area (TPSA),
described to be a predictive indicator of membrane penetration,
is also found to be positive for these potential drugs.
3.5. Discussion
3.4. Docking calculations
A novel series of coumarin derivatives presenting on their struc-
ture a common planar NAC@O framework, represented by an
Since hA2A and hA3 AR seems to be involved in neurodegenera-
tive pathologies and neuroprotective processes [4,7], in which our
group is interested, docking calculations were performed in order
to understand structural features related to the activity against
these receptors. Compound 2 (bearing an aliphatic amide) as the
only derivative with affinity for the hA2A receptor bound with the
coumarin ring orientated toward the bottom of the A2A receptor
cavity and the acrylamide chain pointing toward the upper region
close to the extracellular loops. The carbonyl oxygen in the cou-
marin ring establishes a hydrogen bond interaction with the amide
moiety of the residue Asn253 (Fig. 1a). This hypothetical binding
mode agrees with previous results showing the importance of
the residue Asn6.55 (BallesterosꢀWeinstein numbering) in ligand
recognition [23,32,33]. However, compound 2 does not establish
strong interaction in the hA3 with the corresponding residue
Asn250. Molecular docking showed a hypothetical binding mode
placed in the extracellular area and establishing a possible hydro-
gen bond with the residue Gln167. This hypothetical binding mode
Table 2
Structural properties of the coumarin derivatives 1–8.a
Comp. logP Molecular weight TPSA
n-OH
acceptors
n-OHNH
donors
Volume
(Å3)
(g/mol)
(Å2)
1
2
3
4
5
6
7
8
1.16 203.20
1.76 215.21
1.47 231.21
2.83 265.27
2.54 281.27
1.60 266.26
2.73 271.30
2.09 255.23
59.31
59.31
79.54
59.31
79.54
72.21
59.31
72.45
4
4
5
4
5
5
4
5
1
1
2
1
2
1
1
1
176.53
187.70
195.72
231.38
239.40
227.22
222.09
212.95
Fig. 1. (a) Hypothetical binding mode for compound 2 (green carbons) calculated
through docking to the hA2A AR. Important residues in the protein–ligand
interaction are shown in tube (gray carbons). Hydrogen bond interaction between
compound 2 and residue Asn253 is colored in yellow. (b) Residue interaction scores
(sum of Coulomb, van der Waals and hydrogen bonding interactions) for compound
2 in the hA2A and the hA3 AR.
a
TPSA, topological polar surface area; n-OH, number of hydrogen bond accep-
tors; n-OHNH, number of hydrogen bond donors. The data was determined with
Molinspiration calculation software [27].

M.J. Matos et al. / Bioorganic Chemistry 61 (2015) 1–6
5
this structure (compound 2) improved the A1 (Ki = 8.95
A2A (Ki = 12.2
l
M) and
l
M) affinity, with a concomitant loss of affinity for
the hA3 AR. Compound 2 was the only derivative of the series with
affinity for the hA2A AR. The introduction of an acryl group on the
amidic scaffold could help on the design of hA2A AR ligands.
Comparing compound 1 with the corresponding benzamide
derivative 4 reveals a high increase on the A3 affinity of compound
4 (Ki = 3.24 lM). In addition, a lack of measurable A1 affinity made
compound 4 the most A3 selective compound of this series. The
replacement of the benzene ring by a pyridine substituent (com-
pound 6) led to loss of affinity for all ARs. On the other hand, the
substitution by a thiophene ring at the same position (compound
7) resulted in a selective hA1 AR ligand (Ki = 16.2
introduction of a furyl ring in place of the thiophene (compound
8) improved the hA1 AR affinity (Ki = 5.18 M) with a concurrent
appearance of a similar A3 affinity (Ki AR = 9.79 M). It is interest-
lM). Finally, the
l
l
ing to note that the affinity for hA1 and hA3 ARs can be modulated
by the choice of heteroatom in the aromatic ring in this position.
The results found for compounds 4,
7 and 8 suggest that
3-amidocoumarin is a promising scaffold to develop ligands with
improved A1, A3, or dual A1/A3 affinity and selectivity. Finally, the
results found for compounds 2, suggest that 3-amidocoumarin is
also a promising scaffold to develop A2A ligands.
All adenosine receptor agonists known to date are adenosine
derivatives with the notable exception of
a
series of
2-aminodicyanopyridine derivatives [6]. We confirmed the antag-
onistic nature of compound 7 at the A1 adenosine receptor in a
GTP-shift experiment (data not shown).
Compounds 1, 4 and 8 present better hA3 affinity than theo-
phylline, used as reference compound. In addition, compound 4
is hA3 selective ligand, being 27 times more active against hA3 than
the reference compound. Regarding hA1 affinity, compounds 2 and
8 present similar Ki to theophylline.
Our research group has previously described different coumar-
ins as potential AR ligands [12–14]. In particular, the activity pro-
file depicted by some previously described amidocoumarins was
the basis for this new study [14]. However, compound 4 of this
new series presented with two fold higher affinity for hA3 AR com-
pared to the best compound of the previous study.
To deeply study the new family of compounds, docking calcula-
tions were now performed to better understand the activity and
selectivity against the four AR. Docking results showed a different
hypothetical binding mode for compound 2 into the hA2A and the
hA3 AR. Different energy contributions of individual residues seem
to be responsible for the observed AR selectivity. Moreover, in the
case of compound 4, the pose determined through docking pre-
sented a lower energy interaction in the hA2A with the residue
Asn253. The calculations showed a reduction in the electrostatic
energy contribution possibly due to the repulsion between the car-
bonyl oxygen of the derivative 4 and the oxygen of the amide in the
residue Asn253. As described previously, this residue plays an
important role in the interaction with different ligands
[23,32,33]. The disruption of this interaction provides a valid
explanation for the reduction of hA2A activity.
Moreover, different residues located in the extracellular area of
the ligand binding domain could be very important for ligand entry
and stabilization of binding [33], thereby constituting essential attri-
butes for AR selectivity. hA3 bears respective hydrophobic residues,
such as Val169 and Leu264 that can favor interactions with
hydrophobic substituents like the phenyl group in compound 4.
The corresponding residues in hA1, the Glu172 and Thr270, showed
hydrophilic characteristics more suitable for interaction with polar
substituents in the ring. On the other hand, some residues in the
Fig. 2. (a) Binding pose calculated through molecular docking for compound 4
(green carbons) in the hA3. Important residues in the interaction between the ligand
and the protein are shown in tube (gray carbons). Hydrogen bonds are represented
in yellow. (b) Hypothetical binding mode for compound 4 (green carbons) in the
hA2A. (c) Residue contributions in the ligand–protein interaction (sum of Coulomb,
van der Waals and hydrogen bonding interactions) for compound 4 in the hA2A and
the hA3 AR.
amidic group at position 3, were studied for their ability to bind the
four ARs. Different alkyl, aromatic or heteroaromatic groups were
attached to the amidic function at position 3 and the effect of these
substitutions on affinity was studied. In addition, the presence or
the absence of a hydroxyl group at position 4 was also explored.
As according to a previous study [14], none of the 4-hydroxy
derivatives (compounds 3 and 5) displayed binding affinity for
any AR subtype. The presence of a hydroxyl function at position
4 of the coumarin skeleton is not tolerated, notwithstanding the
presence of an aliphatic (compound 3), or an aromatic (compound
5) group at position 3.
The structurally simpler derivative of the series (compound 1)
showed affinity for hA1 and hA3 ARs (Ki = 53.9
lM
and
Ki = 7.16 M, respectively). The introduction of a double bond on
l

6
M.J. Matos et al. / Bioorganic Chemistry 61 (2015) 1–6
hA2A, such as Glu169 with a negative charge and hydrophilic charac-
teristics or the positively charged His264, are not present in the hA3.
References
[1] C. St Hilaire, S. Carroll, H. Chen, K. Ravid, J. Cell. Physiol. 218 (2009) 35–44.
[2] P. Baraldi, M. Tabrizi, S. Gessi, P. Borea, Chem. Rev. 108 (2008) 238–263.
[3] G. Hasko, J. Linden, B. Cronstein, P. Pacher, Nat. Rev. Drug Discov. 7 (2008) 759–
770.
4. Conclusion
[4] B. Koscsó, B. Csóka, P. Pacher, G. Haskó, Expert Opin. Investig. Drugs 20 (2011)
757–768.
[5] N. Press, J. Fozard, Expert Opin. Ther. Pat. 20 (2010) 987–1005.
[6] B. Fredholm, A. Ijzerman, K. Jacobson, J. Linden, C. Müller, Pharmacol. Rev. 63
(2011) 1–34.
[7] F. Areias, M. Costa, M. Castro, J. Brea, E. Gregori-Puigjané, M.F. Proença, J.
Mestres, M.I. Loza, Eur. J. Med. Chem. 54 (2012) 303–310.
[8] C.E. Müller, K.A. Jacobson, Biochim. Biophys. Acta 1808 (2011) 1290–1308.
[9] F. Borges, F. Roleira, N. Milhazes, E. Uriarte, L. Santana, Front. Med. Chem. 4
(2009) 23–85.
[10] M.J. Matos, D. Viña, S. Vazquez-Rodriguez, E. Uriarte, L. Santana, Curr. Top.
Med. Chem. 12 (2012) 2210–2239.
[11] F. Borges, F. Roleira, N. Milhazes, L. Santana, E. Uriarte, Curr. Med. Chem. 12
(2005) 887–916.
[12] S. Vazquez-Rodriguez, M.J. Matos, L. Santana, E. Uriarte, F. Borges, S. Kachler,
K.-N. Klotz, J. Pharm. Pharmacol. 65 (2013) 697–703.
[13] M.J. Matos, V. Hogger, A. Gaspar, S. Kachler, F. Borges, E. Uriarte, L. Santana, K.-
N. Klotz, J. Pharm. Pharmacol. 65 (2013) 1590–1597.
[14] M.J. Matos, A. Gaspar, S. Kachler, K.-N. Klotz, F. Borges, L. Santana, E. Uriarte, J.
Pharm. Pharmacol. 65 (2013) 30–34.
[15] C.S. Barnes, M.I. Strong, J.L. Occolowitz, Tetrahedron 19 (1963) 839–847.
[16] E. Klussmann, W. Rosenthal, J. Rademann, F. Christian, PCT Int. Appl. (2006)
(WO2006122546 A120061123).
[17] V. Maddi, S.N. Mamledesai, D. Satyanarayana, S. Swamy, Ind. J. Pharmacol. Sci.
69 (2007) 847–849.
[18] L.D.S. Yadav, S. Singh, V.K. Rai, Tetrahedron Lett. 50 (19) (2009) 2208–2212.
[19] A. De Lean, A.A. Hancock, R. Lefkowitz, J. Mol. Pharmacol. 21 (1982) 5–16.
[20] MOE, version 2011.10, Chemical Computing Group Inc., <www.
chemcomp.com>.
[21] Schrödinger Suite, Schrödinger, LLC, New York, NY, USA, 2011
[22] Glide, version 5.7, Schrödinger, LLC, New York, NY, 2011.
[23] V.P. Jaakola, M.T. Griffith, M.A. Hanson, V. Cherezov, E.Y. Chien, J.R. Lane, A.P.
Ijzerman, R.C. Stevens, Science 322 (2008) 1211–1217.
[24] Prime, version 3.0, Schrödinger, LLC, New York, NY, 2011.
[25] M. Congreve, S.P. Andrews, A.S. Doré, K. Hollenstein, E. Hurrell, C.J. Langmead,
J.S. Mason, I.W. Ng, B. Tehan, A. Zhukov, M. Weir, F.H. Marshall, J. Med. Chem.
55 (2012) 1898–1903.
The 3-amidocoumarin scaffold proved to have potential for the
design of novel AR ligands with distinct selectivity profiles for ARs.
A detailed analysis of the results obtained so far allowed to con-
clude that the affinity and/or selectivity of the coumarins toward
ARs can be modulated by the nature of the substituents attached
to the amidic linker at position 3 of the scaffold. Compound 4 (phe-
nyl derivative) proved to be the best compound of the series, being
considered as a starting point for the design and synthesis of new
coumarins as hA3 AR selective ligands. The profile of compound 8
(furyl derivative) suggests this derivative as a lead for ligands with
dual hA1/hA3 AR selectivity. Compound 7 (thiophenyl derivative)
proved to be the only selective hA1 compound of the studied series,
being the inspiration for the design of A1 selective ligands. The sim-
plicity of the synthetic processes and the decoration capability of
the 3-amidocoumarins make them a privileged structure for the
development of novel AR ligands. The theoretical evaluation of
ADME properties confirms the role of these compounds as promis-
ing hits. Additionally, molecular docking simulations have been
supportive to explain the selectivity of the most potent ligands
for hA2A and hA3 ARs.
Acknowledgments
This work was partially supported by University of Santiago de
Compostela and Spanish researchers personal’s founds. MJM grate-
fully acknowledges Fundação para a Ciência e Tecnologia, POPH
(Programa Operacional Potencial Humano) and QREN (Quadro de
Referência Estratégica Nacional) for her postdoctoral Grant
(SFRH/BPD/95345/2013). SV-R gratefully acknowledges the
Universidade do Porto for her postdoctoral fellowship associated
to the QREN FCUP-CIQ-UP-NORTE-07-0124-FEDER-000065 project.
SV thanks Plan Galego de Investigación, Innovación e Crecemento
2011–2015 (I2C), European Social Fund (ESF) and Angeles
Alvariño Program from Xunta de Galicia (Spain).
[26] V. Katritch, I. Kufareva, R. Abagyan, Neuropharmacology 60 (2011) 108–115.
[27] M
www.molinspiration.com/services/properties.html>.
[28] P. Ertl, B. Rohde, P. Selzer, J. Med. Chem. 43 (2000) 3714–3717.
cheminformatics,
Bratislava,
Slovak
Republic,
<http://
[29] K.-N. Klotz, J. Hessling, J. Hegler, C. Owman, B. Kull, B.B. Fredholm, M.J. Lohse,
Naunyn Schmiedebergs Arch. Pharmacol. 357 (1997) 1–9.
[30] K.-N. Klotz, N. Falgner, S. Kachler, C. Lambertucci, S. Vittori, R. Volpini, G.
Cristalli, Eur. J. Pharmacol. 556 (2007) 14–18.
[31] C.A. Lipinski, F. Lombardo, B.W. Dominy, P.J. Feeney, Adv. Drug Deliv. Rev. 23
(1997) 3–26.
[32] A. Gaspar, J. Reis, S. Kachler, S. Paoletta, E. Uriarte, K.-N. Klotz, S. Moro, F.
Borges, Biochem. Pharmacol. 84 (2012) 21–29.
Appendix A. Supplementary material
[33] O. Lenzi, V. Colotta, D. Catarzi, F. Varano, D. Poli, G. Filacchioni, K. Varani, F.
Vincenzi, P.A. Borea, S. Paoletta, E. Morizzo, S. Moro, J. Med. Chem. 52 (2009)
7640–7652.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bioorg.2015.05.
008.