Substituted pyrazolo[3,4-b]pyridines as human A1 adenosine antagonists: Developments in understanding the receptor stereoselectivity

Article abstract of DOI:10.1039/c0ob01064b

A₁ adenosine receptor antagonists have been proposed to possess an interesting range of potential therapeutic applications. We have already reported the synthesis and the biological characterization of a family of pyrazolo[3,4-b]pyridine deriva

Full text of DOI:10.1039/c0ob01064b

Organic &
Biomolecular
Chemistry
www.rsc.org/obc
Volume 9 | Number 12 | 21 June 2011 | Pages 4389–4696
ISSN 1477-0520
FULL PAPER
Silvia Schenone et al.
Substituted pyrazolo[3,4-b]pyridines
as human A1 adenosine antagonists:
Developments in understanding the
receptor stereoselectivity
1477-0520(2011)9:12;1-W

Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 4448
www.rsc.org/obc
PAPER
Substituted pyrazolo[3,4-b]pyridines as human A1 adenosine antagonists:
Developments in understanding the receptor stereoselectivity†
Tiziano Tuccinardi,a Alessandra Tania Zizzari,c Chiara Brullo,b Simona Daniele,d Francesca Musumeci,b
Silvia Schenone,*b Maria Letizia Trincavelli,d Claudia Martini,d Adriano Martinelli,a Gianluca Giorgie and
Maurizio Bottac
Received 7th December 2010, Accepted 9th February 2011
DOI: 10.1039/c0ob01064b
A1 adenosine receptor antagonists have been proposed to possess an interesting range of potential
therapeutic applications. We have already reported the synthesis and the biological characterization of a
family of pyrazolo[3,4-b]pyridine derivatives as A1 adenosine ligands endowed with an antagonistic
profile. In the present work, we report the LC separation of enantiomers of our most active A1
antagonists together with the determination of their absolute configuration by means of X-ray crystal
structure analysis. Biological assays confirmed a different activity for the two enantiomers, with the R
one showing the higher human A1AR affinity. We also developed a homology model of this receptor
subtype in order to suggest a binding disposition of the ligands into the hA1AR. All of the obtained
data suggest that the compound’s chirality plays a key role in A1 affinity.
adenylyl cyclase activity, while activation of A2A and A2BAR causes
Introduction
a stimulation of adenylyl cyclase.3 The physiological significance
The interaction of a drug with its biological target is extremely
and functions of adenosine have been extensively studied and there
specific because it depends on the three-dimensional disposition of
has also been research to improve the understanding of ligand-
groups able to determine this specificity. During the development
A1 receptor interactions. In the last two decades, a large number
of a drug, it is useful to consider the advantages in terms of
of selective A1AR ligands have been developed, both xanthine-
action specificity and safety of using homochiral compounds;
derivatives and different compounds, usually represented by bi- or
furthermore chirality could represent a precious source to study a
drug’s mechanism of action in biological systems.
Adenosine is an endogenous neuromodulator distributed in
tricyclic fused heterocyclic compounds, including purines, deaza-
purines, pyrazolopyridines, imidazotriazines, thienopyridazines,
naphthyridines, quinoxalines, pyridopyrimidines, pyrazoloquino-
a wide variety of tissues, in both the periphery and the central
lines and pyrimidoindoles. Also, some examples of monocyclic A1
nervous system.1,2 The effects elicited by adenosine are mediated
antagonists have been reported, including thiazoles, thiadiazoles
and pyrimidines.4 Different therapeutic applications have been
identified in preclinical and clinical studies for A1AR antagonists,
which are effective as potassium-sparing diuretic agents with
kidney-protecting properties.5 This type of compound is also
being tested in the treatment of bradyarrhythmias associated
with inferior myocardial infarction, cardiac arrest and cardiac
transplant rejection, and could be useful in the treatment of
chronic heart diseases.6 A1AR antagonists may also offer a
therapeutic opportunity for chronic lung diseases such as asthma,
chronic obstructive pulmonary disease and pulmonary fibrosis.7
The observation of the effects of caffeine, a classical non-selective
adenosine antagonist, on the CNS, such as improvement of aware-
ness and learning, encouraged the search of selective antagonists
endowed with central activity. Selective A1AR antagonists induce
cognition enhancement, leading to a general improvement in
memory performance, and these actions are potentially useful
in the treatment of dementia and anxiety disorders.8 Moreover,
Trevitt and colleagues recently reported that treatment with
the A1AR antagonist CPT (8-cyclopentyl-1,3-dimethylxanthine)
by its interactions with four receptor subtypes named A1, A2A,
A2B, A3. These receptors belong to the superfamily of G protein-
coupled receptors. The stimulation of adenosine receptors (AR)
regulates several effector systems, such as the enzyme adenylyl
cyclase. Activation of A1 and A3AR leads to an inhibition of
aDipartimento di Scienze Farmaceutiche, Universita` di Pisa, Via Bonanno 6,
56126, Pisa, Italy
bDipartimento di Scienze Farmaceutiche, Universita` degli Studi di Genova
Viale Benedetto XV 3, 16132, Genova, Italy. E-mail: schensil@unige.it;
Fax: (+ 39) 010-353-8358
cDipartimento Farmaco Chimico Tecnologico, Universita` degli Studi di Siena,
Via De Gasperi 2, 53100, Siena, Italy
dDipartimento di Psichiatria, Neurobiologia, Farmacologia e Biotecnologie
Universita` degli Studi di Pisa Via Bonanno 6, 56126, Pisa, Italy
eDipartimento di Chimica, Universita` degli Studi di Siena, Via De Gasperi
2, 53100, Siena, Italy
† Electronic supplementary information (ESI) available: CD spectra of the
enantiomers of compounds 2–5, Ramachandran plot of the starting and
energy minimized average structure of the human A1AR model. CCDC
reference number 763003. For ESI and crystallographic data in CIF or
other electronic format see DOI: 10.1039/c0ob01064b
4448 | Org. Biomol. Chem., 2011, 9, 4448–4455
This journal is
The Royal Society of Chemistry 2011
©

Table 1 Structures, specific rotation, and experimental affinity (Ki) on
human A1AR of the analyzed compounds
in a model of Parkinson’s disease produced a dose dependent
improvement in locomotion, suggesting that, although the role of
A1AR in Parkinson’s disease is still unclear, the A1AR antagonism
may produce therapeutic effects, particularly at the beginning of
treatment.9
Recently, we synthesized and tested a family of pyrazolo[3,4-
b]pyridine derivatives as A1AR ligands.10,11 These compounds
have shown to be potent and selective antagonists for the bovine
A1 adenosine receptors; furthermore, some of them also showed
a promising human A1 affinity. Extensive computational work
has also been carried out showing a good correlation with the
available experimental data. In addition, one racemic compound
was separated into the two enantiomers, and the biological affinity
showed an interesting behaviour with an enantiomeric selectivity
that encouraged further analysis of this aspect. These preliminary
results prompted us to search for a direct method of enantiomeric
separation in order to evaluate the activity of other enantiomeri-
cally pure compounds. For this purpose, four high human A1AR
affinity compounds (1, 3, 4, 5) and one (2) characterized by a poor
affinity have been chosen. Direct LC separation into enantiomers,
together with the determination of their absolute configuration, is
reported. Finally, molecular modelling studies have been carried
out with the aim of investigating the interactions of the compounds
into the human A1 receptor model herein developed.
Ki (nM) or %
20
Cpds
R1
R2
R3
[a]D
inhibition at 10 mM
(R,S)-1
(R)-1
(S)-1
(R,S)-2
(R)-2
(S)-2
(R,S)-3
(R)-3
(S)-3
(R,S)-4
(R)-4
(S)-4
–OCH3
–OCH3
–OCH3
–OCH3 –Cl –CH2CH3
–OCH3 –Cl –CH2CH3
–OCH3 –Cl –CH2CH3
–OCH3
–OCH3
–OCH3
–OCH3
–OCH3
–OCH3
–F
H
H
H
–CH3
–CH3
–CH3
/
94 7.0a
31 3.2a
435 42a
15%
-3.45
+3.46
/
-17.60 410 40
+20.80 10%
H
H
H
H
H
H
H
H
H
–CH(CH3)2
–CH(CH3)2
–CH(CH3)2
–CH2CH3
–CH2CH3
–CH2CH3
–CH2CH3
–CH2CH3
–CH2CH3
/
18 1.1
11.4 0.35
51 2.2
27 3.3
+6.80
-4.52
/
+14.41 6.1 1.0
-10.73 36 3.2
(R,S)-5
(R)-5
(S)-5
/
17 1.0
12 1.2
–F
–F
-17.5
+14.03 50 4.3
a Ref. 11
Results and discussion
Enantiomeric separation, specific rotation and circular dichroism
(CD) spectra
assignment is coherent with the biological data that showed for all
the R enantiomers a common higher hA1AR affinity with respect
to the corresponding S compounds.
Previously described racemic bovine A1AR antagonists 2–5 have
been resynthesized, as previously described by us.11 The enan-
tiomeric resolution of compounds 2–5 was achieved by using a
Chiralcel HPLC OD column and an isocratic elution. Afterwards,
a Chiralpak AS column was used for semipreparative purposes,
by means of a rapid and practical isocratic method. The collected
fractions were then concentrated by rotary evaporation to provide
a sizeable amount of each enantiomer. Analytical HPLC re-runs of
the elutes were done indicating an enantiomeric purity higher than
95% for each enantiomer. Specific rotation for all the enantiomers
was measured and reported in Table 1.
Biological assays
The enantiomers were tested for their affinity toward human A1
CHO transfected cells.11 Binding affinities expressed as affinity
constant values (Ki) or percent inhibition of the radioligand
binding, are reported in Table 1. Activities on human A1AR have
also been remeasured for the racemic compounds. As expected, all
of the R enantiomers showed higher hA1AR affinity with respect
to the corresponding S and racemic compounds.
CD spectra of the compounds were measured in ethanol. After
subtraction of the solvent baseline, they resulted in mirror images
of each other, indicating their enantiomeric nature (see ESI†).
Molecular modelling
The human A1AR model was generated using the crystal structure
of human A2AAR bound to ZM241385 as a template (3EML).16
The sequence alignment was studied on the four human adenosine
receptor subtypes, the bovine rhodopsin (3C9L),17 human b2
adrenergic (2RH1)18 and turkey b1 adrenergic receptors (2VT4).19
These last three sequences were added to the analysis since crystal
structures of these complexes are deposited in the Protein Data
Bank.20 As shown in Fig. 2, the alignment was guided by the
highly conserved amino acid residues, including the asparagine
residues N1.50, the LA-AD (L2.46, A2.47, A2.49, and D2.50),
and D/ERY-V (D/E3.49, R3.50, Y3.51, and V3.54) motif, the
highly conserved tryptophane W4.50, the two prolines P4.59 and
P6.50, and the NPXXY motif in TM7 (N7.49, P7.50, and Y7.53).21
Following the reported alignment, an A1 receptor model was
built and subjected to a simulated annealing protocol by means of
Assignment of absolute configuration
We previously deduced the absolute configuration of compound
1 through CD spectra comparison with a structurally related
compound already reported in the literature.11,12 In the present
study, the enantiomerically pure compound (-)-1 (blue) has
been crystallized and, as shown in Fig. 1, its X-ray crystal
structure confirmed our previously R configuration hypothesis. It
is interesting to note the change in sign of the optical rotation
of the R and S isomers of compounds 3 and 4, compared
with those for compounds 1, 2 and 5. Even if the absolute
configuration of the stereogenic center could be assigned using
powerful modern ab initio methods,13–15 we obtained it through CD
spectra comparison of the stereoisomers (see ESI†). The reported
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 4448–4455 | 4449
©

Table 2 MM-PBSA results for the (R)-4-A1AR and (S)-4-A1AR com-
plexes. DEMM is the sum of the electrostatic (Ele), van der Waals (VdW) and
internal (Eint) energy, DPBSolv is the sum of the polar (PB) and non-polar
(PBSur) solvation free energy. Finally DGPBSA is the sum of the molecular
mechanical and solvation free energy. Data are expressed as kcal mol-1
HYPO1(R) HYPO2(R) HYPO3(R) HYPO4(R) HYPO3(S)
Ele
VdW
Eint
- 0.85
- 60.32
0.00
- 5.37
- 59.80
0.00
- 11.05
- 68.28
0.00
- 17.19
- 55.62
0.00
- 6.79
- 62.61
0.00
DEMM
PBSur
PB
DPBSolv
DGPBSA
- 59.48
- 5.67
36.21
30.54
- 28.94
- 65.17
- 5.80
37.97
32.17
- 32.99
- 79.33
- 5.65
39.42
28.38
- 45.56
- 72.81
- 5.40
46.29
40.88
- 31.93
- 69.40
- 5.90
37.26
31.35
- 38.05
a fully hydrated phospholipid bilayer environment made up of
palmitoyloleoylphosphatidylcholine (POPC) molecules solvated
by TIP3 water molecules, as described in the Experimental
section. The system contained 230 POPC molecules, 23 415 water
molecules, 12 chlorine ions, the A1AR and compound 4, for a
total of 106 121 atoms. The so obtained four MD trajectories
were further analyzed through the MM-PBSA method25 that
has shown to accurately estimate the ligand–receptor energy
interaction.26,27 This approach averages the contributions of gas-
phase energies, solvation free energies, and solute entropies
calculated for snapshots of the complex molecule as well as the
unbound components, extracted from MD trajectories, according
to the procedure fully described in the Experimental section.
As shown in Table 2, the MM-PBSA results strongly suggested
that the third ligand disposition hypothesis (HYPO3) is the most
reliable, as it showed an interaction energy (DGPBSA = - 45.56 kcal
mol-1) of approximately 12 kcal mol-1 stronger than the other three
ligand binding hypotheses.
The stability of the MD simulations was evaluated by calculating
the total energy of the systems and analyzing the root-mean-square
deviation (RMSD) of all of the a carbons of the TM helices and
the heavy atoms from the starting A1 model structure. Fig. 3
shows the results of this analysis for the (R)-4-A1AR complex
(HYPO3). After about 0.5 ns of MD, the system reached an
equilibrium, since the total energy for the last 4.5 ns remained
approximately constant (Fig. 3, first plot). Analyzing the RMSD
of all a carbons of the TM helices from the starting A1 model
structures, we observed that, after 600 ps, it remained between 1.2
Fig. 1 CD spectra of the enantiomer (-)-1 and relative X-ray crystal
structure. Ortep view of one of the four molecules constituting the
asymmetric unit with atom labelling.
the Modeller program.22 The best scoring structure was energy
minimized and its backbone conformation was evaluated by
PROCHECK23 (see the Experimental section for details). An
analysis of the Psi/Phi Ramachandran plot of the resulting model
indicated that only one amino acid of the second extracellular
loop had a disallowed geometry (see Fig. S2 in ESI†). The good
results obtained from the Psi/Phi Ramachandran plot suggested
that the molecular model created of the A1 receptor could be
used for further studies. Compound 4, which showed the highest
A1 potency, was then docked using the GOLD software24 into
the A1 receptor model. The binding site of the ligand, defined
by taking the interaction of ZM241385 in the X-ray structure of
A2AAR into account, was individuated in the region between TM2,
TM3, and TM5–TM7. The clusterization of the docking results
highlighted four different hypothetical ligand dispositions. In
order to define the most probable ligand disposition, the four A1-
ligand complexes were used as starting points for 5 ns of molecular
dynamic simulations (MD). We carried out the simulations in
˚
and 1.5 A (Fig. 3, second plot), suggesting that the system was
quite stable during the remaining time of MD simulations. This
hypothesis was also confirmed by the analysis of the RMSD of all
heavy atoms that, after an initial increase, after 600 ps, it remained
˚
between 2.6 and 3.2 A (Fig. 3, second plot). The analysis of
the Psi/Phi Ramachandran plot of the energy minimized average
structures of the MD’s last 4.4 ns indicated that no residue had
a disallowed geometry, and only L61, V152 and W156 were in
a generously allowed region (Fig. S3 in ESI†). Fig. 4 shows the
energy minimized average structures of the last 4.4 ns of the A1AR
model complexed with compound (R)-4. The 1H-pyrazolo[3,4-
b]pyridine central scaffold shows a p–p interaction with F171,
a lipophilic interaction with V87 and L250, and a hydrogen
bond with N254. The ethyl ester chain interacted with T270 and
the carbonyl function formed an intramolecular hydrogen bond
4450 | Org. Biomol. Chem., 2011, 9, 4448–4455
This journal is
The Royal Society of Chemistry 2011
©

Fig. 2 Alignment of the human adenosine receptors with bovine rhodopsin, human b2 adrenergic and turkey b1 adrenergic receptors amino acid
sequences. The highly conserved patterns are marked with black background. The other identical residues are in bold characters. In the first and second
lines of the alignment scheme are reported the A1AR numeration, while the TM domains of A2AAR are reported in gray background.
with the amino group at C4. The 4-methoxyphenethyl substituent
interacted with I274 and formed a hydrogen bond with Y251.
Finally the R-2-chloro-2-phenethyl substituent interacted with
L88, M180 and showed a p–p interaction with H251. The main
available experimental data are in agreement with this binding
hypothesis; in particular, an important role for F171 and N254
has already been highlighted by the analysis of the X-ray structure
of the A2A-ZM241385 complex,16 furthermore mutagenesis studies
have suggested an important role for V87, L88 and H251.28
In order to verify if the receptor model was also able to dis-
criminate between the active R and less active S enantiomers, the
MD simulation procedure applied for the (R)-4-A1AR complexes
was also applied for the S enantiomer. As a starting geometry,
we used the third docking hypothesis (HYPO3) of (R)-4 inverting
the chirality of the compound. The analysis of the MD trajectory
confirmed the loss of affinity of this compound with respect to
the R enantiomer. Fig. 4 shows the energy minimized average
structures of the last 4.4 ns of the A1AR model complexed with
compound (S)-4; the inversion of the chiral centre determined
mainly determined the loss of the hydrogen bonds of the central
scaffold with N254 and of the 4-methoxyphenethyl substituent
with Y271. Furthermore, the different disposition of the S-2-
chloro-2-phenethyl substituent also determined the loss of the p–p
interaction with H251.
As shown in Table 2, the qualitative analysis is also confirmed by
the quantitative one because the analysis of the MM-PBSA results
confirmed a greater A1AR affinity for the R enantiomer, with an
energy difference between the interactions of the two enantiomers
of about 7 kcal mol-1.
Conclusions
The previously reported racemic mixtures of the selected A1AR
antagonists were separated into the corresponding enantiomers
through a semipreparative HPLC approach. Absolute config-
uration was determined for (-)-1 enantiomer by means of X-
ray crystal structure analysis, then the CD spectra comparison
allowed the characterization of the stereochemistry of the other
reported compounds. The biological assays confirmed a selective
˚
a shift of the molecule of about 2 A, and this movement
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 4448–4455 | 4451
©

docking/MD simulation approach, a binding disposition for the
(R)-4 compound was provided. Furthermore, the comparison of
the interaction of the (R)- and (S)-4 enantiomers confirmed the
higher affinity of the first one, which was also supported by an
MM-PBSA analysis. All of these data suggest that for such high
A1 affinity, the compound’s chirality plays a key role that must
be taken into consideration in the design of highly potent A1AR
antagonists, and to our knowledge, this is one of the first reports
for the A1 adenosine receptor in this field.29,30
Experimental section
Instrumentation
The chiral separation studies were carried out on a Varian
Prostar HPLC system (Varian Analytical Instruments, USA)
equipped with a binary pump with a manual injection valve
and model Prostar 325 UV-VIS Detector. The CD detection was
achieved on a Jasco CD-815 spectropolarimeter circular dichroism
detector (Jasco Corporation, Tokyo, Japan). Optical rotations
were determined with a Perkin–Elmer Mod 343 polarimeter at
589 nm, using a 10-1 dm microcell. Concentrations are expressed
as g mL-1.
Enantioselective columns and chemicals
The polysaccharide-derived columns were cellulose tris-3,5-
dimethylphenylcarbamate (250 mm ¥ 4.6 mm, Chiralcel OD) and
amylose tris 5-a methylbenzyl carbamate (250 mm ¥ 10 mm,
Chiralpak AS), both coated on 10 mm silica gel. All of the above
columns were obtained from Daicel (Tokyo, Japan). All of the
solvents and reagents were from Sigma Aldrich Srl (Milan, IT).
Fig. 3 Analysis of the MD simulation of the (R)-4 (HYPO3) complex
with A1AR. In the first plot, the total energy of the system vs. the time
is reported; the second plot shows the RMSD of the a carbons of the
TM helices and the heavy atoms of the receptor from the starting model
structure during the simulation.
LC conditions
All separations were carried out at ambient temperature using
various mobile phases constituted by n-hexane and 2-propanol
mixture. Detection was carried out at 300 nm. The injection
volumes were 20 mL and 200 mL for analytical and semipreparative
purpose, respectively.
profile for all of the other enantiomers, with the R one showing
the highest human A1AR affinity. In order to suggest a binding
disposition of the ligands into the human A1AR, an homology
model of this receptor subtype was developed and, using a mixed
Fig. 4 MD simulation results for the (R)-4-A1AR (light blue) and (S)-4-A1AR (magenta) complex.
4452 | Org. Biomol. Chem., 2011, 9, 4448–4455 This journal is The Royal Society of Chemistry 2011
©

Enantioseparation of racemic mixture through a semipreparative
method
supplied by K.-N. Klotz, Wurzburg University, Germany).33
All compounds were routinely dissolved in dimethyl sulfoxide
(DMSO) and diluted in an assay buffer so that the DMSO
never exceeded 2%. The compounds were first of all screened
at 100 nM and 10 mM concentration. For the compounds that
did not display a significant percentage of binding inhibition
(at 10 mM concentrations), Ki value were not determined. For
the active compounds, a dose-response curve using at least
six different compound concentrations was perfomed and the
IC50 value determined. IC50 values, computer-generated using a
nonlinear regression formula on a computer program (GraphPad,
San Diego, CA), were converted to Ki values, knowing the
Kd values of radioligand and using the Cheng and Prusoff
equation.34
For semipreparative purposes, Chiralpak AS (250 ¥ 10 mm) was
used. The binary solvent system consisted of n-hexane (solvent
B) and 2-propanol (solvent A). Samples of racemic mixture were
dissolved with the solvents used for mobile phase at the same
percentage (v/v). The separation was achieved using the isocratic
method. The separation of 2 was achieved using a mobile phase
made up of 95% of n-hexane and 5% of 2-propanol at a flow rate
of 2.0 mL min-1. The separation of 3 was achieved using 90% of n-
hexane and 10% of 2-propanol at a flow rate of 2.0 mL min-1, while
the separations of 4 and 5 were achieved using mobile phase made
of 85% of n-hexane and 15% of 2-propanol at a flow rate of 2.0
and 3.0 mL min-1, respectively. Analytical re-runs of each eluate
were done on Chiralcel OD column with a mobile phase made up
of 95% of n-hexane and 5% of 2-propanol at a flow rate of 1.0 mL
min-1 for compounds 2, 4, 5 and 0.8 mL min-1 for compound 3.
Isolation of single enantiomers of compound 1 has been already
reported.11
Homology modelling
The crystal structure of human A2AAR bound to ZM241385
was taken from the Protein Data Bank,20 while all the primary
sequences were obtained from the SWISS-PROT protein sequence
database.35 The sequential alignment was performed by means
of CLUSTAL W,36 using the Blosum series as a matrix, with a
gap open penalty of 10 and a gap extension penalty of 0.05. The
human A1 and A2AAR show about 45% of identical residues with
a CLUSTAL W alignment score of 48. The human A1 was con-
structed directly from the coordinates of the corresponding amino
acids in A2AAR by means of the Modeller program.22 Starting
from this receptor, 10 structures were generated by means of the
“very slow MD annealing” refinement method, as implemented
in Modeller, and on the basis of the DOPE (discrete optimized
protein energy) assess method, the best receptor model was chosen.
The backbone conformation of the resulting receptor structure
was evaluated by inspection of the Psi/Phi Ramachandran plot
obtained from PROCHECK analysis.23
CD conditions
CD spectra were acquired on a Jasco J-815 dichroism spectrometer
with linear data array, two accumulations and with scanning
speed of 100 nm min-1. A 1.0 mm path-length quartz cell
was used and CD spectra were recorded at room temperature.
CD spectra obtained from compounds eluted from the racemic
mixture separation were acquired in the 190–500 nm range. Pure
enantiomers were dissolved in ethanol to obtain 0.001 mol L-1
solutions. Three scans were averaged and blank-substracted to
obtain the CD spectrum.
Crystallization conditions
Crystals of 1-(B) suitable for single-crystal X-ray diffraction
structure analysis were obtained by slow evaporation of a solution
of 1-(B) in n-hexane/2-propanol, 70 : 30 (v/v) at room temperature
(23 ◦C).
Docking of (R)-4
The ligand was built using Maestro37 and was subjected to
a conformational search (CS) of 1000 steps, using a water
environment model (generalized-Born/surface-area model) by
means of Macromodel.38 The algorithm used was based on
the Monte Carlo method with the MMFFs force field and a
distance-dependent dielectric constant of 1.0. The ligand was
then energy minimized using the conjugated gradient method
X-ray crystal structure determination
A single crystal of 1-(B) was submitted to X-ray data collections
with a Bruker–Nonius FR591 rotating anode diffractometer
with graphite monochromated Mo-Ka radiation (l = 0.71073
˚
A) at 120(2)K. The structure was solved by direct methods
˚
until a convergence value of 0.05 kcal/(A mol) was reached,
implemented in the SHELXS-97 program31 and the refinement
was carried out by full-matrix anisotropic least-squares on F2
for all reflections for non-H atoms, by using the SHELXL-97
program. The compound crystallizes in the Monoclinic crystal
system, P21 space group, in agreement with its chiral nature. The
absolute configuration at C(11) has been determined by the Flack
method32 and it resulted to be R. CCDC-763003 contains the
supplementary crystallographic data. These data can be obtained
free of charge from the Cambridge Crystallographic Data Centre
via www.ccdc.cam.ac.uk/data_request/cif.
using the same force field and parameters used for the CS.
Automated docking was carried out by means of the GOLD
program,24 version 4.1.1. The region of interest used by GOLD
˚
was defined in order to contain the residues within 10 A from
the original position of ZM241385 in the A2AAR X-ray structure.
The “allow early termination” option was deactivated, while the
possibility of clusterizing the results was activated (the RMSD
˚
distance clustering was set to 1.5 A). The remaining GOLD default
parameters were used, and the ligand was submitted to 30 genetic
algorithm runs by applying the GoldScore fitness function. The
clusters showing a GoldScore fitness scoring value that differed less
than 10 units with respect to the best docked conformation were
taken into account; as a result, four docking conformations were
considered.
Human A1AR binding assay
The analyzed compounds were evaluated for their affinity to-
ward human A1AR stably expressed in CHO cells (kindly
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 4448–4455 | 4453
©

MD simulations
the energetic interactions of the two enantiomers of the same
ligand into the receptor, we took only the first three terms of eqn
(1) into account, considering the entropic value as approximately
constant.
All simulations were performed using AMBER 10.39 The five com-
plexes (four (R)-4-A1AR and one (S)-4-A1AR systems) were em-
bedded into a phospholipid bilayer made up of POPC molecules.
The creation of the phospholipid bilayer and the insertion of the
receptor-ligand complexes were carried out using VMD.40 MD
simulations were carried out using the modified parm94 force field
at 300 K. An explicit solvent model of TIP3P water was used. The
system was solvated on the “extracellular” and “intracellular” side
Acknowledgements
National Interest Research Projects (PRIN_2008_5HR5JK) is
gratefully acknowledged.
˚
with a 12 A water cap. Chlorine ions were added as counterions
to neutralize the system. Prior to MD simulations, three steps of
energy minimization were carried out. In the first stage, we kept
References
1 K. A. Jacobson, P. J. van Galen and M. Williams, J. Med. Chem., 1992,
35, 407–422.
the protein and phospholipids fixed with a position restraint of
-2
100 kcal mol-1
and we just energy minimized the positions
˚
A
2 J. W. Daly, J. Med. Chem., 1982, 25, 197–207.
of the water molecules. In the second stage, we energy minimized
the phospholipid–water system, applying a position restraint of
3 S. A. Poulsen and R. J. Quinn, Bioorg. Med. Chem., 1998, 6, 619–641.
4 S. Schenone, C. Brullo, F. Musumeci, O. Bruno and M. Botta, Curr.
Top. Med. Chem., 2010, 10, 878–901.
5 P. S. Modlinger and W. J. Welch, Curr. Opin. Nephrol. Hypertens., 2003,
12, 497–502.
-2
100 kcal mol-1
A
on the protein. Finally, we applied a restraint
only on the a carbons of the receptor. The
˚
of 30 kcal mol-1
-2
˚
A
three energy minimization stages consisted of 10 000 steps. The
first 1000 steps were Steepest Descent, and the last 9000 were
Conjugate Gradient. Molecular dynamics trajectories were run
using the energy minimized structure as the input, and particle
mesh Ewald electrostatics41 and periodic boundary conditions
were used in the simulation. The time step of the simulations
6 R. H. Shah and W. H. Frishman, Cardiol. Rev., 2009, 17, 125–131.
7 C. N. Wilson, A. Nadeem, D. Spina, R. Brown, C. P. Page and S. J.
Mustafa, Handb. Exp. Pharmacol., 2009, 193, 329–362.
8 T. Maemoto, M. Tada, T. Mihara, N. Ueyama, H. Matsuoka, K.
Harada, T. Yamaji, K. Shirakawa, S. Kuroda, A. Akahane, A. Iwashita,
N. Matsuoka and S. Mutoh, J. Pharmacol. Sci., 2004, 96, 42–52.
9 J. Trevitt, C. Vallance, A. Harris and T. Goode, Pharmacol., Biochem.
Behav., 2009, 92, 521–527.
10 F. Manetti, S. Schenone, F. Bondavalli, C. Brullo, O. Bruno, A. Ranise,
L. Mosti, G. Menozzi, P. Fossa, M. L. Trincavelli, C. Martini, A.
Martinelli, C. Tintori and M. Botta, J. Med. Chem., 2005, 48, 7172–
7185.
˚
was 2.0 fs with a cutoff of 12 A for the non-bonded interaction.
SHAKE was employed to keep all bonds involving hydrogen
atoms rigid. A constant-volume was carried out for 200 ps, during
which time the temperature was raised from 0 to 300 K (using the
Langevin dynamics method). Then, 4800 ps of constant-pressure
MD were carried out at 300 K. In the first 400 ps, all of the
a carbons of the receptor were blocked with a harmonic force
11 T. Tuccinardi, S. Schenone, F. Bondavalli, C. Brullo, O. Bruno, L. Mosti,
A. T. Zizzari, C. Tintori, F. Manetti, O. Ciampi, M. L. Trincavelli, C.
Martini, A. Martinelli and M. Botta, ChemMedChem, 2008, 3, 898–
913.
12 F. Da Settimo, G. Primofiore, C. La Motta, S. Taliani, F. Simorini, A.
M. Marini, L. Mugnaini, A. Lavecchia, E. Novellino, D. Tuscano and
C. Martini, J. Med. Chem., 2005, 48, 5162–5174.
constant. This was decreased during the 400 ps from 10 to 1 kcal
-1
mol-1 A . In the last 4400 ps, there were no constraints. The final
˚
structure was obtained as the average of the last 4400 ps of MD
13 R. K. Kondru, P. Wipf and D. N. Beratan, Science, 1998, 282, 2247–
2250.
energy minimized with the CG method until a convergence of 0.05
14 R. K. Kondru, P. Wipf and D. N. Beratan, J. Phys. Chem. A, 1999, 103,
6603–6611.
-1
˚
A
kcal mol-1
was reached. The backbone conformation of the
resulting receptor structure was evaluated by inspection of the y/f
Ramachandran plot obtained from PROCHECK analysis.23 The
General Amber Force Field (GAFF) parameters were assigned to
POPC molecules. The partial charges were calculated using the
AM1-BCC method, as implemented in the Antechamber suite of
AMBER 10.
15 P. L. Polavarapu, Chirality, 2002, 14, 768–781.
16 V. P. Jaakola, M. T. Griffith, M. A. Hanson, V. Cherezov, E. Y. Chien,
J. R. Lane, A. P. Ijzerman and R. C. Stevens, Science, 2008, 322, 1211–
1217.
17 R. E. Stenkamp, Acta Crystallogr., Sect. D: Biol. Crystallogr., 2008,
D64, 902–904.
18 V. Cherezov, D. M. Rosenbaum, M. A. Hanson, S. G. Rasmussen, F.
S. Thian, T. S. Kobilka, H. J. Choi, P. Kuhn, W. I. Weis, B. K. Kobilka
and R. C. Stevens, Science, 2007, 318, 1258–1265.
19 T. Warne, M. J. Serrano-Vega, J. G. Baker, R. Moukhametzianov, P. C.
Edwards, R. Henderson, A. G. Leslie, C. G. Tate and G. F. Schertler,
Nature, 2008, 454, 486–491.
Energy evaluation
We extracted from the last 4400 ps of MD of the ligand–receptor
complexes, 88 snapshots (at time intervals of 50 ps) for each species
(complex, receptor and ligand). The various MM-PBSA energy
terms in equation 1 were computed as follows.
20 H. M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T. N. Bhat, H.
Weissig, I. N. Shindyalov and P. E. Bourne, Nucleic Acids Res., 2000,
28, 235–242.
21 A. Martinelli and T. Tuccinardi, Expert Opin. Drug Discovery, 2006, 1,
459–476.
22 A. Fiser, R. K. Do and A. Sali, Protein Sci., 2000, 9, 1753–1773.
23 R. A. Laskowski, M. W. Macarthur, D. S. Moss and J. M. Thornton,
J. Appl. Crystallogr., 1993, 26, 283–291.
(1)
G = Gpolar + Gnonpolar + Emm - TS
Electrostatic, van der Waals and internal energies (Emm) were
obtained using the SANDER module in AMBER 10. Polar
energies (Gpolar) were obtained from the PBSA module of the
AMBER 10 program (using the Poisson–Boltzman method)
applying dielectric constants of 1 and 80 to represent the gas
and water phases, respectively. Nonpolar energies (Gnonpolar) were
determined using the MOLSURF program. In order to compare
24 M. L. Verdonk, J. C. Cole, M. J. Hartshorn, C. W. Murray and R. D.
Taylor, Proteins: Struct., Funct., Bioinf., 2003, 52, 609–623.
25 P. A. Kollman, I. Massova, C. Reyes, B. Kuhn, S. Huo, L. Chong, M.
Lee, T. Lee, Y. Duan, W. Wang, O. Donini, P. Cieplak, J. Srinivasan, D.
A. Case and T. E. Cheatham, 3rd, Acc. Chem. Res., 2000, 33, 889–897.
26 T. Tuccinardi, S. Bertini, A. Martinelli, F. Minutolo, G. Ortore, G. Pla-
canica, G. Prota, S. Rapposelli, K. E. Carlson, J. A. Katzenellenbogen
and M. Macchia, J. Med. Chem., 2006, 49, 5001–5012.
4454 | Org. Biomol. Chem., 2011, 9, 4448–4455
This journal is
The Royal Society of Chemistry 2011
©

27 T. Tuccinardi, F. Manetti, S. Schenone, A. Martinelli and M. Botta, J.
Chem. Inf. Model., 2007, 47, 644–655.
28 A. Martinelli and T. Tuccinardi, Med. Res. Rev., 2008, 28, 247–277.
29 S. A. Poulsen, D. J. Young and R. J. Quinn, Bioorg. Med. Chem. Lett.,
2001, 11, 191–193.
30 J. R. Pfister, L. Belardinelli, G. Lee, R. T. Lum, P. Milner, W. C. Stanley,
J. Linden, S. P. Baker and G. Schreiner, J. Med. Chem., 1997, 40, 1773–
1778.
35 E. Gasteiger, A. Gattiker, C. Hoogland, I. Ivanyi, R. D. Appel and A.
Bairoch, Nucleic Acids Res., 2003, 31, 3784–3788.
36 J. D. Thompson, D. G. Higgins and T. J. Gibson, Nucleic Acids Res.,
1994, 22, 4673–4680.
37 Maestro, version 9.0, Schro¨dinger Inc: Portland, OR, 2009.
38 Macromodel, version 9.7, Schro¨dinger Inc: Portland, OR, 2009.
39 D. A. Case, T. A. Darden, T. E. Cheatham III, C. L. Simmerling, J.
Wang, R. E. Duke, R. Luo, M. Crowley, R. C. Walker, W. Zhang, K.
M. Merz, B. Wang, S. Hayik, A. Roitberg, G. Seabra, I. Kolossva´ry, K.
F. Wong, F. Paesani, J. Vanicek, X. Wu, S. R. Brozell, T. Steinbrecher,
H. Gohlke, L. Yang, C. Tan, J. Mongan, V. Hornak, G. Cui, D. H.
Mathews, M. G. Seetin, C. Sagui, V. Babin, and P. A. Kollman, 2008,
AMBER 10, University of California: San Francisco, CA, .
40 W. Humphrey, A. Dalke and K. Schulten, J. Mol. Graphics, 1996, 14(33–
38), 27–38.
31 G. M. Sheldrick and T. R. Schneider, Methods Enzymol., 1997, 277,
319–343.
32 H. D. Flack, Acta Crystallogr., Sect. A: Found. Crystallogr., 1983, A39,
876–881.
33 O. Lenzi, V. Colotta, D. Catarzi, F. Varano, G. Filacchioni, C. Martini,
L. Trincavelli, O. Ciampi, K. Varani, F. Marighetti, E. Morizzo and S.
Moro, J. Med. Chem., 2006, 49, 3916–3925.
34 Y. Cheng and W. H. Prusoff, Biochem. Pharmacol., 1973, 22, 3099–
3108.
41 U. Essmann, L. Perera, M. L. Berkowitz, T. Darden, H. Lee and L. G.
Pedersen, J. Chem. Phys., 1995, 103.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 4448–4455 | 4455
©