Design and synthesis of pyrazolo[3,4-d]pyrimidinone derivatives: Discovery of selective phosphodiesterase-5 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2020.127337

A novel series of 1,6-disubstituted pyrazolo[3,4-d]pyrimidin-7-one derivatives, 2a-h, 4a-d, 5 and 6, were successfully synthesized, which showed promising, and potent inhibition of phosphodiesterase 5 (PDE5). The inhibitory activities of 5, 4b, 2a, 2d, 2f

Full text of DOI:10.1016/j.bmcl.2020.127337

Journal Pre-proofs
Design and Synthesis of Pyrazolo[3,4-d]pyrimidinone Derivatives: Discovery
of Selective Phosphodiesterase-5 Inhibitors
Mohamed A. Shaaban, Yaseen A.M.M. Elshaier, Ali H. Hammad, Nahla A.
Farag, Haredy Hassan Haredy, Ahmed A. AbdEl-Ghany, Khaled O.
Mohamed
PII:
S0960-894X(20)30448-0
DOI:
Reference:
https://doi.org/10.1016/j.bmcl.2020.127337
BMCL 127337
To appear in:
Bioorganic & Medicinal Chemistry Letters
Received Date:
Revised Date:
Accepted Date:
6 May 2020
1 June 2020
6 June 2020
Please cite this article as: Shaaban, M.A., Elshaier, Y.A.M., Hammad, A.H., Farag, N.A., Hassan Haredy, H.,
AbdEl-Ghany, A.A., Mohamed, K.O., Design and Synthesis of Pyrazolo[3,4-d]pyrimidinone Derivatives:
Discovery of Selective Phosphodiesterase-5 Inhibitors, Bioorganic & Medicinal Chemistry Letters (2020), doi:
https://doi.org/10.1016/j.bmcl.2020.127337
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
©
2020 Published by Elsevier Ltd.

Design and Synthesis of Pyrazolo[3,4-d]pyrimidinone Derivatives: Discovery of
Selective Phosphodiesterase-5 Inhibitors
a
b,c
d
e
Mohamed A. Shaaban , Yaseen A. M. M. Elshaier , Ali H. Hammad , Nahla A. Farag ,
f
g,h
a,i
Haredy Hassan Haredy , Ahmed A. AbdEl-Ghany , Khaled O.Mohamed .
aPharmaceutical Organic Chemistry Department, Faculty of Pharmacy, Cairo University,
1
1562,Egypt.
bPharmaceutical Organic Chemistry Department, Faculty of Pharmacy, Sadat University,
3
2958, Sadat City, Menoufia, Egypt.
dPharmaceutical Organic Chemistry Department, Faculty of Pharmacy (Boys), Al-Azahar
University, Cairo, 11884,Egypt.
ePharmaceutical Chemistry Department, Faculty of Pharmacy, Misr International
University, Cairo Egypt, 11431, Egypt.
fDepartment of Pharmacology, Faculty of Medicine, Al-Azher University, 71524, Assiut,
Egypt
gBiochemistry Department, Faculty of Pharmacy, Al-Azher University, 71524, Assiut,
Egypt.
hBiochemistry Department, Faculty of Pharmacy, Nahda University, Benisuif, Egypt
i Corresponding authors.
E-mail address: khaled.mohamed@pharma.cu.edu.eg
Tel.: +201145822851
1

cCorresponding co-auther.
E-mail address: yaseenelshaier@azher.edu.eg
Tel.: +201028582031
Abstract
A novel series of 1,6-disubstituted pyrazolo[3,4-d]pyrimidin-7-one derivatives,2a-h, 4a-
d, 5 and 6, were successfully synthesized, which showed promising, and potent inhibition
of phosphodiesterase 5 (PDE5).The inhibitory activities of 5, 4b, 2a, 2d, 2f, 4d and 4a
against PDE5 were similar to that of sildenafil (100%). These compounds exhibited
potent relaxant effects on isolated rat cavernosum tissue with pEC values ranging from
5
0
8
.31-5.16 µM. Pyrazolo[3,4-d]pyrimidin-7-one scaffolds have been rationally designed
via consecutive molecular modelling studies prior to their synthesis and biological
evaluation. In addition, the results of the pharmacophore-based virtual screening revealed
that 1v0p_PVB might have promising activity as a PDE-5 inhibitor.
Keywords
pyrazolo[3,4-d]pyrimidin-7-one,phosphodiesterase
5
(PDE5),
sildenafil,
cavernosum tissue, molecular modelling, PDE-5 inhibitors.
1
. Introduction
Several orally available phosphodiesterase 5 (PDE5) inhibitors have been developed and
are prescribed as first-line therapies for the treatment of erectile dysfunction (ED) [1].
2

The PDEs are a superfamily contain 11 subfamilies (PDE1-11) and each PDE plays an
important role in various tissues [2-6].
Treatment of ED using PDE5 inhibitors occurs via catalyse the inactivation of
intracellular mediators of signal transduction such as cAMP and cGMP. PDEs cleave the
3
′,5′-cyclicphosphate moiety of cAMP and/or cGMP to produce the corresponding 5′-
nucleotide (inactive forms). PDEs are critical modulators of cellular levels of
cAMPand/or cGMP based on many stimuli [6,7].
Erection results when the muscarinic receptors are activated by acetylcholine, which
leads to increased production, release and diffusion of nitric oxide into vascular smooth
muscle cells activates guanylate cyclase, resulting in increased synthesis of cyclic
guanosine monophosphate (cGMP), leading to muscle relaxation and vasodilatation of
the penis tissues [8,9] (Figure 1).
Figure 1.Schematic representation of the structure and function of the α1β1 isoform of
soluble guanylate cyclase (sGC).
3

PDE5 inhibitors protect cGMP from degradation by PDE5, which is specific to cGMP, in
the corpus cavernosum. Nitric oxide (NO) in the corpus cavernosum of the penis binds to
guanylate cyclase receptors, which results in increased levels of cGMP, decreasing the
intracellular calcium concentration, leading to smooth muscle relaxation (vasodilatation)
of the intimal cushions of the helicine arteries. This vasodilatation increases the inflow of
blood into the sponge tissue of the penis, causing an erection. Sildenafil is a potent and
selective inhibitor of cGMP-specific PDE5, which is responsible for the degradation of
cGMP in the corpus cavernosum [8, 9].
The molecular structure of sildenafil is similar to that of cGMP, and it acts as a
competitive binding inhibitor of PDE5 in the corpus cavernosum, resulting in more
cGMP and better erections. Without sexual stimulation and therefore no activation of the
NO/cGMP system (Figure 1), sildenafil should not cause an erection. Other drugs that
®
®
operate by the same mechanism include vardenafil (Levitra ) and tadalafil (Cialis )
Figure 2).
(
O
O
O
H
N
N
NH
N
NH
N
O
O
N
N
N
N
N
N
N
H
O
SO2
SO2
O
N
N
O
Sildenafil (Viagra®)
Vardenafil (Levitra®)
Tadalafil(Cialis ®)
Figure2.Structures of PDE5 inhibitors.
4

Despite its potency, sildenafil has many clinical side effects, such as headache, blurred
vision, facial flashing, visual disturbances, back pain and cardiovascular side effects [2,
1
0-12]; therefore, discovery of a novel PDE5 inhibitor with an improved safety profile
would be a welcome contribution to ED management. [13-15].
In this study, we report the discovery of novel and potent PDE5 inhibitors to potentiate
the relaxant effect. Three series of pyrazolo[3,4-d]pyrimidin-7-ones bearing 1,6-
disubstituted aryl moieties were designed using consecutive computer-aided protocols
and then synthesized. The details of the computational design protocols, synthesis and
structure-activity relationships are described. Moreover, the in vitro PDE inhibitory
activities of these compounds were evaluated using sildenafil as a reference drug, and
their relaxant effects on isolated rat corpus cavernosum tissue were determined.
2
. Results and discussion
.1. Molecular Modeling Studies
Computer-aided drug design (CADD) is the science of using computers to optimize
2
commercially available drugs [lead compounds] to identify new ligands with potential
biological activities against certain biological targets. CADD is based on quantum
mechanics and molecular modelling techniques. The traditional method of discovering
new drugs is a very hectic process that consumes substantial amounts of effort, money
and time [16-18]. Computer-aided programs help to narrow the infinite number of
options for lead optimization. Therefore, only the drug compounds with promising
activities will be synthesized.
5

In the current study, we designed a pyrazolo[3,4-d]pyrimidin-7-one scaffold as a novel
PDE5 inhibitor via consecutive protocols of molecular modelling. First, we studied the
docking of sildenafil in the binding site of phosphodiesterase 5 and then generated
predictive pharmacophore models using a 3D quantitative structure-activity relationship
(3D-QSAR) pharmacophore generation protocol from a set of ligands involving 24
compounds with known activities on our biological target (PDE-5). The generated
pharmacophores were validated to select the best model for the virtual screening of
different databases for searching of additional sildenafil analogues. Then, the resulting
hits were analysed. The hits were then docked into the binding site of the crystal structure
of PDE-5 obtained in complex with sildenafil as an inhibitor (protein data bank code
(2H42) to clarify the key structural features required in the design of novel candidates of
this class. The compatibility between the fit values of the pharmacophore model and its
docking affinity indicates the biological binding mode, offering a relevant basis for the
development of new PDE-5 inhibitors.
2
.1.1. Docking study of sildenafil
Molecular docking was performed using Accelrys DiscoveryStudio 4.0 software using
the Dock Ligands (CDOCKER) protocol, which is an implementation of the CDOCKER
algorithm. CDOCKERis a grid-based molecular docking method that employs a
CHARMm-based molecular dynamics (MD) scheme to dock ligands into a receptor-
binding site. Random ligand conformations are generated using high-temperature MD.
This method allows a refinement docking of any number of ligands with a single protein
receptor. Various scoring functions were applied to the ligands, including -
CDOCKER_ENERGY (CHARMm energy: interaction energy plus ligand strain) and -
6

CDOCKER_INTERACTION_ENERGY (interaction energy only). The poses were
sorted by CHARMm energy, and the top scoring poses were calculated.
The molecular docking was based on the crystal structures of PDE-5 (PDB ID: 2H42)
[19], which is a crystal structures of PDE-5 complexes with sildenafil (5-[2-ethoxy-5-[(4-
methylpiperazin-1-yl)sulfonyl]phenyl]-1-methyl-3-propyl-1H,6H,7H-pyrazolo[4,3-
d]pyrimidin-7-one) as the ligand. The protein was prepared using a prepare protein
parameter that cleans up common problems in the input protein structure in preparation
for further processing; it inserts and minimizes missing atoms in residue as hydrogen, and
it also removes alternate conformations. All of the water molecules are removed.
Sildenafil was prepared using the ligand preparation tool, which adds hydrogen's, fix bad
valences and generates 3D coordinates using a catalyst. Docking was performed using the
CDOCKER protocol, where conformational search tools used in discovery studio
software for docking of small flexible to a macromolecule target of known structure is
simulated annealing algorithms for conformation searching with a rapid grid‐based
method of energy evaluation. The pose cluster radius was set to 0.5 Å, the 10 top hits
were used, and other docking parameters were kept at their default values. The results
were analysed according to their -CDOCKER_ENERGY values. The docking protocol
itself was validated through the re-docking of the co-crystallized structure of the
reference ligand (Sildenafil). Then, the resulting docking results/poses are compared to
that of the crystallized reference ligand. This comparison was carried out by calculating
the RMSD (root mean square deviation), and a RMSD of 1.144 was found, confirming
the validity of the CDOCKER protocol.
7

The docking score of sildenafil is -12.41, it showed a CDOCKER_Interaction_Energy of
56.04, and it forms four hydrogen bonds: two hydrogen bonds to the C=O groups of the
-
pyrimidine moiety, one between the CH of the ethoxy moiety and glutamine 817, and one
hydrogen bond between the N-methyl of the piperidine ring and asparagine 662.
Additionally, the planar regions of pyrazolopyrimidine form hydrophobic interactions
with the hydrophobic pocket formed by amino acids valine 782, phenylalanine 786, and
phenylalanine 820; the phenyl ring forms another van der Waals interaction with leucine
8
04, alanine 783, and methionine 816; and the CH of the ethoxy moiety forms
hydrophobic interactions with isoleucine 813.These interactions resembled the reported
binding mode of sildenafil with PDE5 (2H42) but with additional hydrogen bonds
(Figure 3).
8

Figure 3. Docking of Sildenafil inside binding site of PDE5 (PDB code 2H42) in
3
D style
9

2
.1.2. 3D-QSAR pharmacophore model generation and validation
2
.1.2.1. Generation of the PDE5 inhibitor library
A set of 24 PDE-5 inhibitors with known activities (IC values ranging from 0.9–54 µM)
5
0
were collected from the literature [20] (Figure 4). The ligands in the library were
prepared using Accelrys Discovery Studio 4.0 software. Ligand preparation is essential
for protonation, minimization of energy, fixing bad valences and converting the 2D
structures of the ligands into 3D structures.
N
N
N
N
N
NH
NH
NH
NH
O
N
O
N
O
N
O
N
HN
N
N
HN
N
N
HN
N
N
HN
N
N
Cl
Cl
Cl
Cl
OH
OH
O
O
OH
OH
O
O
T1 = 6.5
T2 = 1.2
T3 = 54
T4 = 63
OH
O
N
N
N
N
NH
NH
NH
NH
O
O
N
O
N
O
N
N
HN
N
N
HN
N
N
HN
N
N
Cl
HN
N
N
Cl
Cl
Cl
OH
O
OH
OH
O
O
OH
O
T5 = 25
T6 = 52
T7 = 6.6
T8 = 0.9
1
0

HO
O
N
O
N
O
N
NH
NH
NH
N
O
N
NH
N
O
N
HN
N
O
N
HN
N
N
Cl
O
N
HN
N
O
Cl
O
Cl
HN
N
N
O
OH
O
Cl
OH
O
T9 = 6.1
T10 = 1.8
T11 = 12
T12 = 15
OH
HO
HO
O
N
NH
NH
O
N
NH
NH
O
N
N
HN
N
O
O
N
O
N
N
HN
N
O
Cl
O
N
N
HN
N
O
Cl
O
HN
N
O
Cl
Cl
O
O
T13 = 5.6
T14 = 13
T15 = 12
T16 = 3.5
N
N
O
N
N
O
N
HN
HN
O
N
O
N
O
N
O
N
OH
N
N
HN
N
N
N
HN
N
O
HN
N
O
HN
N
O
Cl
O
Cl
O
Cl
Cl
O
O
T17 = 4.2
T18 = 2.1
T19 = 2.7
S1 = 32
N
N
N
N
HN
HN
HN
HN
O
N
O
N
O
N
N
O
N
N
N
O
HN
N
N
Cl
O
Cl
O
HN
N
N
HN
N
N
OH
Cl
O
Cl
O
O
O
HO
OH
S2 = 32
S3 = 59
S4 = 100
S5 = 85
1
1

Figure 4.Training set T1-T19, test set S1-S5, and their PDE-5 inhibitory activity inµM
for 3D-QSAR pharmacophore model generation.
2
.1.2.2. 3D-QSAR pharmacophore model generation
The term pharmacophore refers to the functional groups essential for the biological
activity of a compound [21]; therefore, it is a model generated depending on the
relationship between the structure and the activity of a compound (3D-quantitative
structure-activity relationship pharmacophore model). The generation of
a
pharmacophore model can be either ligand-based or structure-based. Ligand-based
pharmacophores are generated according to structural features that are common between
active compounds in a library, and common features of inactive compounds are omitted.
In the present study, a ligand-based pharmacophore model was generated using Accelrys
Discovery Studio 4.0 software using a HypoGen algorithm. The collected library of 24
PDE5 inhibitors with known activities [20] was divided into T 1-19 compounds and S 1-5
(Figure 4). The chemical features selected for developing the protocol were hydrogen
bond acceptors (HBAs), hydrogen bond donors (HBDs), hydrophobic moieties (HYPs),
positive ionizable groups (PIs) and aromatic rings (ARs). The uncertainty was set as1.5
instead of the default value of 3, and IC was selected as the measure of activity.
5
0
1
2

2
.1.2.3. Validation of the pharmacophore models
The generated pharmacophore model must be validated. The validity of any model is
based on its statistical significance and ability to estimate the biological activity of
unknown compounds. The model must be validated based on cost difference, the
accuracy of the predicted activity and the mapping of a reported reference compound.
The HypoGen algorithm generates different hypotheses, and the best hypothesis is
identified based on cost difference [21]. The total cost of each hypothesis is the
summation of three factors: weight, error and configuration cost. The fixed cost, the
minimum possible cost of a model that fits perfectly, and the null cost, the maximum cost
of a model that has no features, were also calculated by the algorithm.
Cost difference = (null cost –fixed cost) - (total cost – fixed cost)
For the model to be statistically significant, the difference between the null and fixed
costs must be greater than that between the total and fixed costs. In addition, the cost
difference between the null and total costs must be between 40 and 60. The hypothesis
with the highest cost difference is the one that will be chosen (Table 1).
Table 1. Ten generated hypotheses with their Maximum fit, total cost, Cost difference,
RMS, and Correlation coefficient.
Hypothesis
Maximum Fit Total Cost
Cost
RMS
Correlatio
n
difference
Coefficient
0.955074
1
7.44472
73.6690
45.808
0.84102
1
3

2
3
4
5
6
7
8
9
6.78785
7.22863
5.38059
7.30776
6.41761
8.46594
9.25339
5.73834
6.93117
85.9444
88.7377
91.5160
92.7914
93.3148
94.2733
94.3107
95.9774
97.0297
33.533
30.739
27.961
26.686
26.162
25.204
25.166
23.500
22.447
1.44232
1.55108
1.64170
1.68426
1.70233
1.73268
1.52816
1.73347
1.80526
0.857448
0.832749
0.810324
0.799066
0.794165
0.785713
0.857580
0.786242
0.764812
1
0
The pharmacophore model also calculates an estimated activity for the compounds in the
training set (Table 2). Additionally, for the model to be valid, the estimated activity of
each compound in the training set must be close to its reported experimental activity (IC50
value).
Further validation of the pharmacophore model can be carried out by mapping a
reference ligand (sildenafil) into the generated model and comparing its fit value to that
of the tested library. The reference ligand must have a known biological activity on the
same target.
Table 2. Training set with their experimental IC value, their estimated activity and fit
5
0
value by the valid hypothesis 1.
1
4

Training set
T8
Experimental IC50
Estimated IC50
0.932396
1.27609
2.27806
2.28145
2.30912
3.33562
5.38998
5.79737
5.84478
6.10227
7.30281
9.04331
9.25853
10.2543
13.3901
18.2587
21.8454
26.4795
52.7211
Fit value
6.8514
0.9
1.2
5.6
2.1
1.8
2.7
4.2
3.5
6.1
6.6
5.2
6.5
12
T2
6.71512
6.46343
6.46279
6.45755
6.29782
6.08941
6.05777
6.05423
6.03551
5.95751
5.86467
5.85446
5.81009
5.69422
5.55953
5.48164
5.39809
5.09902
T13
T18
T10
T19
T17
T16
T9
T7
T6
T1
T11
T15
T12
T14
T5
12
15
13
25
T3
54
T4
63
1
5

Ten hypotheses were generated, and the analysis of these hypotheses revealed that the
first model was the best. This model showed the highest cost difference of 45.808, which
indicates the statistical significance of the model with a high predictive probability of 75-
9
0% and that it is well correlated with the data. Its RMS value of 0.841and correlation
coefficient 0.955 confirming that this model is suitable for predicting the activity of novel
compounds (Table 1).
In addition to cost analysis, the model was validated by comparing the activity predicted
for each compound by the model to their experimental IC values. As shown in Table 2,
5
0
the difference between the estimated and reported IC50 values were quite small,
confirming the validity of the model. The valid model has one hydrogen bond acceptors
(HBA), and two ring aromatic features (RA), (Figure 5a, Table 3).
(a)
(b)
Figure 5a. The best-generated pharmacophore model with the features considered
constraint distances and angles between features, considered hydrogen bond donor
HBA) colored in green, ring aromatic (RA) colored in orange, 5b.Sildenafil viewed as fit
reference compound with fit value of 6.5, estimated IC of 2.08 µM.
(
5
0
1
6

Table 3. The valid model of three pharmacophoric features one hydrogen bond acceptors
HBA), and two ring aromatic features (RA) with interfeature distances in angestrum and
(
angles.
Constraint distances (Ao)
Constraint angles (o)
(HBA_1) – (RA_2), 3.064;
(HBA_1) – (RA_3), 7.406
(RA_2)-(RA_3), 5.150
(HBA_1) – (RA_3) – (RA_2), 126.85
2
.1.2.4. Pharmacophore mapping of sildenafil
To further validate the model, an active reference drug (sildenafil) was mapped inside the
generated model using the ligand-mapping protocol of Accelrys Discovery Studio
4
.0software, and scored a fit value of 6.5 and an estimated IC of 2.08 µM were obtained
50
(Figure 5b).
2
.1.3. Virtual screening
A virtual screening was conducted to find novel and diverse virtual leads suitable for
further optimization as PDE-5 inhibitors. This screening was carried out using the data
base searching protocol of Accelrys Discovery Studio 4.0 software with three databases,
Drug like diverse, MiniMaybridge, and scPDB, using the validated pharmacophore
1
7

model. The significant hits that satisfied the features of the pharmacophore model were
then filtered according to their fit values, which yielded 3617 compounds. Compounds
with high fit values and estimated activities higher than that of sildenafil were chosen.
Five compounds were of particular interest because they fit the pharmacophore model
(7.25-7.03) better than mapped sildenafil 6.5, and these five compounds were
1
v0p_PVB, LIF172331, SPB 02245, CDI685470, and 3pmw_G69 (Table-4).
Table 4. Comparison between reference Sildenafil and 5 selected top HITs from virtual
screening with their pharmacophore fit values, and docking scores, considered
pharmacophore features, hydrogen bond donor (HBA) colored in green, ring aromatic
(
RA) colored in orange.
Selected 2D structure of 5 top Mapping on the
top HITs HITs
Estimat Vit
-
- Cdocker
validated
e
valu Cdocker_ _Interactio
pharmacophore
model Hypothesis 1
activity
µM
e
E
n_Energy
59.25
1
v0p_PV
0.36
7.25 42.57
HO
NH
B
O
N
N
OH
N
N
H
N
Cl
F
LIF17233
1
0.61
7.03 35.97
53.92
N
O
S
HN
O
O
HN
1
8

SPB
0.58
0.27
7.05 21.90
37.75
52.44
N
O
0
2245
NH2
F
O
S
N
F
F
CDI68547
0
7.37 15.04
N
N
O
N
O
NH
O
O
3
9
pmw_G6
0.36
2.08
7.26 9.33
46.70
56.04
HO
S
O
F
HN
F
F
N
N
O
Sildenafil
6.5
12.41
N
O
HN
O
O
NH
S
N
N
N
The top 5 selected hits from the virtual screening were then docked in the binding site of
the prepared PDE5 crystal structure (PDB code 2H42) [18]. The docking results of the
five selected hits from the pharmacophore-based virtual screening were analysed and
filtered according to their binding affinity and binding mode.
1
9

2
2
.1.4. Docking studies of top hits from the virtual screening
.1.4.1. Docking Study of 1V0P_PVB
docking score of 1V0P_PVB
The
was
-42.57,
it
showed
a
CDOCKER_Interaction_Energy of -59.25, and its forms six hydrogen bonds; three
hydrogen bonds between the NH and OH moieties of the side chain to glutamine 817, one
from the carboxylic acid moiety to asparagine 662, and another two hydrogen bonds with
arginine 667. Additionally, the planar regions of imidazolopyrimidine form hydrophobic
interactions with the hydrophobic pocket formed by the amino acids valine 782,
phenylalanine 786, and phenylalanine 820, and the phenyl ring forms a van der Waals
interaction with leucine 725, arginine 667, and isoleucine 665. Additionally, the alkyl
side chains form hydrophobic interactions with alanine 783, alanine 779, and isoleucine
8
13, which may increase the affinity of the compound for the binding site of PDE5,
resulting in a higher predicted inhibitory activity (Figure 6a).
(a)
(b)
2
0

(c)
(d)
(e)
Figure 6a. 3D image of compound 1v0p_PVB inside 2H42 using Accelrys Discovery
Studio 4.5; with docking score of- 42.5735, 6b. Docking score of LIF172331 is of -
3
5.97, 6c. Docking score of SPB 02245is of -21.90. 6d. Docking score of CDI685470 is
of -15.04, 6e. Docking score of 3pmw_G69 is of -9.33all figures reveal the hydrogen
bonds and hydrophobic interaction with the key residues of the receptor. (For
interpretation using colors, the reader is referred to the web version of this article).
2
.1.4.2.
Docking Study of LIF172331.
The docking score of LIF172331 is-35.97, it showed a CDOCKER_Interaction_Energy
of -53.92, and it forms four hydrogen bonds: one hydrogen bond between the oxygen
atom of a carbonyl with glutamine 817, one from the p-fluoro moiety to histidine 685,
and two additional hydrogen bonds to threonine723. Additionally, the planar regions of
indole form hydrophobic interactions with the hydrophobic pocket generated by leucine
7
25, histidine 685, threonine723, and the phenyl ring forms another van der Waals
2
1

interaction with valine 782, phenylalanine 786, isoleucine 813 and methionine 816. These
interactions may increase the affinity of the compound for the binding site of PDE5,
resulting in a high predicted inhibitory activity (Figure 6b).
2
.1.4.3.
Docking Study of SPB 02245
The docking score of SPB 02245 is of -21.90, it showed a CDOCKER
Interaction_Energy of -37.75, and it forms two hydrogen bonds; one hydrogen bond
between the trifluoromethyl moiety and serine 663, and one between the N of the
pyridine moiety with histidine 613.Additionally, the phenyl ring forms hydrophobic
interactions with the hydrophobic pocket formed by leucine 804and isoleucine 665; the
imidazole moiety forms a van der Waals interaction with phenylalanine 786 and leucine
7
25; and the pyridine ring forms another van der Waals interaction with leucine 725.
These interactions may increase the affinity of the compound for the binding site of
PDE5, resulting in a high predicted inhibitory activity (Figure 6c).
2
.1.4.4.
Docking Study of CDI685470
The docking score of CDI685470 is -15.04, it showed a CDOCKER_Interaction_Energy
of -52.44, and it forms seven hydrogen bonds; four hydrogen bonds between the carbonyl
of amide with asparagine 662, and another three hydrogen bonds with histidine 657 and
histidine 685. Additionally, the pyrazolopyridine and the phenyl ring form hydrophobic
interactions with the hydrophobic pocket formed by phenylalanine 820, leucine 804,
methionine 816, and phenylalanine 786.These interactions may increase the affinity of
the compound for the binding site of PDE5, resulting in a high predicted inhibitory
activity (Figure 6d).
2
2

2
.1.4.5.
Docking Study of 3pmw_G69
The docking score of 3pmw_G69 is of -9.33, it showed a CDOCKER Interaction_Energy
of -46.70, and it forms four hydrogen bonds; three hydrogen bonds between the O of the
sulfonyl with asparagine 662, leucine 725, and arginine 667, and one between the
hydroxyl group and valine 782. Additionally, the phenyl ring of the planar portion of 2,3-
dihydro-1H-indene forms a van der Waals interaction with the hydrophobic pocket
formed by phenylalanine 786 and phenylalanine 820, and the pyrazole ring forms another
van der Waals interaction with valine 782, leucine804, and methionine 816. These
interactions may increase the affinity of the compound for the binding site of PDE5,
resulting in a high predicted inhibitory activity (Figure 6e).
2
.1.5. Pharmacophore mapping of the proposed analogues
The proposed analogues were mapped into the validated pharmacophore model to
determine their fitting and docked into the crystal structure of PDE5 in complex with a
natural inhibitor (sildenafil, protein data bank code 2H42) to elucidate the key structural
features required for the inhibition of PDE5 Table 4. The compatibility between their fit
values with the pharmacophore model and their docking affinity indicates the biological
binding mode, offering a relevant basis for the further development of new PDE-5
inhibitors.
2
.1.6. Docking studies of the proposed analogues
In general, a docking study on proposed compounds 2a-h,4a-d,5 and 6 revealed
CDOCKER _Energy values from -11.41 to -2.28, CDOCKER_Interaction_Energy values
from -56.04 to-39.79 (Table 5), and hydrogen bonds between the C=O and NH moieties
2
3

of the pyrimidine with glutamine 817. Additionally, the planar region of
pyrazolopyrimidine formed hydrophobic interactions with the hydrophobic pocket
generated by valine 782, phenylalanine 786, and phenylalanine 820; the phenyl ring
forms a van der Waals interaction with leucine 804, alanine 783, and methionine 816and
a hydrophobic interaction with isoleucine 813.
Table 5. Mapping of sildenafil and the proposed analogues to pharmacophore hypothesis
1
, and their docking scores in the binding site of PDE-5 crystal structure.
Cpd No
-
CDocker_
Energy
- CDocker_Interaction
_Energy
Fit value by mapping of the
proposed analogues to
Estimated activity with
µM using
pharmacophore hypothesis 1
pharmacophore
hypothesis 1
Sildenafil
12.41
11.41
8.87
8.50
7.98
7.06
5.87
5.87
5.53
5.18
4.82
4.30
3.53
56.04
56.20
49.92
44.44
44.46
47.38
44.76
46.95
46.42
43.65
41.20
43.81
43.89
6.50
5.49
5.48
5.63
5.63
5.48
5.49
5.48
6.06
5.64
5.54
6.04
5.48
2.80
21.40
21.60
15.46
15.46
21.64
21.38
21.51
5.76
2
h
2
4
c
c
4
d
2
f
2
e
2
g
5
4
b
15.04
18.86
5.99
4
a
6
2
b
21.47
2
4

2
d
3.36
2.28
44.08
5.48
5.48
21.46
21.57
2
a
39.798
2
.1.6.1. Docking studies of proposed analogue 2h
The docking score of compound 2h
is
-11.41,
it
showed
a
CDOCKER_Interaction_Energy of -56.20, and it forms four hydrogen bonds: three
hydrogen bonds between the C=O and NH moieties of the pyrimidine moiety with
glutamine 817, and one between the N of the pyrazole ring and tyrosine 612.
Additionally, the planar region of pyrazolopyrimidine forms hydrophobic interactions
with the hydrophobic pocket formed by valine 782, phenylalanine 786, alanine 783,
phenylalanine 820, and alanine 767; the N-phenyl ring forms a van der Waals interaction
with leucine 765 and additional interactions with histidine 613; and the p-bromo moiety
interacts with leucine 725. The 6-phenyl moiety forms hydrophobic interactions with the
hydrophobic pocket generated by phenylalanine 786, leucine 804, methionine 816.In
addition, the CH of the butoxy moiety forms hydrophobic interactions with isoleucine
8
7
2
6
13and interactions with phenylalanine786, phenylalanine787, and alanine783 (Figure
a).The docking study showed that the binding mode of 2h inside PDE5 (protein code
H42) resembles that reported for sildenafil with additional hydrogen bonds with tyrosine
12 due to variations in the ring. The replacement of the propyl moiety with an aromatic
ring strengthens the interactions with the hydrophobic pocket generated by leucine 765
and histidine613. Moreover, the replacement of the 2-alkoxy moiety with a long chain
substituent strengthens the hydrophobic interactions with isoleucine 813and the
interactions with phenylalanine786, phenylalanine787, and alanine783.
2
5

(a)
(b)
Figure 7a. 3D image of compound 2hinside 2H42usingAccelrys Discovery Studio 4.5;
with Cdocker energy of -11.41, 7b. Docking score of 5 is of -5.53, all figures reveal the
hydrogen bonds and hydrophobic interaction with the key residues of the receptor. (For
interpretation using colors, the reader is referred to the web version of this article).
2
.1.6.2.
Docking studies of the designed proposed analogue 5
The docking score of compound
5
is of -5.53, it showed
a
CDOCKER_Interaction_Energy of -46.42, and it forms five hydrogen bonds; two
hydrogen bonds between the C=O and NH groups of the pyrimidine moiety, one between
the H of the side chain with glutamine 817, one between the N of the pyrazole ring with
tyrosine 612, and two from the CH group of the ester with asparagine 662. Additionally,
3
the planar portion of the pyrazolopyrimidine forms hydrophobic interactions with the
hydrophobic pocket formed by valine 782, phenylalanine 786, alanine 783, phenylalanine
8
7
20, and alanine 767; the N-phenyl ring forms a van der Waals interaction with leucine
65 and an additional interaction with histidine 613; and the 6-phenyl moiety forms
2
6

hydrophobic interactions with the hydrophobic pocket generated by leucine 804 (Figure
b).
7
The docking study showed that the binding mode of compound 5 inside PDE5 (2H42)
resembles that reported of sildenafil along with an additional hydrogen bond with
tyrosine 612 due to variations in the ring. The replacement of the propyl moiety with an
aromatic ring strengthens the interactions with the hydrophobic pocket formed by leucine
7
65, histidine613, and leucine 725.In addition, the 2-methyl ester forms two hydrogen
bonds with asparagine 662.
.2. Design Strategy
2
Different protocols in Accelrys Discovery Studio software version 4.0 were used to
design new scaffolds of PDE5 inhibitors and estimate their activities prior to their
synthesis. In the current study, we reported molecular modelling studies on PDE5
inhibitors using the 3D-quantitative structure-activity relationship (3D-QSAR)
pharmacophore generation protocol, which can be used to generate predictive
pharmacophore models from a set of compounds with known activities on PDE5[18].The
generated pharmacophore hypotheses were validated to select the best model for the
virtual screening of different databases and to predict the activity of the proposed
structures to be synthesized.
According to our molecular modelling studies, sildenafil was modified to design new
potent PDE5 inhibitors (Figure 8, 9).
a. The ring was varied from pyrazolo[3,4-d]pyrimidine-4-one to pyrazolo[4,3-
d]pyrimidine-7-one to study the effect of changing the site of fusion where it
2
7

forms three hydrogen bonds between the C=O and NH groups of the pyrimidine
moiety with glutamine 817 and an additional hydrogen bond between the N of the
pyrazole ring and tyrosine 612. Additionally, planar portion of the
pyrazolopyrimidine forms hydrophobic interactions with the hydrophobic pocket
generated by valine 782, phenylalanine 786, alanine 783, phenylalanine 820, and
alanine 767.
b. The propyl moiety in the pyrazole ring was replaced with an aryl moiety, resulting
in stronger hydrophobic interactions with the hydrophobic pocket generated by
histidine 613, leucine 765, leucine 725, and tyrosine.
c. The 2-ethoxymoiety was varied to strengthen the hydrophobic interactions with
the hydrophobic pocket generated by phenylalanine 786, phenylalanine 787,
leucine 804, isoleucine 813, and methionine 816.
d. Additionally, to study the effect of different conformational isomers and their
ligand receptor interactions, a linker was inserted at position 6 of the
pyrazolo[4,3-d]pyrimidine-7-one, and an aminomethyl moiety (CH NH) or its
2
bioisosteric methoxy (CH O) was installed on the phenyl moiety.
2
e. In addition, the 2-methyl benzoate ester forms extra hydrogen bonds with
asparagine 662.
Therefore, we designed and synthesized a new series of pyrazolo[3,4-d]pyrimidinones as
sildenafil analogues for the treatment of ED (Figure 8, 9).
2
8

Pharmacophoremodel generation
Virtual screening ofdata bases
The best-generatedpharmacophore model with the features
consideredconstraint distances and angles between features,
Training set T1-T19, test set S1-S5, and their PDE-5 inhibitory
activity inµM for 3D-QSAR pharmacophore model generation.
O
Mappingof sildenafil to
pharmacophoremodel
N
N
NH
O
N
O
N
Docking
studies and
mapping
hydrogen bond donor (HBA), two ring aromatic (RA) .Sildenafil
viewed as fit referencecompound with fit value of 6.5
R1
O
NH
N
SO2
N
N
N
R
R= H, Br
R1= H, n-propyl, isopropyl,n-butyl
Sildenafil
O
N
O
Synthesis
N
N
NH
O
R2
NH
H
N
N
N
N
S
N
O
O
R
R= H, Br
2= OH, NH2
R
Docking of Sildenafil inside binding
site of PDE5 (PDB code 2H42)
N
O
PDE5 enzyme
activity assay
Sildenafil
NH
R3
N
N
N
O
R
R3= H, Br
COOC2H5, COOH
R
=
Ex-vivo Study
relaxant effect
Sildenafil structure and design strategyof target compounds
Figure 8.Schematic representation to rational,and design of novel pyrazolo[3,4-
d]pyrimidin-7-ones scaffold; synthesis and biological evaluation.
2
9