Ligand-based modelling followed by synthetic exploration unveil novel glycogen phosphorylase inhibitory leads

Article abstract of DOI:10.1016/j.bmc.2011.06.086

Glycogen phosphorylase (GP) is a valid anti-diabetic target. Accordingly, we applied a drug discovery workflow to unveil novel inhibitory GP leads via combining pharmacophore modeling, QSAR analysis and in silico screening, followed by synthetic exploration of active hits. Virtual screening identified six low micromolar inhibitory leads from the National Cancer Institute (NCI) list of compounds. The most potent hits exhibited anti-GP IC₅₀ values of 3.2 and 4.1 μM. Synthetic exploration of hit 59 (IC₅₀ = 4.1 μM) yielded 25 lead inhibitors with the best illustrating IC₅₀ of 3.0 μM. Interestingly, we prepared several novel mixed oxalyl amide anti-GP leads employing new chemical reaction involving succinic acid-based adducts.

Full text of DOI:10.1016/j.bmc.2011.06.086

Bioorganic & Medicinal Chemistry 19 (2011) 4746–4771
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Ligand-based modelling followed by synthetic exploration unveil
novel glycogen phosphorylase inhibitory leads
Maha Habash a, Mutasem O. Taha b,
⇑
a Department of Pharmaceutical Chemistry and Pharmacognosy, Faculty of Pharmacy, Applied Science University, Amman, Jordan
b Drug Discovery Unit, Department of Pharmaceutical Sciences, Faculty of Pharmacy, University of Jordan, Amman, Jordan
a r t i c l e i n f o
a b s t r a c t
Article history:
Glycogen phosphorylase (GP) is a valid anti-diabetic target. Accordingly, we applied a drug discovery
workflow to unveil novel inhibitory GP leads via combining pharmacophore modeling, QSAR analysis
and in silico screening, followed by synthetic exploration of active hits. Virtual screening identified six
low micromolar inhibitory leads from the National Cancer Institute (NCI) list of compounds. The most
Received 19 May 2011
Revised 28 June 2011
Accepted 29 June 2011
Available online 7 July 2011
potent hits exhibited anti-GP IC50 values of 3.2 and 4.1
l
M. Synthetic exploration of hit 59 (IC50 = 4.1
lM)
yielded 25 lead inhibitors with the best illustrating IC50 of 3.0
novel mixed oxalyl amide anti-GP leads employing new chemical reaction involving succinic acid-based
adducts.
l
M. Interestingly, we prepared several
Keywords:
Glycogen phosphorylase
Pharmacophore
QSAR
Ó 2011 Elsevier Ltd. All rights reserved.
In silico screening
In vitro validation
Synthetic exploration
1. Introduction
glucose-6-P, (4) the purine nucleoside site, which binds heterocy-
clic compounds such as caffeine and flavopiridol, (5) the glycogen
Glycogen phosphorylase (GP) plays key role in the regulation of
blood glucose levels.1,2 Normally, GP mediates glycogenolysis,
which contributes 40–50% of overnight glucose production. In type
storage site, and (6) the dimer interface site, which binds indole
derivatives. Four of these sites are known to be inhibitor binding
sites: the catalytic site, the AMP allosteric site, the inhibition site
and the dimer interface site.5
2
diabetic patients, despite increased serum glucose levels,
phosphorylase-mediated glycogenolysis continues to contribute
40–50% of overnight glucose production, increasing the problem
of hyperglycemia.1,2
GP exists in two interconvertible forms (a and b); the propor-
tion of each form depends on the level of phosphorylation. The b
form is transformed by phosphorylation into a form.2,5,6 GPa, also
known as relaxed state, has higher substrate affinity compared to
GPb (Tense state).3,7
Inhibition of GPa limits glycogen degradation and reduces he-
patic glucose production, and therefore, should provide potential
for the treatment of type 2 diabetes.7 Accordingly, GP inhibitors
have received great recent interest as potential treatments for type
2 diabetes. For example, PSN-357 (Fig. 1) has been in phase I clin-
ical trials since July 2005 as potential treatment for type 2 diabe-
tes.9 Other interesting GPa inhibitors were developed by major
pharmaceutical companies, some of which are shown in Figure 2.3
The main focus of recent efforts towards the development
of new GPa inhibitors concentrate on structure-based ligand
design,2,3,5,6,8,10,1c,69 3D-QSAR studies (i.e., COMFA and
COMSIA)10–12,14 and high throughput screening.13
GP catalyzes the hepatic glycogenolysis process by converting
glycogen to glucose-1-phosphate, which is subsequently broken
down by glucolysis into pyruvate that enters into citric acid cycle.3
In humans, there are three isozymes of GP named according to
the tissues in which they are expressed: liver, muscle and brain.
The muscle and brain isozymes that provide their specified tissues
with its glucose demand, while the liver isozyme meets the glyce-
mic demands of the body as a whole. Several published reports
suggest that GP inhibition can lower blood glucose in diabetic
models, thus promoting GP as a potential therapeutic target for
treatment of type 2 diabetes.4,5
GP contains at least six potential regulatory sites: (1)
the Ser14-phosphate recognition site, (2) the catalytic site that
binds glycogen and glucose-1-P as well as glucose and glucose ana-
logues, (3) the AMP allosteric site, which binds AMP, IMP, ATP and
To date, 22 human liver GPa X-ray complexes are documented
in the Protein Data Bank (PDB codes: 3DD1, 3DDS, 3DDW, 2ZB2,
3CEH, 3CEJ, 3CEM, 2QLL, 2ATI, 1WVO, 1EXV, 1EM6, 1L7X, 1L5R,
1WV1, 1WUY, 1WUT, 1XOI, 1L5Q, 1L5S, 1FCO and 1FA9 with reso-
lution range: 1.80–3.3 Å). However, although considered the most
⇑
Corresponding author. Tel.: +962 6 5355000x23305; fax: +962 6 5339649.
E-mail address: mutasem@ju.edu.jo (M.O. Taha).
0968-0896/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2011.06.086

M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
4747
HO
The continued interest in designing new GP inhibitors1c com-
bined with the drawbacks of structure-based design prompted us
to explore the possibility of developing ligand-based three-
dimensional (3D) pharmacophore(s) integrated within self-
consistent QSAR model. This approach avoids the pitfalls of
structure-based techniques; furthermore, the pharmacophore mod-
el(s) can be used as 3D search queries to discover new GP inhibitory
scaffolds. We previously reported the use of this innovative ap-
proach towards the discovery of new inhibitory leads against glyco-
gen synthase kinase-3,20 bacterial MurF,21 protein tyrosine
phosphatase,22 DPP IV,23 hormone sensitive lipase,24 b-secretase,25
influenza neuraminidase,26 cholesteryl ester transfer protein,27
N
O
Cl
HN
F
N
N
O
H
CDK128, b- -galactosidase29 and Heat Shock Protein 90a.30
D
Figure 1. PSN-357 GPa inhibitor.
We employed the HYPOGEN module from the CATALYST
software package to construct numerous reasonable binding
hypotheses for GP inhibitors. Subsequently, genetic function algo-
rithm (GFA) and multiple linear regression (MLR) analyses were
employed to search for an optimal QSAR that combine high-quality
binding pharmacophores with other molecular descriptors and
capable of explaining bioactivity variation across a collection of
GP inhibitors. The optimal pharmacophores were further validated
reliable structural information that can be used for drug design,
crystallographic structures are limited by inadequate resolution15
and crystallization-related artifacts of the ligand–protein
complex.16–18 Moreover, crystallographic structures generally
ignore structural heterogeneity related to protein anisotropic
motion and discrete conformational substrates.19
OH
HO
O
N
N
O
O
Cl
Cl
HN
NH
F
N
O
NH
O
H
Pf
izer
Pfizer
O
NH
NH
NH
N
OH
O
OH
O
O
OH
NH
NH
Cl
Cl
Pfizer
Merck
Cl
S
N
N
O
H
N
S
F
Cl
N
H2N
N
H
H
O
O
N
H2N
AstraZeneca
SIMM
Figure 2. Some selected recently developed GPa inhibitors together with originating companies.3

4748
M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
by evaluating their abilities to successfully classify a list of com-
pounds as active or inactive by assessing their receiver–operating
characteristic (ROC) curves. The best pharmacophore was
employed to screen the national caner institute (NCI) list of com-
pounds. One of the most active hits was used to guide the synthesis
of two new series of potential GPa inhibitors based on oxalyl and
malonyl scaffolds. Two of the synthesized compounds displayed
low micromolar potencies.
CATALYST models drug–receptor interaction using information
derived only from the drug structure. HYPOGEN identifies a 3D
array of a maximum of five chemical features common to active
training molecules, which provides a relative alignment for each
input molecule consistent with their binding to a proposed com-
mon receptor site. The chemical features considered can be hydro-
gen bond donors and acceptors (HBD and HBA), aliphatic and
aromatic hydrophobes (Hbic), positive and negative ionizable
(PosIon and NegIon) groups and aromatic planes (RingArom). The
conformational flexibility of training ligands is modeled by creat-
ing multiple conformers, judiciously prepared to emphasize repre-
used as starting points for conformational analysis and in the deter-
mination of various molecular descriptors for QSAR modeling.
The conformational space of each inhibitor was extensively
sampled utilizing the poling algorithm employed within the
CONFIRM module of CATALYST.35 Conformational coverage was
performed employing the ‘Best’ module to ensure extensive sam-
pling of conformational space. Efficient conformational coverage
guarantees minimum conformation-related noise during pharma-
cophore generation and validation stages. Pharmacophore genera-
tion and pharmacophore-based search procedures are known for
their sensitivity to inadequate conformational sampling within
the training compounds.36
2.2. Exploration of GP pharmacophoric space
CATALYST-HYPOGEN enables automatic pharmacophore con-
struction by using a collection of at least 16 molecules with bioac-
tivities spanning over 3.5 orders of magnitude.37–42
Accordingly, since we have a list of 37 GP inhibitors of evenly
spread bioactivities over several orders of magnitude, we were
prompted to employ HYPOGEN to identify possible pharmaco-
phoric binding modes assumed by GP inhibitors within their corre-
sponding GP binding site (see Section S1 in Supplementary data).39
It was decided to explore the pharmacophoric space of GP
inhibitors under reasonably imposed ‘boundaries’ through twelve
HYPOGEN automatic runs (Table S2 in Supplementary data) and
employing three carefully selected training subsets (i.e., from the
collected compounds): subsets I, II and III, shown in Table 1. The
training compounds in these subsets were selected in such a way
to guarantee maximal 3D diversity and continuous bioactivity
spread over more than 3.5 logarithmic cycles. Furthermore, the
training inhibitors were selected in such a way that differences
in their anti-GP bioactivities are primarily attributable to the pres-
ence or absence of pharmacophoric features (e.g., HBA or HBD or
hydrophobic or Ring Aromatic) rather than steric shielding and/
or bioactivity-enhancing or -reducing auxiliary groups (e.g., elec-
tron donating or withdrawing groups). Special emphasis was given
to the 3D diversity of the most active compounds in each training
subset (Table 1) because these have significant influence on the ex-
tent of the evaluated pharmacophore space via the Constructive
Phase of HYPOGEN algorithm (see Section S1 in Supplementary
data).
sentative coverage over
a specified energy range. CATALYST
pharmacophores have been used as 3D queries for database
searching and in 3D-QSAR studies.21,31–33
2. Results and discussion
CATALYST-based automatic pharmacophore construction was
performed using a collection of molecules with biological activities
ranging from 3.5 to 4.0 logarithmic. CATALYST pharmacophores
(hypotheses) explain the variability of activity of the molecules
with respect to the geometric localization of the chemical features
present in the molecules used to build it. Different hypotheses
were generated for a series of GP inhibitors. A total of 37 com-
pounds were used in this study (1–37, Table S1 in Supplementary
data). Three training subsets were selected from the collection
(Table 1). Genetic algorithm and multiple linear regression statisti-
cal analysis were subsequently employed to select an optimal com-
bination of complementary pharmacophores capable of explaining
bioactivity variations among all inhibitors. Bioactive hits were used
to guide synthetic exploration for new GP inhibitors.
2.1. Data mining and conformational coverage
Guided by our reasonably restricted pharmacophoric explora-
tion concept, the software was restricted to explore pharmaco-
phoric models incorporating from zero to three features of
selected feature types (i.e., HBA, HBD, hydrophobic and Ring Aro-
matic), that is, instead of the default range of zero to five. Further-
more, to further limit the investigated pharmacophoric space, only
four and five-featured pharmacophores were explored. Three- and
two-featured pharmacophores are rather promiscuous as 3D
search queries and not adequate descriptions of ligand–GP binding.
Table S2 (in Supplementary data) shows the training sets and CAT-
ALYST run parameters employed in exploring GP pharmacophoric
space.
The literature was extensively surveyed to identify as many re-
ported structurally diverse GP inhibitors as possible. A dataset of
37 dihydroquinolinones (1–37, Table S1 in Supplementary data)
was used for pharmacophore modeling and subsequent QSAR anal-
ysis (1–37, see Table S1in Supplementary data). However, we were
forced to select this relatively limited list of compounds because
they satisfied two important prerequisites for pharmacophore
and QSAR modeling: (i) they were all bioassayed by a single proce-
dure. Consistency in bioassay is a major requirement for QSAR
modeling as it is not possible to model bioactivity data generated
via more than one bioassay procedure. (ii) They exhibit wide
bioactivity range, that is, over 4 logarithmic cycles.
In each run, the resulting binding hypotheses were automati-
cally ranked according to their corresponding ‘total cost’ value:
defined as the sum of error cost, weight cost and configuration cost
The 2D structures of the inhibitors were imported into CATALYST
and converted automatically into plausible 3D single conformer
representations. The resulting single conformer 3D structures were
Table 1
Training subsets employed in exploring the pharmacophoric space of GP inhibitors, numbers correspond to compounds in Table S1 in Supplementary data
Training set
Most active subseta
Intermediate subset
Least active subsetb
No. of compounds
I
II
III
7, 13, 15, 17, 31, 32, 33, 37
4, 7, 9, 17, 24, 31, 36, 37
4, 7, 9, 17, 24, 31, 32, 37
2, 12, 18, 19, 28
1, 2, 8, 12, 14, 21
1, 2, 12, 14, 18, 19, 21, 28
22, 25, 29
22, 25
22, 25, 29
16
16
20
a
Potency categories as defined by Eq. S1 in Supplementary data.
Potency categories as defined by Eq. S2 in Supplementary data.
b

M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
4749
Logð1=IC50Þ¼6:27þ0:22ðHypo12=7Þꢀ0:20logðZÞ1:35j2a
(see Section S2 in Supplementary data). Error cost provides the
r230 ¼0:69; F-statistic¼18:70; rB2S ¼0:69; r2LOO ¼0:58; rP2RESSð7Þ ¼0:91
highest contribution to total cost and it is directly related to the
capacity of the particular pharmacophore as 3D-QSAR model, that
is, in correlating the molecular structures to the corresponding bio-
logical responses. HYPOGEN also calculates the cost of the null
hypothesis, which presumes that there is no relationship in the
data and that experimental activities are normally distributed
about their mean. Accordingly, the greater the difference from
the null hypothesis cost (residual cost, Table 2), the more likely
that the hypothesis does not reflect chance correlation.37–42 Table
2 summarizes the pharmacophoric features and success criteria
of best GP representative hypotheses (see Section S2 in Supple-
mentary data).
Eventually, 120 pharmacophore models emerged from 12 auto-
matic HYPOGEN runs, out of which 84 models illustrated confidence
levels P85% (based on Fisher’s randomization test).43,61 These suc-
cessfulmodels were clustered and the best representatives (28 mod-
els) were used in subsequent QSAR modeling, as in Table 2.
Clearly from Table 2, all representative models shared compara-
ble features and acceptable statistical success criteria. Emergence
of several comparable pharmacophore models suggests the ability
of GP ligands to assume multiple pharmacophoric binding modes
within the binding pocket. Therefore, it is quite challenging to se-
lect any particular pharmacophore hypothesis to represent the
binding process.
ð1Þ
where r230 is the correlation coefficient, F is Fisher statistic, n is
the number of observations, r2BS is the bootstrapping regression
coefficient, r2LOO is the leave-one-out correlation coefficient and
r2PRESS is the predictive r2 determined for 7 randomly selected test
compounds.37,45 Log (Z) is natural logarithm of Hosoya connectiv-
ity index, while ja2 is the second order Kier’s alpha-modified shape
index.44 Hypo12/7 represents the fit values of the training com-
pounds against the 7th pharmacophore model generated in the
12th modeling runs, as in Tables 3 and 4. Figure 4 shows
Hypo12/7 and how it maps the most potent training inhibitor 9
(IC50 = 25 nM), while Table S3 (in Supplementary data) shows the
X, Y, and Z coordinates of the Hypo12/7.
Emergence of connectivity and shape indices in Eq. 1 illustrates
certain role played by the ligands’ topology in the binding process.
However, despite their predictive significance, their information
content is quite obscure.
On the other hand, emergence of Hypo12/7 in QSAR Eq. 1
suggests the existence of a single corresponding binding mode
assumed by inhibitors within the binding pocket of GPa. To further
validate this conclusion we compared this binding model with the
bound conformation of the co-crystallized ligand CP-403700 with-
in GP (PDB code: 1L5R, resolution 2.1 Å).
2.3. QSAR modeling
Figure 5 shows the chemical structure of CP-403700 and com-
pares its GP complex with the way it maps Hypo12/7 employing
rigid mapping, that is, by fitting the ligand’s bound state against
corresponding pharmacophore without conformational adjust-
ments. Fitting the carboxylate fragment of CP-403700 against
HBA feature in Hypo12/7 (Fig. 5b) corresponds to hydrogen-
bonding interaction connecting this fragment with H2O481, which
is inturn hydrogen-bonded to phenolic hydoxyl of Tyr226, as in
Figure 5a. Similarly, fitting one of the ligand’s carbonyls against a
HBA feature in Hypo12/7 corresponds to hydrogen-bonding inter-
action connecting the same carbonyl with the ammonium side
chain of Lys191 (Fig. 5a and b). While, mapping the benzyl ring
of CP-403700 against Hbic feature in Hypo12/7 (Fig. 5b) correlates
with placing this ring in a hydrophobic pocket comprised of the
side chains of His57, Phe53 and Pro188. On the other hand, map-
ping of the chlorinated indole ring of CP-403700 ligand against
two hydrophobic features in Hypo12/7 (Fig. 5b) agrees with posi-
tioning this fragment in a hydrophobic pouch consisting of the side
chains of Trp67, Val40, Val64, Pro229 and the butylenes linker of
Lys191 and (Fig. 5a).
Pharmacophoric hypotheses are important tools in drug design
and discovery as they provide excellent insights into ligand-
macromolecule recognition. Moreover, they can be used as 3D
search queries to mine for new biologically interesting scaffolds.
However, their predictive value as 3D-QSAR models is usually
limited by steric shielding and bioactivity-enhancing or -reducing
auxiliary groups.42 This point combined with the fact that phar-
macophore modeling of GP inhibitors furnished several binding
hypotheses of comparable statistical criteria (Table 2) prompted
us to employ classical QSAR analysis to search for the best com-
bination of pharmacophore(s) and other 2D descriptors capable of
explaining bioactivity variation across the whole list of collected
inhibitors (1–37, Table S1
in Supplementary data). Genetic
function approximation and multiple linear regression QSAR
(GFA–MLR-QSAR) analysis were employed to search for an
optimal QSAR equation(s).
GFA–MLR-QSAR selects optimal descriptor combinations based
on the Darwinian concept of genetic evolution whereby the
statistical criteria of regression models from different descriptor
combinations (chromosomes) are employed as fitness criteria.44
GFA–MLR-QSAR analysis was employed to explore various combi-
nations of pharmacophores and other structural descriptors and to
evaluate their statistical properties as predictive QSAR models.
The fit values obtained by mapping the 28 representative
hypotheses against all collected GP inhibitors (1–37, Table S1 in
Supplementary data) were enrolled as independent variables
(genes) in a cycle of GFA–MLR-QSAR analysis over 30,000 itera-
tions employing Friedman’s LOF fitness criterion (see Section
4.1.3).37,45 However, since it is essential to access the predictive
power of the resulting QSAR models on an external set of inhibi-
tors, 7 molecules were randomly selected and used as external test
molecules for validating the QSAR models (Table S1 in Supplemen-
tary data, see Section 4.1.3). Moreover, all QSAR models were
cross-validated automatically using the leave-one-out cross-
validation in CERIUS2.37,45
Clearly from the above discussion, Hypo12/7 represents a valid
binding mode assumed by ligand within GP. Furthermore, this
model points to limited number of critical interactions required
for high ligand–GP affinity in the binding mode. In contrast, crys-
tallographic complexes reveal many bonding interactions without
highlighting critical ones. Incidentally, Figure 5a only shows inter-
actions corresponding to pharmacophoric features while other
binding interactions were hidden for clarity.
2.4. Addition of exclusion volumes
Although ligand-based pharmacophores serve as excellent tools
to probe ligand/macromolecule recognition and can serve as useful
3D-QSAR models and 3D search queries, they suffer from a major
drawback: They lack steric constrains necessary to define the size
of the binding pocket. This liability might render pharmacophoric
models promiscuous. Therefore, we decided to complement our
QSAR-selected pharmacophore with exclusion spheres employing
HipHop-Refine module within CATALYST. Excluded volumes
Equation 1 shows the details of the optimal QSAR model. Figure
3 shows the corresponding scatter plots of experimental versus
estimated bioactivities for the training and testing inhibitors.

4750
M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
Table 2
The success criteria of representative GP pharmacophore hypotheses
f
Training Seta
Runb
Hypothesesc
Pharmacophoric features in
generated hypotheses
Total
cost
Config
cost
Cost of null
hypothesis
Residual
cost
Re
F-Statistic
Cat-scramble
(%)g
d
I
1
2
3
6
8
9
2
3
4
5
7
10
2
3
2
6
7
10
3
1
8
10
1
4
8
9
9
HBA, 3ꢁHbic, 5ꢁEV
73.9
75.4
80.9
81.3
81.6
73.6
79.1
79.4
80.6
81.5
85.3
75.6
76.5
76.4
79.5
79.5
79.9
73.6
90.4
101
15.8
15.8
15.8
15.8
15.8
14.1
14.1
14.1
14.1
14.1
14.1
15.8
15.8
14.0
14.0
14.0
14.0
13.2
15.8
15.8
15.8
14.02
14.02
14.02
14.02
13.3
10.124
10.124
100.9
100.9
100.9
100.9
100.9
100.9
100.9
100.9
100.9
100.9
100.9
94.6
94.6
94.6
94.6
94.6
94.6
94.6
120.9
120.9
120.9
120.9
120.9
120.9
120.9
120.9
120.9
120.9
27.0
25.5
20.0
19.6
19.3
27.3
21.8
21.5
20.3
19.4
15.6
19.0
14.1
18.0
15.1
15.1
14.7
21.0
30.5
19.9
18.9
32.0
21.0
18.8
18.5
23.8
29.5
28.8
0.98
0.98
0.88
0.90
0.91
0.96
0.92
0.91
0.94
0.86
0.82
0.94
0.93
0..92
0.87
0.87
0.86
0.93
0.94
0.83
0.82
0.94
0.83
0.81
0.80
0.85
0.88
0.87
26.9
2.9
19.2
34.7
26.6
4.9
1.3
1.8
17.1
1.3
1.9
9.3
8.7
4.3
5.1
2.2
6.6
3.5
2.1
15.7
5.6
1.9
6.7
0.0
85
85
95
95
95
85
85
85
85
85
90
95
85
95
85
85
85
90
85
95
95
85
90
90
90
95
95
95
HBA, 3ꢁHbic, RingArom, 6ꢁEV
HBA, 3ꢁHbic, RingArom, 5ꢁEV
HBD, 3ꢁHbic, 3ꢁEV
HBD, HBA, 3ꢁHbic, 1ꢁEV
HBD, 2ꢁHbic, RingArom, 9ꢁEV
HBD, 2ꢁHbic, RingArom, 5ꢁEV
HBA, 2ꢁHbic, RingArom, 7ꢁEV
HBA, 3ꢁHbic, 5ꢁEV
2
HBA, 2ꢁHbic, RingArom, 8ꢁEV
HBA, 2ꢁHbic, RingArom, 7ꢁEV
HBD, 2ꢁHbic, RingArom, 5ꢁEV
HBD, 3ꢁHbic, 5ꢁEV
II
5
6
HBD, 2ꢁHbic, RingArom,
2HBA, 2ꢁHbic
HBD, HBA, Hbic, RingArom
HBD, HBA,2 Hbic
7
9
HBD, 3ꢁHbic, RingArom, 6ꢁEV
III
2ꢁHBD, 3ꢁHbic, 7ꢁEV
HBD, 3ꢁHbic, RingArom, 6ꢁEV
3ꢁHbic, RingArom, 4ꢁEV
HBD, Hbic, 2ꢁ RingArom, 7ꢁEV
HBA, 3ꢁHbic
102
10
88.9
99.9
102.1
102.4
97.1
91.4
92.1
2ꢁ HBD, Hbic, RingArom, 4ꢁEV
2ꢁ HBA, Hbic, RingArom, 6ꢁEV
2HBA, 3ꢁHbic, 4ꢁEV
2.0
5.2
2.1
4.5
11
12
6
HBD, HBA, 3ꢁHbic, 3ꢁEV
2HBA, 3ꢁHbic, 3ꢁEV
7h
a
b
c
d
e
f
Training sets as in Table 1.
Run number as in Table S2 in Supplementary data.
Serial numbers given by CATALYST for individual hypotheses and reported in the log book of the corresponding automatic run.
The difference between the total cost and the cost of the corresponding null hypotheses.
Correlation coefficient for given by CATALYST for individual hypotheses against the particular training subset only.
Fisher-statsitic calculated from the correlation between pharmacophore-based bioactivity estimates (calculated from Eqs. S1 and S2 in Supplementary data) against the
bioactivities of all collected compounds.
g
Fisher confidence level calculated employing the cat. scramble methods (see Section S2 in Supplementary data).
This pharmacophore appeared in the best QSAR equation (bolded).
h
resemble sterically inaccessible regions within the binding
site.20–22,24,28,32,37,46
refinement undermined the true positive rate (TPR) by ca. 25%,
which suggest that although steric-refinement improved the clas-
sification power of the pharmacophore (AUC), it caused some loss
in the number of captured active hits (true positives). This might
be because exclusion volumes impose strict restrictions on cap-
tured hits causing the pharmacophore to inappropriately miss
some active hits.
HipHop-Refine requires a list of training compounds together
with two qualitative descriptors that characterize the way by
which each training compound contributes in defining the exclu-
sion space, that is, Principal and MaxOmitFeat (see Section S3 in
Supplementary data for more details). Table S4 (in Supplementary
data) lists the collected compounds together with their HipHop-
Refine parameters. Figure 4c and d show the sterically-refined ver-
sion of Hypo12/7 (32 added exclusion volumes) and how it maps 9
(IC50 = 25 nM).
2.6. In silico screening of databases
Sterically-refined Hypo12/7 was used as 3D search query
against the national cancer institute list of compounds (NCI, in-
cludes 238,819 compounds).36 Hits are defined as those com-
pounds that have their chemical moieties spatially overlap (map)
with corresponding features in the pharmacophoric query. The
search culminated in 7739 hits that were subsequently filtered
based on Lipinski’s47 and Veber’s rules48 to yield 146 drug-like hits.
Surviving hits were fitted against Hypo12/7 (see the fit Eq. 6 un-
der Experimental) and their fit values were substituted in QSAR Eq. 1
to determine their predicted bioactivities. However, in order to min-
imize the impact of any possible extrapolatory QSAR-based predic-
tion errors on decisions regarding which hits merit subsequent
in vitro testing, Log (1/IC50) predictions were employed merely to
rank the corresponding hits. Subsequently, the highest ranking 52
hits were acquired for in vitro evaluation. Table 3 shows the highest
ranking hits (48–99) based on QSAR Eq. 1, their QSAR-based predic-
tions, as well as their experimental in vitro anti-GP bioactivities.
The tested hits illustrated diverse in vitro activities against GPa
2.5. Receiver operating characteristic (ROC) curve analysis and
shape constraints
To further validate the resulting models (both QSAR and phar-
macophore), we subjected Hypo12/7 to receiver-operating curve
(ROC) analysis. In ROC analysis, the ability of a particular pharma-
cophore model to correctly classify a list of compounds as actives
and inactives is indicated by the area under the curve (AUC) of
the corresponding ROC in addition to other parameters, namely,
overall accuracy, overall specificity, overall true positive rate and
overall false negative rate (see Section 4.1.4 for more details). Table
S5 (in Supplementary data) and Figure 6 show the ROC results of
Hypo12/7, as well as its sterically-refined version.
Clearly from the Figure 6 and Table S5 (in Supplementary data),
steric refinement significantly improved the performance of
Hypo12/7 as reflected by ROC-AUC values, which shifted from
70% to 90% upon addition of exclusion spheres. Nevertheless, steric
ranging from zero to 65% inhibition at 10 lM concentrations.

M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
4751
of GPa. It was hypothesized that introducing two well-positioned
central HBA features separating a-two aromatic ring-based scaffold
should maximize ligand–GPa binding by allowing the ligands to
tightly map Hypo12/7.
1
0
A
-6
-5
-4
-3
-2
-1
0
1
-1
-2
-3
-4
-5
-6
Accordingly, we prepared several malonyl- and oxalyl-diamide
products by reaction of some selected aromatic amines with mal-
onyl dichloride, or succinic anhydride/oxalyl chloride mixture,
respectively, as in Schemes 1 and 2. The aromatic amines were se-
lected in such a way to maximize electronic and steric diversity
among prepared compounds, that is, the aromatic substituents
ranged from electron donating (e.g., methyl, aromatic rings, meth-
oxy groups) to electron withdrawing groups (e.g., chloro), and from
small substituents (e.g., methylbenzene) to fairly large groups (e.g.,
naphthyl), as can be seen in Table 4. However, it must be men-
tioned that we failed to access any nitro-substituted derivatives
probably because of the weak nucleophilic properties of nitro-
substituted aromatic amines.
Although the preparation procedure of the malonyl derivatives
is well known, that is, through nucleophilic amine attack on malo-
nyl dichloride (Scheme 1),49 our synthetic approach to the oxalyl
diamides is completely novel. In this approach, succinic anhydride
was employed as amine donor to allow mono-substitution of the
oxalyl chloride as shown in Scheme 2. Scheme 3 shows the pro-
posed mechanism of this approach: the mixed anhydride interme-
diate of succiniamide (A in Scheme 3) undergoes azalactonization
reaction to yield cyclic isoimide intermediate (B in Scheme 3) in a
similar fashion to that reported by Gerhard and Rene50 albeit med-
iated by oxalylchloride instead of perchloric acid. Subsequently, the
isoimide intermediate (B in Scheme 3) apparently acts as imine
nucleophile and attacks a second oxalyl-chloride molecule to form
isoimidium-oxalylchloride adduct (C in Scheme 3). Apparently,
intermediate C (in Scheme 3) is immune to second nucleophilic at-
tacks by isoimide B (in Scheme 3), that is, to yield symmetrical
oxalamides, probably because (i) imine nitrogens of intermediate
B are weak nucleophiles due to their sp2 hybridization and the steric
bulk of isoimide, and (ii) stabilization of the anti-bonding orbital
(LUMO) of the acyl chloride carbon by the electronic density do-
nated from the lone pair of electrons of the cyclic isoimide oxygen,
as in Figure 7.
Experimental log (1/IC50)
1
0
B
-6
-5
-4
-3
-2
-1
0
1
-1
-2
-3
-4
-5
-6
Experimental log (1/IC50)
Figure 3. Experimental versus fitted (A, 30 compounds, r2LOO ¼ 0:583) and predicted
(B, 7 compounds, r2PRESS ¼ 0:907) bioactivities calculated from the best QSAR model
(Eq. 1). The solid lines are the regression lines for the fitted and predicted
bioactivities of training and test compounds, respectively, whereas the dotted lines
indicate the 1.0 log point error margins.
However, six hits achieved P40% GPa inhibition at 10 lM
concentration prompting subsequent assessment of their IC50 val-
ues against GPa. Figure S1 (in Supplementary data) shows the
dose–inhibition plots of hit compounds.
However, the isoimidium intermediate (C) seems to readily un-
dergo nucleophilic attacks by the added aromatic amines to yield
isoimidium-oxalyl-amide adducts (intermediate D in Scheme 3),
which is hydrolyzed to oxalyl diamide products and succinic anhy-
dride. Interestingly, similar N-phenyl-isoimides were reported to
act as amine donors to aldehydes leading to imines.50
Clearly from Table 3 and Figure S1 (in Supplementary data) that
there is significant deviation between QSAR-predicted bioactivities
and experimentally determined IC50 values. This is not unexpected
as there is significant structural dissimilarity between training
compounds and hits (see Section 2.8). Furthermore, the number
of training compounds is rather limited to allow effective pharma-
cophore and QSAR analysis (only 37 compounds, Table S1 in
Supplementary data). Both factors seem to explain the difference
between predicted and experimental inhibitory profiles of
captured hits. Nevertheless, hit 59 (Table 3) exhibited the closest
predicted-to-experimental IC50 difference, suggesting that QSAR
equation 1 and the associated pharmacophore model (Hypo12/7)
exhibit optimal predictive potential for this scaffold, which
Upon evaluating the resulting oxalyl- and malonyl-diamides
against GPa at 10
lM concentrations, they show diverse anti-GPa
bioactivities. Table
4
shows the synthesized compounds
(100–111, 120–132), their QSAR-based predictions and experi-
mental IC50 values.
Clearly from Table 4, a significant difference is seen between
predicted IC50 values of prepared compounds and their experimen-
tal bioactivities. Probably, this behavior is due to the significant
physicochemical differences between training and prepared
compounds (see Section 2.8).
together with its superior anti-GPa IC50 (4.1 lM), prompted us to
use this structure as guide for subsequent synthetic exploration
towards more potent GPa inhibitors.
Surprisingly, although 120 and 131 failed to fully fit Hypo12/7
(see Figs. 9b and 8b), they illustrated comparatively significant
anti-GPa IC50 values, that is, 5 and 3 lM, respectively. In the case
2.7. Synthesis and bioactivities of pharmacophore-guided GP
Inhibitors
of 131, this behavior is probably explainable by the existence of
an additional interaction within the 131–GPa complex extra to
those suggested in Hypo12/7. To probe this point further, we
docked 131 into the binding pocket of the GPa (PDB code 1L5R, Res-
olution = 2.1 Å) employing Ligandfit docking engine and consensus
scoring based on 6 docking-scoring functions, namely, JAIN,51 Lig-
Score1, LigScore2,52,53 PLP1,54 PLP255 and PMF.56–58 Figure 8a shows
the highest-ranking docked pose of 131 in GPa compared with
The facts that Hypo12/7 exhibited two adjacent HBA features
combined with the presence of two adjacent amidic carbonyls in
potent training compounds (e.g., 9, see Fig. 4) and in the well-
predicted potent NCI hit 59 prompted us to synthetically explore
several diverse oxalyl and malonyl diamides as potential inhibitors

Table 3
In silico hits, their fit values against (Hypo12/7), corresponding QSAR estimates from Eq. 1 and in vitro anti-GP bioactivities
Hit Compound
NCI code
Structure
Fit value against Hypo12/7a
QSAR Predictions
Log (1/IC50 IC50
Experimental
% Inhibition at 10
Mb
)
(lM)
l
IC50
3.2
(l
M)b
Cl
O
N
48
NSC90313
5.4
ꢀ1.847
ꢀ3.738
0.07
5.47
65
N
O
O
49
NSC174281
4.8
30
NDc
O
O
O
O
S
50
51
NSC280876
NSC404052
9.0
6.6
ꢀ1.511
ꢀ1.626
0.03
0.04
49
32
16.5
NDc
O
H
N
O
Cl
O
O
O
O
NDc
NDc
O
52
53
NSC682567
NSC66430
6.9
3.7
ꢀ3.112
ꢀ1.995
1.29
0.10
24
27
O
O
S
O
N
O
N
H
Cl
O
H
NDc
Br
54
NSC176370
7.7
ꢀ1.475
0.03
24
N
O
NDc
NDc
O
55
56
NSC195114
NSC327175
7.5
6.2
ꢀ2.4
0.25
0.01
23
29
5
O
O
O
Cl
S
ꢀ0.696
N
N
N
Cl
H

Cl
H
N
O
57
58
NSC681174
NSC686774
4.6
7.9
ꢀ2.087
ꢀ2.687
0.12
0.49
45
42
6.9
7.5
S
N
N
O
H
S
H
N
Cl
N
O
S
N
N
O
H
S
H
-O
O
N+
O
59
NSC61623
9.3
ꢀ2.966
0.92
60
4.1
O
O
O
N
O
N
60
61
62
NSC88165
NSC164459
NSC216623
8.1
8.6
9.1
ꢀ3.574
ꢀ3.083
ꢀ2.754
3.75
1.21
0.57
39
38
36
NDc
NDc
NDc
Cl
H
H
Cl
O
O
N
H
H
N
O
Cl
H
O
N
H
N
H
S
H
O
N
H
O
O
N
N
-O
63
64
NSC302323
NSC673994
8.7
9.2
ꢀ2.633
ꢀ3.127
0.43
1.34
32
53
NDc
N+
O
Cl
H
H
S
O
N
N
S
N
O
N
5.9
F
Cl
F
Cl
F
(continued on next page)

Table 3 (continued)
Hit Compound
NCI code
Structure
Fit value against Hypo12/7a
QSAR Predictions
Log (1/IC50 IC50
Experimental
% Inhibition at 10
Mb
)
(lM)
l
IC50
NDc
(l
M)b
H
O
O
65
66
NSC11437
8.8
ꢀ2.603
ꢀ3.316
0.40
2.07
25
N
Cl
O
H
H
NSC16743
9.0
8.0
NDc
O
O
H
O
-O
Cl
N+
O
O
Cl
67
68
NSC29159
NSC50690
8.8
9.8
ꢀ4.003
ꢀ2.693
10.07
0.49
3
NDc
N
O
O
H
N+
-O
H
N
O
O
22
NDc
O
H
H
O
O
N
H
69
70
NSC58325
NSC88898
10.3
9.5
ꢀ4.075
ꢀ4.323
11.89
21.04
26
11
NDc
Cl
O
N
O
N
NDc
O
H
H
H
O
O
S
NDc
NDc
S
71
72
NSC134224
NSC141980
9.4
ꢀ2.721
ꢀ3.159
0.53
1.44
34
15
O
Br
O
S
N
N
10.3
O
O
H
S

O
S
H
N
N
73
74
NSC141981
NSC144076
10.4
9.8
ꢀ3.134
ꢀ1.693
1.36
0.05
9.0
NDc
NDc
O
O
H
S
O
N
O
S
O
18
H
O
O
O
75
NSC146837
9.4
ꢀ2.472
0.30
34
NDc
O
O
O
O-
O
N+
H
H
O
76
77
NSC151098
NSC157495
8.00
ꢀ2.838
ꢀ2.97
0.69
0.93
25
17
NDc
N
O
Cl
H
N
7.9
8.8
NDc
N
O
N+
-O
O
-O
Cl
NDc
N+
78
NSC164588
ꢀ3.442
2.77
25
12
O
N
H
H
O
O
N
NDc
NDc
O
6
79
80
NSC191949
NSC202468
8.5
ꢀ1.99
0.10
1.04
N
S
O
O
O
S
N
10.0
ꢀ3.018
2.0
S
7
O
H
N
N
O
81
82
NSC203848
NSC203972
8.7
8.4
ꢀ3.814
ꢀ3.928
6.52
8.47
12
18
NDc
H
O
O
-O
O
H
N
H
N
N+
F
F
NDc
F
O
Cl
Cl
(continued on next page)

Table 3 (continued)
Hit Compound
NCI code
Structure
Fit value against Hypo12/7a
QSAR Predictions
Log (1/IC50 IC50
Experimental
% Inhibition at 10
Mb
)
(
lM)
l
IC50
NDc
(l
M)b
H
O
O
N
Cl
N
83
84
85
NSC204090
8.8
ꢀ3.329
ꢀ2.986
ꢀ3.768
2.13
0.97
5.86
7.0
H
H
O
O
O
H
N
O
NSC204442
NSC205486
8.3
9.6
17
NDc
S
O
H
O
H
N
O
N
7.0
NDc
O
O
H
O
N
N
86
87
88
NSC205499
NSC205520
NSC205686
8.0
9.3
ꢀ2.848
ꢀ3.757
ꢀ3.067
0.70
5.71
1.17
26
NDc
NDc
NDc
H
O
H
O
H
O
O
O
N
N+ O-
9.0
9.0
O
Cl
10.4
O
O
O
O
O
N+
O-
N
H
-O
H
H
N
N+
N
89
90
NSC205703
NSC205705
10.7
8.6
ꢀ3.094
ꢀ2.055
1.24
0.11
16
21
NDc
O
S
Cl
O
-O N+
H
NDc
N
O
Cl
N

O
O
H
N
H
N
91
NSC205797
8.7
ꢀ4.056
11.38
7.0
NDc
O
S
H
H
O
N
92
93
NSC205827
NSC211982
9.1
9.1
ꢀ2.765
ꢀ3.633
0.58
4.30
17
NDc
N
N
O
O
H
S
H
Cl
F
S
Cl
O
H
16
20
NDc
N
O
O
O
H
O
O
NDc
S
94
NSC280880
8.5
ꢀ2.585
0.38
O
H
O
O
O
NH
95
96
NSC370919
NSC379459
8.9
9.7
ꢀ2.395
ꢀ3.570
0.25
3.72
24
32
NDc
NDc
HN
O
Cl
O-
O
N+
H
N
O
O
Cl
H
N
O
O
O
O
S
97
NSC648606
10.7
ꢀ3.209
1.62
34
NDc
N
H
N
N
O
H
Cl
Cl
O
O
O
98
99
NSC682580
NSC658713
8.3
9.6
ꢀ1.374
ꢀ2.972
0.02
0.94
26
15
NDc
NDc
HN
F
F
F
O
O
O
NH
F
HN
a
b
c
Best-fit values calculated by Eq. S2 in Supplementary data.
Bioactivity values are the average of at least duplicate (twice) measurements
ND: not determined

4758
M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
mapping 131 against Hypo12/7. Clearly, mapping the two carbon-
yls of 131 against two HBA features in Hypo12/7 agrees with hydro-
gen-bonding interactions connecting the two groups with the
guanidine of Arg60 and ammonium of Lys191. In fact, the later
interaction is reinforced by a network of hydrogen bonds involving
H2O34 and the peptidic NH of Ser192. Similarly, mapping the ben-
zylic ring against a hydrophobic feature in Hypo12/7 agrees with
positioning this group into a hydrophobic pocket comprised of
the side chains of Trp67, Pro229 and Val64 (Fig. 8a). Similarly, map-
ping the chlorophenyl group against a hydrophobic feature in
Hypo12/7 corresponds to hydrophobic stacking involving this
group and the aromatic ring of Phe53.
However, clearly from Figure 8b, 131 misses one hydrophobic
feature in Hypo12/7 corresponding to interaction with Val64 (see
Fig. 5). Still, the close proximity of the o-chloro substituent in
131 to H2O34 suggests it is involved in electrostatic and hydro-
gen-bonding interactions with this water molecule, which is pin-
ned by hydrogen-bonding interactions with the peptidic NH of
Ser192. This extra interaction probably compensates for the lost
hydrophobic feature in Hypo12/7.
Likewise, 120 seems to dock into GPa in a similar manner to the
way it maps Hypo12/7 (see Fig. 9): fitting the two central HBAs
against the oxalyl carbonyls correspond to hydrogen-bonding
interactions connecting these carbonyls with the side chains of
Lys192 and Arg60, as in Figure 9. On the other hand, fitting the
aromatic methyls against two hydrophobic features agrees with
docking these methyls into hydrophobic pockets comprised of
the side chains of Val40, Val64 and Leu39. Nevertheless, the same
set of interactions can be also noticed in the docked complexes of
other oxalamides despite their inferior bioactivities, for example,
124, 126 and 132. We believe that the superior bioactivity of 120
is related to the close analogy between its bound conformer and
the lowest energy conformer of its unbound structure. The two
ortho-methyl groups force the phenyl substituents to assume per-
pendicular conformation to that of the oxalamide linker in the un-
bound structure, which is although electronically disfavoured
because it breaks the electronic delocalization of the two phenyl
groups across the oxalamide linker, it closely resembles the docked
conformer of 120 (see Fig. 9). This suggests that 120–GPa binding
requires minimal energy expenditure to strain 120 within the
binding pocket, which is apparently not the case in the less potent
oxalamides. In other words, the energy difference between un-
bound conformers and bound counterparts is rather significant in
poorly-potent oxalamides suggesting significant energetic cost in
straining the ligand molecules to fit within the binding site. This
should reduce the free energy gains of binding explaining the
apparent reduction in bioactivity of poorly potent oxalamide
analogues despite their theoretically tight fit against Hypo12/7,
for example, 121, 122 and 130 (see Fig. 10).
unique occurrences of molecular rings, ring assemblies, bridge
assemblies, chains, Murcko assemblies, or any combination of
these. Murcko assemblies are contiguous ring systems plus chains
that link two or more rings.59 The two libraries are compared using
a Tanimoto similarity of the assemblies based on the fragments
that are common and unique to each library.60 On the other hand,
in Bayesian Model approach, two Bayesian models were built, one
to learn library A and one to learn library B. Finally, it scores all
ligands using both models. A distance is computed as Eq. 2:
Distance ¼ ScoreAA þ ScoreBB ꢀ ScoreAB ꢀ ScoreBA
ð2Þ
where ScoreAA is the average score of library A molecules scored by
the Bayesian model that learned library A molecules, while ScoreBB
is the average score of library B molecules scored by the Bayesian
model that learned library B. ScoreAB and ScoreBA are the average
scores of libraries A and B molecules scored by the Bayesian models
that learned libraries B and A, respectively. The higher the distance,
the more dissimilar the libraries are.60
Finally, the Global Fingerprint comparison algorithm generates
a global fingerprint for all ligands in the training list and all ligands
in the hits list and then computes a Tanimoto similarity coefficient
between the two libraries.60
Table S6 (in Supplementary data) shows the results of the three
similarity/diversity assessment procedures. Clearly, the three
methods suggest minimal structural similarity between known
GP inhibitors and our captured hits, synthesized compounds,
which explains the observed differences between predicted and
experimental bioactivities.
3. Conclusion
GP inhibitors are currently considered as potential treatments
for diabetes. The pharmacophoric space of GP inhibitors was ex-
plored via three sets of inhibitors and using CATALYST-HYPOGEN
to identify high quality binding model. Subsequently, genetic algo-
rithm and multiple linear regression analysis were employed to ac-
cess optimal QSAR model capable of explaining anti-GP bioactivity
variation across 37 collected GP inhibitors (r230 ¼ 0:69, F = 18.70,
n = 30, r2BS ¼ 0:69, rL2OO ¼ 0:58, rP2RESS against 7 external test inhibi-
tors = 0.91). One pharmacophore model emerged in the QSAR
equation suggesting the existence of a major binding mode acces-
sible to ligands within GP binding pocket. The QSAR equation and
the associated pharmacophoric model were used to guide syn-
thetic exploration of a new series of GP inhibitors that resulted
in several novel low micromolar GP inhibitors.
4. Experimental
4.1. Molecular modeling
2.8. Similarity analysis between training compounds versus
active prepared compounds and in silico hits
4.1.1. Software and hardware
The following software packages were utilized in the present
research.
In an attempt to explain the observed discrepancies between
predicted and experimental bioactivities of screening hits and syn-
thesized compounds (e.g., 54, 56 and 124) we hypothesized that
these differences are because of significant structural differences
between training compounds compared to captured hits and syn-
thesized compounds. Accordingly, to test this hypothesis, we em-
ployed three library comparison methods (implemented in
Discovery Studio 2.5) to assess the structural similarity/diversity
between the modeled compounds (1ꢀ37, Table S1 in Supplemen-
tary data)34 (library A) compared to captured hits and synthesized
compounds (48–111, 120–132, Table 4, library B), namely, Murcko
Assemblies, Bayesian Model and Global Fingerprints. In Murcko
Assemblies, the algorithm breaks the ligands of each library into
ꢂ
ꢂ
CATALYST (Version 4.11), Accelrys Inc. (www.accelrys.com), USA.
CERIUS2 (Version 4.10), Accelrys Inc. (www.accelrys.com), USA.
ꢂ CS ChemDraw Ultra (Version 11.0), Cambridge Soft Corp., USA.
(http://www.cambridgesoft.com), USA.
CATALYST and CERIUS2 software suites installed on a Silicon
Graphics Octane2 desktop workstation equipped with a dual
600 MHz MIPS R14000 processor (1.0 GB RAM) running the Irix
6.5 operating system. Structure drawing was performed employing
ChemDraw Ultra 6.0 installed on a Pentium 4 PC.

M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
4759
Table 4
Synthesized compounds with their fit values against (Hypo12/7), corresponding QSAR estimates from Eq. 1 and in vitro anti-GP bioactivities
No.a
Structure
Fit value against Hypo12/7b
QSAR Predictions
Experimental
Novelty
c
Log (1/IC50
)
IC50
(
lM)
% Inhibition at 10
2
l
Mc
IC50 (lM)
H
H
N
N
100
1.9
ꢀ4.724
52.9
NDd
Reported70
O
O
O
O
H
N
H
N
101
102
103
7.7
7.7
5.2
ꢀ3.053
ꢀ3.273
ꢀ0.108
1.2
1.9
1.4
4
3
NDd
NDd
NDd
Reported70
Reported70
Reported70
H
N
H
N
O
O
O
O
H
N
H
N
40
O
O
O
H
N
H
N
104
105
8.7
8.1
ꢀ4.931
ꢀ4.387
85.3
24.4
28
25
NDd
NDd
Reported70
O
O
O
H
N
H
N
New
O
O
Cl
Cl
H
N
H
N
106
5.4
ꢀ5.029
106.9
5
NDd
Reported70
O
O
O
O
107
108
109
110
0.5
9.0
1.8
9.0
ꢀ6.3
216
53.8
198
53.8
1
24
23
24
NDd
NDd
NDd
NDd
Reported71
Reported72
Reported70
Reported73
N
H
N
H
H
N
H
Cl
Cl
N
Cl
Cl
ꢀ4.731
ꢀ5.297
ꢀ4.731
O
O
Cl
Cl
H
N
H
N
O
O
H
N
H
N
O
O
Cl
Cl
O
O
H
N
H
N
111
9.8
ꢀ5.532
340
9
NDd
New
O
O
O
O
Reported74
Reported75
7.9e
7.4
120
121
ꢀ1.988
ꢀ2.056
0.10
55
36
5.4
NH
O
HN
O
0.11
NDd
NH
HN
O
O
O
O
122
123
8.4
6.3
ꢀ4.474
ꢀ3.376
29.8
24
19
NDd
NDd
Reported76
NH
HN
O
O
NH
HN
7.8
New
Cl
(continued on next page)

4760
M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
Table 4 (continued)
No.a
Structure
Fit value against Hypo12/7b
QSAR Predictions
Experimental
% Inhibition at 10 IC50 (lM)
Mc
Novelty
New
c
Log (1/IC50
)
IC50
(
l
M)
l
O
O
NH
HN
124
ꢀ2.745
0.56
3.22
20
NDd
7.9e
O
O
O
O
NDd
New
New
NH
HN
125
126
4.04
ꢀ3.508
ꢀ3.902
23
15
O
O
NDd
NH
HN
7.2
7.98
O
O
NH
HN
127
128
129
6.5
6.7
ꢀ2.848
ꢀ2.739
ꢀ3.748
0.70
0.55
5.60
30
23
24
NDd
NDd
NDd
Reported77
O
O
NH
HN
New
O
O
O
7.9
9.3
New
NH
HN
Cl
O
O
O
130
131
132
ꢀ3.994
ꢀ2.933
ꢀ2.545
9.87
0.86
0.35
4.0
64
5
NDd
3.0
New
NH
HN
O
Cl
O
O
8.1e
8.1e
New
HN
NH
O
O
O
NDd
Reported78
N
NH
HN
a
b
c
Compound numbers as in Schemes 1 and 2.
Best-fit values calculated by Eq. S2 in Supplementary data.
Bioactivity values are the average of at least duplicate (twice) measurements.
ND: not determined.
d
e
Underlined fit values were calculated by allowing one omitted pharmacophoric feature.
4.1.2. Pharmacophore modelling
being higher than 20,000 nM (e.g., 27, 30, 34 and 35), they were as-
sumed to be 100,000 nM. This assumption is necessary to allow
statistical correlation and QSAR analysis. The logarithmic transfor-
mation of IC50 values should minimize any potential errors result-
ing from this assumption.
The two-dimensional (2D) chemical structures of the inhibitors
were sketched using ChemDraw Ultra and saved in MDL-mol file
format. Subsequently, they were imported into CATALYST, con-
verted into corresponding standard 3D structures and energy min-
imized to the closest local minimum using the molecular
The structures of 37 GPa inhibitors (1–37, Table S1 in
Supplementary data) were collected from published literature.34
The in vitro bioactivities of the collected inhibitors were expressed
as the concentration of the test compound that inhibited the
activity of GPa by 50% (IC50). Table S1 in Supplementary data
shows the structures and IC50 values of the considered inhibitors.
The logarithm of measured IC50 (nM) values were used in pharma-
cophore modeling and QSAR analysis, thus correlating the data lin-
ear to the free energy change. In cases where IC50 is expressed as

M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
4761
A
B
C
D
O
N
E
HN
O
Cl
NH
Figure 4. (A) QSAR-selected Hypo12/7, (B) Hypo12/7 mapped against 9 (Table S1 in Supplementary data, IC50 = 25 nM), (C) sterically-refined Hypo12/7, (D) sterically-refined
Hypo12/7 mapped against 9. Light blue spheres represent Hbic features, green vectored spheres encode for HBAs and grey spheres represent exclusion volumes and (E)
chemical structure of 9.
mechanics CHARMm force field implemented in CATALYST. The
resulting 3D structures were utilized as starting conformers for
conformational analysis.
from each training set were clustered separately, for example, all
84 pharmacophore models generated from training subset A were
clustered into 28 groups. To select best representative pharmaco-
phores from each cluster, each group members were fitted against
the collected compounds (1–37, Table S1 in Supplementary data)
employing the ‘best’ fit option implemented in CATALYST, and
their fit values (calculated from Eq. S4 in Supplementary data)
were regressed against the corresponding experimental bioactivi-
ties. The highest-ranking models, based on their F-statistic, were
selected to represent their corresponding clusters in subsequent
QSAR modeling. Table 2 (see Section 2) shows the pharmacophoric
features of selected representative binding hypotheses, combined
with their CATALYST success criteria.
The molecular flexibilities of the collected compounds were ta-
ken into account by considering each compound as a collection of
conformers representing different areas of the conformational
space accessible to the molecule within a given energy range.
Accordingly, the conformational space of each inhibitor (1–37,
Table S1 in Supplementary data) was explored adopting the ‘best
conformer generation’ option within CATALYST, which is based
on the generalized CHARMm force field implemented in the
program. Default parameters were employed in the conformation
generation procedure, that is, a conformational ensemble was
generated with an energy threshold of 20 kcal/mol from the local
minimized structure which has the lowest energy level and a
maximum limit of 250 conformers per molecule. This search pro-
cedure will probably identify the best 3D arrangement of chemical
functionalities explaining the activity variations among the train-
ing set.37
Subsequently, the collected compounds were used to explore
the pharmacophoric space of GPa binding pocket (see detailed
experimental procedures under Sections S1 and S2 in Supplemen-
tary data).
Afterward, the generated binding hypotheses were clustered as
follows: optimal models were clustered into 28 groups (every
three models were clustered together) utilizing the hierarchical
average linkage method available in CATALYST. Models generated
4.1.3. QSAR modeling
A subset of 30 compounds from the total list of inhibitors (1–37,
Table S1 in Supplementary data) was utilized as a training set for
QSAR modeling. However, since it is essential to access the predic-
tive power of the resulting QSAR models on an external set of
inhibitors, the remaining 7 molecules (ca. 20% of the dataset) were
employed as an external test subset for validating the QSAR
models. The test molecules were selected as follows: The 37 inhib-
itors (1–37, Table S1 in Supplementary data) were ranked accord-
ing to their IC50 values, and then every fifth compound was
selected for the test set starting from the high-potency end. This
selection considers the fact that the test molecules must represent
a range of biological activities similar to that of the training set.

4762
M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
A
C
B
OH
O
H
N
O
O
N
N
H
Cl
Figure 5. Co-crystallized ligands of GPa X-ray structure 1L5R (resolution = 2.10 Å). (A) Binding pocket of 1L5R complexed with its corresponding co-crystallized ligand (CP-
403700), (B) mapping Hypo12/7 against the co-crystallized ligands of 1L5R (rigid mapping) and (C) the chemical structures of the co-crystallized ligand.
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
A
B
0
0.1 0.2
0.3 0.4
0.5 0.6 0.7
0.8 0.9
1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
FALSE POSITIVE RATE
FALSE POSITIVE RATE
Figure 6. Receiver operating characteristic curves (ROCs) of (A) Hypo12/7 and (B) sterically-refined Hypo12/7.
The logarithm of measured 1/IC50 (nM) values was used in QSAR,
thus correlating the data linear to the free energy change. The
chemical structures of the inhibitors were imported into CERIUS2
as standard 3D single conformer representations in SD format. Sub-
sequently, different descriptor groups were calculated for each
compound employing the C2.DESCRIPTOR module of CERIUS2.
The calculated descriptors included various simple and valence
connectivity indices, electro-topological state indices and other
molecular descriptors (e.g., logarithm of partition coefficient, polar-
izability, dipole moment, molecular volume, molecular weight,
molecular surface area, etc.).44 Furthermore, the training com-
pounds were fitted (using the Best-fit option in CATALYST) against
the representative pharmacophores (28 models, Table 2), and their
fit values were added as additional descriptors. The fit value for any
compound is obtained automatically via Eq. S4 in Supplementary
data.37
Genetic function approximation (GFA) was employed to search
for the best possible QSAR regression equation capable of correlat-
ing the variations in biological activities of the training compounds
with variations in the generated descriptors, that is, multiple linear

M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
4763
H
H
O
N
N
a
O
Cl
O
R
R
O
NH2
2 R
NH
O
O
OH
R1
Cl
100 R= 2-methylphenyl
101 R= 3-methylphenyl
102 R= 4-methylphenyl
103 R= 2-methoxyphenyl
104 R= 4-methoxyphenyl
105 R= 5-chloro-2-methylphenyl
106 R= 1-naphthyl
O
R1
R1
O
Cl
107 R= benzyl
O
O
108 R= 3,4-dichlorophenyl
109 R= 2-chlorophenyl
110 R= 4-chlorophenyl
111 R= 2-methoxy-4-methylphenyl
O
N
N
O
O
H
(A)
(B)
Cl
O
R1
Scheme 1. Preparation of malonyl diamides 100–111. Reagents and conditions: (a)
Malonyl dichloride in dioxane at 0 °C.
O
Cl
O
O
O
N
R1
(C)
O
O
Cl
R2
O
O
O
N
H2N R1
NH2
a
b, c
O
O
R1 NH
H2N R2
O
(D)
(In pyridine)
R1
O
O
s
i
HN
OH
s
y
l
3
o
r
d
112 R1= 2-methylphenyl
113 R1= 3-methylphenyl
114 R1= 2-methoxy-4-methylphenyl
115 R1= 1-naphthyl
y
H
R2
116 R1= 5-chloro-2-methylphenyl
117 R1= 4-methoxyphenyl
118 R1= 2-chlorophenyl
119 R1= 2-pyridyl
O
HO
O
O
O
R1 NH
O
O
NH
H
y
d
r
o
l
y
s
i
s
O
O
HN R2
120 R1=R2= 2-methylphenyl
121 R1=R2= 3-methylphenyl
122 R1=R2= 2-methoxy-4-methylphenyl
HN
O
OH
123 R1= 1-naphthyl, R2= 5-chloro-2-methylphenyl
124 R1= 1-naphthyl, R2= 2-methylphenyl
125 R1= 1-naphthyl, R2= 2-methoxylphenyl
126 R1= 1-naphthyl, R2= 2-methoxyl-4-methylphenyl
127 R1= 1-naphthyl, R2= 4-methylphenyl
R2
Scheme 3. Proposed mechanism of N,N-di-oxalamide derivatives synthesis.
128 R1= 1-naphthyl, R2= 3-methylphenyl
129 R1= 5-chloro-2-methylphenyl, R2= 2-methoxy-4-methylphenyl
130 R1= 4-methoxylphenyl , R2=2-methoxyl-4-methylphenyl
131 R1= benzyl, R2= 2-chlorophenyl
r2PRESS ¼ SD-PRESS=SD
ð3Þ
132 R1= 2-methoxyphenyl, R2= 2-pyridyl
Scheme 2. Preparation of oxalyl diamides. Reagents and conditions: (a) Succinic
anhydride in dioxane under reflux conditions, (b) 2 equiv oxalyl chloride (neat) at
0 °C followed by, (c) particular aniline derivative in pyridine.
where SD is the sum of the squared deviations between the bio-
logical activities of the test set and the mean activity of the training
set molecules, PRESS is the squared deviations between predicted
and actual activity values for every molecule in the test set.
regression modeling (MLR). GFA techniques rely on the evolution-
ary operations of ‘crossover and mutation’ to select optimal combi-
nations of descriptors (i.e., chromosomes) capable of explaining
bioactivity variation among training compounds from a large pool
of possible descriptor combinations, that is, chromosomes popula-
tion. However, to avoid overwhelming GFA–MLR with large
number of poor descriptor populations, the lowest-variance
descriptors (20%) were removed prior to QSAR analysis.
Each chromosome is associated with a fitness value that reflects
how good it is compared to other solutions. The fitness function
employed herein is based on Friedman’s ‘lack-of-fit’ (LOF).44
Our preliminary diagnostic trials suggested the following opti-
mal GFA parameters: explore linear, quadratic and spline equa-
tions at mating and mutation probabilities of 50%; population
size = 500; number of genetic iterations = 30,000 and lack-of-fit
(LOF) smoothness parameter = 1.0. However, to determine the
optimal number of explanatory terms (QSAR descriptors), it was
decided to scan and evaluate all possible QSAR models resulting
from 3 to 6 explanatory terms.
4.1.4. Receiver operating characteristic (ROC) curve analysis
Successful pharmacophore models (Hypo12/7 and refined phar-
macophore) were further validated by assessing their abilities to
selectively capture diverse GP active compounds from a large list
of decoys employing ROC analysis.
Therefore, it was necessary to prepare valid evaluation struc-
tural database (testing set) that contains an appropriate list of de-
coy compounds in combination with diverse list of known active
compounds. The decoy list was prepared as described by Verdonk
and co-workers.62,63 Briefly, the decoy compounds were selected
based on three basic one-dimensional (1D) properties that allow
the assessment of distance (D) between two molecules (e.g., i
and j): (1) the number of hydrogen-bond donors (NumHBD); (2)
number of hydrogen-bond acceptors (NumHBA) and (3) count of
nonpolar atoms (NP, defined as the summation of Cl, F, Br, I, S
and C atoms in a particular molecule). For each active compound
in the testing set, the distance to the nearest other active com-
pound is assessed using their euclidean distance (Eq. 4):
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
All QSAR models were validated employing leave one-out cross-
validation (rL2OO), bootstrapping (rB2S) and predictive r2 (rP2RESS) calcu-
lated from the test subsets. The predictive r2PRESS is defined as:
2
2
2
Dði; jÞ ¼ ðNumHBDi ꢀ NumHBDjÞ þ ðNumHBAi ꢀ NumHBAjÞ þ ðNPi ꢀ NPjÞ
ð4Þ

4764
M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
A
B
C
O
N+
O
Cl
O
O
Figure 7. (A) The chemical structure of the proposed isoimidium-oxalylchloride adduct, (B) energy-minimized structure of the adduct showing an overlap between the
oxygen unshared pair of electrons and the lowest unoccupied molecular orbital (LUMO) of the acyl chloride and (C) CPK presentation of the adduct showing the close
proximity of the isoimidium oxygen and the carbon of the acyl chloride fragment.
A
B
C
O
H
O
N
H
Cl
N
Figure 8. Prepared compound 131 (IC50 = 3 lM, Table 4). (A) Compound 131 docked into the binding pocket of GPa (PDB code: 1L5R, resolution 2.10 Å), (B) 131 fitted against
Hypo12/7 and (C) chemical structure of 131.
The minimum distances are then averaged over all active com-
pounds (Dmin). Subsequently, for each active compound in the
testing set, 33 decoys were randomly chosen from the ZINC data-
base.64 The decoys were selected in such a way that they did not
exceed Dmin distance from their corresponding active compound.
Moreover, to further diversify the actives members, that is, to
avoid close similarity among actives in the testing set, any active
compound having zero distance (Dði; jÞ) from other active com-
pound(s) in the testing set were excluded. Active testing com-
pounds were defined as those possessing GP inhibition ranging

M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
4765
A
B
C
H
N
O
N
H
O
Figure 9. Synthesized compound 120 (IC50 = 5 lM, Table 3). (A) Compound 120 docked into the binding pocket of GPa (PDB code: 1L5R, resolution 2.10 Å), (B) 120 fitted
against Hypo12/7 and (C) chemical structure of 120.
from 25 to 6400 nM. The testing set included 28 active compounds
and 909 ZINC compounds.
with the molecule that got the lowest score, most of the actives
will have a higher score than the decoys. Since some of the actives
will be scored lower than decoys, an overlap between the distribu-
tion of active molecules and decoys will occur, which will lead to
the prediction of false positives and false negatives.62,65 The selec-
tion of one score value as a threshold strongly influences the ratio
of actives to decoys and therefore, the validation of a VS method.
The ROC curve method avoids the selection of a threshold by con-
sidering all Se and Sp pairs for each score threshold.68 A ROC curve
is plotted by setting the score of the active molecule as the first
threshold. Afterwards, the number of decoys within this cutoff is
counted and the corresponding Se and Sp pair is calculated. This
calculation is repeated for the active molecule with the second
highest score and so forth, until the scores of all actives are consid-
ered as selection thresholds.
The ROC curve representing ideal distributions, where no over-
lap between the scores of active molecules and decoys exists, pro-
ceeds from the origin to the upper-left corner until all the actives
are retrieved and Se reaches the value of 1. Thereafter, only decoys
can be found using the VS method. Thus, the ideal ROC curve con-
tinues as a horizontal straight line to the upper-right corner where
all actives and all decoys are retrieved, which corresponds to Se = 1
and Sp = 0. In contrast to that, the ROC curve for a set of actives and
decoys with randomly distributed scores tends towards the Se = 1-
Sp line asymptotically with increasing number of actives and de-
coys.65 The success of a particular virtual screening workflow can
be judged from the following criteria:
The testing set (937 compounds) was screened by each particu-
lar pharmacophore (for ROC analysis) employing the ‘Best flexible
search’ option implemented in CATALYST, while the conforma-
tional spaces of the compounds were generated employing the
‘Fast conformation generation option’ implemented in CATALYST.
Compounds missing one or more features were discarded from
the hit list. The in silico hits were scored employing their fit values
as calculated by Eq. S4 in Supplementary data.
The ROC curve analysis describes the sensitivity (Se or true po-
sitive rate, Eq. 5) for any possible change in the number of selected
compounds (n) as a function of (1-Sp). Sp is defined as specificity or
true negative rate (Eq. 6).62,65
Number of selected actives
Total number of actives
Number of discarded inactives
TP
TP þ FN
Se ¼
Sp ¼
¼
ð5Þ
ð6Þ
TN
¼
Total number of inactives
TN þ FP
where TP is the number of active compounds captured by the vir-
tual screening method (true positives), FN is the number of active
compounds discarded by the virtual screening method, TN is the
number of discarded decoys (presumably inactives), while FP is
the number of captured decoys (presumably inactives).
If all molecules scored by a virtual screening (VS) protocol with
sufficient discriminatory power are ranked according to their score
(i.e., fit values), starting with the best-scored molecule and ending

4766
M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
A
B
O
H
N
N
H
O
O
O
C
D
H
O
N
NH
O
E
F
O
O
O
NH
HN
O
Figure 10. (A) Compound 130 mapped against Hypo12/7, (B) chemical structure of 130, (C) 121 mapped against Hypo12/7, (D) chemical structure of 121, (E) 122 mapped
against Hypo12/7 and (F) chemical structure of 122.
(1) Area under the ROC curve (AUC).65 In an optimal ROC curve
(3) Overall specificity (SPC): describes the percentage of dis-
an AUC value of 1 is obtained; however, random distribu-
carded inactives by the particular virtual screening work-
tions cause an AUC value of 0.5. Virtual screening that per-
flow. Inactive test compounds are assigned a binary score
forms better than a random discrimination of actives and
value of zero (compound not captured) or one (compound
decoys retrieve an AUC value between 0.5 and 1, whereas
captured) regardless to their individual fit values.62,66,67
an AUC value lower than 0.5 represents the unfavorable case
(4) Overall true positive rate (TPR or overall sensitivity):
of a virtual screening method that has a higher probability to
describes the fraction percentage of captured actives from
assign the best scores to decoys than to actives.62,65
the total number of actives. Active test compounds are
(2) Overall accuracy (ACC): describes the percentage of correctly
assigned a binary score value of zero (compound not cap-
classified molecules by the screening protocol (equation 7).
tured) or one (compound captured) regardless to their indi-
Testing compounds are assigned a binary score value of
vidual fit values.
zero (compound not captured) or one (compound
(5) Overall false negative rate (FNR or overall percentage of
captured).62,66,67
discarded actives): describes the fraction percentage of
active compounds discarded by the virtual screening
TP þ TN
A
N
A
ꢃ Se þ ð1 ꢀ Þ ꢃ Sp
N
ACC ¼
¼
ð7Þ
method. Discarded active test compounds are assigned a
binary score value of zero (compound not captured) or
one (compound captured) regardless to their individual fit
values.
N
where N is the total number of compounds in the testing data-
base, A is the number of true actives in the testing database.

M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
4767
4.1.5. In silico screening for new GP inhibitors
(2H, CH2, D-exchange), 7.09 (d, J = 8 Hz, 4H), 7.47 (d J 8, 4H),
10.10 ppm (br s, 2H, NH) ppm; 13C NMR (100 MHz, DMSO-d6): d
23.07 (2 ꢁ CH3), 48.34 (CH2), 121.68 (4 ꢁ CH), 131.80 (4 ꢁ CH),
135.02 (2 ꢁ C), 138.96 (2 ꢁ C), 167.82 (2 ꢁ C@O) ppm; HRMS-ESI
m/z [M+Na]+ calcd for C17H18N2NaO2: 305.12659, found
305.12605. Anal. Calcd for C17H18N2O2: C, 72.32; H, 6.43. Found:
C, 69.38; H, 6.28.
Refined Hypo12/7 was employed as 3D search queries against
the NCI using the ‘Best Flexible Database Search’ option imple-
mented within CATALYST. NCI hits were filtered based on Lipinski’s
and Veber’s rules.47,48 The remaining hits were fitted against
Hypo12/7 using the ‘best-fit’ approach implemented within CATA-
LYST. Subsequently, the fit values together with other relevant
molecular descriptors were substituted in QSAR Eq. 1 to predict
anti-GP IC50 values. The highest-ranking hits were subsequently
tested in vitro.
4.2.1.4. N1,N3-Bis-(2-methoxy-phenyl)-malonamide (103). Yield
(1.5 g, 60%): mp: 160–163C (cryst. water); mmax (KBr): 3367,
3292, 1676 cmꢀ1 1H NMR (400 MHz, DMSO-d6): d 3.44 (2H, CH2,
;
4.2. Synthetic procedures
D-exchange), 3.70 (s, 6H, OCH3), 6.93 (m, 2H), 7.05 (m, 4H), 8.02
(m, 2H), 9.70 ppm (br s, 2H, NH) ppm; 13C NMR (100 MHz,
Melting points were measured using Gallenkampf melting point
apparatus and are uncorrected. 1H NMR and 13C NMR spectra were
collected on a Bruker NMR400 spectrometer. High resolution mass
spectrometry was performed using LC Mass Bruker Apex-IV mass
spectrometer utilizing an electrospray interface. Infrared spectra
were recorded using WQF-520 FTIR spectrometer. The samples
were analyzed as KBr pellets. Analytical thin layer chromatography
(TLC) was carried out using pre-coated alumina plates and visual-
ized by UV light (at 254 and/or 360 nm) using different combina-
tions of mobile phase to ensure purity. Elemental analysis was
performed using EuroVector elemental analyzer. Chemicals and
solvents were used without further purification.
DMSO-d6):
d
47.0 (CH2), 58.37 (2 ꢁ OCH3), 113.78 (2 ꢁ CH),
122.96 (2 ꢁ CH), 124.01 (2 ꢁ CH), 127.18 (2 ꢁ CH), 129.57
(2 ꢁ C), 151.89 (2 ꢁ C), 168.50 (2 ꢁ C@O) ppm; HRMS-ESI m/z
[M+Na]+ calcd for C17H18N2NaO4: 337.11642, found 337.11588.
Anal. Calcd for C17H18N2O4: C, 64.96; H, 5.77. Found: C, 62.41; H,
5.55.
4.2.1.5. N1,N3-Bis-(4-methoxy-phenyl)-malonamide (104). Yield
(2 g, 74%): mp: 180–184 °C (cryst. water); mmax (KBr): 3280,
1682 cmꢀ1 1H NMR (400 MHz, DMSO-d6): d 3.41 (2H, CH2, D-ex-
;
change), 3.70 (s, 6H, OCH3),, 6.87 (d, J = 8 Hz, 4H), 7.50 (d,
J = 8 Hz, 4H), 10.03 ppm (br s, 2H, NH) ppm; 13C NMR (100 MHz,
DMSO-d6): d 48.20 (CH2), 57.79 (2 ꢁ OCH3), 116.52 (4 ꢁ CH),
123.19 (4 ꢁ CH), 134.68 (2 ꢁ C), 157.93 (2 ꢁ C), 167.57 (2 ꢁ C@O)
ppm; HRMS-ESI m/z [M+Na]+ calcd for C17H18N2NaO4: 337.11642,
found 337.11588. Anal. Calcd for C17H18N2O4: C, 64.96; H, 5.77.
Found: C, 63.02; H, 5.77.
4.2.1. Preparation of malonyl diamide derivatives (100–111)
Malonyl diamide derivatives were prepared as described ear-
lier.49 Briefly: To a magnetically stirred ice bath-cooled solution
of the selected amine (14 mmol), and triethylamine (14 mmol,
1.4 g) in dioxane (20 ml), malonyl dichloride (1.0 g, 7.0 mmol)
solution in dioxane (20 ml) was added dropwise over 15 min.
The reaction mixture was stirred at 0 °C for one hour. Subse-
quently, HCl (2 N, 50 ml) was added to the reaction suspension
and stirred for 30 minutes. Finally, the mixture was filtered to af-
ford the products as whitish to gray solids that were further
washed by stirring in water (50 mL) for 30 min (Scheme 1).
Using the same general procedure, the following compounds
were prepared.
4.2.1.6.
(105). Yield (1.8 g, 60%): mp: 157–158 °C (cryst. water); mmax
(KBr): 3280, 1655 cmꢀ1 1H NMR (400 MHz, DMSO-d6): d 2.20 (s,
N1,N3-Bis-(5-chloro-2-methylphenyl)-malonamide
;
6H, CH3), 3.55 (2H, CH2, D-exchange), 7.11 (m, 2H), 7.22 (m, 2H),
7.68 (m, 2H), 9.76 ppm (br s, 2H, NH) ppm; 13C NMR (100 MHz,
DMSO-d6):
d
19.87 (2 ꢁ CH3), 46.60 (CH2), 125.97 (2 ꢁ CH),
127.26 (2 ꢁ CH), 132.06 (2 ꢁ C), 132.62 (2 ꢁ C), 134.45 (2 ꢁ CH),
139.84 (2 ꢁ C), 168.64 (2 ꢁ C@O) ppm; HRMS-ESI m/z [M+Na]+
calcd for C17H16N2Cl2NaO2: 373.04865, found 373.04810. Anal.
Calcd for C17H16Cl2N2O2: C, 58.13; H, 4.59; N, 7.98. Found: C,
57.98; H, 4.47; N, 7.87.
4.2.1.1. N1,N3-Di-o-tolylmalonamide (100). Yield (1.4 g, 56%):
mp: 192–194 °C (cryst. water); mmax (KBr): 3273, 1687 cmꢀ1 1H
;
NMR (400 MHz, DMSO-d6): d 2.20 (s, 6H, CH3), 3.42 (s, 2H, CH2,
D-exchange), 7.08 (m, 2H), 7.20 (m, 4H), 7.52 (m, 2H), 9.67 ppm
(br s, 2H, NH) ppm; 13C NMR (100 MHz, DMSO-d6): d 20.40
(2 ꢁ CH3), 46.56 (CH2), 126.94 (2 ꢁ CH), 127.81 (2 ꢁ CH), 128.68
(2 ꢁ CH), 132.98 (2 ꢁ CH), 133.72 (2 ꢁ C), 138.59 (2 ꢁ C), 168.38
(2 ꢁ C@O) ppm; HRMS-ESI m/z [M+Na]+ calcd for C17H18N2NaO2:
4.2.1.7. N1-(Naphthalen-1-yl)-N3-(naphthalen-1-yl)malonamide
(106). Yield (1.8 g, 60%): mp: 220–223 °C (cryst. water); mmax
(KBr): 3300, 1640 cmꢀ1 1H NMR (400 MHz, DMSO-d6): d 3.55
;
(2H, CH2, D-exchange), 7.55 (m, 6H), 7.80 (m, 4H), 7.94 (m, 2H),
8.18 (m, 2H), 10.36 ppm (br s, 2H, NH) ppm; 13C NMR (100 MHz,
305.12659, found 305.12605. Anal. Calcd for
C
17H18N2O2: C,
DMSO-d6):
d
47.10 (CH2), 124.11 (2 ꢁ CH), 125.26 (2 ꢁ CH),
72.32; H, 6.43; N, 9.92. Found: C, 69.99; H, 6.73; N, 9.23.
128.13 (2 ꢁ CH), 128.28 (2 ꢁ CH), 128.71 (2 ꢁ CH), 130.22
(2 ꢁ CH), 130.84 (2 ꢁ CH), 135.82 (2 ꢁ C), 135.95 (2 ꢁ C), 136.38
(2 ꢁ C), 169.25 (2 ꢁ C@O) ppm; HRMS-ESI m/z [M+Na]+ calcd for
4.2.1.2. N1,N3-Di-m-tolylmalonamide (101). Yield (1.4 g, 56%):
mp: 150–153 °C (cryst. water); mmax (KBr): 3273, 1687ꢀ1
;
1H
C
C
23H18N2NaO2: 377.12659, found 377.12605. Anal. Calcd for
23H18N2O2: C, 77.95; H, 5.12; N, 7.90. Found: C, 76.70; H, 4.76;
NMR (400 MHz, DMSO-d6): d 2.27 (s, 6H, CH3), 3.45 (2H, CH2, D-
exchange), 6.86 (m, 2H), 7.18 (m, 2H), 7.37 (m, 2H), 7.43 (m, 2H),
10.10 ppm (br s, 2H, NH) ppm; 13C NMR (100 MHz, DMSO-d6): d
23.83 (2 ꢁ CH3), 48.45 (CH2), 118.86 (2 ꢁ CH), 122.19 (2 ꢁ CH),
126.78 (2 ꢁ CH), 131.27 (2 ꢁ CH), 140.62 (2 ꢁ C), 141.39 (2 ꢁ C),
167.95 (2 ꢁ C@O) ppm; HRMS-ESI m/z [M+Na]+ calcd for
C17H18N2NaO2: 305.12659, found 305.12605. Anal. Calcd for
N, 7.73.
4.2.1.8. N1,N3 -Dibenzyl-malonamide (107). Yield (1.4 g, 56%):
mp: 139–142 °C (cryst. water); mmax (KBr): 3284, 1655 cmꢀ1 1H
;
NMR (400 MHz, DMSO-d6): d 3.42 (2H, CH2, D-exchange), 4.24
(m, 4H, CH2), 7.28 (m, 10H), 8.51 ppm (br s, 2H, NH) ppm; 13C
C
17H18N2O2: C, 72.32; H, 6.43. Found: C, 70.19; H, 6.35.
NMR (100 MHz, DMSO-d6):
d
44.87 (CH2), 46.12 (2 ꢁ CH2),
129.43 (2 ꢁ CH), 129.86 (4 ꢁ CH), 131.15 (4 ꢁ CH), 141.84
4.2.1.3. N1,N3-Di-p-tolylmalonamide (102). Yield (1.5 g, 60%):
mp: 242–245 °C (cryst. water); mmax (KBr): 3273, 1674, 1651
(2 ꢁ C), 169.52 (2 ꢁ C@O) ppm; HRMS-ESI m/z [M+Na]+ calcd for
C
C
17H18N2NaO2: 305.12659, found 305.12605. Anal. Calcd for
17H18N2O2: C, 72.32; H, 6.43. Found: C, 69.35; H, 5.62.
cmꢀ1 1H NMR (400 MHz, DMSO-d6): d 2.23 (s, 6H, CH3), 3.41
;

4768
M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
4.2.1.9. N1,N3 -Bis(3,4-dichlorophenyl)malonamide (108). Yield
(1.7 g, 61%): mp: 214–216 °C (cryst. water); mmax (KBr): 3276,
4.2.2.2. N-m-Tolyl-succinamic acid (113). Yield (1.4 g, 70%): mp:
136–140 °C (cryst. diethylether); mmax (KBr): 3300–2700, 1693 cm
ꢀ1
1655 cmꢀ1
;
1H NMR (400 MHz, DMSO-d6): d 3.51 (2H, CH2, D-
.
1H NMR (400 MHz, DMSO-d6): d 1.80 (s,3H, CH3), 2.58 (m, 2H,
exchange), 7.47 (m, 2H), 7.54 (m, 2H), 7.98 (m, 2H), 10.50 ppm
(br s, 2H, NH) ppm; 13C NMR (100 MHz, DMSO-d6): d 48.62
(CH2), 121.78 (2 ꢁ CH), 122.86 (2 ꢁ CH), 127.59 (2 ꢁ C), 133.37
(2 ꢁ CH), 133.71 (2 ꢁ C), 141.48 (2 ꢁ C) 168.20 (2 ꢁ C@O) ppm;
HRMS-ESI m/z [M+Na]+ calcd for C15H10Cl4N2NaO2: 412.93940,
found 412.93886. Anal. Calcd for C15H10Cl4N2O2: C, 45.95; H,
2.57. Found: C, 45.23; H, 2.82.
CH2), 2.73 (m,2H, CH2), 7.51 (m, 3H), 8.00 (m, 1H), 10.21 (br s,
1H, NH), 12.00 (br s, 1H, OH) ppm; 13C NMR (100 MHz, DMSO-
d6): d 23.72 (CH3), 31.21 (CH2), 33.66 (CH2), 116.71 (CH), 121.90
(CH), 125.27 (CH), 131.43 (CH), 134.51 (C), 140.89 (C), 172.52
(C@O), 176.31 (C@O) ppm; HRMS-ESI m/z [MꢀH]+ calcd for
C11H12NO3: 206.08172, found 206.08227.
4.2.2.3. N-4-Methyl-o -anisidine-succinamic acid (114). Yield
(1.4 g, 70%): mp: 147–150 °C (cryst. diethylether); mmax (KBr):
4.2.1.10. N1,N3 -Bis(2-chlorophenyl)malonamide (109). Yield
(1.8 g, 14 mmol) and triethylamine (1.4 g, 14 mmol) to yield the ti-
tle compound as white powder (1.5 g, 53%): mp: 168–170 °C (cryst.
3300–2650, 1701 cmꢀ1 1H NMR (400 MHz, DMSO-d6): d 2.01
.
(s,3H, CH3), 2.35 (m, 2H, CH2), 2.72 (m,2H, CH2), 3.98 (s, 3H,
OCH3), 7.03 (m, 1H), 7.18 (m, 1H), 7.71 (m, 1H), 9.83 (br s, 1H,
NH), 12.14 ppm (br s, 1H, OH) ppm; 13C NMR (100 MHz, DMSO-
d6): d 23.51 (CH3), 31.01 (CH2), 33.521 (CH2), 57.97 (OCH3),
116.42 (CH), 123.08 (CH), 127.94 (CH), 135.15 (C), 157.62 (C),
161.44 (C), 172.17 (C@O), 176.53 (C@O) ppm; HRMS-ESI m/z [M-
H]+ calcd for C12H14NO4: 236.9228, found 236.09283.
water); mmax (KBr): 3276, 1655 cmꢀ1 1H NMR (400 MHz, DMSO-
;
d6): d 3.57 (2H, CH2, D-exchange), 7.19 (m, 2H), 7.34 (m, 2H),
7.49 (m, 2H), 7.88 (m, 2H), 10.10 ppm (br s, 2H, NH) ppm; 13C
NMR (100 MHz, DMSO-d6): 46.5 (CH2), 127.61 (2 ꢁ CH), 128.05
(2 ꢁ CH), 128.88 (2 ꢁ CH), 130.23 (2 ꢁ CH), 132.15 (2 ꢁ C),
137.05 (2 ꢁ C) 168.61 (2 ꢁ C@O) ppm; HRMS-ESI m/z [M+Na]+
calcd for C15H12Cl2N2NaO2: 345.01735, found 345.01680. Anal.
Calcd for C15H12Cl2N2O2: C, 55.75; H, 3.74. Found: C, 54.55; H, 3.83.
4.2.2.4. N-Naphthylamine-succinamic acid (115). Yield (1.43 g,
0.01 mmol) to yield the title compound as faint violet powder
(1.4 g, 70%): mp: 147–150 °C (cryst. diethylether); mmax (KBr):
4.2.1.11. N1,N3 -Bis(4-chlorophenyl)malonamide (110). Yield
(2 g, 61%): mp: 255–258 °C (cryst. water); mmax (KBr): 3276,
3300–2600, 1714 cmꢀ1 1H NMR (300 MHz, DMSO-d6): d 2.55 (t J
.
1655 cmꢀ1 1H NMR (400 MHz, DMSO-d6): d 3.52 (2H, CH2, D-ex-
;
6,2H, CH2), 2.70 (t, J 6, 2H, CH2), 7.38 (m, 4H), 7.60 (m, 1H), 7.70
(m, 1H), 7.85 (m, 1H), 8.20 (br s, 1H, NH), 9.95 ppm (br s, 1H,
OH); 13C NMR (75 MHz, DMSO-d6): d 29.49 (CH2), 31.06 (CH2),
122.00 (CH), 123.39 (CH), 125.62 (CH), 126.05 (CH), 126.23
(2 ꢁ CH), 126.49 (CH), 128.56 (C), 134.03 (C), 134.18 (C), 171.27
(C@O), 174.39 (C@O) ppm. HRMS-ESI m/z [MꢀH]+ calcd for
change), 7.34 (d, J = 8 Hz, 4H), 7.62 (d, J = 8 Hz, 4H), 10.33 ppm
(br s, 2H, NH) ppm; 13C NMR (100 MHz, DMSO-d6): d 48.54
(CH2), 123.22 (4 ꢁ CH), 123.31 (2 ꢁ C), 129.69 (4 ꢁ CH), 140.42
(2 ꢁ C), 168.06 (2 ꢁ C@O) ppm; HRMS-ESI m/z [M+Na]+ calcd for
C
C
15H12Cl2N2NaO2: 345.01735, found 345.01680. Anal. Calcd for
15H12Cl2N2O2: C, 55.75; H, 3.74; N, 8.67. Found: C, 57.57; H,
C14H12NO3: 242.08172, found 242.08227.
3.93; N, 8.65.
4.2.1.12. N1,N3 -Bis(2-methoxy-4-methylphenyl) malonamide
(111). Yield (1.7 g, 58%): mp: 160–163 °C (cryst. water); mmax
4.2.2.5. N-5-Chloro-2-methylaniline-succinamic acid (116).
Yield (1.42 g, 0.01 mmol) to yield the title compound as white
powder (1.4 g, 70%): mp: 163–165 °C (cryst. diethylether); mmax
(KBr): 3280, 1640 cmꢀ1 1H NMR (400 MHz, DMSO-d6): d 2.20 (s,
;
(KBr): 3300–2650, 1701 cmꢀ1 1H NMR (300 MHz, DMSO-d6): d
.
6H, CH3), 3.45 (2H, CH2, D-exchange), 3.67 (s, 6H, OCH3), 6.88 (m,
4H), 7.85 (m, 2H), 9.61 ppm (br s, 2H, NH) ppm; 13C NMR
2.14 (s, 3H, CH3), 2.46 (m, 2H, CH2), 2.57 (m, 2H, CH2), 7.16 (m,
2H), 7.55 (m, 1H), 9.35 (br s, 1H, NH), 12.00 ppm (br s, 1H, OH)
ppm; 13C NMR (75 MHz, DMSO-d6): d 17.78 (CH3), 29.38 (CH2),
30.99 (CH2), 124.21 (CH), 124.80 (CH), 130.15 (C), 132.22 (CH),
138.09 (C), 138.22 (C), 170.95 (C@O), 174.28 (C@O) ppm. HRMS-
ESI m/z [MꢀH]+ calcd for C11H11ClNO3: 240.04275, found
240.04329.
(100 MHz, DMSO-d6):
d
23.14 (2 ꢁ CH3), 47.09 (CH2), 58.43
(2 ꢁ OCH3), 113.65 (2 ꢁ CH), 124.76 (2 ꢁ CH), 127.33 (2 ꢁ CH),
129.24 (2 ꢁ C), 131.68 (2 ꢁ C), 149.88 (2 ꢁ C), 168.40 (2 ꢁ C@O)
ppm; HRMS-ESI m/z [M+Na]+ calcd for C19H22N2NaO4: 365.14772,
found 365.14718. Anal. Calcd for C19H22N2O4: C, 66.65; H, 6.48.
Found: C, 63.88; H, 6.28
4.2.2. Preparation of succinic monoamide derivatives
4.2.2.6. N-p-Anisidine-succinamic acid (117). Yield (1.4 g, 70%):
mp: 167–170 °C (cryst. diethylether); mmax (KBr): 3300–2700,
To a magnetically stirred solution of the selected aromatic amine
(0.01 mol) in anhydrous dioxane (30 ml), succinic anhydride
(0.01 mol) was added and the resulting suspension was refluxed
for 24 h. Subsequently, the reaction mixture was cooled to room
temperature. Subsequently, petroleum ether (200 mL) was added
to the reaction mixture to yield the succinic monoamide derivatives
(112–119, see Scheme 2) as white to grey solids. Using the same
general procedure, the following compounds were synthesized.
1701 cmꢀ1 1H NMR (300 MHz, DMSO-d6): d 2.48 (m, 2H, CH2),
.
3.01 (m, 2H, CH2), 3.70 (s, 3H, OCH3), 6.85 (m, 2H), 7.45 (m, 2H),
10.00 (br s, 1H, NH), 11.10 ppm (br s, 1H, OH) ppm; 13C NMR
(75 MHz, DMSO-d6): d 55.82 (OCH3), 44.14 (CH2), 46.18 (CH2),
114.56 (2 ꢁ CH), 121.27 (2 ꢁ CH), 132.67 (C), 156.01 (C), 164.22
(C@O), 168.47 (C@O) ppm. HRMS-ESI m/z [MꢀH]+ calcd for
C11H12NO4: 222.07663, found 222.07718.
4.2.2.1. N-o-Tolyl-succinamic acid (112). Yield (1.07 g,
0.01 mmol) to yield the title compound as white powder (1.4 g,
70%): mp: 144–146 °C (cryst. diethylether); mmax (KBr): 3300–
4.2.2.7. N-2-Chloroaniline-succinamic acid (118). Yield (1.4 g,
70%): mp: 145–148 °C (cryst. diethylether); mmax (KBr): 3300–
2650, 1701 cmꢀ1 1H NMR (400 MHz, DMSO-d6): d 3.49 (m,2H,
.
2700, 1655 cmꢀ1 1H NMR (400 MHz, DMSO-d6): d 1.78 (s,3H,
.
CH2), 3.61 (m,2H, CH2), 7.16 (m, 1H), 7.32 (m, 1H), 7.48 (m, 1H),
7.74 (m, 1H), 9.53 (br s, 1H, NH), 12.18 ppm (br s, 1H, OH) ppm;
13C NMR (100 MHz, DMSO-d6): d 31.57 (CH2), 33.26 (CH2), 128.64
(CH), 128.83 (CH), 129.94 (CH), 130.04 (CH), 137.65 (C), 155.11 (C),
173.21 (C@O), 176.44 (C@O) ppm; HRMS-ESI m/z [MꢀH]+ calcd for
CH3), 2.62 (m, 2H, CH2), 2.76 (m,2H, CH2), 7.48 (m, 3H), 8.00 (m,
1H), 10.27 (br s, 1H, NH), 12.13 (br s, 1H, OH) ppm; 13C NMR
(100 MHz, DMSO-d6): d 23.82 (CH3), 31.24 (CH2), 33.71 (CH2),
118.77 (CH), 122.10 (CH), 126.27 (CH), 131.13 (CH), 133.61 (C),
140.92 (C), 172.62 (C@O), 176.52 (C@O) ppm. HRMS-ESI m/z [M-
H]+ calcd for C11H12NO3: 206.08172, found 206.08227.
C10H9NClO3: 226.02710, found 226.02722.

M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
4769
4.2.2.8. N-2-Aminopyridine-succinamic acid (119). Yield (1.4 g,
70%): mp: 134–137 °C. (cryst. diethylether); mmax (KBr): 3300–
NMR (400 MHz, DMSO-d6): d 2.26 (s, 3H, CH3), 7.31 (m, 2H), 7.66
(m, 5H), 7.94 (m, 3H), 10.40 (br s, 1H, NH), 11.0 (br s, 1H, NH)
ppm; 13C NMR (100 MHz, DMSO-d6): d 19.76 (CH3), 125.60 (CH),
126.93 (CH), 128.16 (2 ꢁ CH), 128.52 (CH), 128.89 (2 ꢁ CH),
129.47 (CH), 130.79 (CH), 131.00 (C), 132.71 (CH), 133.84 (C),
134.61 (C), 134.96 (C), 136.34 (C), 139.06 (C), 161.32 (C@O),
161.91 (C@O) ppm; HRMS-ESI m/z [MꢀH]+ calcd for C19H14ClN2O2:
337.07438, found 337.07493. Anal. Calcd for C19H15ClN2O2: C,
67.36; H, 4.46; N, 8.27. Found: C, 68.28; H, 4.52; N, 8.19.
2650, 1701 cmꢀ1 1H NMR (400 MHz, DMSO-d6): d 2.81 (m,2H,
.
CH2), 3.40 (m,2H, CH2), 6.56 (m, 1H), 6.94 (m, 1H), 7.19 (m, 1H),
7.47 (m, 1H), 9.93 (br s, 1H, NH), 12.16 (br s, 1H, OH) ppm; 13C
NMR (100 MHz, DMSO-d6): d 31.43 (CH2), 32.46 (CH2), 115.82
(CH), 128.53 (CH), 132.25 (CH), 140.18 (CH), 160.63 (C), 172.87
(C@O), 176.23 (C@O) ppm; HRMS-ESI m/z [MꢀH]+ calcd for
C9H9N2O3: 193.06132, found 193.06187.
4.2.3. Preparation of N,N-di-oxalamide derivatives
4.2.3.5. N-Naphthalen-1-yl-N0-o-tolyl-oxalamide (124). Pre-
pared from 115 (1.0 g, 4 mmol) and commercially available
o-toluidine (0.44 g, 4 mmol) to yield the title compound as faint
off-white powder (0.3 g, 21%): mp: 125–129 °C (cryst. acetone);
To a magnetically stirred ice-bath cooled solution of the se-
lected succinic monoamide derivative (4 mmol) in anhydrous diox-
ane (20 ml) neat oxalyl chloride was added (8 mmol). After stirring
over 30 min the reaction mixture was warmed to room tempera-
ture. Two hours later, the reaction mixture was re-cooled to 0 °C
in ice bath followed by drop wise addition of the particular aro-
matic amine (4 mmol) in pyridine (10 mL) to the reaction mixture.
After stirring over 30 min the reaction mixture was warmed to
room temperature and allowed to stir over 24 h. Subsequently,
the reaction was carefully quenched with saturated sodium bicar-
bonate solution (200 mL). Finally, the mixture was filtered to offer
the products as whitish to gray solids that were re-crystallized
from acetone (see Scheme 2).
mmax (KBr): 3280, 1666, 1585 cmꢀ1 1H NMR (400 MHz, DMSO-
;
d6): d 2.25 (s, 3H, CH3), 7.24 (m, 4H), 7.56 (m, 4H), 7.88 (m, 3H),
10.30 (br s, 1H, NH), 10.90 (br s, 1H, NH) ppm; 13C NMR
(100 MHz, DMSO-d6): d 20.25 (CH3), 125.94 (CH), 125.63 (CH),
127.39 (CH), 127.50 (CH), 128.19 (CH), 128.87 (CH), 129.40 (CH),
130.80 (CH), 130.99 (CH), 133.10 (CH), 134.84 (CH), 134.92 (C),
135.07 (C), 136.34 (C), 137.67 (C), 137.75 (C), 161.18 (C@O),
162.23 (C@O) ppm. Anal. Calcd for C19H16N2O2: C, 74.98; H, 5.30;
N, 9.20. Found: C, 72.06; H, 5.41; N, 9.16.
4.2.3.6. N-(2-Methoxy-phenyl)-N0-naphthalen-1-yl-oxalamide
(125). Prepared from 115 (1.0 g, 4 mmol) and commercially avail-
able o-anisidine (0.51 g, 4 mmol) to yield the title compound as
faint off-white powder (0.25 g, 17%): mp: 176–179 °C (cryst. ace-
4.2.3.1. N,N0-Di-o -tolyl-oxalamide (120). Prepared from 112
(1.0 g, 5 mmol) and commercially available o-toluidine (0.56 g,
5 mmol) to yield the title compound as faint off-white powder
(0.3 g, 20%): mp: 208–210 °C (cryst. acetone); mmax (KBr): 3288,
tone); mmax (KBr): 3344, 3303, 1678, 1668 cmꢀ1
;
1H NMR
1666 cmꢀ1
;
1H NMR (400 MHz, DMSO-d6): d 2.30 (s, 6H, CH3),
(400 MHz, DMSO-d6): d 3.90 (s, 3H, OCH3), 7.18 (m, 3H), 7.57 (m,
4H), 7.91 (m, 3H), 8.27 (m, 1H), 9.88 ppm (br s, 1H, NH), 11.05
(br s, 1H, NH) ppm; 13C NMR (100 MHz, DMSO-d6): d 58.68
(OCH3), 113.93 (CH), 122.15 (CH), 122.29 (CH), 123.36 (CH),
125.64 (CH), 125.92 (C), 128.14 (2 ꢁ CH), 128.37 (C), 128.92
(2 ꢁ CH), 129.60 (CH), 130.75 (CH), 131.06 (C), 134.96 (C), 136.32
(C), 151.64 (C@O), 160.00 (C@O) ppm; HRMS-ESI m/z [M+Na]+
calcd for C19H16N2NaO3: 43.10586, found 343.11188. Anal. Calcd
for C19H16N2O3: C, 71.24; H, 5.03; N, 8.74. Found: C, 70.96; H,
5.20; N, 8.75.
6.97 (m, 2H), 7.24 (m, 2H), 7.64 (m, 4H), 10.68 ppm (br s, 2H,
NH) ppm; 13C NMR (100 MHz, DMSO-d6): d 23.86 (2 ꢁ CH3),
120.23 (2 ꢁ CH), 123.56 (2 ꢁ CH), 128.02 (2 ꢁ CH), 131.25
(2 ꢁ CH), 140.05 (2 ꢁ C), 140.66 (2 ꢁ C), 161.16 (2 ꢁ C@O) ppm.
Anal. Calcd for C16H16N2O2: C, 71.62; H, 6.01; N, 10.44. Found: C,
68.50; H, 5.89; N, 9.95.
4.2.3.2. N,N0-Di-m -tolyl-oxalamide (121). Prepared from 113
(1.0 g, 5 mmol) and commercially available m-toluidine (0.56 g,
5 mmol) to yield the title compound as faint off-white powder
(0.35 g, 23%): mp: 132–135 °C (cryst. acetone); mmax (KBr): 3292,
1668 cmꢀ1; 1H NMR (400 MHz, DMSO-d6): d 2.30 (s, 6H, CH3), 6.97
(m, 2H), 7.24 (m, 2H), 7.64 (m, 4H), 10.67 ppm (br s, 2H, NH) ppm;
13C NMR (100 MHz, DMSO-d6): d 23.83 (2 ꢁ CH3), 120.28 (2 ꢁ CH),
123.56 (2 ꢁ CH), 127.99 (2 ꢁ CH), 131.24 (2 ꢁ CH), 140.15 (2 ꢁ C),
140.63 (2 ꢁ C), 161.22 (2 ꢁ C@O) ppm. Anal. Calcd for C16H16N2O2:
C, 71.62; H, 6.01. Found: C, 71.40; H, 6.87.
4.2.3.7. N-(2-Methoxy-4-methyl-phenyl)-N 0-naphthalen-1-yl-
oxalamide (126). Prepared from 115 (1.0 g, 4 mmol) and commer-
cially available 4-methyl-o-anisidine (0.55 g, 4 mmol) to yield the
title compound as faint off-white powder (0.35 g, 23%): mp: 160–
162 °C (cryst. acetone); mmax (KBr): 3357, 1699, 1682 cmꢀ1 1H
;
NMR (400 MHz, DMSO-d6): d 2.26 (s, 3H, CH3), 3.86 (s, 3H, OCH3),
7.00 (m, 2H), 7.57 (m, 4H), 7.91 (m, 2H), 7.97 (m, 1H), 8.10 (s, 1H),
9.80 (br s, 1H, NH), 11.0 (br s, 1H, NH) ppm; 13C NMR (100 MHz,
DMSO-d6): d 23.23 (CH3), 58.69 (OCH3), 113.73 (CH), 122.72 (CH),
125.51 (CH), 125.78 (CH), 127.96 (C), 128.16 (CH), 128.25 (CH),
128.96 (2 ꢁ CH), 129.61 (CH), 130.79 (CH), 130.98 (C), 132.19 (C),
134.74 (C), 136.31 (C), 149.49 (C), 159.74 (C@O), 161.92 (C@O)
ppm; HRMS-ESI m/z [M+Na]+ calcd for C20H18N2NaO3: 357.12151,
found 357.12096.. Anal. Calcd for C20H18N2O3: C, 71.84; H, 5.43; N,
8.38. Found: C, 69.61; H, 5.55; N, 8.34.
4.2.3.3.
N,N 0-Bis-(2-methoxy-4-methyl-phenyl)-oxalamide
(122). Prepared from 113 (1.0 g, 4 mmol) and commercially avail-
able 4-methyl-o-anisidine (0.55 g, 4 mmol) to yield the title com-
pound as faint off-white powder (0.3 g, 19%): mp: Decompose at
180 °C (cryst. acetone); mmax (KBr): 3361, 1687 cmꢀ1 1H NMR
;
(400 MHz, DMSO-d6): d 2.27 (s, 6H, CH3), 3.87 (s, 6H, OCH3), 7.01
(m, 4H), 8.00 (m, 2H), 9.80 ppm (br s, 2H, NH) ppm; 13C NMR
(100 MHz, DMSO-d6):
d
23.16 (2 ꢁ CH3), 58.76 (2 ꢁ OCH3),
113.81 (2 ꢁ CH), 122.84 (2 ꢁ CH), 122.99 (2 ꢁ CH), 128.47
(2 ꢁ C), 132.20 (2 ꢁ C), 149.57 (2 ꢁ C), 159.68 (2 ꢁ C@O) ppm;
HRMS-ESI m/z [M+Na]+ calcd for C18H20N2NaO4: 351.13207, found
351.13152.
4.2.3.8. N-Naphthalen-1-yl-N0-p-tolyl-oxalamide (127). Pre-
pared from 115 (1.0 g, 4 mmol) and commercially available
p-toluidine (0.44 g, 4 mmol) to yield the title compound as faint
off-white powder (0.2 g, 14%): mp: 166–170 °C (cryst. acetone);
mmax (KBr): 3340, 3302, 1703, 1674 cmꢀ1 1H NMR (300 MHz,
;
4.2.3.4.
N-(5-Chloro-2-methyl-phenyl)-N 0-naphthalen-1-yl-
DMSO-d6): d 2.26 (s, 3H, CH3), 7.58 (m, 1H), 7.65 (m, 4H), 7.76
(m, 2H), 7.87 (m, 4H), 10.75 (br s, 1H, NH), 10.93 (br s, 1H, NH)
ppm; 13C NMR (75 MHz, DMSO-d6): d 21.06 (CH3), 120.93 (CH),
123.44 (CH), 123.56 (CH), 123.63 (CH), 126.06 (CH), 126.72
(2 ꢁ CH), 126.77 (CH), 127.24 (CH), 128.63 (CH), 128.91(2 ꢁ C),
oxalamide (123). Prepared from 116 (1.0 g, 4 mmol) and com-
mercially available 1-naphthylamine (0.60 g, 4 mmol) to yield the
title compound as faint off-white powder (0.2 g, 13%): mp: 162–
165 °C (cryst. acetone); mmax (KBr): 3276, 1674, 1666 cmꢀ1 1H
;

4770
M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
129.66 (CH), 133.02 (C), 134.20 (C), 135.69 (C), 158.93 (C@O),
160.21 (C@O) ppm. Anal. Calcd for C19H16N2O2: C, 74.98; H, 5.30;
N, 9.20. Found: C, 71.35; H, 5.07; N, 8.83.
(m, 3H), 7.96 (m, 1H), 8.27 (m, 1H), 8.42 (m, 2H), 9.86 (br s, 1H,
NH), 10.27 (br s, 1H, NH) ppm; 13C NMR (100 MHz, DMSO-d6): d
57.83 (OCH3), 113.0 (CH), 115.69 (2 ꢁ CH), 121.47 (CH), 122.94
(CH), 123.12 (2 ꢁ CH), 127.60 (CH), 132.71 (CH), 135.21 (C),
136.17 (C), 141.0 (C), 161.26 (C@O), 164.58 (C@O) ppm; HRMS-
ESI m/z [M+Na]+ calcd for C14H13N3O3: 272.10352, found
272.10297. Anal. Calcd for C14H13N3O3: C, 61.99; H, 4.83; N,
15.49. Found: C, 60.77; H, 4.77; N, 14.79.
4.2.3.9. N-Naphthalen-1-yl-N0-m-tolyl-oxalamide (128). Pre-
pared from 115 (1.0 g, 4 mmol) and commercially available
m-toluidine (0.44 g, 4 mmol) to yield the title compound as faint
off-white powder (0.25 g, 17%): mp: 127–130 °C (cryst. acetone);
mmax (KBr): 3340, 3265, 1687, 1666 cmꢀ1 1H NMR (300 MHz,
;
DMSO-d6): d 2.28 (s, 3H, CH3), 7.58 (m, 1H), 7.65 (m, 4H), 7.76
(m, 2H), 7.87 (m, 4H), 10.75 (br s, 1H, NH), 10.93 (br s, 1H, NH)
ppm; 13C NMR (75 MHz, DMSO-d6): d 21.73 (CH3), 117.97 (CH),
118.06 (CH), 121.30 (CH), 123.31(CH), 125.80 (CH), 126.01 (CH),
126.71 (CH), 126.75 (CH), 127.20 (CH), 128.60 (CH), 128.80 (C),
129.07 (CH), 132.80 (C), 134.17 (C), 137.90 (C), 138.49 (C),
158.95 (C@O), 160.26 (C@O) ppm. Anal. Calcd for C19H16N2O2: C,
74.98; H, 5.30; N, 9.20. Found: C, 72.65; H, 5.47; N, 9.61.
4.3. In vitro experimental studies
4.3.1. Chemicals of bioassay
Materials were purchased from corresponding companies and
were used in the experimentation without further purification.
GPa rabbit enzyme, dimethyl sulfoxide (DMSO), Hepes buffer,
Glucose-1-phosphate, glycogen, magnesium chloride, KCl, mala-
chite green, ammonium molybdate, EGTA. All were obtained from
(Sigma–Aldrich, USA). NCI hit were kindly donated by the National
Caner Institute.
4.2.3.10.
N-(5-Chloro-2-methyl-phenyl)-N’-(2-methoxy-4-
methyl-phenyl)-oxalamide (129). Prepared from 116 (1.0 g,
4 mmol) and commercially available 4-methyl-o-anisidine
(0.0.57 g, 4 mmol) to yield the title compound as faint off-white
powder (0.35 g, 22%): mp: 160–163 °C (cryst. acetone); mmax
4.3.2. In vitro GP enzyme inhibition assay
GPa (from rabbit muscle), glycogen, glucose-1-phosphate, mal-
achite green, and ammonium molybdate were purchased from the
Sigma–Aldrich Corporation (St. Louis, MO, USA). Reagents and sol-
vents were obtained from commercial suppliers and used without
further purification.
The enzymatic inhibition of phosphorylase activity was moni-
tored using microplate reader (BioTek, USA) based on published
methods.4 In brief, GPa activity was measured in the direction of
glycogen synthesis by the release of phosphate from glucose-1-
phosphate. Each compound was dissolved in DMSO and diluted
(KBr): 3361, 1687 cmꢀ1 1H NMR (400 MHz, DMSO-d6): d 2.26 (s,
;
6H, 2 ꢁ CH3), 3.87 (s, 3H, OCH3), 7.00 (m, 3H), 8.00 (8, 3H), 9.81
(br s, 1H, NH), 9.91 (br s, 1H, NH) ppm; 13C NMR (100 MHz,
DMSO-d6): d 22.24 (2 ꢁ CH3), 57.84 (OCH3), 112.73 (CH), 112.81
(CH), 113.73 (CH), 116.52 (CH), 121.96 (CH), 122.52 (CH), 127.59
(C), 128.73 (C), 130.68, (C), 131.77 (C), 132.17 (C), 149.32 (C),
159.16 (C@O), 161.56 (C@O) ppm; HRMS-ESI m/z [M+Na]+ calcd
for C17H17ClN2NaO3: 355.08254, found 355.08199.
appropriately. The enzyme was added to the 100 lL buffer contain-
4.2.3.11. N-(2-Methoxy-4-methyl-phenyl)-N’-(4-methoxy-phe-
nyl)-oxalamide (130). Prepared from 117 (1.0 g, 4 mmol) and
commercially available 4-methyl-o-anisidine (0.0.58 g, 4 mmol)
to yield the title compound as faint off-white powder (0.30 g,
19%): mp: 152–156 °C (cryst. acetone); mmax (KBr): 3334, 3302,
ing 50 mM Hepes (pH 7.2), 100 mM KCl, 2.5 mM MgCl2, 0.5 mM
glucose-1-phosphate, and 1 mg/mL glycogen together with tested
compounds in 96-well microplates. After the addition of 150 lL
of 1 M HCl containing 10 mg/mL ammonium molybdate and
0.38 mg/mL malachite green, reactions were run at 22 °C for
20 min, then the phosphate absorbance was measured at 620 nm.
Caffeine was used as a standard inhibitor for GP enzyme.68
1666, 1660 cmꢀ1 1H NMR (400 MHz, DMSO-d6): d 2.27 (s, 3H,
;
CH3), 3.74 (s, 3H, OCH3), 3.86 (s, 3H, OCH3), 6.98 (m, 4H), 7.75
(m, 2H), 8.02 (m, 1H), 9.84 ppm (br s, 1H, NH), 10.08 ppm (br s,
1H, NH) ppm; 13C NMR (100 MHz, DMSO-d6): d 23.24 (CH3),
57.87 (OCH3), 58.71 (OCH3), 113.73 (CH), 116.52 (2 ꢁ CH), 122.59
(C), 122.73 (CH), 124.68, (CH), 124.77 (2 ꢁ CH), 128.17 (C),
132.12 (C), 133.10 (C) 149.55 (C), 158.96 (C@O), 160.42 (C@O)
ppm. Anal. Calcd for C17H18N2O4: C, 64.96; H, 5.77; N, 8.91. Found:
C, 63.02; H, 5.57; N, 8.43.
Acknowledgments
This project was partially sponsored by the Faculty of Graduate
Studies (This work is part of Ph.D. Thesis of Rand Shahin). The
authors thank the Deanship of Scientific Research and Hamdi-
Mango Center for Scientific Research at the University of Jordan
for their generous funds.
4.2.3.12. N-Benzyl-N ’-(2-chloro-phenyl)-oxalamide (131). Pre-
pared from 118 (1.0 g, 4 mmol) and commercially available benzyl-
amine (0.47 g, 4 mmol) to yield the title compound as faint off-white
powder (0.30 g, 20%): mp: 139–143 °C (cryst. water); mmax (KBr):
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2011.06.086.
3286, 1668, 1655 cmꢀ1 1H NMR (400 MHz, DMSO-d6): d 4.4 (2H,
;
CH2), 7.23 (m, 2H), 7.31 (m, 4H), 7.40 (m, 1H), 7.55 (m, 1H), 8.00
(m, 1H), 9.70 (br s, 1H, NH), 10.16 (br s, 1H, NH) ppm; 13C NMR
(100 MHz, DMSO-d6): d 45.44 (CH2), 126.26 (CH), 126.50 (CH),
129.49 (CH), 129.70 (CH), 130.15 (2 ꢁ CH), 130.58 (2 ꢁ CH),
130.99 (CH), 132.21 (C), 136.17 (C), 141.0 (C), 160.73 (C@O),
162.20 (C@O) ppm. Anal. Calcd for C15H13ClN2O2: C, 62.40; H,
4.52; N, 9.70. Found: C, 62.59; H, 4.46; N, 9.78.
References and notes
1. (a) Ercan-Fang, N.; Taylor, R. M.; Treadway, L. J.; Levy, B. C.; Genereux, E. P.;
Gibbs, M. E.; Rath, L. V.; Kwon, Y.; Gannon, C. M.; Nuttall, Q. F. Am. J. Physiol.
Endocrinol. Metab. 2005, 289, E366; (b) Praly, J.-P.; Vidal, S. Mini-Rev. Med. Chem.
2010, 10, 1102; (c) Hayes, J. M.; Leonidas, D. D. Mini-Rev. Med. Chem. 2010, 10,
1156.
2. Tsirkone, V. G.; Tsoukala, E.; Lamprakis, C.; Manta, S.; Hayes, J. M.; Skamnaki, V.
T.; Drakou, C.; Zographos, S. E.; Komiotis, D.; Leonidas, D. D. Bioorg. Med. Chem.
2010, 15, 3413.
3. Oikonomakos, N. G. Curr. Protein Pept. Sci. 2002, 3, 561.
4. Martin, W. H.; Hoover, D. J.; Armento, S. J.; Stock, I. A.; Mcpherson, R. K.; Danley,
D. E.; Stevenson, R. W.; Barrett, E. J.; Tredway, J. L. Proc. Natl. Acad. Sci. U.S.A.
1998, 95, 1776.
5. (a) Deng, Q.; Lu, Z.; Bohn, J.; Ellsworth, K. P.; Myers, R. W.; Geissler, W. M.;
Harris, G.; Willoughby, C. A.; Chapman, K.; McKeever, B.; Mosley, R. J. Mol.
4.2.3.13.
N-(2-Methoxy-phenyl)-N’-pyridin-2-yl-oxalamide
(132). Prepared from 119 (1.0 g, 5 mmol) and commercially avail-
able o-anisidine (0.63 g, 5 mmol) to yield the title compound as
faint off-white powder (0.35 g, 21%): mp: 132–136 °C (cryst.
water); mmax (KBr): 3383, 3357, 1716, 1691 cmꢀ1
;
1H NMR
(400 MHz, acetone-d6): d 4.00 (s, 3H, OCH3), 7.07 (m, 1H), 7.27

M. Habash, M. O. Taha / Bioorg. Med. Chem. 19 (2011) 4746–4771
4771
Graphics Modell. 2005, 23, 457; (b) Oikonomakos, N. G.; Somsak, L. Curr. Opin.
Investig. Drugs 2008, 9, 379; (c) Somsak, L.; Czifrak, K.; Toth, M.; Bokor, E.;
Chrysina, E. D.; Alexacou, K. M.; Hayes, J. M.; Tiraidis, C.; Lazoura, E.; Leonidas,
D. D.; Zographos, S. E.; Oikonomakos, N. G. Curr. Med. Chem. 2008, 15, 2933; (d)
Henke, B. R.; Sparks, S. M. Mini-Rev. Med. Chem. 2006, 6, 845.
6. Zographos, S. E.; Oikonomakos, N. G.; Tsitsanou, K. E.; Leonidas, D. D.; Chrysina,
E. D.; Skamnaki, V. T.; Bischoff, H.; Goldmann, S.; Watson, K. A.; Johnson, L. N.
Structure 1997, 5, 1413.
40. Poptodorov, K.; Luu, T.; Langer, T.; Hoffmann, R. In Methods and Principles in
Medicinal Chemistry. Pharmacophores and Pharmacophores Searches; Hoffmann,
R. D., Ed.; Wiley-VCH: Weinheim, Germany, 2006; Vol. 2, pp 17–47.
41. Sutter, J.; Güner, O.; Hoffmann, R.; Li, H.; Waldman, M. In Pharmacophore
Perception, Development, and Use in Drug Design; Güner, O. F., Ed.; International
University Line: La Jolla, CA, 2000; p 501.
42. Bersuker, I. B.; Bahçeci, S.; Boggs, J. E. In Pharmacophore Perception,
Development, and Use in Drug Design; Güner, O. F., Ed.; International
University Line: La Jolla, CA, 2000; p 457.
7. Agius, L. Mini-Rev. Med. Chem. 2010, 10, 1175.
8. Chen, L.; Li, H.; Liu, J.; Zhang, L.; Liu, H.; Jiang, H. Bioorg. Med. Chem. 2007, 15,
6763.
9. (a) Chen, J.; Liu, J.; Zhang, L. Y.; Wu, G.; Hua, W.; Wu, X.; Sun, H. Bioorg.
Med. Chem. Lett. 2006, 16, 2915; (b) Bradley, S. E.; Jeevaratnam, R. P.;
Krulle, T. M.; Procter, M. J.; Rowley, R. J.; Thomas, G. H.; Valdes,
WO2005085245, 2005.
43. Fisher, R. The Principle of Experimentation Illustrated by a Psycho-Physical
ExpeHafner Publishing Co, 8th ed.; Hafner Publishing: New York, 1966.
44. CERIUS2, QSAR Users’ Manual, version 4.10; Accelrys Inc.: San Diego, CA, 2005, pp
43–88, 221–235.
45. Ramsey, L. F.; Schafer, W. D. The Statistical Sleuth, 1st ed.; Wadsworth
Publishing Company: Belmont, CA, 1997.
10. Prathipati, P.; Pandey, G.; Saxena, K. A. J. Chem. Inf. Model. 2005, 45, 136.
11. Gohlke, H.; Klebe, G. J. Med. Chem. 2002, 45, 4153.
46. Sprague, P. W.; Hoffmann, R.; Van de Waterbeemd, H.; Testa, B.; Folkers, G.
Curr. Tools Med. Chem. 1997, 230.
12. Sutherland, J. J.; Nandigam, R. K.; Erickson, J. A.; Vieth, M. J. Chem. Inform.
Model. 2007, 47, 2293.
47. Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J. Adv. Drug Delivery Rev.
2001, 46, 3.
13. Evans, K. A.; Li, Y. H.; Coppo, F. T.; Graybill, T. L.; Cichy-Knight, M.; Patel, M.;
Gale, J.; Li, H.; Thrall, S. H.; Tew, D.; Tavares, F.; Thomson, S. A.; Weiel, E. J.;
Boucheron, J. A.; Clancy, D. C.; Epperly, A. H.; Golden, P. L. Bioorg. Med. Chem.
Lett. 2008, 18, 4068.
14. Peng, Z.; ZhiLiang, L. Sci. China, Ser. B 2007, 50, 568.
15. Beeley, N. R. A.; Sage, C. Targets 2003, 2, 19.
16. Klebe, G. Drug Discovery Today 2006, 11, 580.
17. Steuber, H.; Zentgraf, M.; Gerlach, C.; Sotriffer, C. A.; Heine, A.; Klebe, G. J. Mol.
Biol. 2006, 363, 174.
18. Stubbs, M. T.; Reyda, S.; Dullweber, F.; Moller, M.; Klebe, G.; Dorsch, D.;
Mederski, W.; Wurziger, H. ChemBioChem 2002, 3, 246.
19. DePristo, M. A.; de Bakker, P. I. W.; Blundell, T. L. Structure 2004, 12, 831.
20. Taha, M. O.; Bustanji, Y.; Al-Ghussein, M. A. S.; Mohammad, M.; Zalloum, H.; Al-
Masri, I. M.; Atallah, N. J. Med. Chem. 2007, 51, 2062.
21. Taha, M. O.; Atallah, N.; Al-Bakri, A. G.; Paradis-Bleau, C.; Zalloum, H.; Younis,
K.; Levesque, R. C. Bioorg. Med. Chem. 2008, 16, 1218.
22. Taha, M. O.; Bustanji, Y.; Al-Bakri, A. G.; Yousef, M.; Zalloum, W. A.; Al-Masri, I.
M.; Atallah, N. J. Mol. Graphics Modell. 2007, 25, 870.
48. Veber, D. F.; Johnson, S. R.; Cheng, H. Y.; Smith, B. R.; Ward, K. W.; Kopple, K. D.
Bioorg. Med. Chem. 2002, 45, 2615.
49. Mai, A.; Cheng, D.; Bedford, M. T.; Valente, S.; Nebbioso, A.; Perrone, A.; Brosch,
G.; Sbardella, G.; Bellis, F. D.; Miceli, M.; Altucci, L. J. Med. Chem. 2008, 51, 2279.
50. Gerhard, B. V.; Rene, M. L. J. Chem. Soc. 1: Org. Bio-Org. Chem. 1978, 11, 1338.
51. Jain, A. N. J. Comput. Aided Mol. Des. 1996, 10, 427.
52. Venkatachalam, C. M.; Jiang, X.; Oldfield, T.; Waldman, M. J. Mol. Graphics
Modell. 2003, 21, 289.
53. Krammer, A.; Kirchhoff, P. D.; Jiang, X.; Venkatachalam, C. M.; Waldman, M. J.
Mol. Graphics Modell. 2005, 23, 395.
54. Gehlhaar, D. K.; Verkhivker, G. M.; Rejto, P. A.; Sherman, C. J.; Fogel, D. B.; Fogel,
L. J.; Freer, S. T. Chem. Biol. 1995, 2, 317.
55. Gehlhaar, D. K.; Bouzida, D.; Rejto, P. A. In Rational Drug Design: Novel
Methodology and Practical Applications; Parrill, L., Reddy, M. R., Eds.; American
Chemical Society: Washington, DC, 1999; pp 292–311.
56. Muegge, I.; Martin, Y. C. J. Med. Chem. 1999, 42, 791.
57. Muegge, I. Perspect. Drug Discovery Des. 2000, 20, 99.
58. Muegge, I. J. Comput. Chem. 2001, 22, 418.
23. Al-masri, I. M.; Mohammad, M. K.; Taha, M. O. Chem. Med. Chem 2008, 3, 1763.
24. Taha, M. O.; Dahabiyeh, L. A.; Bustanji, Y.; Zalloum, H.; Saleh, S. J. Med. Chem.
2008, 51, 6478.
25. Al-Nadaf, A.; Abu Sheikha, G.; Taha, M. O. Bioorg. Med. Chem. 2010, 18, 3088.
26. Abu-Hammad, A.; Zalloum, W. A.; Zalloum, H.; Abu-Sheikha, G.; Taha, M. O.
Eur. J. Med. Chem. 2009, 44, 2583.
59. Bemis, G. W.; Murcko, M. A. J. Med. Chem. 1996, 39, 2887.
60. Discovery Studio 2.5.5 User Guide, 2010.
61. Krovat, E. M.; Langer, T. J. Chem. Inf. Comput. Sci 2004, 44, 1123.
62. Kirchmair, J.; Markt, P.; Distinto, S.; Wolber, G.; Langer, T. J. Comput. Aided Mol.
Des. 2008, 22, 213.
63. Verdonk, M. L.; Marcel, L.; Berdini, V.; Hartshorn, M. J.; Mooij, W. T. M.; Murray,
C. W.; Taylor, R. D.; Watson, P. J. Chem. Inform. Comput. Sci. 2004, 44, 793.
64. Irwin, J. J.; Shoichet, B. K. J. Chem. Inf. Comput. Sci. 2005, 45, 177.
65. Triballeau, N.; Acher, F.; Brabet, I.; Pin, J.-P.; Bertrand, H.-O. J. Med. Chem. 2005,
48, 2534.
66. Gao, H.; Williams, C.; Labute, P.; Bajorath, J. J. Chem. Inf. Comput. Sci. 1999, 39, 164.
67. Jacobsson, M.; Liden, P.; Stjernschantz, E.; Bostroem, H.; Norinder, U. J. Med.
Chem. 2003, 46, 5781.
27. Abu Khalaf, R.; Abu Sheikha, G.; Bustanji, Y.; Taha, M. O. Eur. J. Med. Chem. 2010,
45, 1598.
28. Al-Sha’er, M. A.; Taha, M. O. Eur. J. Med. Chem. 2010, 45, 4316.
29. Abdula, A.; Abu Khalaf, R.; Mubarak, M.; Taha, M. J. Comput. Chem. 2010, 26, 463.
30. Al-Sha’er, M. A.; Taha, M. O. J. Chem. Info. Model. 2010, 50, 1706.
31. Singh, J.; Chuaqui, C.; Sjodin, P.; Lee, W.; Pontz, T.; Corbley, M.; Cheung, H.;
Arduini, R.; Mead, J.; Newman, M.; Papadatos, J.; Bowes, S.; Josiah, S.; Ling, L.
Bioorg. Med. Chem. Lett. 2003, 13, 4355.
32. Taha, M. O.; Qandil, A. M.; Zaki, D. D.; AlDamen, M. A. Eur. J. Med. Chem. 2005,
40, 701.
33. Garino, C.; Tomita, T.; Pietrancosta, N.; Laras, Y.; Rosas, R.; Herbette, G.;
Maigret, B.; Quéléver, G.; Iwatsubo, T.; Kraus, J. J. Med. Chem. 2006, 49,
4275.
68. Loughlin, A. W.; Pierens, K. G.; Petersson, J. M.; Hendersona, C. L.; Healya, C. P.
Bioorg. Med. Chem. 2008, 16, 6172.
69. Birch, A. M.; Kenny, P. W.; Oikonomakos, N. G.; Otterbein, L.; Schofield, P.; Paul,
R. O.; Whittamore; Whalley, D. P. Bioorg. Med. Chem. Lett. 2007, 17, 394.
70. Georgiyants, V. A. Klinichna Farmatsiya 2003, 7, 63.
71. Georiyants, V. A.; Bezuglyi, P. A. Russ. J. Org. Chem. 2002, 38, 1079.
72. Shahnawaaz, M.; Naqvi, A.; Rao, A. V.; Seth, D. S. Orient. J. Chem. 2008, 24, 701.
73. Moradi, R.; Enayat, O. Chin. Chem. Lett. 2010, 21, 542.
74. Pelly, S. C.; Parkinson, C. J.; Van Otterlo, W. A. L.; De Koning, C. B. J. Org. Chem.
2005, 70, 10474.
34. Rosauer, K. G.; Ogawa, A. K.; Willoughby, C. A.; Ellsworth, K. P.; Geissler, W. M.;
Myers, W. R.; Chapman, K. T.; Harris, G.; Moller, D. E. Bioorg. Med. Chem. 2003,
13, 4385.
35. CATALYST 4.11 Users’ Manual; Ligandfit, 2005.
36. Sheridan, R. P.; Kearsley, S. K. Drug Discovery Today 2002, 7, 903.
37. CATALYST 4.11 Users’ Manual; Accelrys Software Inc.: San Diego, CA, 2005.
38. Kurogi, Y.; Guner, O. F. Curr. Med. Chem. 2001, 8, 1035.
39. Li, H.; Sutter, J.; Hoffmann, R. In Pharmacophore Perception, Development, and
Use in Drug Design; Güner, O. F., Ed.; International University Line: La Jolla, CA,
2000; p 173.
75. Kubicova, L.; Waisser, K.; Kunes, J.; Kralova, K.; Odlernova, Z.; Slosarek, M.;
Janota, J.; Svoboda, Z. Molecules 2000, 5, 714.
76. Biland, Hans R.; Duennenberger, M.; Luethi, C. Oxalic Acid Diarylamides. U.S.,
1975, p 31.
77. CIBA-GEIGY, 1968.
78. Goldfarb, D.S. U.S. Pat. Appl. Publ., 2009, 57.