QSAR, in silico docking and in vitro evaluation of chalcone derivatives as potential inhibitors for H1N1 virus neuraminidase

Article abstract of DOI:10.1007/s00044-016-1636-5

Thirty three chalcones were synthesized and tested on viral H1N1 neuraminidase activity by using MUNANA assay [2′-(4-methylumbelliferyl)-α-d-N-acetylneuraminic acid] assay with DANA (2,3-didehydro-2-deoxy-N-acetylneuraminic acid) was used as standard. 2D and 3D-quantitative structure?activity relationship models have been successfully developed with a good correlative and predictive ability for quantitative structure?activity relationships of these chalcone derivatives. Result from the 2D-quantitative structure?activity relationship model indicates that electrostatic parameter enhanced bioactivity of the chalcones while steric substituents diminished their potency as H1N1 neuraminidase inhibitors. 3D-quantitative structure?activity relationship model showed the importance of the position of the hydroxyl group in chalcone derivatives which can influence on hydrophobicity, hydrogen bond donor and aromatic ring features that enhance the biological activity. Finally, docking studies showed that chalcones MC8 and MC16 with low C docker interaction energies and higher numbers of hydrogen bonding have better inhibitory activity against viral H1N1 neuraminidase.

Full text of DOI:10.1007/s00044-016-1636-5

MEDICINAL
CHEMISTRY
RESEARCH
Med Chem Res
DOI 10.1007/s00044-016-1636-5
ORIGINAL RESEARCH
QSAR, in silico docking and in vitro evaluation of chalcone derivatives as
potential inhibitors for H1N1 virus neuraminidase
Marzieh Yaeghoobi1,2 Neni Frimayanti1,2 Chin Fei Chee1,2 Kusaira K. Ikram3
●
●
●
●
Belal O. Najjar3 Sharifuddin M. Zain1 Zanariah Abdullah1 Habibah A. Wahab3
Noorsaadah Abd. Rahman1,2
●
●
●
●
Received: 8 April 2015 / Accepted: 27 June 2016
© Springer Science+Business Media New York 2016
●
●
●
Abstract Thirty three chalcones were synthesized and
tested on viral H1N1 neuraminidase activity by using
MUNANA assay [2′-(4-methylumbelliferyl)-α-D-N-acet-
ylneuraminic acid] assay with DANA (2,3-didehydro-2-
deoxy-N-acetylneuraminic acid) was used as standard. 2D
and 3D-quantitative structure−activity relationship models
have been successfully developed with a good correlative
and predictive ability for quantitative structure−activity
relationships of these chalcone derivatives. Result from the
2D-quantitative structure−activity relationship model indi-
cates that electrostatic parameter enhanced bioactivity of the
chalcones while steric substituents diminished their potency
as H1N1 neuraminidase inhibitors. 3D-quantitative struc-
ture−activity relationship model showed the importance of
the position of the hydroxyl group in chalcone derivatives
which can influence on hydrophobicity, hydrogen bond
donor and aromatic ring features that enhance the biological
activity. Finally, docking studies showed that chalcones
MC8 and MC16 with low C docker interaction energies
and higher numbers of hydrogen bonding have better
inhibitory activity against viral H1N1 neuraminidase.
Keywords Neuraminidase Chalcone Pharmacophore
2D-QSAR 3D-QSAR
●
Introduction
Influenza virus belongs to the orthomyxoviridae, a family of
RNA viruses which can cause infection in birds and
mammalian cells with the help of hemagglutinin (H) and
neuraminidase (NA) of the influenza virus. Neuraminidase
is involved in the initiation of the influenza infection by
promoting the release of the virus from the host cell. This
enzyme is a crucial part of influenza replication. Thus,
finding suitable inhibitors to block the function of NA is
possibly an effective way to restrain influenza (Babu et al.,
2000).
In the last decade, a number of investigations have been
conducted on phytomedicines for their potential as neur-
aminidase inhibitor. Flavonoids, especially, are considered
promising compounds and are widely found in traditional
herb based medicines for influenza. For example, a
series of flavonoids isolated from Sophoraflavescens,
Glycyrrhizauralensis, cudraniatricuspidata and Rhodio-
larosea were found to show moderate activities for NA
inhibition (Jeong et al., 2009; Ryu et al., 2008, 2009, 2010).
However, most work have been conducted on naturally
occurring flavonoids, and amongst all the subtypes of fla-
vonoids studied, chalcones have not been extensively
investigated for their potency as NA inhibitors (Ryu et al.,
2008, 2009; Gao et al., 2011; Mercader and Pomilio, 2010).
Thus far, there is no report on quantitative structure–activity
relationship for the chalcones as NA inhibitors. This study
was conducted to investigate quantitative structure–activity
Electronic supplementary material The online version of this article
(doi:10.1007/s00044-016-1636-5) contains supplementary material,
which is available to authorized users.
* Neni Frimayanti
nenifrimayanti@yahoo.com
1
Department of Chemistry, Faculty of Science, University of
Malaya, Lembah Pantai, 50603 Kuala Lumpur, Malaysia
2
Drug Design and Development Research Group, University of
Malaya, 50603 Kuala Lumpur, Malaysia
3
Pharmaceutical Design and Simulation Laboratory, School of
Pharmaceutical Sciences, Universiti Sains Malaysia, 11800
Penang, Malaysia

Med Chem Res
relationships for the chalcone compounds with the aid of 2D
and 3D-quantitative structure−activity relationship (QSAR)
models.
were obtained from this calculation. These descriptors were
then reduced to a smaller set of descriptors. These set should
be information rich but as small as possible.
To ascertain the influence of different functional groups
on NA inhibition, 33 chalcones with different functional-
ities were synthesized through Claisen–Schmidt condensa-
tion reaction from the corresponding acetophenones and
benzaldehydes (Marais et al., 2005). All synthetic com-
pounds were purified by chromatographic method and
characterized by using various spectroscopic techniques.
The H1N1 neuraminidase inhibitory potency of the syn-
thetic chalcones was tested by using in vitro MUNANA
assay with DANA (2, 3-didehydro-2-deoxy-N-acetylneur-
aminic acid) as standard inhibitor (Potier et al., 1979).
Correlation matrix was applied to select the best subset
of descriptors to be included in the model by eliminating
descriptors that are highly correlated with each other.
The next step involved the scaling of descriptors which
requires thorough manipulation since there may be under-
lying relationship between these descriptors and it may not
be possible to foresee the effects of this process.
The selected descriptors were used to build the QSAR
model. In this study, QSAR models were developed using
multiple linear regression analysis (MLRA) technique. In
this technique, the values for F-to-enter and F-to-leave were
set to 4. Cross-validation analysis was performed using the
leave-one-out (LOO) method where one compound is
removed from the dataset and its activity is calculated using
the model derived from the rest of the dataset. The cross-
validated r2(CV) and conventional r2 that resulted in the
lowest error of prediction were chosen (Wermuth, 1998).
Unless otherwise stated, the default values for the other
QSAR parameters were used.
Material and methods
General procedure for the synthesis of chalcones
All chalcones were prepared via standard Claisen–Schmidt
reaction. To a solution of the corresponding acetophenone
(1 eq) in ethanol (2.5 mL/mmol), sodium hydroxide (3 eq)
was added. After 10 min, benzaldehyde (1.2 eq) was added
and the solution was stirred at room temperature overnight.
After cooling with ice, the reaction mixture was neutralized
carefully using 1N hydrochloric acid. The crude mixture
was extracted with ethyl acetate, washed with water and
brine afforded chalcones (yield 32–92 %) after column
chromatography.
3D-QSAR model and pharmacophore generation
The 3D-QSAR model was carried out using MOE software
packed (Chemical Computing Group Inc). This model was
developed using 33 compounds in the data set. QSAR
model was developed using partial least squares (PLS)
technique. The structure of each molecule was sketched
using the molecule builder tool in the MOE software and
then minimized using MMFF94× force field to a gradient
0.00001 kcal/mol/Å.
Data set preparation
A data set of 33 chalcone-based compounds (Table 1) and
percentage of NA inhibition for chalcones studied at
100 mg/mL in Fig. 1 expressed as % inhibition in the range
of 6.80–54.0 %. Among these 33 compounds, 22 com-
pounds were used in the training set for model development
and 11 compounds were used as a test set for model
validation.
Fifteen pharmacophore descriptors were generated and at
the same time pharmacophore of the ligands was performed.
Feature of the pharmacophore was carried out using phar-
macophore query editor tool and the hypothesis for phar-
macophore alignment. In this study, the best alignments of
pharmacophore were generated using three features,
hydrogen bond donor, hydrogen bond acceptor and hydro-
phobic atom.
2D QSAR modeling
The features of this pharmacophore were then used to
ensure that 3D-QSAR model is applicable to molecules
with the same properties as the pharmacophore alignment,
and can be used to predict the biological activities of the
unknown compounds.
The 2D molecular structures of the ligands were sketched
using the Chem Draw 6.0 software (Cambridge Scientific
Computing) while Corina in TSAR 3.3 (Accelrys) software
packages used to convert the structures into their 3D con-
formation. The geometries of these molecules were opti-
mized using the Cosmic module of TSAR. The calculation
was terminated when the energy difference or the energy gra-
dient becomes smaller than 1 × 10−5 and 1 × 10−10 kcal/mol,
respectively. Molecular descriptors were generated using
TSAR 3.3 (Accelrys) for each compound. 316 descriptors
H1N1 viral neuraminidase MUNANA assay
All chalcones were tested on H1N1 viral neuraminidase by
using MUNANA [2'-(4-methylumbelliferyl)-α-D-N-acet-
ylneuraminic acid] assay with DANA was used as standard
inhibitor. Fresh stock solution (1 mg/mL) of the sample was

Med Chem Res
Table 1 Structure and activity data of chalcones
R2
A
O
B
R3
R1
Entry Compd no
R1
R2
R3
% of
Entry Compd no R1
R2
R3
% of
inhibitiona
inhibitiona
1
2
MC1
MC2
H
H
6.9
18
19
MC20
MC22
OCH3 OCH3 12.90
H
H
70.1
OCH3 OCH3 21.90
3
4
MC3
MC4
H
H
H
H
39.60
92.10
20
21
MC23
MC24
OCH3 OCH3 11.50
OCH3 OCH3 15.40
OCH3 OCH3 19.10
5
6
MC5
MC6
OH
OH
H
H
12.90
60.88
22
23
MC25
MC26
OCH3
H
20.90
7
8
MC7
MC8
OH
OH
H
H
20.63
96.50
24
25
MC33
MC38
OCH3
NH2
H
H
19.50
34.90
9
MC9
OH
OH
OH
H
H
H
22.80
20.20
12.90
26
27
28
MC39
MC40
MC41
NH2
NH2
NH2
H
H
H
63.00
26.40
29.40
10
11
MC10
MC11
12
13
MC12
MC13
OH
OH
H
H
19.50
17.00
29
30
MC42
MC43
NH2
H
H
6.80
NH2 74.40
NH2 41.20
14
MC14
OH
H
15.20
31
MC44
H

Med Chem Res
Table 1 continued
Entry Compd no
R1
R2
R3
% of
Entry Compd no R1
R2
R3
% of
inhibitiona
inhibitiona
15
16
MC15
MC16
OH
H
16.40
32
33
MC45
MC46
H
NH2 36.60
NH2 54.00
OH
OCH3 93.40
H
17
MC19
OCH3 OCH3 30.50
DANA
98.1
a
Percentage of inhibition at 1 mg/mL
H1N1 NA % INHIBITION
Fig. 1 Percentage of NA inhi-
bition for chalcones studied at 1
mg/mL
120
100
80
60
40
20
0
COMPOUNDS
prepared in 2.5 % DMSO. Briefly, 25 µL H1N1 viral
neuraminidase (SINOBIO) was added to 25 µL of sample
solution mixed with buffer in a 96-microplate well. A
volumn of 50 µL of substrate (MUNANA (SIGMA,
M8639) in 32.5 mM MES (SIGMA, M8250) buffer (pH
6.5) was then added and the mixture was incubated at 37 °
C. After one hour, formation of 4-methylumbelliferone was
immediately quantified fluorometrically on a Modulus
Microplate Reader (Turner Biosystem, USA). The excita-
tion wavelength was set at 365 nm and the emission
wavelength at 450 nm. Percentage of inhibition was
obtained by fitting experimental data to the logistic graph.
Docking of chalcones to H1N1 viral neuraminidase
The docking of all 33 chalcones onto the neuraminidase of
A/Breving Mission/1/1918 H1N1 strain in complex with
zanamivir which downloaded from PDB data bank (www
.pdb.org, PDB ID: 3B7E) was achieved using Discovery
studio 2.5 software packages (Accelrys). Hydrogen atoms
were added to the protein and its backbone was minimized.
All ligands were minimized.
Docking was performed by using the Cdocker protocol; a
grid based molecular docking method that employs
CHARMM forcefields. The protein was firstly held rigid

Med Chem Res
and the ligands were allowed to flex during the refinement.
Two hundred ligand conformations were then generated
from the initial ligand structure though high temperature
molecular dynamic followed by random rotation, refine-
ment with grid based (GRID 1) simulated annealing and a
final grid based or full force field minimization. Upon
completion of the docking proses, conformations with the
lowest cdocker energy were then chosen and compared with
the standard DANA (active agents against neuramininidase
H1N1 with Cdocker energy equal to −46.11 kcal/mol).
Generally, QSAR model is accepted if it has an r2greater
than 0.6 and r2(CV) greater than 0.5 (Medina-Franco et al.,
2005; Frimayanti et al., 2013). This model, therefore,
exhibited high prediction value as can be seen from r2 and
r2(CV) value.
The cross-validated coefficient r2(CV) defines the
goodness of prediction whereas the non-cross-validated
correlation coefficient (r2) indicates the goodness of fit of a
QSAR model. The F test value stands for the degree of
statistical confidence. The statistical output of the MLRA
model is presented in Table 2. A cross-validated coefficient
of 0.65 was obtained using the ‘leave-one-out’ cross vali-
dation procedure. This value indicates very good internal
predictive capability of the developed model.
The model also exhibited a non-cross-validated correla-
tion coefficient of 0.81. The high value of this parameter
adds to its usefulness as a predictive tool. The statistical
significance of this parameter is listed in Table 3.
Results and discussion
2D QSAR modeling
The best QSAR model developed for the H1N1 inhibition
using MLRA technique has an r2 value of 0.81 and an
r2 (CV) value of 0.65. The equation is:
A plot of experimental vs. predicted of log 1/% inhibition
is shown in Fig. 2. There is no outlier observed in the plot.
Hence, the developed QSAR model could be considered to
be stable. The calculated inhibition of compounds in the test
set re shown in Table 4.
Log 1=%inhibition ¼ ꢀ 1:06 ꢁ kier chiV5 ðpathÞ
þ 0:98 ꢁ VAMP HOMO þ 0:79
ꢁ VAMP octupole ZZZ þ 2:10
ð1Þ
Based on the QSAR model described above, it could be
inferred that percentage of inhibition will improve with the
increase of the electrostatic parameter (i.e., VAMP octupole
ZZZ and VAMP HOMO). The electrostatic parameters are
properties of a molecule which are related to its electron
affinity and demonstrate its susceptibility towards attack by
nucleophiles (Sharma and Kohli, 2014). In this study,
VAMP octupole ZZZ and VAMP HOMO correlates well
with the percentage of inhibition. As an example, the pre-
sence of dimethylamine at R1; hydroxyl group at R2 and a
methoxy group at R3 of compound MC16 (93.4 % NA
inhibitory activity) increased the electrostatic parameter
which promotes the percentage of inhibition.
Table 2 Statistical output and F test value of the multiple linear
regression analysis (MLRA) model
Statistical output
Value
Non-cross validated r2
Cross validation r2 (CV)
F value
0.81
0.65
27.54
1.23e−007
0.18
F-probability
Standard error of estimate (SEE)
Residual sum of square (RSS)
Predictive sum of square (PRESS)
Topological descriptor (kier ChiV5) calculates the Kier
and Hall kappa molecular shape indices which are important
descriptors used to define the steric influence of substituents
in the interactions of organic compounds with
0.62
1.16
Table 3 Statistical significance
Descriptors
Regression coefficienta Jackknife SEb
Covariance SEc tvalued tProbabilitye
of parameter in QSAR equation
Kier ChiV5 (path)
VAMP HOMO
−1.06
0.20
0.049
0.40
0.12
0.17
0.20
−8.59 5.73e−008
5.53 2.43e−005
−3.84 0.001
0.98
VAMP octupole ZZZ
0.79−
a
The regression coefficient for each variable in the equation
b
Standard error of each regression coefficient derived from a Jackknife procedure on the final regression
model
c
Estimate of the standard error of each regression coefficient derived from the covariance matrix
d
e
Significance of each variable included in the final model
Statistical significance for t values

Med Chem Res
component analysis (PCA) was applied to reduce the
dimensionality of set of molecular descriptors by linearly
transforming the data. 3D plot of PCA1, PCA2 and PCA3 is
shown in Fig. 3.
3D-QSAR model was then generated using the partial
least squares (PLS) technique where the percentage of
inhibition served as the dependent variable and the phar-
macophore descriptors were used as the independent vari-
able. The resulting model gave a root mean squares error
(RMSE) and coefficient correlation (r2) of 0.204 and 0.71,
respectively.
Fig. 2 Plot of actual value vs. predicted value of log 1/% inhibition of
A plot of the actual value of the percentage of inhibition
vs. the predicted value of the percentage of inhibition is
shown in Fig. 4. The distribution of compounds is observed
to be in surrounding region of the linear regression, except
for three chalcones (MC46, MC15 and MC5) which were
seen as outliers (circled in Fig. 4).
compounds in the training set
Table 4 Calculated value of Log 1/% inhibition of chalcone based
compounds in the test set based on 2D QSAR model
Entry Compound No. Experimental log 1/ % Predicted log 1/%
inhibition
inhibition
The 3D-QSAR model was then validated by predicting
the percentage of inhibition of 11 compounds in the test set.
The predicted value and actual value of compounds in the
test is shown in Table 5. Overall, this 3D-QSAR model is
able to predict the biological activity of compounds in the
test set with r2 prediction of 0.51.
The suggested descriptors (in the QSAR model) were
then validated using pharmacophore alignment. The train-
ing set was aligned on the template using the fit and sta-
tistically significant pharmacophore hypothesis for chalcone
MC8 (96.5 % NA inhibition) (Fig. 5). The pink sphere in
Fig. 5 is featured for hydrogen bond donor; green sphere for
hydrophobic and yellow sphere represented the aromatic
ring feature.
1
MC45
MC38
MC19
MC41
MC40
MC9
1.56
1.54
1.48
1.46
1.42
1.36
1.34
1.32
1.3
1.54
1.55
1.45
1.38
1.41
1.32
1.29
1.36
1.30
1.28
1.29
2
3
4
5
6
7
MC22
MC26
MC10
MC12
MC33
8
9
10
11
1.29
1.29
A pharmacophore is the ensemble of steric and electronic
features that is necessary to ensure the optimal supra-
molecular interactions with a specific biological target and
to trigger (or block) its biological response (Wermuth,
1998). In addition, the pharmocophore also shows the
importance of the hydrophobic, hydrogen bond donor and
aromatic ring features in enhancing biological activity. As it
shown in Fig. 5, hydrogen bond donor (pink) and hydro-
phobic sphere (green) were observed in the vicinity of
ring A. Thus, it is presumed that compound such as MC8,
the highest percentage of NA inhibition may be due to the
presence of hydroxyl groups in the R1 and R2 positions.
From the result above, it is reasonable to conclude that
NA inhibitory action of chalcones can be regulated by
multiple mechanisms. To find the possible structure
requirement for NA inhibitory activity, meaningful 2D and
3D-QSAR models were derived. 2D-QSAR model
indicated that electrostatic parameter may enhance the
bioactivity of the chalcone such as MC16 while steric
influence of substituent and probably lack of flexibility such
as in chalcone MC42 may diminish its potency as NA
inhibitor.
macromolecular drug receptor (Praphat et al., 2011).
Increasing the value of kier ChiV descriptor will decrease
the percentage of inhibition. For example, the bulky naph-
talene ring in R1 of compound MC42 increased the kier
ChiV5 values but showed a diminished NA inhibitory
activity (6.8 %).
The percentage of inhibition of the compounds in the test
set (i.e., 11 chalcone-based compounds) was predicted
using this QSAR model (Eq. 1). The calculated percentage
of inhibition values of the compounds in the test set are
listed in Tables 4.
The correlation coefficient (r2) between predicted and
experimental value for the QSAR model was also calcu-
lated. A predictive correlation coefficient r2 value (test set)
of 0.94 was obtained for the developed QSAR model.
3D QSAR modeling
Pharmacophore features in this study generated 15 types of
descriptors. The feature selection was used to select the best
descriptors that could be used in the QSAR model. Principle

Med Chem Res
Fig. 3 3D plot of principle
component analysis (PCA)
Fig. 4 Plot of predicted log 1/%
inhibition vs. actual log 1/%
inhibition of compounds in the
training set
3D-QSAR model also showed the importance of the
position of the hydroxyl group in A ring which can influ-
ence the hydrophobic, hydrogen bond donor and aromatic
ring features of the compounds and presumably enhance
their biological activities.
interaction of chalcones with NA protein in order to
ascertain potential directions for synthetic lead-optimization
studies.
Docking results
Amongst all chalcones, MC8 showed the best NA
inhibitory activity with the percentage of inhibition almost
comparable with DANA. The present investigation provides
useful information on the structure requirements for the
In silico docking studies were performed to evaluate the
effects of chalcones against neuraminidase. The Cdocker
energy reflects the interaction energy for the ligand-protein

Med Chem Res
complex. The lower the Cdocker energy means the inter-
action is more stable. The Cdocker energy is presented in
Table 6.
The calculated Cdocker energy for MC8 and MC16
showed relatively low value which seems to corroborate
with the experimental results. In addition, MC5 and MC15
which showed relatively higher Cdocker interaction energy
did not show good H1N1 inhibition activities.
MC8 showed to have four hydrogen bonds (blue dashed
line) with residue Arg118, Glu227, Arg292 and Arg371. In
addition, a π interaction was also found between ligand and
residue Arg152 and Arg371. Residues Asp151, Arg152,
Glu119, Glu277 and Asn294 showed interaction with the
ligand through van der Waals interaction (green cycle)
suggesting the importance of these five residues in the
formation of van der waals binding pocket. Likewise, in the
case of MC16, three hydrogen bonds were observed
between the ligand and residues Arg118D, Arg292 and
Arg374. This ligand displayed π interaction with residues
Arg152 and Arg374. The higher the number of the
hydrogen bond, may accounts for ligand is more active
(Mahto et al., 2012).
However, MC5 with only one hydrogen bond (between
hydroxyl group of the ligand and residue Arg152) and a π
interaction with the residue Arg292 showed relatively high
Cdocker energy and less inhibition. The binding interaction
for MC5, MC8 and MC16 are illustrated in Fig. 6
Combined inhibitory activity and docking studies sug-
gested that in the presence of hydroxyl groups in the R1 and
R2 positions in MC8 and MC 16 may fill better into the
adjunct pockets resulting more number of hydrogen bond-
Table 6 The Cdocker interaction energies of chalcone derivatives
with neuraminidase
Entry Compound (-) Cdocker Entry Compound (-) Cdocker
interaction
energy(kcal/
mol)a
interaction
energy(kcal/
mol)a
1
MC1
28.25
31.80
35.69
29.74
28.08
36.94
33.04
36.94
35.35
38.65
43.25
35.41
32.54
35.37
29.25
43.05
38.39
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
MC20
MC22
MC23
MC24
MC25
MC26
MC33
MC38
MC39
MC40
MC41
MC42
MC43
MC44
MC45
MC46
DANA
37.21
38.58
39.28
38.12
35.31
35.31
33.82
31.74
31.60
35.85
35.78
31.19
30.32
27.82
33.26
35.29
46.11
2
MC2
3
MC3
4
MC4
Table 5 The actual value and predicted value of log 1/% inhibition of
compounds in the test set based on 3D QSAR model
5
MC5
6
MC6
Entry Compound Actual value of log 1/ Predicted value of log
7
MC7
% inhibition
1/% inhibition
8
MC8
1
MC45
MC38
MC19
MC41
MC40
MC9
1.56
1.54
1.48
1.46
1.42
1.36
1.34
1.32
1.3
1.61
1.78
1.06
1.68
1.61
1.52
1.27
1.08
1.41
1.54
0.58
9
MC9
2
10
11
12
13
14
15
16
17
a
MC10
MC11
MC12
MC13
MC14
MC15
MC16
MC19
3
4
5
6
7
MC22
MC26
MC10
MC12
MC33
8
9
10
11
1.29
1.29
Calculation performed using Cdocker (Discovery Studio 2.5,
Accelrys)
Fig. 5 The best pharmacophore
hypothesis for compound MC8,
distance between
pharmacophore features is
reported in Angstrom.
Pharmacophores are color coded
with yellow for aromatic ring,
pink for hydrogen bond donor
and green for hydrophobic

Med Chem Res
Fig. 6 The binding interaction
of MC5, MC8 and MC16
MC5 binding site
MC8 binding site
MC16 binding site

Med Chem Res
discovery of a novel, highly potent, orally active, and selective
influenza neuraminidaseinhibitor through structure-based drug-
design. J Med Chem 43:3482–3486
ing and π interactions with the relative residue as Arg118,
Arg227, Arg292, Arg371, Arg152, Arg371, Asp151,
Arg152, Glu119, Glu277, Asn294 and Arg118D, Arg292,
Arg374, Arg152, Arg374, respectively, which can pre-
sumably enhance the biological activity of these chalcones.
Frimayanti N, Lee VS, Zain SM, Wahab HA, Rahman NS (2013)
QSAR and pharmacophore studies on thiazolidine-4-carboxilic
acid derivatives as neuraminidase inhibitors in H3N2 influenza
virus. Med Chem Res 23(3):1447–1453
Gao L, Zu M, Wu S, Liu AL, Du GH (2011) 3D QSAR and
docking study of flavone derivatives as potent inhibitors of
influenza H1N1 virus neuraminidase. Bioorg Med Chem Lett 21:
5964–70
Conclusion
Jeong HJ, Ryu YB, Park SJ, Kim JH, Kwon HJ, Park KH, Rho MC,
Lee WS (2009) Neuraminidase inhibitory activities of
flavonols isolated from Rhodiolarosea roots and their in
vitro anti-influenza viral activities. Bioorg Med Chem 17:
6816–6823
Mahto MK, Raj VK, Bhaskar PM, Divya R (2012) ADMET and
molecular docking studies of novel zanamivir analogs as neur-
aminidase inhibitors. Int J Pharm Sci Rev Res 13(1):91–94
Medina-Franco JL, Golbraikh A, Oloff S, Castillo R (2005) Quanti-
tative structure activity relationship analysis of pyridinone HIV-1
reverse transcriptase inhibitors using the k nearest neighbor
method and QSAR-based database mining. J Comput Aided Mol
Des 19:229–242
The 2D and 3D-QSAR model has been successfully
developed with a good correlative and predictive ability for
predicting neuraminidase inhibitory activity of this chalcone
series. Based on 2D-QSAR model, the electrostatic prop-
erties can enhance the inhibitory activity. 3D-QSAR model
showed that the increasing of the biological activity will be
influenced by the hydrogen bond donor, hydrophobic
properties and aromatic ring features of the compound.
Docking studies were performed to evaluate the effects of
chalcones against neuraminidase.
Docking study showed the binding affinity of chalcone
derivatives to be within the enzyme binding pockets with
relatively less Cdocker interaction energies and higher
numbers of hydrogen bonding which validated them as
potential candidates for second generation drug discovery.
Bioassay studies of all chalcones indicated that the com-
pounds such as MC8 and MC16 bearing hydroxyl groups in
the R1 and R2 positions have better biological activities
against neuraminidase, suggesting that they are promising as
potential inhibitors for H1N1 virus Neuraminidase.
Marais JP, Ferreira D, Slade D (2005) Stereoselective synthesis of
monomeric flavonoids. Phytochem 66:2095–2176
Mercader AG, Pomilio AB (2010) QSAR study of flavonoids and
biflavonoids as influenza H1N1 virus neuraminidase inhibitors.
Eur J Med 45:1724–1730
Potier M, Mameli L, Belisle M, Dallaire L, Melancon SB (1979)
Fluorometric assay of neuraminidase with a sodium (4-methy-
lumbelliferyl-α-Image -N-acetylneuraminate) substrate. Anal
Biochem 94:287–296
Praphat K, Preety S, Singh JS (2011) Topological based QSAR study
of benzamine as inhibitor of trombin. J Chem Pharm Res
3(4):396–303
Ryu YB, Curtis-Long MJ, Kim JH, Jeong SH, Yang MS, Lee KW, Le
WS, Park KH (2008) Pterocarpans and flavanones from
Sophoraflavescens displaying potent neuraminidase inhibition.
Bioorg Med Chem Lett 18:6046–6049
Ryu YB, Curtis-Long MJ, Lee JW, Ryu HW, Kim JY, Lee WS, Park
KH (2009) Structural characteristics of flavanones and flavones
from Cudraniatricuspidata for neuraminidase inhibition. Bioorg
Med Chem Lett 19:4912–4915
Acknowledgments We thank University of Malaya for the financial
support through the project Grant Scheme 53-02-03-1059 and
Malaysian Ministry of Science, Technology and Innovation for the
H1N1 grant 09-05-IFN-MEB 004.
Compliance with ethical standards
Ryu YB, Kim JH, Park SJ, Chang JS, Rho MC, Bae KH et al. (2010)
Inhibition of neuraminidase activity by polyphenol compounds
isolated from the roots of Glycyrrhiza uralensis. Bioorg Med
Chem Lett 20:971–974
Conflict of interest The authors declare that they have no conflict of
interest.
Sharma MC, Kohli D (2014) Insight into the structural requirement of
substituted quinazolinone biphenyl acylsulfonamides derivatives
as Angiotensin II AT1 receptor antagonist: 2D and 3D QSAR
approach. J Saudi Chem Soc 18(1):35–45
Wermuth CG (1998) Glossary of term used in medicinal chemistry.
Pure Appl Chem 70:1129–1143
Reference
Babu YS, Chand P, Bantia S, Kotian P, Dehghani A, El- Kattan Y, Lin
TS, Hutchison TL, Elliott AJ, Parker CD, Ananth SL, Horn LL,
Laver GW, Montgomery JA (2000) BCX-1812 (RWJ-270201):