Design and synthesis of a novel mPGES-1 lead inhibitor guided by 3D-QSAR CoMFA

Article abstract of DOI:10.1016/j.molstruc.2019.07.004

The search of novel mPGES-1 inhibitors has recently intensified probably due to the superior safety in comparison to existing anti-inflammatory drugs. Although two mPGES-1 inhibitors have entered clinical trials, none has yet reached the market. In this study, we performed modifications guided by 3D-QSAR CoMFA on 2, which is an unsymmetrical curcumin derivative with low binding affinity towards mPGES-1. To counter the PAINS properties predicted for 2, the diketone linker was replaced with a pyrazole ring. On the other hand, both prenyl and carboxylate ester groups were introduced to improve the activity. When tested in vitro, 11 suppressed PGE₂ biosynthesis in activated macrophages and showed promising human mPGES-1 inhibition in microsomes of interleukin-1β-stimulated A549 cells. Altogether, 11 has been identified as a potential mPGES-1 inhibitor and could be a promising lead for a novel class of mPGES-1 inhibitors.

Full text of DOI:10.1016/j.molstruc.2019.07.004

Accepted Manuscript
Design and synthesis of a novel mPGES-1 lead inhibitor guided by 3D-QSAR CoMFA
Mohd Fadhlizil Fasihi Mohd Aluwi, Kamal Rullah, Andreas Koeberle, Oliver Werz, Nur
Sakinah Abdul Razak, Leong Sze Wei, Fatimah Salim, Nor Hadiani Ismail, Ibrahim
Jantan, Lam Kok Wai
PII:
S0022-2860(19)30838-5
DOI:
https://doi.org/10.1016/j.molstruc.2019.07.004
MOLSTR 26757
Reference:
To appear in: Journal of Molecular Structure
Received Date: 26 February 2018
Revised Date: 30 April 2019
Accepted Date: 1 July 2019
Please cite this article as: M.F. Fasihi Mohd Aluwi, K. Rullah, A. Koeberle, O. Werz, N.S. Abdul Razak,
L.S. Wei, F. Salim, N.H. Ismail, I. Jantan, L.K. Wai, Design and synthesis of a novel mPGES-1 lead
inhibitor guided by 3D-QSAR CoMFA, Journal of Molecular Structure (2019), doi: https://doi.org/10.1016/
j.molstruc.2019.07.004.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Graphical Abstract
To create your abstract, type over the instructions in the template box below.
Fonts or abstract dimensions should not be changed or altered.
Design and synthesis of a novel mPGES-1 lead
inhibitor guided by 3D-QSAR CoMFA
Leave this area blank for abstract info.
a,*
b
c
c
Mohd Fadhlizil Fasihi Mohd Aluwi , Kamal Rullah , Andreas Koeberle , Oliver Werz , Nur
d
e
f
f
g
Sakinah Abdul Razak , Leong Sze Wei , Fatimah Salim , Nor Hadiani Ismail , Ibrahim Jantan , Lam
d,∗
Kok Wai
OMe
Ester
group
O
O
O
N
N
MeO
Bulky
group
O
compound 2
DESIGNED HIT
low binding affintiy towards mPGES-1
compound 11
mPGES-1 inhibition: 37% (1 µM)
non-PAINS

ACCEPTED MANUSCRIPT
Journal of Molecular Structure
journal homepage: www.elsevier.com
Design and synthesis of a novel mPGES-1 lead inhibitor guided by 3D-QSAR
CoMFA
a,*
b
c
c
Mohd Fadhlizil Fasihi Mohd Aluwi , Kamal Rullah , Andreas Koeberle , Oliver Werz , Nur Sakinah
d
e
f
f
g
d,∗
Abdul Razak , Leong Sze Wei , Fatimah Salim , Nor Hadiani Ismail , Ibrahim Jantan , Lam Kok Wai
a
Faculty of Industrial Science & Technology, Universiti Malaysia Pahang, Lebuhraya Tun Razak, 26300, Gambang, Pahang, Malaysia
Department of Pharmaceutical Chemistry, Kuliyyah of Pharmacy, International Islamic University Malaysia, Bandar Indera Mahkota, 25200, Kuantan,
b
Pahang, Malaysia
c
Department of Pharmaceutical/Medicinal Chemistry, Institute of Pharmacy, Friedrich-Schiller-University Jena, 07743 Jena, Germany
Drug and Herbal Research Centre, Faculty of Pharmacy, Universiti Kebangsaan Malaysia, Jalan Raja Muda Abdul Aziz, 50300, Kuala Lumpur, Malaysia
Department of Microbiology, Faculty of Biotechnology and Biomolecular Sciences, Universiti Pura Malaysia, 43400, Serdang, Selangor, Malaysia
d
e
f
Atta-ur-Rahman Institute for Natural Product Discovery, Universiti Teknologi MARA, Kampus Puncak Alam, 42300, Puncak Alam, Selangor, Malaysia
g
School of Pharmacy, Taylor University Lakeside Campus, No 1, Jalan Taylor’s, 47500, Subang Jaya, Selangor Darul Ehsan, Malaysia,
ARTICLE INFO
ABSTRACT
Article history:
Received
Revised
Accepted
Available online
The search of novel mPGES-1 inhibitors has recently intensified probably due to the superior
safety in comparison to existing anti-inflammatory drugs. Although two mPGES-1 inhibitors
have entered clinical trials, none has yet reached the market. In this study, we performed
modifications guided by 3D-QSAR CoMFA on 2, which is an unsymmetrical curcumin
derivative with low binding affinity towards mPGES-1. To counter the PAINS properties
predicted for 2, the diketone linker was replaced with a pyrazole ring. On the other hand, both
prenyl and carboxylate ester groups were introduced to improve the activity. When tested in vitro,
Keywords:
1
2
1 suppressed PGE biosynthesis in activated macrophages and showed promising human
3
D-QSAR CoMFA
mPGES-1 inhibition in microsomes of interleukin-1β-stimulated A549 cells. Altogether, 11 has
been identified as a potential mPGES-1 inhibitor and could be a promising lead for a novel class
of mPGES-1 inhibitors.
PGE
2
mPGES-1
PAINS
2
009 Elsevier Ltd. All rights reserved.
fever and pain but also stimulates proliferation, differentiation
and angiogenesis [3], without suppressing the biosynthesis of
other prostanoids downstream of PGH2.
Introduction
Microsomal prostaglandin E2 synthase (mPGES)-1 has
garnered an extraordinary amount of interest as a potential
therapeutic target for the treatment of inflammation-related
diseases. It is a terminal enzyme that is part of the membrane-
associated protein (MAPEG) superfamily, which is involved in
eicosanoid and glutathione metabolism [1]. In general, this
protein family comprises 5-lipoxygenase-activating protein
Curcumin and its derivatives are known to possess a wide
range of biological activities such as anti-inflammatory, anti-
tyrosinase and immunomodulatory functions [4-10]. More
recently, our group had synthesized a series of unsymmetrical
monocarbonyl 1 [11] and dicarbonyl curcumin analogues 2 and
demonstrated that they could suppress the PGE production level
(
FLAP), leukotriene C4 synthase (LTC S), and microsomal
2
4
in both LPS-induced human and murine monocytic cell lines. In
spite of this, our efforts to develop the molecules had been
hampered by the growing evidences that curcumin and its
derivatives might be categorized under pan-assay interference
substances (PAINS), which are defined by their ability to show
activity across a range of assay platforms and against a range of
proteins. Many PAINS, including the Michael acceptor
curcumin, function as reactive chemicals rather than
discriminating between drugs with defined structure activity
relationship. Others give false-positive readouts in a variety of
ways such as metal chelation, chemical aggregation, redox
glutathione transferases (MGST1, MGST2, and MGST3). While
mPGES-1 is constitutively expressed in lung, kidney and
reproductive organs, it is normally found at lower levels
elsewhere in the body although it can be strongly upregulated by
inflammatory stimuli such as interleukin-1β [2]. Since mPGES-1
is critical to the production of PGE a proinflammatory signaling
molecule found in inflammation conditions, it appears to be an
attractive drug target for osteoarthritis, rheumatoid arthritis and
cancer. This is further supported by scientific evidences that
selective mPGES-1 inhibitors could alleviate the effects of
2,
excessive production of PGE , which promotes inflammation,
2
—
——
∗
Corresponding author. Tel.: +603-92897031; e-mail: david_lam@ukm.edu.my (Lam Kok Wai); Tel.: +609-5492399; fasihi@ump.edu.my (Mohd Fadhlizil Fasihi Mohd Aluwi)

activity, compound fluorescence, cysteine oxidation or
CoMFA has been widely used to study steric potential in the
form of a Lennard-Jones function and electrostatic potential in
the form of a Coulomb function surrounding a set of biologically
characterized molecules. Whereas molecular docking has been
established as the main approach for recognizing the most likely
bioactive conformation of compounds in the receptor binding
site. Thus, combining these two methods would result in a more
thorough protein-ligand binding interaction analysis. To start
with, two different conformational sampling methods were
adopted in this study: a structural-based alignment (method A)
and a docking-based alignment (method B) (Figures 2a and b).
For the first method, the top-ranked docked ligand 3 (Figure 1a)
was initially used to build a hypothesis and predefine the binding
conformations of the other compounds. In the latter method, all
the compounds were docked to the binding site and their
conformations were used without reallignment for further
analysis. Both methods accurately predicted the mPGES-1
inhibtory activities of the MK-886 derivatives (Table 1).
However, only the results from method B were discussed here as
it has been proven to be more successful in term of predicting the
ligand binding mode and describing the contribution of the region
of interest surrounding the ligands at the protein target site
accurately.
ACCEPTED MANUSCRIPT
promiscuous binding. Moreover, curcumin is not suitable for
clinical application due to its poor oral availability, low aqueous
stability and low circulating concentrations in plasma. Whilst the
debate is still actively on-going of whether curcumin and its
derivatives are suitable to be developed as drugs, we took the
challenge to improving the properties of these molecules. Guided
by a structural based drug design approach and CoMFA, we
made several rational structural modifications to compound 2
(Figure 1a) and tested the target compound on mPGES-1
activity.
Interestingly, a quick glance through the list of some of the
mPGES-1 inhibitors found in nature (Figure 1b) revealed that
each of them comprises at least a prenyl group in their structure.
It could be a strong indicator that this functional group might
play an important role for the inhibitory activity. In fact, recent
reports have shown that curcumin and its prenylated analogues
also displayed mPGES-1 inhibition [12, 13]. It is known that the
presence of a prenyl group in a small molecule not only enhances
lipophilicity but also bestows a strong affinity for biological
membranes [14]. Nevertheless, it could improve the
pharmacokinetic profile of a given compound leading to a
broader range of interesting pharmacological activities.
OMe
OMe
O
(a)
(a)
HO
OH
OMe
OH
MeO
O
O
1
OH
Curcumin
O
O
O
Cl
N
MeO
F
(b)
2
3
(b)
O
O
OH
OH
O
HO
HO
O
O
O
O
OH
HO
OH
O
OH
O
Garcinol
Azranol
Hyperforin
Figure 2: Structural alignment of the MK-886 derivatives by the template-
based method according to the core specified by the top-docked conformation
of 3.
OMe
OMe
O
OH
O
O
In the best model, 29 reported MK-886 derivatives were
retrieved, in which 19 compounds were randomly selected as a
training set, and the remaining 10 compounds were used as a test
set. As stated in Table 1, the best model produced was highly
Prenylated curcumin
Figure 1. (a) Curcumin, monocarbonyl (1) and dicarbonyl (2) curcumin
derivatives and a MK886 derivative (3). (b) Selected structures of mPGES-1
inhibitors obtained from natural product isolation and by chemical semi-
synthesis.
2
predictive with a non-cross-validated correlation coefficient (r )
of 0.990 and
a leave-one-out cross-validated correlation
2
coefficient (q ) of 0.437 using three principle components. The
Results and discussion
external predictive ability of the model was further validated with
2
a high q value of 0.759 (Figure 3).
3
D-QSAR CoMFA
To gain insights into the contour features of mPGES-1 ligand
binding site, a set of MK-886 derivatives known to inhibit
mPGES-1 activity were retrieved from literature [15] and
subjected to docking and 3D-QSAR CoMFA (comparative
molecular field analysis) studies. To initiate that, first the
mPGES-1 protein crystal structure (PDB ID 4AL1: 1.95 Å) was
retrieved from RCSB Protein Data Bank and prepared according to
the standard protocol implemented in Discovery Studio 3.1. The
reported inhibitors were prepared, minimized and docked to the
putative binding site; the complex then underwent in situ ligand
minimization in a flexible receptor environment to refine the final
ligand binding conformation.
Figure 3: Trend of the observed versus calculated pIC50 values of the
training-set and test-set compounds.

The steric maps include green contour (Figure 4a) that
The red contour (Figure 4c) indicates region where positive
electrostatic potential is favorable for activity, and blue contour
(Figure 4d) indicates region where an increase in negative
electrostatic potential enhances activity. Superposition of the
contour maps over the binding site of mPGES-1 reveals that blue
contours predominantly occupy the region in helix IV suggesting
that negative electrostatic functional groups are highly favorable
for activity at this position while the red regions are mainly
located alongside helix I’. Several blue contour maps are also
observed in the positions adjacent to the side chains of Arg38’
and Arg52’. The latter residue is known to form an important salt
bridge interaction with the carboxylate moiety of a MK-886
derivative important for inhibition of mPGES-1.
ACCEPTED MANUSCRIPT
indicates region where the steric bulk groups favor inhibition
while the yellow contour (Figure 4b) indicates region where
steric bulk does not favor activity. Superposition of the contour
maps over the binding site of mPGES-1 reveals that the green
contours are mainly located at the edge of the shallow groove,
which comprises Leu132, Ala133, Gln134, Leu135, Pro136 and
Cys137 from helix IV and Val30’, Ala31’ and Ile32’ from helix
I’ suggesting that sterically bulky substituents are highly favored.
On the other hand, the yellow contours can be seen occupying
largely the position where cofactor GSH is bound. This suggests
that further modifications beyond the phenyl ring at this position
is highly unlikely to improve the activity.
Table 1 Statistics of the comparative molecular field analysis models generated using different alignment methods.
Leave one
out-cross-validation
External predictive
Non cross-validation
Grid
Spacing (Å)
q-squared
Model
2 a
RMS residual
error
2 b
c
Mean Absolute
2 d
RMS
error
r
q
N
q
Error
Method A (Structural alignment)
1
2
3
4
5
6
7
8
9
0.932
-
-
-
0.3062
-
-
-
0.213
0.201
0.202
0.292
0.299
0.300
0.384
0.392
0.394
1
0.898
0.897
0.899
0.848
0.844
0.844
0.785
0.774
0.772
0.828
-
-
-
0.613
-
-
-
1.0
1.5
2.0
3
4
3
4
5
3
4
5
-
-
-
-
1.000
-
-
0.0055
-
-
0.793
-
-
0.539
-
-
1.000
0.0100
0.869
0.510
Method B (Docking-based alignment)
1
1
1
1
1
1
1
1
1
0
1
2
3
4
5
6
7
8
0.922
-
-
0.867
-
-
0.990
-
-
0.3276
-
-
0.4283
-
-
0.1191
-
-
0.483
0.480
0.480
0.376
0.361
0.360
0.437
0.436
0.435
1
3
4
1
3
4
3
4
5
0.773
0.797
0.798
0.730
0.771
0.770
0.688
0.691
0.687
0.671
-
-
0.695
-
-
0.759
-
-
0.591
-
-
0.572
-
-
0.501
-
-
1.0
1.5
2.0
a
b
c
Non cross-validated correlation coefficient.
Leave one out-cross-validated correlation coefficient.
Optimum number of components obtained from cross-validated PLS analysis.
Correlation coefficient for test set predictions.
d
Pro136
Cys137
(
a)
(b)
Leu135
Gln134 Ala133
Leu132
Ile32’
Val30’
Ala31’
Cofactor GSH
binding site
Helix IV
(
c)
(d)
Helix I’
Arg38’
Arg52’
Figure 4: Superposition of the steric ((a) and (b)) and electrostatic ((c) and (d)) contour maps within the putative binding site of human mPGES-1.

Design and synthesis of 11
to obtain the corresponding crude product of 2-
benzoylcyclohexanone 8.
ACCEPTED MANUSCRIPT
Figure 5 depicts the design strategy adopted behind the
modifications of compound 2. Considering that bulky and
carboxylic acid moieties critically contribute to the binding
affinity of 3 to mPGES-1 as supported in the CoMFA study, we
opted to introduce two major functional groups into the designed
compound including i) a prenyl group that could occupy the top
cavity at the binding groove formed by Leu132, Ala133, Gln134,
Leu135, Pro136 and Cys137 from helix IV and Val30’, Ala31’
and Ile32’ from helix I’ ii) and a carboxylate ester that could
form a salt bridge interaction with Arg38’ and Arg52’. With the
introduction of these two functional groups, we predict that the
inhibitory activity and the binding affinity should be greatly
improved.
O
H
O
Br
H
HO
4
O
5
2 3
Scheme 1: Reagent and condition: acetone, K CO , reflux, 20 h.
Subsequent purification by flash chromatography yielded the
pure product with a conversion rate of 30-40%. The purified 8
was then reacted with an equivalent amount of B O to give the
2
3
boron complex. The complex was then coupled with the
prenylated benzaldehyde 5 in the presence of tributyl borate and
n-butylamine in ethyl acetate. The reaction was carried out under
reflux condition overnight. The crude product was purified using
column chromatography to obtain the unsymmetrical dicarbonyl
curcumin derivative 9. Cyclization of the purified 9 was achieved
via Michael addition by reacting with hydrazine to give pyrazole
10. Finally, the carboxylate ester group was introduced to the
compound via N-alkylation with methyl 2-bromoacetate in the
To initiate the synthesis, the prenyl group was coupled to 4-
hydroxybenzaldehyde (4) by treatment with 1-chloro-3-methyl-2-
butene in the presence of K CO in acetone under reflux
2
3
condition. The desired product 5 was obtained in good yield
80%) (Scheme 1) and was used as a building block for the
(
synthesis of 11 as depicted in Scheme 2. Cyclohexanone 6 (48
mmol) was reacted with pyrrolidine (48 mmol) in the presence of
p-toluenesulfonic acid (1 mmol) to afford N-(1-
cyclohexenyl)pyrrolidine 7, an essential enamine intermediate.
Next, 7 (1 mmol) was reacted with benzoic anhydride (2 mmol)
presence of Cs CO in DMF as solvent. The crude product was
2 3
then purified using column chromatography to obtain the final
product 11 in satisfactorily yield (55%) (Scheme 2).
Arg52'
Arg38'
Arg52'
Arg38'
HO
O
Negative
electrostatic
group
Design
strategy
Cl
Leu132
Leu135
O
Leu132
Leu135
O
N
Cys137
Cys137
Ile32'
MeO
Ile32'
Ala31'
Bulky
group
F
2
Ala31'
Leu142
Leu142
Val30'
Val30'
Ile25'
Ile25'
MK-886 derivative, 3
Figure 5: A general design strategy behind the modifications of 2.
O
O
O
O
O
O
a
N
7
b
c
d
B
O
6
8
NH
N
OH O
e
O
O
10
9
f
OMe
O
N
N
O
11
°
Scheme 2: Reagents and conditions: (a) pyrolidine, p-TSOH, toluene, reflux under Dean and Stark apparatus. (ii) Distilled until it reached 108-110 C; (b)
°
benzoic anhydride, toluene, r.t. overnight; (c) B
2
O
3
, ethyl acetate, refllux, 70 C, 3 h; (d) (i) respective benzaldehyde 5, tributylborate, 30 min, r.t., (ii) n-
°
butylamine, reflux, 70 C, overnight, (iii) 1
M
HCl, reflux, 30 min; e) hydrazine, EtOH, reflux, 1h.; f) methyl 2-bromoacetate, Cs CO , DMF, r.t.
2
3

Effects of 11 on PGE2 production level in RAW264.7
macrophages and human cell-free mPGES-1
Binding interactions of 11 with mPGES-1
ACCEPTED MANUSCRIPT
As shown in Figure 7, the RMSD of the simulated protein
backbone converged to equilibrium state after approximately 4 ns
while 11 showed a stabilizing trend after 19 ns. As anticipated in
the early simulation, the carboxylate group of 11 binds to the
putative binding site by forming an important hydrogen bond
with Arg52’ while the prenyl group occupies the cavity at the top
of the binding groove with a series of hydrophobic interactions
with amino acids from helix IV (Gln134, Leu135 and Ala138)
and helix I’ (Tyr28’ and Ile32’). The central core of 11 resides
above the GSH binding site while both the phenyl rings interact
with the adjacent protein hydrophobic residues (Figure 8).
However, the carboxylate ester group of 11 failed to sustain its
hydrogen bonding interaction with the guanidium side chain of
Arg52’ after the 19 ns mark. Without this important interaction, it
is expected that 11 would possess a lower binding affinity. In
contrary, the prenyl group forms a hydrophobic π-alkyl
interaction with Ile25, which was evident over the entire time
frames. Furthermore, the flexibility of the prenyl chain around
the hydrophobic cavity could play a key role for its dual activity
on both rat and human mPGES-1. This is further supported since
mPGES-1 exhibits different amino acid sequence at this position
resulting in a smaller edge [16].
To verify whether the designed compound could interfere
with mPGES-1 activity, we investigated its effect on mPGES-1-
dependent PGE2 biosynthesis in murine RAW264.7
macrophages. As expected, the compound almost completely
suppressed PGE2 formation (>90%) in LPS-activated
macrophages at 25 µM. However, it is known that there is a
difference in the mPGES-1 structure between human and
murine/rat species. To further confirm whether 11 could act on
human mPGES-1, we carefully assessed its inhibitory activity
against human cell-free mPGES-1 using microsomes of
interleukin-1β-stimulated A549 cells. At 1 µM, 11 was found to
inhibit mPGES-1 activity by 37% (Figure 6). MK-886 (Figure
S-1) (10 µM) was used as a reference inhibitor and suppressed
the PGE synthesis as expected. Since 11 was designed from the
2
lead compound 2, which represents a derivative of curcumin, we
evaluated its PAINS character using in silico tools (FAF-Drug4).
Overall, 11 was not categorized as potential PAINS or covalent
inhibitor while 2 and curcumin were identified as covalent
inhibitors. Thus, 11 might represent a valuable lead for novel
mPGES-1 inhibitors that lacks the pan-assay interference
activities.
Figure 6: Effect of 11 (1 µM) and the reference inhibitor MK-886 (10 µM)
on cell-free mPGES-1 activity. Data as relative units are paired for
independent experiments, n = 3, paired t-test based on logarithmized values.
Figure 7: RMSD plots of mPGES-1 protein backbone (black line) and 11
(red line) as a function of simulation time.
0
ns
30 ns
50 ns
Figure 8: Binding interactions of 11 with the adjacent residues in the mPGES-1 binding site at time frame of 0 ns, 30 ns and 50 ns.

tetrazolium bromide (MTT, 5 mg/mL in phosphate buffered
ACCEPTED MANUSCRIPT
Experimental
saline (PBS) to formazan salts. After treatment, the supernatant
were removed followed by addition of 20 µL of MTT reagents
into each well. The mixture of culture media and MTT were
removed after incubated at 37°C for 4 h, and the formazan salts
were dissolved by adding 100% DMSO. The absorbance was
then measured at 570 nm on a SpectraMax Plus microplate reader
(Molecular Devices Inc., Sunnyvale, CA, USA) at room
temperature.
Chemical synthesis and spectral characterizations
General procedure for the preparation of 4-((3-methylbut-2-en-
-yl)oxy)benzaldehyde (5). 10 mmol of 4-hydroxylbenzaldehyde
was treated with 10 mmol of 1-chloro-3-methyl-2-butene in the
presence of K CO in acetone under reflux condition for 20 h.
1
2
3
The reaction mixture was extracted with ethyl acetate and water,
subsequently dried over anhydrous magnesium sulphate. The
desired product 5 was purified by column chromatography and
was used for the next steps. General procedure for the
preparation of 11. A catalytic amount of p-toluenesulphonic acid
was added into a mixture of cyclohexanone (20 mmol) and
pyrrolidine (20 mmol) in 30 mL of toluene and the mixture was
then refluxed on a Dean & Stark apparatus for 2 hours to prepare
enamine. Upon completion, 20 mmol of benzoic anhydride in 20
mL of toluene was added dropwise into the reaction solution and
stirred for 24 hours. Distilled water (10 ml) was then added and
further refluxed for 30 min. The resulting reaction mixture was
Determination of PGE . The cell culture supernatants were
2
collected and analyzed for PGE secretion using PGE EIA kits
2
2
(Cayman Chemical, Ann Arbor, MI, USA). The protocols
provided by the manufacturers were followed to the detail. The
data was obtained using a SpectraMax Plus microplate reader
(Molecular Device, Sunnyvale, CA, USA). The concentration of
PGE2 for each sample was calculated from their respective
standard curves.
Determination of human mPGES-1 activity: Microsomes were
isolated from interleukin-1β-stimulated A549 cells and mPGES-1
activity determined as previously reported [17]. In brief,
microsomal membranes (2.5-5 µg total protein) were diluted in
0.1ꢀM potassium phosphate buffer, pHꢀ7.4, containing 2.5ꢀmM
glutathione and pre-incubated with vehicle (DMSO, 2%) or test
extracted with 3
M HCl and with 20 mL water. The toluene layer
was dried over anhydrous magnesium sulphate and concentrated
in vacuo to give crude compound. The crude compound was
purified by column chromatography to afford 2-
benzoylcyclohexanones. Boron trioxide (1.5 g) was added into a
flask containing 10 mL of ethyl acetate in the presence of 2-
compounds for 15 min at 4°C. PGE
volume of 100 µl) was initiated by addition of PGH
final concentration) and terminated after 1 min by addition of 100
µl FeCl (40 mM), citric acid (80 mM) and 11β-PGE (10 µM) as
was extracted and analysed by RP-HPLC
synthesis (in a reaction
2
(20 µM
2
benzoylcyclohexanone and stirred at 70°C for
3 hours.
Compound 6 (20 mmol) and tributyl borate (20 mmol) were then
added into the solution and the mixture was stirred for 30
minutes. Catalytic amount of n-butylamine was added dropwise
2
2
internal standard. PGE
as described [17].
2
and stirred for overnight followed by reflux with 10 mL of 1
M
HCl for 30 minutes. The reaction mixture was extracted with
ethyl acetate and dried over anhydrous magnesium sulphate. The
desired product was purified by column chromatography. Methyl
In silico modeling
Molecular Docking. Docking studies were performed on the
crystal structure of mPGES-1 (pdb ID: 4AL1; 1.95 Å) retrieved
from the Protein Data Bank using the Cdocker protocol under the
(E)-2-(7-(4-((3-methylbut-2-en-1-yl)oxy)benzylidene)-3-phenyl-
4
,5,6,7-tetrahydro-2H-indazol-2-yl)acetate (11). White crystal
®
1
receptor–ligand interaction section in Discovery Studio 3.1
(32%); m.p: 114-115°C. H NMR (600 MHz, CDCl ) δ 7.73-7.74
3
(Accelrys, San Diego, USA). Cdocker is a grid-based molecular
(d, J = 6 Hz, 2H), 7.40-7.42 (t, J = 6 Hz, 2H), 7.31-7.33 (t, J = 6
docking method that employs CHARMM force fields. During
refinement, the receptor was held rigid while the ligands were
allowed to flex. High-temperature molecular dynamics, followed
by random rotations, refinement by grid-based (GRID 1)
simulated annealing, and a final grid-based or full force-field
minimization [18] generated 200 random ligand conformations
from the initial ligand structure. In this experiment, the ligand
was heated to a temperature of 700 K in 2000 steps and cooled by
Hz, 1H), 7.22-7.23 (d, J = 6 Hz, 2H), 6.91-6.93 (d, J = 12 Hz,
2
4
H), 6.47 (s, 1H), 5.50-5.52 (t, J = 6 Hz, 1H), 5.17 (s, 2H), 4.53-
.54 (d, J = 6 Hz, 2H), 3.81 (s, 3H), 2.84-2.86 (m, 2H), 2.73-2.75
13
(m, 2H), 1.87-1.89 (m, 2H), 1.81 (s, 3H), 1.79 (s, 3H). C NMR
(
151 MHz, CDCl ) δ 168.97, 157.93, 148.23, 140.39, 138.42,
3
1
1
2
33.73, 130.58, 129.25, 128.56, 128.44, 127.45, 127.06, 123.83,
19.55, 116.92, 114.48, 64.84, 53.24, 52.66, 27.91, 25.85, 24.68,
2.66, 18.23.
3
00 K in 5000 steps. The grid extension was set to 14 Å.
Hydrogen atoms were added to the structure and all ionizable
residues were set at their default protonation states at a neutral
pH. The ten top ligand-binding poses were ranked according to
their CDOCKER energies, and the predicted binding interactions
were analyzed. The best pose was retrieved and minimized in situ
according to the standard protocol implemented in Discovery
Studio before being subjected to binding-energies and binding-
scores calculations.
Determination of biological activity
Cell Culture. Murine macrophages. The RAW264.7 cells line
®
™
from the (ATCC TIB-71 ) cells line from the American Type
Culture Collection (ATCC) (Manassas, VA, United States) were
grown in DMEM containing 10% FBS, 1% (v/v) penicillin
°
G/streptomycin in 5% CO at 37 C. RAW264.7 cells (80–90%
2
confluent) were detached and centrifuged at 1000 RPM at 4°C
for 10 min. The viability of cultured cells used in the assay was
always >95% as determined by trypan blue dye exclusion. Cell
stimulation and treatment. RAW264.7 (5 × 10 cells/well) were
seeded into a tissue culture grade 96-well plate and incubated for
3
D QSAR (CoMFA). To initiate the study, 29 reported MK-886
derivatives [15] that exhibited IC50 of less than 10 µM have been
selected for in silico studies while 3 compounds that exhibited
IC50 more than 10 µM have been omitted. A template-based
alignment method was first applied to the molecules in the data
set, in which the most active compound, i.e., 3, was chosen as the
template for the alignment of the other compounds. Nineteen
compounds were categorized as the training set and the
remaining ten compounds were used as the test set. The ratio of
training set to test set compounds is ~7:3 and the training set
5
2
4 h at 37°C, 5% CO for cell attachment. The attached cells
2
were stimulated in 100 U/mL of recombinant IFN-γ and 2 µg/mL
of LPS in presence of vehicle (DMSO, 0.1%) or test compound at
a final volume of 100 µL/well. Cells were then incubated at
3
7°C, 5% CO for 17–20 h. Cell Viability. The cytotoxicity of
2
test compound on cultured cells was determined by assaying the
reduction of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-

holds the most active and most inactive compounds. Similarly,
the test-set compounds should cover a range of biological
activities and enable evaluation of the quality of the generated 3D
QSAR method. The inhibitory activity (IC ) values of the
Kebangsaan Malaysia for the funds provided under the Research
ACCEPTED MANUSCRIPT
University Grant UKM-DIP-2014-16.
References
50
[1] C. Jegerschold, S.C. Pawelzik, P. Purhonen, P. Bhakat, K.R.
studied compounds were converted into pIC50 using pIC = -log
50
Gheorghe, N. Gyobu, K. Mitsuoka, R. Morgenstern, P.J. Jakobsson,
H. Hebert, Structural basis for induced formation of the
inflammatory mediator prostaglandin E2, Proceedings of the
National Academy of Sciences of the United States of America
105(32) (2008) 11110-5.
[2] S. Chandrasekharan, N.A. Foley, L. Jania, P. Clark, L.P. Audoly,
B.H. Koller, Coupling of COX-1 to mPGES1 for prostaglandin E2
biosynthesis in the murine mammary gland, Journal of lipid research
46(12) (2005) 2636-48.
IC50 and used as dependent variables in the 3D QSAR
calculations. To accelerate the analysis and reduce noise, column
filtering was set at three different energies, i.e., 2.0, 1.5 and 1.0
kcal/mol, such that only the higher steric and electrostatic
energies in each case were considered in the partial least-squares
(
PLS) analysis. The highest values of the cross-validated
2
2
correlation coefficient (q ) and non-cross-validation (r ) were
used to judge the quality of the models and avoid over-fitting.
The cross-validated result quantifies the predictive ability of the
model and is determined by the leave-one-out (LOO) procedure
of cross-validation, in which each compound is successively
removed from the model derivation. Finally, the predictive
quality of the “best” correlation model was determined.
[
3] H.H. Chang, E.J. Meuillet, Identification and development of
mPGES-1 inhibitors: where we are at?, Future medicinal chemistry
(15) (2011) 1909-34.
4] C.L. Tham, C.Y. Liew, K.W. Lam, A.-S. Mohamad, M.K. Kim,
3
[
Y.K. Cheah, Z.-A. Zakaria, M.-R. Sulaiman, N.H. Lajis, D.A. Israf,
A synthetic curcuminoid derivative inhibits nitric oxide and
proinflammatory cytokine synthesis, European Journal of
Pharmacology 628(1–3) (2010) 247-254.
Molecular dynamics simulation. The molecular dynamics
simulation input files were prepared using GROMACS v2016.3
with the united-atom GROMOS96 54a7 force field [19]. The
parameterization of 11 was carried out using ATB (Automated
Topology Builder) server [20]. The POPC bilipid layer
membrane structure along with the Berger Lipid parameters were
retrieved from Biocomputing Group website [21]. The protein-
ligand complex structure was then packed with the lipids using
the InflateGRO methodology. Next, the membrane protein
system was solvated with water molecules. To avoid random
water molecules from filling up the gaps in the lipid acyl chains,
the van der Waals radii of the carbon atoms in the system were
artificially enlarged by changing the value of 0.15 to 0.375. The
final system contained three protein chains, 123 POPC moieties
and 4545 water molecules, (a total of 24683 atoms in a 70 Å × 70
Å × 71 Å simulation box). The ionic strength was adjusted by the
[5] K.-H. Lee, F.H. Ab. Aziz, A. Syahida, F. Abas, K. Shaari, D.A.
Israf, N.H. Lajis, Synthesis and biological evaluation of curcumin-
like diarylpentanoid analogues for anti-inflammatory, antioxidant
and anti-tyrosinase activities, European Journal of Medicinal
Chemistry 44(8) (2009) 3195-3200.
[6] C.L. Tham, K.W. Lam, R. Rajajendram, Y.K. Cheah, M.R.
Sulaiman, N.H. Lajis, M.K. Kim, D.A. Israf, The effects of a
synthetic
curcuminoid
analogue,
2,6-bis-(4-hydroxyl-3-
methoxybenzylidine)cyclohexanone on proinflammatory signaling
pathways and CLP-induced lethal sepsis in mice, European Journal
of Pharmacology 652(1–3) (2011) 136-144.
[7] S.N.A. Bukhari, G. Lauro, I. Jantan, G. Bifulco, M.W. Amjad,
Pharmacological evaluation and docking studies of α,β-unsaturated
carbonyl based synthetic compounds as inhibitors of secretory
phospholipase
proinflammatory cytokines, Bioorganic & Medicinal Chemistry
2(15) (2014) 4151-4161.
8] S. Leong, S. Faudzi, F. Abas, M. Aluwi, K. Rullah, L. Wai, M.
A2,
cyclooxygenases,
lipoxygenase
and
+
-
addition of 0.1
The complex was subjected to energy minimization, followed by
ns of NVT equilibration and 5 ns of NPT equilibration. The
M of NaCl with a total of 40 Na and 63 Cl ions.
2
[
5
Bahari, S. Ahmad, C. Tham, K. Shaari, N. Lajis, Synthesis and Sar
Study of Diarylpentanoid Analogues as New Anti-Inflammatory
Agents, Molecules 19(10) (2014) 16058.
production run was performed using the LINCS algorithm to
constrain bond lengths and periodic boundary conditions were
applied in all directions [22]. Long range electrostatic forces will
be treated using the Fast Particle-Mesh Ewald method (PME)
[9] S.M. Mohd Faudzi, S.W. Leong, F. Abas, M.F.F. Mohd Aluwi,
K. Rullah, K.W. Lam, S. Ahmad, C.L. Tham, K. Shaari, N.H. Lajis,
Synthesis, biological evaluation and QSAR studies of
diarylpentanoid analogues as potential nitric oxide inhibitors,
MedChemComm 6(6) (2015) 1069-1080.
[
23]. Van der Waals forces and Coulomb potential were treated
using a cut-off of 1.2 nm and the simulation time step was set to
fs. An initial velocity obtained according to a Maxwell
2
distribution at 310 K is given to all the atoms. Nosé-Hoover
thermostat and semiisotropic Parinello-Rahman pressure
coupling were set at 310 K and 1 bar with a coupling time of τT
[10] S.W. Leong, S.M. Mohd Faudzi, F. Abas, M.F.F. Mohd Aluwi,
K. Rullah, K.W. Lam, M.N. Abdul Bahari, S. Ahmad, C.L. Tham, K.
Shaari, N.H. Lajis, Nitric oxide inhibitory activity and antioxidant
evaluations of 2-benzoyl-6-benzylidenecyclohexanone analogs, a
novel series of curcuminoid and diarylpentanoid derivatives,
Bioorganic & Medicinal Chemistry Letters 25(16) (2015) 3330-
=
0.5 ps and τP = 2 ps, respectively [24, 25]. The run was set for
5
0 ns at constant pressure and temperature conditions.
Conclusion
3
337.
[
11] M.F.F. Mohd Aluwi, K. Rullah, B.M. Yamin, S.W. Leong,
Compound 11 was successfully synthesized as a mPGES-1
inhibitor guided by a 3D-QSAR CoMFA approach. The
compound is predicted to lack the pan-assay interference
properties due to removal of the diketone linker. Apart from that,
both carboxylic acid and prenyl groups have been identified to
play a vital role in the activity. Future study shall address the
structure-activity relationships of this novel class of mPGES-1
inhibitor in more detail.
M.N. Abdul Bahari, S.J. Lim, S.M. Mohd Faudzi, J. Jalil, F. Abas,
N. Mohd Fauzi, N.H. Ismail, I. Jantan, K.W. Lam, Synthesis of
unsymmetrical monocarbonyl curcumin analogues with potent
inhibition on prostaglandin E2 production in LPS-induced murine
and human macrophages cell lines, Bioorganic & Medicinal
Chemistry Letters.
[12] A. Koeberle, H. Northoff, O. Werz, Curcumin blocks
prostaglandin E2 biosynthesis through direct inhibition of the
microsomal prostaglandin E2 synthase-1, Mol Cancer Ther 8(8)
(2009a) 2348-55.
Acknowledgement
[13] M. Iranshahi, M.G. Chini, M. Masullo, A. Sahebkar, A.
This work was financially supported by ScienceFund (02-01-02-
SF00665), Ministry of Science, Technology & Innovation,
Malaysia and FRGS (FRGS/2/2014/ST01/UKM/02/3), Ministry
of High Education, Malaysia. Authors also thank Universiti
Javidnia, M. Chitsazian Yazdi, C. Pergola, A. Koeberle, O. Werz, C.
Pizza, S. Terracciano, S. Piacente, G. Bifulco, Can Small Chemical
Modifications of Natural Pan-inhibitors Modulate the Biological

Selectivity? The Case of Curcumin Prenylated Derivatives Acting as
ACCEPTED MANUSCRIPT
HDAC or mPGES-1 Inhibitors, J Nat Prod 78(12) (2015) 2867-79.
14] K. Rullah, M.F.F. Mohd Aluwi, B.M. Yamin, M.N. Abdul
[
Bahari, L.S. Wei, S. Ahmad, F. Abas, N.H. Ismail, I. Jantan, L.K.
Wai, Inhibition of prostaglandin E2 production by synthetic minor
prenylated chalcones and flavonoids: Synthesis, biological activity,
crystal structure, and in silico evaluation, Bioorganic & Medicinal
Chemistry Letters 24(16) (2014) 3826-3834.
[15] D. Riendeau, R. Aspiotis, D. Ethier, Y. Gareau, E.L. Grimm, J.
Guay, S. Guiral, H. Juteau, J.A. Mancini, N. Methot, J. Rubin, R.W.
Friesen, Inhibitors of the inducible microsomal prostaglandin E2
synthase (mPGES-1) derived from MK-886, Bioorganic & medicinal
chemistry letters 15(14) (2005) 3352-5.
[16] S.C. Pawelzik, N.R. Uda, L. Spahiu, C. Jegerschold, P.
Stenberg, H. Hebert, R. Morgenstern, P.J. Jakobsson, Identification
of key residues determining species differences in inhibitor binding
of microsomal prostaglandin E synthase-1, J Biol Chem 285(38)
(2010) 29254-61.
[17] A. Koeberle, U. Siemoneit, U. Buhring, H. Northoff, S. Laufer,
W. Albrecht, O. Werz, Licofelone suppresses prostaglandin E2
formation by interference with the inducible microsomal
prostaglandin E2 synthase-1, The Journal of pharmacology and
experimental therapeutics 326(3) (2008) 975-82.
[18] G. Wu, D.H. Robertson, C.L. Brooks, 3rd, M. Vieth, Detailed
analysis of grid-based molecular docking: A case study of
CDOCKER-A CHARMm-based MD docking algorithm, Journal of
computational chemistry 24(13) (2003) 1549-62.
[19] D. Van Der Spoel, E. Lindahl, B. Hess, G. Groenhof, A.E.
Mark, H.J. Berendsen, GROMACS: fast, flexible, and free, J
Comput Chem 26(16) (2005) 1701-18.
[20] A.K. Malde, L. Zuo, M. Breeze, M. Stroet, D. Poger, P.C. Nair,
C. Oostenbrink, A.E. Mark, An Automated Force Field Topology
Builder (ATB) and Repository: Version 1.0, Journal of Chemical
Theory and Computation 7(12) (2011) 4026-4037.
[21] O. Berger, O. Edholm, F. Jähnig, Molecular dynamics
simulations of a fluid bilayer of dipalmitoylphosphatidylcholine at
full hydration, constant pressure, and constant temperature,
Biophysical Journal 72(5) (1997) 2002-2013.
[22] B. Hess, H. Bekker, H.J.C. Berendsen, J.G.E.M. Fraaije,
LINCS: A linear constraint solver for molecular simulations, Journal
of Computational Chemistry 18(12) (1997) 1463-1472.
[23] T. Darden, D. York, L. Pedersen, Particle mesh Ewald: An
N⋅log(N) method for Ewald sums in large systems, The Journal of
Chemical Physics 98(12) (1993) 10089-10092.
[
24] S. Nosé, A molecular dynamics method for simulations in the
canonical ensemble, Molecular Physics 52(2) (1984) 255-268.
25] M. Parrinello, A. Rahman, Crystal Structure and Pair Potentials:
A Molecular-Dynamics Study, Physical Review Letters 45(14)
1980) 1196-1199.
[
(

ACCEPTED MANUSCRIPT
FakultiFarmasi
Faculty of Pharmacy
30 April 2019
Professor Jan Lundell, PhD
Editor
Journal of Molecular Structure
Dear Sir,
Submission of Manuscript Entitled ‘Design and synthesis of a novel mPGES-1 lead
inhibitor guided by 3D-QSAR CoMFA’
I hereby include a copy of the revised manuscript entitled ‘Design and synthesis of a novel
mPGES-1 lead inhibitor guided by 3D-QSAR CoMFA’ for consideration to be published in
your respective journal.
This paper describes the design and synthesis of unsymmetrical pyrazolocurcumin
derivative as a novel mPGES-1 lead inhibitor guided by in silico 3D-QSAR CoMFA
studies of MK-886 derivatives. The biological and computational results presented in this
study will also allow the readers to understand the plausible protein-ligand complex
interactions involved. We are of the opinion that the publication of this paper in your
journal will provide the scientific community a wider understanding on the application of
modeling studies in the field of drug discovery.
Looking forward to hearing from you.
Thank you.
Yours sincerely,
Lam Kok Wai, PhD
Drug and Herbal Research Centre
Faculty of Pharmacy
Universiti Kebangsaan Malaysia
Jalan Raja Muda Abdul Aziz
5
0300 Kuala Lumpur, Malaysia
Tel : +60392897031
Email: david_lam@ukm.edu.my
FakultiFarmasi, UniversitiKebangsaan Malaysia, Jalan Raja Muda Abdul Aziz, 50300 Kuala Lumpur, Malaysia
Telefon: +603-9289 7074 Faksimili: +603-2698 3271 Laman Web: http://www.ukm.my/farmasi

ACCEPTED MANUSCRIPT
FakultiFarmasi
Faculty of Pharmacy
Highlights
•
•
We have successfully designed and synthesized a novel mPGES-1 lead inhibitor guided by in silico
D-QSAR CoMFA studies of MK-886 derivatives.
3
The biological and modeling results presented provide in-depth understanding on the plausible
binding interactions between 11 and mPGES-1 binding site.
FakultiFarmasi, UniversitiKebangsaan Malaysia, Jalan Raja Muda Abdul Aziz, 50300 Kuala Lumpur, Malaysia
Telefon: +603-9289 7074 Faksimili: +603-2698 3271 Laman Web: http://www.ukm.my/farmasi