Active site characterization and structure based 3D-QSAR studies on nonredox type 5-lipoxygenase inhibitors
Zaheer Ul-Haq, Naveed Khan, Syed Kashif Zafar, Syed Tarique Moin
PII:
DOI:
Reference:
S0928-0987(16)30068-9
doi: 10.1016/j.ejps.2016.03.014
PHASCI 3514
To appear in:
Received date:
Revised date:
Accepted date:
26 October 2015
11 February 2016
12 March 2016
Please cite this article as: Ul-Haq, Zaheer, Khan, Naveed, Zafar, Syed Kashif, Moin, Syed
Tarique, Active site characterization and structure based 3D-QSAR studies on non-redox
type 5-lipoxygenase inhibitors, (2016), doi: 10.1016/j.ejps.2016.03.014
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.
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
Active Site Characterization and Structure Based 3D-QSAR
MA
Studies on Non-redox Type 5-Lipoxygenase Inhibitors
Dr. Panjwani Center for Molecular Medicine and Drug Research, International Center for Chemical
PT
a
ED
ZaheerUl-Haqa,*, Naveed Khana, Syed KashifZafara and Syed TariqueMoinb
HEJ Research Institute of Chemistry,International Center for Chemical and Biological Sciences,
AC
b
CE
and Biological Sciences, University of Karachi, Karachi-75270, Pakistan.
University of Karachi, Karachi-75270, Pakistan.
*Towhomcorrespondenceshouldbeaddressed.
E-mail:zaheer.qasmi@iccs.edu;
ACCEPTED MANUSCRIPT
Abstract
Structure-based 3D-QSAR study was performed on a class of 5-benzylidene-2-phenylthiazolinones
non-redox type 5-LOX inhibitors. In this study, binding pocket of 5-Lipoxygenase (pdb id 3o8y) was
NU
SC
RI
PT
identified by manual docking using 15-LOX (pdb id 2p0m) as a reference structure. Additionally, most
of the binding site residues were found conserved in both structures. These non-redox inhibitors were
then docked into the binding site of 5-LOX. To generate reliable CoMFA and CoMSIA models, atom fit
data base alignment method using docked conformation of the most active compound was employed.
The q2cv and r2ncv values for CoMFA model were found to be 0.549 and 0.702, respectively. The q2cv
and r2ncv values for the selected CoMSIA model comprised four descriptors steric, electrostatic,
hydrophobic and hydrogen bond donor fields were found to be 0.535 and 0.951, respectively. Obtained
results showed that our generated model was statistically reliable. Furthermore, an external test set
MA
validates the reliability of the predicted model by calculating r2pred i.e.0.787 and 0.571 for CoMFA and
CoMSIA model, respectively. 3D contour maps generated from CoMFA and CoMSIA models were
ED
utilized to determine the key structural features of ligands responsible for biological activities. The
Introduction
CE
1.
PT
applied protocol will be helpful to design more potent and selective inhibitors of 5-LOX.
Lipoxygenases (LOXs) has grasped so much attention of the researchers during last three decades
AC
because of their catalytic products, such as Leukotrienes (LTs) and Lipoxins[1] and [2]. Lipoxygenases
is the non-heme iron containing dioxygenase that oxidizes 1,4-cis-cis-pentadienes containing poly
unsaturated fatty acids. Classification of mammalian LOXs is based on regio and stereo-selective
oxidation of their natural substrates for instance, arachidonic acid (AA) [3] and [4]. Among several
LOXs, 5-LOX has been proved as the key enzyme for biosynthesis of LTs which are important
mediators of inflammation and allergic responses. Earlier, it was reported that LTs also caught up in
Acute Lung Injury (ALI) [5]. 5-LOX is also involved in gastroesophageal reflux disease, rheumatoid
arthritis and atherosclerosis [6], [7] and [8]. Prostate and other cancer cell lines found high expression
of 5-LOX [9]. Lots of experimental along with computational studies have been carried out on 5-LOX.
Recently, different classes of 5-LOX inhibitors were reported in literature [10], [11] and [12]. So far, on
the basis of mechanism of action, four different types of 5-LOX inhibitors have been identified. Among
them three are direct inhibitors for 5-LOX (one form chelate with Iron, 2nd type reduces the binding site
Iron and third one are non-redox type), according to our knowledge, all the direct inhibitors were bind
ACCEPTED MANUSCRIPT
at the same binding site. The remaining one is indirect inhibitor which blocks the functional interaction
between 5-LOX and 5-Lipoxygenase activating protein (FLAP) [13] and [14]. In spite of all these
efforts,, Zileuton is the only direct 5-LOX inhibitor as a drug available in the market for the treatment
of asthma with some therapeutic drawbacks [15]. Therefore, selective inhibitors of 5-LOX are an
NU
SC
RI
PT
exigency that can be obtained by using different rational approaches.
Three dimensional quantitative structure activity relationships (3D-QSAR) is an extensively used
technique among all Computational Aided Drug Design (CAAD) techniques. This method predicts the
biological activity of known and unknown compounds by using statistical techniques and optimizing
new lead molecules [16]. Especially, Comparative Molecular Field Analysis (CoMFA) [17] and
Comparative Molecular Similarity Indices Analysis (CoMSIA) [18] are widely used as 3D-QSAR
methods. CoMFA describing two fields such as steric and electrostatic fields and correlates molecular
MA
interaction fields with the experimental biological activities via partial least square (PLS) [19].
CoMSIA is somewhat similar to CoMFA but it has other different fields like steric, electrostatic,
hydrophobic, hydrogen bond donor and hydrogen bond acceptor. Crystal structure of ligand-free 5-
ED
LOX was reported in Protein Data Bank (pdb id 3o8y) [20] but no complex structure is reported so far.
Therefore, ambiguity still remains in binding site of 5-LOX. Here, we characterized binding site of 5-
PT
LOX (pdb id 3o8y) through docking that came out to be same as described in an earlier study [21].
Molecular docking and structure based 3D-QSAR using CoMFA and CoMSIA are carried out on a set
CE
of 27 thiazolinone inhibitors identified through virtual screening and synthesized by Bettina Hofmann
and co-workers [22]. Our study will provide an insight to establish a relationship between experimental
AC
biological activity and structural attributes of 5-LOX inhibitors that may also guide to design selective
and potent inhibitors of 5-LOX.
2.
Material and Methods
A class of thiazolinone with 2-phenyl moeity reported as 5-LOX inhibitors was selected as data set
from the past article along with their IC50 values from the cell free S100 assay [22]. The data set
contain 27, IC50 values were reformed into pIC50 value by taking the negative logarithm (-log10) of
IC50, out of 27, six inhibitors were selected as test set which includes the whole range of biological
activities and diverse structures as present in training set. Structure of compounds is shown in Table 1
along with their IC50and pIC50 values.
All molecular modeling were performed on a dual core processor (Intel®XenonTM, Santa Clara, CA,
ACCEPTED MANUSCRIPT
USA), CPU 3.00 GHz, openSUSE 11.4 on Linux platform.
2.1.
Molecular Modeling of Ligand
were calculated for all inhibitors by using
NU
SC
RI
PT
Structures of the inhibitors were built and hydrogens were added. Furthermore, partial atomic charges
gasteiger-huckle charge scheme [23] as well as
minimization of each structure was carried out by utilizing standard tripos molecular mechanics force
field (MMFF) [24] with a 0.05 kcal/mol A0 energy gradient convergence criteria with 1000 iteration as
implemented in Sybyl7 [25].
Similarity Checking Between 5-LOX and 15-LOX
MA
2.2.
From extensive literature survey, we have come to know that among several LOX homologs, 5-LOX is
highly identical with 15-LOX at its catalytic domain due to which 15-LOX has been chosen to
ED
characterize the binding site of 5-LOX [26]. Initially, sequence similarity between 5-LOX (ligand-free)
[20] and 15-LOX (co-crystallized with RS7 inhibitor, (2E)-3-(2-oct-1-yn-1-yl-phenyl) acrylic acid pdb-
PT
id: 2p0m, chain B) [27] were checked.
These sequences were submitted for multiple sequence
alignment utilizing ClustalW[28]. The aligned sequences were then uploaded to Ident and Sim web
CE
server in order to check the similarity between the proteins as shown in Figure S1. Results show that 5-
2.3.
AC
LOX and 15-LOX were found within similarity range obtained as 44.46%.
Molecular Docking Simulation
Coordinates of the 5-LOX and the 15-LOX were retrieved from the databank. 15-LOX was prepared
by deletion of water molecules and adding hydrogens to the protein subsequently, two proteins (15LOX and 5-LOX) were superimposed taking 15-LOX as a reference structure shown in Figure 1. The
superposition led to modify the coordinates of the target protein (5-LOX) same as of the reference
protein. Thereafter, coordinates of the RS7 were transferred from 15-LOX to 5-LOX, water molecules
were deleted and partial atomic charges were assigned to whole protein except metal ion with formal
charge of +2. Structure was minimized by applying constraint to the co-ordinate residues of Iron using
Standard Tripos MMFF [24] and Powell energy minimization algorithm [29] with 0.05 kcal/molÅ
gradient convergence with 1000 iteration. This whole procedure was done on Sybyl7.3 [25].To further
validate the reliability of the model and docking protocol, two different docking software MOE and
ACCEPTED MANUSCRIPT
GOLD were utilized.
2.4.
Alignment of Database
NU
SC
RI
PT
Alignment is one of the most fundamental step for generation of the best QSAR model. Here, two
alignment strategies are exercised, receptor based alignment, afterward, database alignment. The
validated model for docking was used as a complex structure for docking the whole dataset of
thiazolinone inhibitors. GOLD [30] was utilized to find out the most favorable conformation for the
whole dataset. The protocol applied for the docking is as follows: Default algorithm was selected and
binding site of 10Å was chosen around the centroid of the ligand docked incorporated in the 5-LOX.
Default metal geometry was adopted, and number of conformations generated for each docking run
MA
was set to 10. Early termination is opted if top three poses fall within rmsd of 1.5 Å. The highest ranked
GOLDSCORE conformation of the most active compound i.e. compound 18 was selected as the
CoMFA Set-up
PT
2.5.
ED
template for aligning other twenty-six inhibitors through database alignment tool (Figure 2).
CE
CoMFA describes two descriptors steric (Lennard Jones potential) and electrostatic (Columbic
interactions). These fields were measured by setting a grid spacing of 2Å in (x,y,z) directions and
AC
placing all the aligned molecules inside the grid. The grid region was automatically generated as a 3Dcubic lattice that extends 4A0 beyond the volume for all investigated molecules; a sp3 carbon atom is
used as a probe atom having +1 charge. A distance-dependent dielectric constant of 1.00 was applied
using standard tripos force field and a cutoff 30 kcal/mol was selected to avoid steric and electrostatic
energies domination.
2.7.
CoMSIA Set-Up:
The inherent deficiencies arising in Leonard Jones and Columbic potential used in CoMFA due to its
functional form can be avoided by CoMSIA having five different physicochemical properties (steric,
electrostatic, hydrophobic, hydrogen bond donor and hydrogen bond acceptor). In CoMSIA lattice box
as set for CoMFA was used to calculate the above-mentioned fields at each lattice intersection. A sp3
hybridized carbon atom was used as a probe atom with +1 charge for all the fields (electrostatic, steric,
ACCEPTED MANUSCRIPT
hydrophobicity, hydrogen bond donor and hydrogen bond acceptor) and a default value of 0.3 was set
as an attenuation factor for each molecule [31]. The CoMSIA method helps to interpret correlation in
terms of field contributions.
PLS:
NU
SC
RI
PT
2.6.
Partial least square (PLS) [19] is a statistical technique which describes the relationship between model
generated by CoMFA and CoMSIA fields and the experimental biological activity. The cross validation
was performed through leave one out method (LOO) [32] in which one molecule is removed from the
data set and its biological activity is predicted with respect to the remaining model. The column
filtering was assigned 2.0kcal/mol. The cross validated co-relation coefficient (q2cv) value was
ED
MA
calculated by the
PT
where,Yobs experimental value and Yprep.predicted value and Ymean average value
CE
Finally non-cross-validated correlation was performed using Optimum Number of Components (ONC)
derived from cross validation analysis and their Standard Error of Estimate (SEE). The field type “PLS
AC
stdev * PLS coeff” was used to visualize the CoMFA and CoMSIA results in contour maps. For the
validation of CoMFA and CoMSIA models, pIC50s were determined for the test set molecules.
3.
Results and Discussion:
3.1.
Similarity between 15-LOX and 5-LOX:
Ligand-free 5-LOX crystal structure from human source has already been reported. A number of
ambiguity concerning binding site remains to be addressed, since there is no complex structure of the
5-LOX. Docking of the RS7 inhibitor in the 5-LOX binding site and intensive visual analysis provided
a meaningful solution to characterize the binding site. Based on the sequence alignment shown in
Figure S1, sequence identity and sequence similarity were assessed (31.37 % identical and 44.47 %
ACCEPTED MANUSCRIPT
similar). Results from the sequence alignment are summarized in Table 2 that shows that out of 695
residues 218 residues were identical and 91 residues were similar. It is conceivable from the alignment
data that based on sequence similarity along with extensive comparative 3D-structure visualization of
the two proteins that the said proteins are structurally closely similar as well. The results obtained
3.2.
NU
SC
RI
PT
would also be worthwhile to characterize the binding pattern of the inhibitor.
Molecular Docking Simulation:
Matching coordinates of the 15-LOX and the 5-LOX via alignment led to the provision of complex
structure of the 5-LOX. After superposing the two proteins the RS7 inhibitor of the 15-LOX was
MA
docked in the binding site of 5-LOX. A careful visual inspection was needed to thoroughly compare the
binding pattern of the inhibitor in the active of both proteins. Binding pattern of the two complexes of
the proteins was depicted in Figure 3a and 3b as 2D representations of the RS7 inhibitor in the 15-
ED
LOX and the 5-LOX, respectively. Figure 4a and 4b displays the binding pattern of the inhibitor with
binding site residues of the 15-LOX and the corresponding residues of the 5-LOX, respectively. A good
PT
correlation was found between these two as most of the interactions were found conserved in both
structures.
CE
It was also obvious from 3D-complex structures of the 5-LOX and the 15-LOX bound with the RS7
inhibitor. The long alkyl chain of RS7 is buried in the hydrophobic pocket in both the proteins as
AC
shown in Figure 4a and 4b. Most of the interactions were retained in both complexes like Ala404 of
15-LOX and the corresponding Ala410 of 5-LOX making interactions with the inhibitors were intact.
Similarly Leu362, Leu408 of 15-LOX and the corresponding residues Leu368 and Leu414 of 5-LOX
formed interactions with the inhibitor. Other residues of 15-LOX include His366, His361, Gln548,
Phe353 and Phe415 which make interactions with the inhibitor and the similar interactions were also
seen in 5-LOX with its corresponding residues such as His372, His367, Gln557, Phe359 and Phe421.
Beside this 5-LOX and 15-LOX have some dissimilarity in their active site too, as they are two
different closely related homologues. In 15-LOX Glu357 shows interactions with RS7, the similar type
of interaction is observed between RS7 and Gln363 in 5-LOX. In addition to this, hydrophobic residues
Met419 and Ile418 are present a little far from benzene ring in 15-LOX while polar Asn425 and an
inactive Pro569 are present at the same position in 5-LOX. Additionally, Leu597 is present at the
cavity opening place in 15-LOX while in 5-LOX Phe177 and Tyr181 are present at the cavity opening
place [26]. Table 1S in the supplementary data highlight residues that are same and shows similar
ACCEPTED MANUSCRIPT
interaction with the ligand RS7 by “*” while residues that are not same but show similar interaction by
“#”. Furthermore, “$” sign indicates the residues that are not shown similar interaction but present at
the same position while “@” represent the residues that are totally different in two proteins. The above
discussion suggests that although the complex structure of 5-LOX obtained from the docking with
NU
SC
RI
PT
some dissimilarity is close to the experimentally determined co-crystallized 15-LOX. The results were
also in good agreement with earlier work in which active site of the 5-LOX (pdb-id: 3O8Y) [20] was
described [21].
3.2.1
Validation of Docking Method
To further validate the reliability of our generated model,a co-crystallize inhibitorRS7was re-docked by
using two different docking approaches,MOE [34]and GOLD [30]. A comparative and reliable re-dock
poses were obtained for both the software. Figure S2 (supplementary data) showthe re-dock
MA
conformations obtained from MOE and GOLD. That suggests the reliability of our model and the
binding site of 5-LOX.
Alignment:
ED
3.3.
PT
For the alignment, initially, all the 27 compounds were docked using GOLD [30] software and ten
conformations were generated for each compound. The highest ranked GOLDSCORE conformation of
CE
each ligand were taken to generate CoMFA and CoMSIA models but the results were not so convincing
to proceed further using the approach . The alignment obtained from the docking of 27 was shown in
AC
the Figure S3 (supplementary data). Afterward the models were generated as twenty-six ligands were
aligned on the highest ranked docked conformation of the most potent compound (c.f. Figure 2) i.e.
compound 18. The pose of the compound 18 in the 5-LOX after docking is shown in Figure 5.
Strong hydrogen bond was formed between the carbonyl oxygen of thiazolinone ring and hydrogen of
amino group of Gln363 (O---HNE2, 1.827 A0), and strong π- πinteractions were present between
benzylidene ring of compound 18 and imidazole ring of His367. Methyl of the methoxy group present
at para position of the benzylidene ring took part in hydrophobic interactions with Ala410 and Leu368.
Phenyl ring of compound 18 was engaged in H-bonding with -OH group of Tyr181 (centroid distance
of OH--- π-electrons ofphenyl ring was 3.529 A0) and with -NH group of Asn425 (centroid
distance of NH--- π electrons of phenyl ring is 3.692 A0). These π-hydrogen bonds were
weaker than the classical hydrogen bond[33] whereas para position of phenyl ring of
compound 18 occupied by chlorine group was involved in hydrophobic interaction with Ala424 and
Trp599 at a distance of 2.470 A0 and 3.816 A0, respectively.
ACCEPTED MANUSCRIPT
3.4.
CoMFA Statistics:
NU
SC
RI
PT
A good 3D-QSAR model would have the cross validated correlation coefficient q2cv value of > 0.5 and
r2ncv ≈ 1. As discussed earlier, the whole dataset was aligned on the docked conformation of the most
active compound 18 for the generation of QSAR model. Statistical analyses of the CoMFA model
along with the selected CoMSIA model are listed in Table 3. The q2cv value of 0.549 was obtained with
an optimal number of component 1 having Standard Error of Prediction value of 0.332 whereas the
non-cross-validated PLS analysis was performed with optimal number of component 1 resulted in r2ncv
value of 0.702 with Standard Error of Estimation value of 0.270 and F-value of 44.803. The
MA
contribution of steric and electrostatic fields was 44.3% and 55.7%, respectively. The r2pred value for the
test set to validate the CoMFA model is found out to be 0.787. The statistical values suggesting that the
QSAR model is in good agreement with a standard CoMFA model and can be employed as an optional
CoMSIA Statistics:
PT
3.5.
ED
tool for predicting the IC50 value of unknown 5-LOX inhibitor.
CE
The statistical parameters of all possible combinations of CoMSIA fields were listed in Table 2Sin the
supplementary data. The maximum possible sets generated for two or more than two combinations of
|P(S)| = 2n
AC
the fields were 26 by applying power set theory expressed as:
Where, S is set of n elements.
The selection criterion for the best CoMSIA model was based on the satisfactory q 2cv, r2ncv and r2pred
values. In Figure 6 q2cv and r2ncv values are used as dependent variables while CoMSIA field
combinations were as independent variable. Here, q2cv and r2ncv values for all the 26 model with
different field combinations were simultaneously compared. It is clear from the graph in Figure 6 that
out of 26 models seven models (i.e. model 1, 5, 11, 17, 18, 21 and 23) lie at the high peak for both q2cv
and r2ncv values which suggest that these models are comparatively better than rest of the others. The
electrostatic field is present in all the seven models that signify the pronounced effect of the said field.
The r2pred values were calculated for external set (i.e. test set) of these seven field combination models.
The r2pred values along with q2cv and r2ncv values were listed for the aforesaid model in Table 4. Out of
the seven models, models 1, 18 and 23 did not provide satisfactory r2pred value while models 17 and 21
ACCEPTED MANUSCRIPT
have relatively high statistical parameter values as compared to model 5 and 11. Model 17 comprises
three fields steric, electrostatic and hydrophobic while model 21 comprises of one additional field i.e.
hydrogen bond donor field. Model 26 with all five fields did not give satisfactory q2cv value. Therefore,
model 21 was selected as the best CoMSIA model. The selected Model has a good cross validated
NU
SC
RI
PT
correlation coefficient q2cv value of 0.535 (>0.5) with optimum number of component 5 with Standard
error of prediction 0.380 and a significant r2ncv of 0.951 was obtained with an standard error of
estimation of 0.124 and F-value of 57.814. Steric, electrostatic, hydrophobic and hydrogen bond
contribution fields were 6.0%, 55.7%, 17.4 and 20.9%, respectively. The r2pred value for the test set is
0.571. The statistical parameter values for the selected model of CoMSIA were also listed in Table 2.
Table 5 and Table 6 showed the actual and predicted pIC50 values of training set and test set along with
their residual values, respectively. The graph between actual and predicted values for CoMFA and
Validation Through CoMFA and CoMSIA contour map:
ED
3.6.
MA
CoMSIA models are shown in Figure 7a and 7b, respectively.
The statistical information derived from 3D-QSAR model could be visualized by generating CoMFA
PT
and CoMSIA contour maps in 3D space around the molecules and notify the increase or decrease in
activity due to the change in different fields electrostatic, steric etc around the inhibitors. The CoMFA
CE
and CoMSIA contour maps not only help us to determine the feature important for activity of the
various substitutions on two phenyl rings but it also provides an idea for those positions which was not
3.7.
AC
considered experimentally to enhance the activity.
CoMFA Contour Map:
The CoMFA steric field and electrostatic fields are presented as contour in Figure 8a and 8b,
respectively. The green color indicated favored steric field (bulky group favorable) 75% and yellow
color disfavored steric field (minor group favorable) 25%. Electrostatic field comprises of red and blue
color, red color donates favored electronegative (electron withdrawing favorable) 12% and blue color
donates favored electropositive (electron donating favorable) 88%.
3.7.1. CoMFA Steric Contour:
Here, the most biologically active compound that is compound 18 is selected as a reference compound.
ACCEPTED MANUSCRIPT
A big green isopleth near para position of benzylidene ring around Ala410 indicates the presence of
steric substituent is favorable at that region, part of methoxy group at para position of benzylidene ring
in compound 18 is found within the steric field that enhance the biological activity, likewise, the
ethoxy ester group of compound 20 at the same position also appears in green isopleth region that's
NU
SC
RI
PT
why it has more potency then compound 07 and compound 14 that occupy methoxy and hydrogen.
3.7.2. CoMFA Electrostatic Contour
A big red contour in between meta and para position of phenyl ring indicates the presence of electron
withdrawing group would be favorable at this position. This proves experimentally by considering the
compounds 18, 24, 25, 26 ( Cl, p-C(O)CH3, m-C(O)CH3 and m-F respectively) with compounds 23
MA
contains electron donating group p-NH2 at the said position.
A blue contour is present near the para position of benzylidene ring which suggest that electron
donation would be favorable, this could be clear from compound 7 and compound 14 that contain
ED
-OCH3 and H substituent.
A big red contour present at meta position of benzylidene ring would suggest that electron donating
PT
group would decrease the activity. This corresponds from compound 12 and compound 13 that contain
-OCH3 and -OC2H5 respectively. Since methoxy group is more electron donating than ethoxy group
CoMSIA Contour Map:
AC
3.8.
CE
that's why compound 13 is more biologically active than compoud12.
The best model selected for CoMSIA studies is model 21 which compose of four fields steric,
electrostatic, hydrophobic and hydrogen bond donor. The steric (Green-favored 80% and yellow
disfavored 20%) and electrostatic (blue-favored 75% and red disfavored 25%) contour maps were
nearly same as in CoMFA contour. Since these fields were already explained in CoMFA contour
therefore these were not explained here, only hydrophobic and hydrogen bond donor fields of CoMSIA
are discussed in this section. Figure 9 a-d shows the CoMSIA steric, electrostatic hydrophobic and
hydrogen bond donor fields, respectively.
3.8.1. CoMSIA Hydrophobic Contour Map:
Figure 9c shows the CoMSIA hydrophobic field where magenta contour shows the
ACCEPTED MANUSCRIPT
hydrophobic favorable region (75%) while white contour represents hydrophobic unfavorable region
(25%). The most potent inhibitor is selected as a reference compound that is compound 18.
A white contour around the meta and para positions of benzylidene ring is present which shows
that hydrophobicity is disfavored in this region since a polar His 367 and a non-polar Leu 368 residues
NU
SC
RI
PT
are lying in this region which is also experimentally clear from compound 03 acquire -H at para
position which is less polar therefore more biological active than compound 05 that occupy -OH that is
highly polar and therefore, less active. This could be more signifies as we compare the same
compounds 03 and 05 from meta position. Compound 03 contain -Cl at meta position while compound
05 contain –H at the same position which is less hydrophobic than -Cl due to this compound 03 is more
active than compound 05.
Two magenta isopleths are appeared, one is near the 2nd meta position of benzylidene ring,
MA
therefore hydrophobic group at that position would be favorable and this corresponds from compound
12 and compound 13 which contain -OCH3 and -OC2H5 substituents respectively, compound 13 is
more active because it contains more hydrophobic group and surrounded by hydrophobic residues
ED
Leu368 and Ile415. The other magenta isopleth is located at para position of phenyl residue This could
be explain by comparing the experimental activity of compound 18 that contain -Cl group with
PT
compound 23 that acquires hydrophilic substituent -NH2 group. Since the hydrophobic part of Trp599
CE
and Ala424 is present around this position.
AC
3.8.2. CoMSIA Hydrogen Bond Donor:
Figure 9d show the contour for hydrogen bond donor field. Cyan contour represents hydrogen bond
donor field favourable zone (75%) while orange shows hydrogen bond donor unfavourable zone
(25%). Compound 18 was selected as reference compound.
Two cyan isopleth a little far from para position of phenyl ring suggest the presence of hydrogen bond
donor group might enhance the activity. -NH2 substituent is present at the said position that although
an electron donating group as already discussed while describing CoMFA electrostatic contour map but
this compound is not much inactive might be due to the reason it make hydrogen bond with carbonyl
oxygen of Leu 420 which lie near it.
Two big cyan contour are also appeared in between meta and para position of benzylidene ring
convincing the hydrogen bond donor group availability for the better activity but no compound in the
whole data set was reported that have hydrogen bond donor substituent present at m-position. However,
compound 5, 12 and 13 contain –Cl and compound 6 contain –NO2 group at the said position.
ACCEPTED MANUSCRIPT
Experimentally their biological activities are less and they are less active in the data set that is
according to our contour map and providing a clue that our model is good and can be utilized for future
prediction.
Conclusion
NU
SC
RI
PT
4.
In this study, molecular docking was performed on the 5-LOX by transfering coordinates of the
inhibitor (RS7) co-crystallized with the 15-LOX. The geomerty of the 5-LOX/RS7 complex had
consistency with its closely related homolog and previously reported findings. It was also evaluated
that the most of the contacts between the inhibitor and amino acid residues were conserved in both the
proteins. In this way, characterization of the 5-LOX binding site was performed that was further
MA
utilized to carry out Molecular docking and 3-D QSAR studies on the non-redox type “thiazolinones”
inhibitors showing inhibitory potency against 5-LOX. Molecular docking and two predictive models
CoMFA and CoMSIA illustrate the useful information about the ligand binding with the 5-LOX taking
ED
the most active compound (compound 18) as template. Electrostatic and hydrophobic interactions are
highly effective in this particular case as revealed by the study. The Key amino acids influencing the
PT
receptor ligand interactions are Gln363, Asn425, Tyr181, Ala410, Leu420, Leu368, Ala424 and
Trp599. Both of the aromatic rings present in the inhibitors are held responsible for their activity
CE
against 5-LOX and in the absence of these rings the activity may be lost as suggested by the contour as
well as proved by the previousexperimentalstudies. The contour also suggested that presence of
AC
electronegative group at the para and meta position of phenyl ring would enhance the activity similarly
meta positions of benzylidene ring also require more electronegative group to increase the inhibitory
potency. The predictive ability of our model is very close to experimental data which also gives a clue
that our results could be reliable. As suggested by this study, a correlation was drawn between
physicochemical properties and biological activities of the 5-LOX inhibitors. This study can be
beneficial for the future designing of suitable selective and more potent inhibitors as well as to
understand the binding pattern of the inhibitors affecting the 5-LOX activity.
Acknowledgment
Authors gratefully acknowledge the Higher Education Commission (HEC) Pakistan for financial
assistance under ‘‘National Research Program for Universities’’ Project No. 20-1329. Special thanks to
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA
NU
SC
RI
PT
Prof. Bernd M. Rode (University of Innsbruck) for his technical support.
References
1.
2.
3.
4.
5.
6.
Jakschik, B.A. and L.H. Lee, Enzymatic assembly of slow reacting substance. Nature, 1980.
287: p. 51-52.
Samuelsson, B., et al., Leukotrienes: a new group of biologically active compounds. Advances
in Prostaglandin Thromboxane Research, 1980. 6: p. 1-18.
Andreou, A. and I. Feussner, Lipoxygenase-structure and reaction mechanism. Phytochemistry,
2009. 70(13): p. 1504-1510.
Boyington, J.C., B.J. Gaffney, and L.M. Amzel, The three-dimensional structure of an
arachidonic acid 15-lipoxygenase. Science, 1993. 260(5113): p. 1482-1486.
VanderMeer, T.J., et al., Acute lung injury in endotoxemic pigs: role of leukotriene B4. Journal
of Applied Physiology, 1995. 78(3): p. 1121-1131.
Hashimoto, A., et al., Differential expression of leukotriene B4 receptor subtypes (BLT1 and
ACCEPTED MANUSCRIPT
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
NU
SC
RI
PT
13.
MA
12.
ED
11.
PT
9.
10.
CE
8.
AC
7.
BLT2) in human synovial tissues and synovial fluid leukocytes of patients with rheumatoid
arthritis. The Journal of Rheumatology, 2003. 30(8): p. 1712-1718.
Julemont, F., et al., New trends in dual 5-LOX/COX inhibition. Current Medicinal Chemistry,
2002. 9(9): p. 941-962.
Mehrabian, M. and H. Allayee, 5-lipoxygenase and atherosclerosis. Current Opinion in
Lipidology, 2003. 14(5): p. 447-457.
Wang, D. and R.N. DuBois, Eicosanoids and cancer. Nature, 2010. 10(3): p. 181-193.
Chini, M.G., et al., Design and synthesis of a second series of triazole-based compounds as
potent dual mPGES-1 and 5-lipoxygenase inhibitors. European Journal of Medicinal Chemistry,
2012. 54: p. 311-323.
Eleftheriou, P., et al., Fragment-based design, docking, synthesis, biological evaluation and
structure activity relationships of 2-benzo/benzisothiazolimino-5-aryliden-4-thiazolidinones as
cycloxygenase/lipoxygenase inhibitors. Europeon Journal of Medicinal Chemistry, 2012. 47: p.
111-124.
Hieke, M., et al., SAR-study on a new class of imidazo [1, 2-a] pyridine-based inhibitors of 5lipoxygenase. Bioorganic Medicinal Chemistry Letter, 2012. 22(5): p. 1969-1975.
Evans, J.F., et al., What's all the FLAP about?: 5-lipoxygenase-activating protein inhibitors for
inflammatory diseases. Trends in Pharmacological Sciences, 2008. 29(2): p. 72-78.
Werz, O. and D. Steinhilber, Development of 5-lipoxygenase inhibitors-lessons from cellular
enzyme regulation. Biochemical Pharmacology, 2005. 70(3): p. 327-333.
Israel, E., et al., Effect of treatment with zileuton, a 5-lipoxygenase inhibitor, in patients with
asthma. The Journal of the American Medical Association, 1996. 275(12): p. 931-936.
Verma, J., V.M. Khedkar, and E.C. Coutinho, 3D-QSAR in drug design-a review. Current Topics
in Medicinal Chemistry, 2010. 10(1): p. 95-115.
Cramer, R.D., D.E. Patterson, and J.D. Bunce, Comparative molecular field analysis (CoMFA).
1. Effect of shape on binding of steroids to carrier proteins. Journal of American Chemical
Society, 1988. 110(18): p. 5959-5967.
Klebe, G. and U. Abraham, Comparative Molecular Similarity Index Analysis (CoMSIA) to
study hydrogen-bonding properties and to score combinatorial libraries. Journal of ComputerAided Molecular Design, 1999. 13(1): p. 1-10.
Ståhle, L. and S. Wold, Partial least squares analysis with cross-validation for the two-class
problem: A Monte Carlo study. Journal of Chemometrics, 1987. 1(3): p. 185-196.
Gilbert, N.C., et al., The structure of human 5-lipoxygenase. Science, 2011. 331(6014): p. 217219.
Reddy, N.P., et al., Structure based drug design, synthesis and evaluation of 4-(benzyloxy)-1phenylbut-2-yn-1-ol derivatives as 5-lipoxygenase inhibitors. Europeon Journal of Medicinal
Chemistry, 2012. 47: p. 351-359.
Hofmann, B., et al., A class of 5-benzylidene-2-phenylthiazolinones with high potency as direct
5-lipoxygenase inhibitors. Journal of Medicinal Chemistry, 2011. 54(6): p. 1943-1947.
Tsai, K.C., et al., A comparison of different electrostatic potentials on prediction accuracy in
CoMFA and CoMSIA studies. Journal of Medicinal Chemistry, 2010. 45(4): p. 1544-1551.
Clark, M., R.D. Cramer, and N. Van Opdenbosch, Validation of the general purpose Tripos 5.2
force field. Journal of Computational Chemistry, 2004. 10(8): p. 982-1012.
SYBYL Molecular Modeling Software, Tripos Associated Ltd., St. Louis, MO.
Charlier, C., et al., Structural insights into human 5-lipoxygenase inhibition: combined ligandbased and target-based approach. Journal of Medicinal Chemistry, 2006. 49(1): p. 186-195.
Choi, J., et al., Conformational flexibility in mammalian 15S-lipoxygenase: Reinterpretation of
the crystallographic data. Proteins, 2008. 70(3): p. 1023-1032.
Li, K.B., ClustalW-MPI: ClustalW analysis using distributed and parallel computing.
ACCEPTED MANUSCRIPT
NU
SC
RI
PT
34.
MA
33.
ED
32.
PT
31.
CE
30.
AC
29.
Bioinformatics., 2003. 19(12): p. 1585-1586.
Fletcher, R. and M.J.D. Powell, A rapidly convergent descent method for minimization. The
Computer Journal, 1963. 6(2): p. 163-168.
Jones, G., P. Willett, and R.C. Glen, Molecular recognition of receptor sites using a genetic
algorithm with a description of desolvation. Journal of Molecular Biology, 1995. 245(1): p. 4353.
Viswanadhan, V.N., et al., Atomic physicochemical parameters for three dimensional structure
directed quantitative structure-activity relationships. 4. Additional parameters for hydrophobic
and dispersive interactions and their application for an automated superposition of certain
naturally occurring nucleoside antibiotics. Journal of Chemical Information and Modeling,
1989. 29(3): p. 163-172.
Tuda, K., et al., Learning to predict the leave-one-out error of kernel based classifiers. Lecture
Notes in Computer Science, 2001. 2130: p. 331-338.
Meyer, E.A., R.K. Castellano, and F. Diederich, Interactions with aromatic rings in chemical
and biological recognition. Angewandte Chemie International Edition, 2003. 42(11): p. 12101250.
Molecular Operating Environment (MOE) , 2012.10 Chemical Computing Group
Inc.1010Sherbooke St. West, Suite #910, Montreal, QC, Canada, H3A 2R7. 2012.
ACCEPTED MANUSCRIPT
Figure Captions:
Fig 1: Superposition of 15-LOX (gray) and 5-LOX (blue), using 15-LOX as reference protein, cocrystal ligand RS7 (black) of 15-LOX and metal ion (brown).
NU
SC
RI
PT
Fig 2: (a) Structure of compound 18 (most active compound), atoms selected for the data base
alignment are highlighted by asteric. (b) Alignment diagram of whole dataset containing 27 inhibitors.
O is shown as red, S in yellow, N in blue, C in silver, X in green, H are removed for the sake of clarity
(for color interpretation see online version).
Fig 3: 2D-Ligand protein interaction diagram. (a) Binding site of interaction of ligand RS7 and 15LOX (b) Binding site interaction of Ligand RS7 and 5-LOX after transferring the co-ordinates of RS7
from 15-LOX to 5-LOX.
MA
Fig 4: 3D-Ligand protein interaction diagram, where brown shows Metal Ion and black represents
ligand RS7. (a) Binding site of interaction of ligand RS7 and 15-LOX (b) Binding site interaction of
Ligand RS7 and 5-LOX after transferring the co-ordinates of RS7 from 15-LOX to 5-LOX.
PT
ED
Fig 5: Binding site interaction of compound 18 and 5-LOX. (a) 2D interaction of compound 18 and 5LOX, green arrow shows strong hydrogen bonding is formed between carbonyl oxygen of thiazolinone
ring and Gln 363, this hydrogen bonding can also be seen in 3D-surface diagram. (b) 3D-Surface
diagram of 5-LOX and compound 18, where green region shows X, blue shows N, red shows O, sand
color shows C and white shows H while C of ligand represented as black, N in blue, S in yellow, O in
red and X in green.
CE
Fig 6: Graphical comparison for all possible CoMSIA field Combination Models. Here, x-axis
comprises CoMSIA model number associated with field combination while y-axis represents statistical
parameters (q2cv and r2ncv values).
AC
Fig 7: The correlation plots of predicted versus actual pIC50 values based on (a) CoMFA and (b)
CoMSIA. Here blue square boxes representing training and red square boxes corresponding test set.
Fig 8: CoMFA STDEV*Coeff contour maps of the most biological active compound (compound 18 in
ball and stick model), some important amino acid also shown in silver color with ball and stick model.
(a) CoMFA steric contour: green isopleth corresponds to bulky group favourable (75% contribution)
and yellow minor group favourable (25% contribution). (b) CoMFA electrostatic contour: blue isopleth
correlated with increase binding afinity for electron withdrawing group (88% contribution) and red
correlated with electron donating group (12% contribution).
Fig 9: CoMSIA STDEV*Coeff contour maps of the most biological active compound (compound 18 in
ball and stick model), some important amino acid also shown in silver color with ball and stick model.
(a) CoMSIA steric contour. (b) CoMFA electrostatic contour. (c) CoMSIA hydrophobic contour:
magenta polyhedron shows the hydrophobic favoured region (75% contribution) while white contour
represents hydrophobic disfavoured region (25% contirbution). (d) CoMSIA hydrogen contour: cyan
color represents hydrogen bond donor favoured region (75% contribution) and orange color shows
hydrogen bond donor disfavoured region (25% contribution).
AC
CE
PT
ED
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
Figure 1
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA
Figure 2
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA
Figure 3
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA
Figure 4
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA
Figure 5
ED
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 6
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA
Figure 7
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA
Figure 8
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
Figure 9
ACCEPTED MANUSCRIPT
Table1: Molecular structures of thiazolinone analogues of 5-LOX inhibitors with actual inhibitory values in
IC50 and pIC50.
R1
S
NU
SC
RI
PT
R2
N
O
EtO
Comp_02
IC50 p-IC50
(µM)
(M)
R2
H
0.50
6.3010
p-CH3
0.30
6.5229
H
0.3
6.5229
CE
PT
Comp_01
AC
ED
R1
MA
Structure
Compounds
Cl
H3CO
Comp_03
Cl
OH
ACCEPTED MANUSCRIPT
H3CO
OH
H3CO
Comp_05
HO
6.2676
H
3.00
5.5229
H
3.00
5.5229
p-CH3
0.30
6.5229
p-CH3
0.13
6.8861
MA
Cl
ED
H3CO
HO
PT
Comp_06
0.54
H
NU
SC
RI
PT
Comp_04
Comp_07
AC
CE
O2N
H3CO
H3CO
Comp_08
OCH3
ACCEPTED MANUSCRIPT
H3CO
OCH3
6.3979
p-CH3
0.98
6.0088
p-CH3
1.30
5.8861
p-CH3
2.70
5.5686
p-CH3
1.25
5.9031
p-CH3
H3CO
OCH3
Comp_10
MA
H3CO
OH
ED
H3CO
CE
PT
Comp_11
NU
SC
RI
PT
Comp_09
0.40
Comp_12
AC
H3CO
HO
Cl
EtO
Comp_13
HO
Cl
ACCEPTED MANUSCRIPT
MA
Comp_15
6.4559
p-CH3
0.23
6.6383
t-Bu
H
0.3
6.5229
0.3
6.5229
PT
ED
Comp_16
0.35
p-CH3
NU
SC
RI
PT
Comp_14
Comp_18
Comp_19
Comp_20
H
AC
CE
Comp_17
H3CO
p-Cl
0.09
7.0458
H3CO
H
0.15
6.8239
0.19
6.7212
H3CO(O)H2CCO
p-CH3
ACCEPTED MANUSCRIPT
H3CO
NU
SC
RI
PT
p-OCH2C(O)OCH3
Comp_21
Comp_22
H3CO
Comp_23
H3CO
Comp_24
H3CO
Comp_25
0.58
6.2366
0.65
6.1871
p-NH2
0.63
6.2007
p-C(O)CH3
0.11
6.9586
H3CO
m-C(O)CH3
0.13
6.8861
Comp_26
H3CO
m-F
0.12
6.9208
Comp_27
HO
p-Cl
0.11
6.9586
AC
CE
PT
ED
MA
p-OH
ACCEPTED MANUSCRIPT
NU
SC
RI
PT
Table 2: ClustalW allignment Result for 3O8Y_B|PDBID|CHAIN|S vs 2P0M_B|PDBID|CHAIN|S.
695
Alignment Length
218
Identical residues
Similar residues
91
31.37%
Percent Identity
44.46%
AC
CE
PT
ED
MA
Percent Similarity
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA
Graphical abstract