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