- Email: [email protected]
Molecular modeling studies of atorvastatin analogues as HMGR inhibitors using 3D-QSAR, molecular docking and molecular dynamics simulations
Accepted Manuscript Molecular modeling studies of atorvastatin analogues as HMGR inhibitors using 3D-QSAR, molecular docking and molecular dynamics simulations Zhi Wang, Liping Cheng, Zhenpeng Kai, Fanhong Wu, Zhuoyu Liu, Minfeng Cai PII: DOI: Reference:
S0960-894X(14)00683-0 http://dx.doi.org/10.1016/j.bmcl.2014.06.055 BMCL 21774
To appear in:
Bioorganic & Medicinal Chemistry Letters
Received Date: Revised Date: Accepted Date:
10 April 2014 17 June 2014 19 June 2014
Please cite this article as: Wang, Z., Cheng, L., Kai, Z., Wu, F., Liu, Z., Cai, M., Molecular modeling studies of atorvastatin analogues as HMGR inhibitors using 3D-QSAR, molecular docking and molecular dynamics simulations, Bioorganic & Medicinal Chemistry Letters (2014), doi: http://dx.doi.org/10.1016/j.bmcl.2014.06.055
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.
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.
(A) MD conformation derived for
compound 20 with HMG the binding site of HMGR (PDB entry code: 2Q1L). (B) The “scorpion” conformation of compound 20 at the binding sit of MHGR. Molecular modeling studies of atorvastatin analogues as HMGR inhibitors using 3D-QSAR, molecular docking and molecular dynamics simulations Zhi Wanga,§, Li Ping Chenga,§,*, Zhen Peng Kaia, Fan Hong Wua,b,*, Zhuo Yu Liua and Min Feng Caia School of Chemical and Environmental Engineering, Shanghai Institute of Technology, Shanghai 201418, China b Key Laboratory for Advanced Materials and the Institute of Fine Chemicals, School of Chemistry and Molecular Engineering, East China University of Science and Technology, Shanghai 200237, China
a
§
These two authors (Zhi Wang and Liping Cheng) contributed equally to this work and should be considered as co-first authors.
1
Bioorganic & Medicinal Chemistry Letters j o ur n al h om e p a g e : w w w . e l s e v i er . c o m
Molecular modeling studies of atorvastatin analogues as HMGR inhibitors using 3DQSAR, molecular docking and molecular dynamics simulations Zhi Wanga,§ , Liping Chenga,§,∗ , Zhenpeng Kaia , Fanhong Wua,b,*, Zhuoyu Liu a and Minfeng Caia a
School of Chemical and Environmental Engineering,Shanghai Institute of Technology, Shanghai 201418, China Key Laboratory for Advanced Materials and the Institute of Fine Chemical, School of Chemistry and Molecular Engineering, East China University of Science and Technology, Shanghai 200237, China b
A R T IC LE IN F O
A B S TR A C T
Article history: Received Revised Accepted Available online
3-hydroxy-3-methylglutaryl coenzyme-A reductase (HMGR) is generally regarded as targets for the treatment of hypercholesterolemia. HMGR inhibitors (more commonly known as statins) are discovered as plasma cholesterol lowering molecules. In this work, 120 atorvastatin analogues were studied using a combination of molecular modeling techniques including three-dimensional quantitative structure-activity relationship (3D-QSAR), molecular docking and molecular dynamics (MD) simulation. The results show that the best CoMFA (comparative molecular field analysis) model has q2=0.558 and r2= 0.977, and the best CoMSIA (comparative molecular similarity indices analysis) model has q2=0.582 and r2=0.919. Molecular docking and MD simulation explored the binding relationship of the ligand and the receptor protein. The calculation results indicated that the hydrophobic and electrostatic fields play key roles in QSAR model. After MD simulation, we found four vital residues (Lys735, Arg590, Asp690 and Asn686) and three hydrophobic regions in HMGR binding site. The calculation results show that atorvastatin analogues obtained by introduction of F atoms or gem-difluoro groups could obviously improve the inhibitory activity. The new HMGR inhibitor analogues design in this paper had been submitted which is being currently synthesized by our laboratories.
Keywords: HMGR statins 3D-QSAR gem-difluoro molecular dynamics simulations
2014 Elsevier Ltd. All rights reserved.
∗ Corresponding author. Tel.: +86 21 60873250; fax: +86 21 60 877231; E-mail: [email protected] (L.P. Cheng) §
These two authors (Zhi Wang and Liping Cheng) contributed equally to this work and should be considered as co-first authors.
1
The diseases caused by atherosclerosis (e.g., Coronary heart disease (CHD) and Cerebral infarction) are one of the major health problems in the contemporary world. The high cholesterol level is the primary reason for causing coronary artery disease.1, 2 HMGR is the rate-controlling enzyme which catalyzes the conversion of HMG-CoA to mevalonate. This reaction is a fourelectron oxidoreduction and it is the committed step to endogenous cholesterol biosynthesis and isoprenoids.3 The HMGR is a tight tetramers protein.4 It has two isoforms, isoform1 and isoform2, which share high sequence homology (87% identity).5 The isoform2 plays an important role in mevalonate path way along with isoform1. HMGR inhibitors (statins) have been considered to be the best drug to cure the CHD and other diseases. Over the past decades, a number of HMGR inhibitors had been synthesized and listed, including Lovastatin, Pravastatin, Fluvastatin, Atorvastatin, Rosuvastatin and Pitavastatin.6 The atorvastatin is the most efficient inhibitor against both the two isoforms of HMGR.5 Novel statins should be not only with the ability of reducing the risk of cardiovascular disease and the LDL cholesterol level, but also is safe to health. Furthermore, the latter is the most important. Although statins are considered to be generally safe since the first statin (lovastatin) listed, skeletal muscle-related toxicity remains an associated adverse effect.7-9 Therefore, it is necessary to develop more and new efficient and safe HMGR inhibitors.
Figure 1 the structure of 9f
In recent years, many papers have reported some progresses in HMGR or its inhibitors in theoretical calculations. For example, da Silva et al.10 used virtual screening, flexible docking and molecular interaction fields to study and design novel HMGR inhibitors. Thilagavathi et al. 11 had developed a 3D-QSAR model for HMGR inhibitors based on a training set of 29 compounds by COMSIA method.12 The built model can be extrapolated to predict novel and more potent molecules. Besides the 29 compounds used by Thilagavathi, Zhang et al.13 added another 17 compounds from the MDDR database to build a new 3D-QSAR model. In their paper, different chemometric tools had been used including comparative molecular field analysis (CoMFA),14 molecular docking and pharmacophore model. However, on one hand, the total number of samples in their data set is less than 50. On the other hand, they did not explore the interactions between the inhibitors and receptor proteins. To obtain more understanding to the structural and chemical properties required for HMGR inhibitory activity, more 3D-QSAR and other computational explorations analyses for comparatively large number of newly synthesized and tested HMGR inhibitors are necessary.In the present work, 120 atorvastatin analogues reported by the same research group15-20 as potent and selective HMGR inhibitors were collected as dataset. A series of studies were performed using some computational methods include 3DQSAR tools like CoMFA and CoMSIA, molecular docking and MD simulation. The results of CoMFA and CoMSIA studies will not only explain the conformation or spatial orientation of atorvastatin analogues but also provide useful information for the design of potent and selective HMGR inhibitor. The introduction of F atom often improves the biological activity of target
compounds21 and it is a common and effective method to structural modifications. 22 Fluorinated compounds play an increasing important role in medicinal chemistry. Our laboratories23-25 had successfully synthesized some compounds by introducing the F atom or fluoride group into the proper position. For example, Wang26 had designed and synthesized a series of gem-difluoro derivatives of statins, in which the compound 9f (Fig. 1) has a good inhibitory activity to HMGR. These results encouraged us to design and synthesize more high inhibitory activity containing F derivatives. Moreover, the docking and MD analysis, performed using HMGR crystal structures, might expound the probable binding modes of theses inhibitors at the HMG bonding site. On the basis of molecular field information of 3D-QSAR tools, molecular docking and MD simulations, a few proposals were offered to design novel molecules with improved activity. In this study, we used the ligand-based alignment rule. The alignment result is shown in Fig. 2. According to the partial least squares (PLS) method,27,28 some statistical parameters were used to evaluate and analysis the stand or fall of these models, including q2, r2, standard error of estimate (SEE) and F-statistic values. Table 1 lists the statistical parameter results of the CoMFA and CoMSIA analyses. From Table 1, the optimal CoMFA and CoMSIA models show good predictive power with the cross-validated q2 of 0.558 and 0.582. The correlation coefficient r2, SEE and F-statistic values are 0.977, 0.087, and 354.857 for CoMFA model and 0.919, 0.161, and 121.009 for CoMSIA model. For the CoMFA model, the contributions of the steric and electrostatic fields are 81.8 and 18.2%. The main contribution in the model was steric field. For the optimal CoMSIA model, the steric, electrostatic, hydrophobic, H-Bond donor and acceptor field contribution was 16.3, 12.7, 27.7, 24.0, and 19.3%, respectively. The H-bond donor and hydrophobic fields were found to be the important contributions in the optimal CoMSIA model.
Figure 2 Alignment of the training set and compound 20 used as a template for alignment, with the common substructure shown in bold, and some substituents including the important R2 R3, R4 and Y substituents shown in dashed red border. Molecules are colored in white for common C, blue for N, red for O, yellow for S, cyan for H, green for F, respectively. Table 1 Statistical results for the CoMFA and CoMSIA models CoMFA q2 0.558 noc 10 2 r 0.977 SEE 0.087 F 354.857 fraction Steric 0.818 Electrostatic 0.182 H-acceptor H-donor Hydrophobic -
CoMSIA 0.582 8 0.919 0.161 121.009 0.163 0.127 0.193 0.240 0.277
Before the validation of the 3D-QSAR models, an initial inspection of the inhibitor molecules is very necessary. The
2
residual values of compounds between the predicted and experimented pIC50 exceeded one logarithm unit29were considered as outliers. According to the above rules, outliers of the optimal CoMFA and CoMSIA models are compound 05, 08 and 96. These outliers are different from the template compound 20 at some substituent position. For example, compared with compound 20, compound 05 is only with the Bn at position R3, but compound 20 with the –CN at Bn substituent, the –CN substituent is very important to inhibitory activity. 30 In compound 09, the R3 and R4 substituent is a ring, which is different from the compound 20. Similar to compound 08, compound 96 has not any substituent at R3 position. The calculation results based on the optimal CoMFA and CoMSIA models are shown in Supplementary data (Table A2). The linear fit between predicted and actual activity of the optimal CoMFA and CoMSIA model is described in Fig. 3. According to the optimal CoMFA and CoMSIA models, the StDev*Coeff contour maps were built. To aid in simplification, compound 20 is used to illustrate all contour maps of the optimal CoMFA and CoMSIA models. From Figs. 4-6, we can insight into the five fields and explain the key structural features required for inhibitory activity.
Figure 3 Plot of experimental activity [log1/IC50] against activity predicted by the best CoMFA (A) and CoMSIA (B) models
Figs. 4 and 5 show the steric and electrostatic contour maps of CoMFA and CoMSIA. From Fig. 4, the red (electronegative groups favored) and blue (electropositive groups) contours indicate the default 20% and 80% level contributions, respectively. Fig. 4A depicts the CoMFA electrostatic contours. A big and medium-sized red contour near the X substituent (pharmacophore) indicates electronegative groups would be favorable. A medium-sized and small red contour is also near the R1 position. Some large blue contours around the ring A (R2R3 substituent) show electropositive groups are favored for inhibitory activity. This is in agreement with the experimental data. For example, the order of some compounds inhibitory activity is: 30 (H) >32 (Me) >33 (Et). Fig. 4B depicts the CoMSIA electrostatic contour maps. Compared with the CoMFA contour map, only one big red contour is near the –COO- of the pharmacophore. There is also a big blue contour near the –COO-. Another big blue contour around the R2R3 substituent position shows the electropositive groups are favor to bioactivity.
Figure 4 Electrostatic contours of the CoMFA (A) and CoMSIA (B) models. The red color shows the favored negative electrostatic area and the blue color shows the favored positive electrostatic area
Figure 5 Steric contours of the CoMFA (A) and CoMSIA (B) models. The green color shows the favored steric area and the yellow color show steric area
From Fig. 5, the green (sterically favorable) and yellow (sterically unfavorable) contours also represent default level contributions. Fig. 5A shows the CoMFA steric contour maps of the best model with compound 20 surround. The contour maps mainly focus on the part of hydrophobic substituents. The green contour region around ring A (R2R3 substituent), B (R4 substituent) and beside the R1 substituent shows that the steric field give a positive effect to improve bioactivity. Some high bioactivity compounds such as 08, 13, 16 and 20 have large functional groups at the R2R3 region around the big green contour. Nevertheless, due to only the small substituent at the R2R3 region, the inhibitory capacity of some compounds (32, 33) is lower than that of compound 20 by 2 logarithm units. In the database set, some compounds are absent from ring B, such as 35-73 and 93111, the inhibitory activity is lower than that of the most potent inhibitor. Every green contour has a yellow contour region at hand, which indicates that the substituent of the above three positions (ring A and B, R1 substituent) with too large group would lead to a decrease of bioactivity, such as compounds 20 and 21. In short, the appropriate bulky or large groups are benefit to inhibitory activity. The contour maps of CoMSIA are described in Fig. 5B. As the Fig. 5B described, a large green (sterically favorable) contour map is in the center of the ring A, and a bulky yellow (sterically unfavorable) contour map surround the R2R3 region of compound 20. Since the CoMSIA steric contour map is similar to that of the CoMFA model, and the analysis of the CoMSIA steric contour map is neglected herein. The hydrophobic contour map of CoMSIA is shown in Fig. 6A. From the Fig, one large white (hydrophilic favorable) contour map is around the ring pyrrole, suggesting that introduction of hydrophilic substituents into the ring pyrrole will be benefit for inhibitory activity. Besides the X substituent (pharmacophore), other positions were replaced by the hydrophobic groups. One large yellow (hydrophobic favorable) contour map crossed the R2R3 substituent, indicating this position side with hydrophilic groups. The –NH2 is the hydrophilic group, compound 30 has this substituent with high bioactivity, but compound 31-35 own hydrophobic groups at the same position with low bioactivity. Compound 20 is used as template molecule to analysis the influence of H-bond donor and acceptor fields on inhibitory activity. As shown in Fig. 6B, a big cyan (H-bond donor favorable) contour map above the ring B suggests that the electropositive H-Bond donors in these regions are favorable for the inhibitory activity. A big purple contour map was observed farther from compound 20, indicating the H-bond acceptor has little effect on inhibitory activity. A big red contour map appears near the 5-OH and 7-COO- of the X substituent implies that Hbond acceptor groups could improve the inhibitory activity. A big magenta contour map near the R1 substituent indicates the Hbond acceptor groups could have a positive influence on the inhibitory activity, and a medium sized red contour map is near the 4-position of ring A.
3
Figure 6 Hydrophobic (A) and Hydrogen bonding (B) contours of the CoMSIA models. The yellow color shows the favored hydrophobic area, the white color shows the disfavored hydrophobic area, the magenta color shows the favored H-acceptor area, the red color shows the disfavored H-acceptor area, the cyan color shows the favored H-donor area, and the purple color represents the disfavored H-donor area
Based on the above theoretical calculation results and the known synthesized gem-difluoro derivatives of statins. We designed and predicted some HMGR inhibitor analogues by introduction of F atoms into proper position of pharmacophore. Compound 20, with the highest inhibitory activity was taken as a template to design new compounds. The predicted values are shown in Table 2. Table A3 shows that the pIC50 values of all designed new analogues are higher than that of 20, indicating their high inhibitory activity. These new designed compounds are being currently synthesized by our laboratories. The corresponding results will be published in further publications. The pIC50 value of the above mentioned 9f was also predicted by our 3D-QSAR model. The predicted pIC50 value is 8.24 (CoMFA) and 8.567 (CoMSIA), which is close to its experimental value (8.051).26 Therefore, the theoretical results are consistent with our experimental results and the model built in the present study is reliable.
Figure 7 Docking result of the compounds in the study. The binding site (2Q1L) surrounding the compound 20. Hydrogen bonds are shown as yellow dashed lines, with distance unit of Å. The inhibitor is also shown as purple stick model. Molecular surface is shown as cyan.
20_2
9.894
9.434
20_3
9.766
9.420
20_4
9.763
9.414
20_5
9.764
9.414
20_6
9.746
9.431
20_7
9.766
9.422
9.767
9.434
8.240
8.567
20_8
OH OH COF
9f
Docking protocols are widely used to explore the binding affinities of a series of ligands.31 In this paper, to study the binding environment in which the inhibitor interacts within the HMGR, molecular docking studies were performed on all 120 compounds. The docking results show that the highest total-score of compound 22 was 16.56, the total-scores of docking studies are shown in Supplementary data (Table A2), from the 120 listed total-scores, the pIC50 of most inhibitors keep in touch with the docking total-scores. Fig.7 describes the dock result of compound 20 and HMGR crystal structure. As shown in Fig.7, it is easily seen that some key amino acids (Arg568, Ser565, Lys 735, Ser684, Lys692, Asp690 and Lys691) interact with the HMGR inhibitors by hydrogen bonds at the binding site. The hydrogen bond distances observed are 2.52 Å (Arg568-HN-H…NC-), 2.01 Å (Ser565-OH…O-), 2.29 Å (Lys735-HN-H… O-), 1.89 Å (Ser684-OH…O-), 2.70 Å (Lys692-HN-H… O-), 1.93 Å (Asp690O…HO-), 1.86 Å (Lys691-HN-H… O-), respectively.
Table 2 Designed molecules and predicted inhibit activity values of HMGR No. R
20 20_1
Skeleton type Predicted activity value CoMFA CoMSIA 9.765 9.420 9.891
9.518
4
Figure 8 (A) Plot of the Total-energy docked complex versus the MD simulation time in the MD-simulated structures.(B) View of superimposed backbone atoms of the lowest energy structure of the MD simulation (green) and the initial structure (pink) for compound 20-2Q1L complex. Compound 20 is represented as carbon-chain in pink for the initial complex and carbonchain in green for the lowest energy complex.
Figure 9 MD conformation derived for compound 20 with HMG the binding site of HMGR (PDB entry: 2Q1L). Hydrogen bonds are shown as yellow dashed lines, with distance unit of Å. The inhibitor is also shown as purple stick model. Molecular surface is shown as cyan.
MD simulation is used to explore the interaction of the receptor protein and ligand. As the highest bioactivity inhibitor, compound 20 was used as the template molecule to explain the MD simulations. In our study, a 5 ns simulation of the ligandreceptor complex was run to energy balance at 2 ns to obtain a stable conformation. The total-energy of ligand-receptor compound ranging from 10200 to 9750 KJ/mol is shown in Fig. 8A. After 2 ns, the total-energy of the complex decreases to 9750 KJ/mol and maintains similar energy value at last 3 ns simulation. The overlay of the stable structure in the whole 5 ns MD simulations and the original docked structure is depicted in Fig. 8B. From the Fig. 8B, the pink and green ribbon represents the original docked structure and the lowest energy structure, respectively. The carbon-chain of compound 20 from two conformation structures also has the same color with the ribbon of the protein. Last but not least, the original docked structure and the overlay of the stable structure are docked into the same binding site, and their pharmacophore structure and other structures are basically similar. After 5 ns MD simulation studies, we found some residues including the Cys561, Leu857, Met657, Arg590 et al. are still essential to the interaction between inhibitor and HMGR. However, compared with the initial molecular docking results, the number of amino acid residues and hydrogen bonds at the binding site decrease. As shown in Fig.9, it has only four amino acid residues, Lys735, Asp690, Asn686 and Arg590 to
form 5 hydrogen bonds with the compound 20. In addition, from the MD results shown in Fig. 8B, the amino acid residues of HMGR binding pocket can form not only hydrogen bonds but also the electrostatic and van der Waals force. Met, Leu and Cys are the hydrophobic amino acids,32-34 they can form powerful interaction with the ring A, B and C. Here the rings A, B, and C are hydrophobic substituents and they may be regarded as three hydrophobic regions. Therefore, we predict that the interaction between the hydrophobic amino acids and the three hydrophobic regions perhaps play key roles in the inhibitory activity. Based on the results of QSAR, docking and MD simulation, we suppose that the inhibitor binds to the HMGR with a “scorpion” conformation depicted in Fig. 10B, where the pyrrole ring form the main brain, the three hydrophobic regions act as the two big pincers and mouth, and the X substituent is the key tail of the scorpion, respectively. Combined with the Fig 10A, firstly, three corners of the main brain of “scorpion” tightly interact with three hydrophobic regions of the ligand with the residues of Leu857, Cys561 and Met657, respectively. The hydrophobic groups are wonderful to inhibitors that form hydrophobic interaction with protein receptor, which correlating well with the previous QSAR and docking studies. The tail of the “scorpion” is the pharmacophore structure interaction with the HMGR by Lys735, Asp690, Asn686 and Arg590, respectively.
Figure 10 (A) MD conformation derived for compound 20 with HMG the binding site of HMGR (PDB entry code: 2Q1L). (B) The “scorpion” conformation of compound 20 at the binding sit of MHGR.
In this study, a series of computer-aided drug design processes, such as 3D-QSAR studies, molecular docking, and MD simulations, were used to explore the feasibility of introduction of F atoms into atorvastatin analogue as HMGR inhibitors. The built 3D-QSAR models show good internal and external predictive ability and can be extrapolated to predict novel and more potent molecules. The contour maps obtained from the CoMFA and CoMSIA analysis could guide to design new chemical entities with high HMGR inhibitory activity. To study the binding modes of inhibitors at the active site of HMGR protein, molecular docking studies of representative compounds was performed. Some key residues such as Lys735, Arg590, Asp690 and Asn686 as well as three hydrophobic regions in HMGR binding site were found. To further confirm the reliability of docking results, MD simulations were carried out for representative compounds. Based on the built QSAR models, several F-containg novel HMGR inhibitors were designed. The results show that the atorvastatin analogues obtained by introduction of F atoms or fluoride substituent groups could improve the inhibitory activity. Acknowledgments Thanks for financial support given by the National Natural Science Foundation of China (No. 21172148), the Scientific Research Innovation Project of Shanghai (No. 14YZ148) and
5
“College Students' (PE2014044).
Innovative
Activities”
of
Shanghai
References and notes 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.
16.
17.
18.
19.
Grundy, S. M. J. Am. Med. Assoc.1986, 256, 2849 Thelle, D.S. Drug Invest. 1990, 2,1 Tabernero, L.; Bochar, D. A.; Rodwell, V. W.; Stauffacher, C. V. Proc. Natl. Acad. Sci. U.S.A.1999, 13, 7167 Istvan, E. S.; Palnitkar, M.; Buchanan, S. K.; Deisenhofer, J. J. EMBO. 2000, 19, 819 Karthik, M. V. K.; Satya Deepak, M. V. K. N.; Shukla, P. Comput. Biol. Med. 2012, 42, 156 Istvan, E.S.; Deisenhofer, J. Science. 2001, 292, 1160 Thompson, P.D.; Clarkson, P.; Karas, R.H. JAMA. 2003, 289, 1681 Ballantyne, C. M.; Corsini, A.; Davidson, M. H.; Holdaas, H.; Jacobson, T. A.; Leitersdorf, E.; Marz, W.; Reckless, J. P.; Stein, E. A. Arch. Intern. Med. 2003,163, 553 Omar, M.A.; Wilson, J. P.; Cox, T.S. Ann. Pharmacother. 2001, 35, 1096 da Silva, V. B.; Taft, C.A.; Silva, C. H. T. P. J. Phys. Chem. A. 2008, 112, 2007 Thilagavathi, R.; Kumar, R.; Aparna; Sobhia, M. E.; Gopalakrishnan, B.; Chakraborti, A. K. Bioorg. Med. Chem. Lett. 2005, 15, 1027 Klebe, G.; Abraham, U.; Mietzner, T. J. Med. Chem. 1994, 37, 4130 Zhang, Q. Y.; Wan, J.; Xu, X.; Yang, G. F.; Ren, Y. L.; Liu, J. J.; Wang, H.; Guo, Y. J. Comb. Chem. 2007, 9, 131 Cramer, R. D. III.; Patterson, D. E.; Bunce, J. D. J. Am. Chem. Soc. 1988, 110, 5959 Pfefferkorn, J. A.; Choi, C.; Song, Y. T.; Trivedi, B. K.; Larsen, S. D.; Askew, V.; Dillon, L.; Hanselman, J. C.; Lin, Z. W.; Lu, G. H.; Robertson, A.; Sekerke, C.; Auerbach, R.; Pavlovsky, A.; Harris, M. S.; Bainbridge, G.; Caspers, N. Bioorg. Med. Chem. Lett. 2007, 17, 4531 Pfefferkorn, J. A.; Song, Y.; Sun, K. L.; Miller, S. R.; Trivedi, B. K.; Choi, C.; Sorenson, R. J.; Bratton, L. D.; Unangst, P. C.; Larsen, S.D.; Poel, T. J.; Cheng, X. M.; Lee, C.; Erasga, N.; Auerbach, B.; Askew, V.; Dillon, L.; Hanselman, J. C.; Lin, Z. W.; Lu, G.; Robertson, A.; Olsen, K.; Mertz, T.; Sekerke, C.; Pavlovsky, A.; Harris, M. S.; Bainbridge, G.; Caspers, N.; Chen, H. F.; Eberstadt, M. Bioorg. Med. Chem. Lett. 2007, 17, 4538 Larsen, S. D.; Poel, T. J.; Filipski, K. J.; Kohrt, J. T.; Pfefferkorn, J.A.; Sorenson, R. J.; Tait, B. D.; Askew, V.; Dillon, L.; Hanselman, J. C.; Lu, G. H.; Robertson, A.; Sekerke, C.; Kowala, M. C.; Auerbach, B. J. Bioorg. Med. Chem. Lett. 2007, 17, 5567 Park, W. K. C.; Kennedy, R.M.; Larsen, S. D.; Miller, S.; Roth, B. D.; Song, Y. T.; Steinbaugh, B. A.; Sun, K.; Tait, B. D.; Kowala, M.C.; Trivedi, B. K.; Auerbach, B.; Askew, V.; Dillon, L.; Hanselman, J. C.; Lin, Z. W.; Lu, G. H.; Robertson, A.; Sekerke, C. Bioorg. Med. Chem. Lett. 2008, 18, 1151 Pfefferkorn, J.A.; Litchfield, J.; Hutchings, R.; Cheng, X. M.; Larsen, S. D.; Auerbach, B.; Bush, M. R.; Lee, C.; Erasga, N.; Bowles, D. M.; Boyles, D. C.; Lu, G. H.; Sekerke, C.; Askew, V.; Hanselman, J.C.; Dillon, L.; Lin ,Z. W.; Robertson, A.; Olsen, K.; Boustany, C.; Atkinson, K.; Goosen, T. C.; Sahasrabudhe, V.; Chupka, J.; Duignan, D. B.; Feng, B.; Scialis, R.; Kimoto, E.; Bi, Y. A.; Lai, Y. R.; El-Kattan, A.; Bakker-Arkema, R.; Barclay, P.; Kindt, E.; Le, V.; Mandema, J. W.; Milad, M.; Tait, B. D.; Kennedy, R.; Trivedi, B. K.; Kowala, M. Bioorg. Med. Chem. Lett.
2011, 21, 2725 20. Bratton, L. D.; Auerbach, B.; Choi, C.; Dillon, L.; Hanselman, J. C.; Larsen, S. D.; Lu, G.; Olsen, K.; Pfefferkorn, J. A.; Robertson, A.; Sekerke, C.; Trivedi, B. K.; Unangst, P. C. Bioorg. Med. Chem. 2007, 15, 5576 21. Kirk, K. L. J. Fluorine Chem. 2006, 127, 1013 22. Mogi, R.; Morisaki, K.; Hu, J.B.; Prakash, G. K.S.; Olah, G. A. J. Fluorine Chem. 2007, 128, 1098 23. Wang, X. G.; Fang, X.; Yang, X. Y.; Ni, M.; Wu, F. H. Chin. J. Chem. 2012, 30, 2767 24. Wang, X. G.; Fang, X.; Xiao, H. Y.; Yin, Y.; Xia, H. M.; Wu, F. H. J. Fluorine Chem. 2012, 133, 178 25. Yang, B.; Shi, L.Y.; Wu, J. J.; Fang, X.; Yang, X. Y.; Wu, F. H. Tetrahedron. 2013, 69, 3331 26. Wang, X.G. East China University of Science and Technology. 2013 27. Stǻhle, L.; Wold, S. J. Chemom. 1987, 1, 185 28. Wold, S.; Albano, C.; Dunn III, W. J.; Edlund, U.; Esbenson, K.; Geladi, P.; Hellberg, S.; Lindburg, W.; Sjostrom, M. Chemometrics. 1984, 138, 17 29. Patel, P.D.; Patel, M. R.; Kaushik-Basu, N.; Talele, T. T. J. Chem. Inf. Model. 2008, 48, 42 30. Bonnichon, F.; Grabner, G.; Guyot, G.; Richard, C. J. Chem. Soc., Perkin. Trans. 2, 1999, 6, 1203 31. Politi, A.; Durdagi, S.; Moutevelis-Minakakis, P.; Kokotos, G.; Mavromoustakos. T. J. Mol. Graph. Model. 2010, 29, 425 32. Lee, B.; Richards, F. M. J. Mol. Biol. 1971, 55, 379 33. Manavalan, P.; Ponnuswamy, P. K. Nature 1978, 275, 673 34. Krigbaum, W. R.; Komoriya, A. Biochim. Biophys. 1979, 576, 204
6
Figure 1 the structure of 9f
Figure 2 Alignment of the training set and compound 20 used as a template for alignment, with the common substructure shown in bold, and some substituents including the important R2R3, R4 and Y substituents shown in dashed red border. Molecules are colored in white for common C, blue for N, red for O, yellow for S, cyan for H, green for F, respectively.
7
Figure 3 Plot of experimental activity [log1/IC50] against activity predicted by the best CoMFA (A) and CoMSIA (B) models
Figure 4 Electrostatic contour of the CoMFA (A) and CoMSIA (B) models. The red color shows the favored negative electrostatic area and the blue color shows the favored positive electrostatic area
Fig. 5 Steric contour of the CoMFA (A) and CoMSIA (B) models. The green color shows the favored steric area and the yellow color show steric area
8
Figure 6 Hydrophobic (A) and Hydrogen bonding (B) contours of the CoMSIA models. The yellow color shows the favored hydrophobic area, the white color shows the disfavored hydrophobic area, the magenta color shows the favored H-acceptor area, the red color shows the disfavored H-acceptor area, the cyan color shows the favored H-donor area, and the purple color represents the disfavored H-donor area
Figure 7 Docking result of the compounds in the study. The binding site (2Q1L) surrounding the compound 20. Hydrogen bonds are shown as yellow dashed lines, with distance unit of Å. The inhibitor is also shown as purple stick model. Molecular surface is shown as cyan.
9
Figure 8 (A) Plot of the Total-energy docked complex versus the MD simulation time in the MD-simulated structures.(B) View of superimposed backbone atoms of the lowest energy structure of the MD simulation (green) and the initial structure (pink) for compound 20-2Q1L complex. Compound 20 is represented as carbon-chain in pink for the initial complex and carbon-chain in green for the lowest energy complex.
10
Figure 9 MD conformation derived for compound 20 with HMG the binding site of HMGR (PDB entry code: 2Q1L). Hydrogen bonds are shown as yellow dashed lines, with distance unit of Å. The inhibitor is also shown as purple stick model. Molecular surface is shown as cyan.
Figure 10 (A) MD conformation derived for compound 20 with HMG the binding site of HMGR (PDB
entry code: 2Q1L). (B) The “scorpion” conformation of compound 20 at the binding sit of MHGR.
11
Table 1 Statistical results for the CoMFA and CoMSIA models CoMFA 0.558 q2 10 noc 0.977 r2 0.087 SEE 354.857 F fraction 0.818 Steric 0.182 Electrostatic H-acceptor H-donor Hydrophobic
CoMSIA 0.582 8 0.919 0.161 121.009
0.163 0.127 0.193 0.240 0.277
Table 2 Designed molecules and predicted inhibit activity values of HMGR No. R
20
Skeleton type Predicted activity value CoMFA CoMSIA 9.765 9.420
20_1
9.891
9.518
20_2
9.894
9.434
20_3
9.766
9.420
20_4
9.763
9.414
20_5
9.764
9.414
20_6
9.746
9.431
20_7
9.766
9.422
12
20_8
9.767
9.434
9f
8.240
8.567
13
(A) MD conformation derived for compound 20 with HMG the binding site of HMGR (PDB entry code:
2Q1L). (B) The “scorpion” conformation of compound 20 at the binding sit of MHGR.
14
S0960-894X(14)00683-0 http://dx.doi.org/10.1016/j.bmcl.2014.06.055 BMCL 21774
To appear in:
Bioorganic & Medicinal Chemistry Letters
Received Date: Revised Date: Accepted Date:
10 April 2014 17 June 2014 19 June 2014
Please cite this article as: Wang, Z., Cheng, L., Kai, Z., Wu, F., Liu, Z., Cai, M., Molecular modeling studies of atorvastatin analogues as HMGR inhibitors using 3D-QSAR, molecular docking and molecular dynamics simulations, Bioorganic & Medicinal Chemistry Letters (2014), doi: http://dx.doi.org/10.1016/j.bmcl.2014.06.055
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.
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.
(A) MD conformation derived for
compound 20 with HMG the binding site of HMGR (PDB entry code: 2Q1L). (B) The “scorpion” conformation of compound 20 at the binding sit of MHGR. Molecular modeling studies of atorvastatin analogues as HMGR inhibitors using 3D-QSAR, molecular docking and molecular dynamics simulations Zhi Wanga,§, Li Ping Chenga,§,*, Zhen Peng Kaia, Fan Hong Wua,b,*, Zhuo Yu Liua and Min Feng Caia School of Chemical and Environmental Engineering, Shanghai Institute of Technology, Shanghai 201418, China b Key Laboratory for Advanced Materials and the Institute of Fine Chemicals, School of Chemistry and Molecular Engineering, East China University of Science and Technology, Shanghai 200237, China
a
§
These two authors (Zhi Wang and Liping Cheng) contributed equally to this work and should be considered as co-first authors.
1
Bioorganic & Medicinal Chemistry Letters j o ur n al h om e p a g e : w w w . e l s e v i er . c o m
Molecular modeling studies of atorvastatin analogues as HMGR inhibitors using 3DQSAR, molecular docking and molecular dynamics simulations Zhi Wanga,§ , Liping Chenga,§,∗ , Zhenpeng Kaia , Fanhong Wua,b,*, Zhuoyu Liu a and Minfeng Caia a
School of Chemical and Environmental Engineering,Shanghai Institute of Technology, Shanghai 201418, China Key Laboratory for Advanced Materials and the Institute of Fine Chemical, School of Chemistry and Molecular Engineering, East China University of Science and Technology, Shanghai 200237, China b
A R T IC LE IN F O
A B S TR A C T
Article history: Received Revised Accepted Available online
3-hydroxy-3-methylglutaryl coenzyme-A reductase (HMGR) is generally regarded as targets for the treatment of hypercholesterolemia. HMGR inhibitors (more commonly known as statins) are discovered as plasma cholesterol lowering molecules. In this work, 120 atorvastatin analogues were studied using a combination of molecular modeling techniques including three-dimensional quantitative structure-activity relationship (3D-QSAR), molecular docking and molecular dynamics (MD) simulation. The results show that the best CoMFA (comparative molecular field analysis) model has q2=0.558 and r2= 0.977, and the best CoMSIA (comparative molecular similarity indices analysis) model has q2=0.582 and r2=0.919. Molecular docking and MD simulation explored the binding relationship of the ligand and the receptor protein. The calculation results indicated that the hydrophobic and electrostatic fields play key roles in QSAR model. After MD simulation, we found four vital residues (Lys735, Arg590, Asp690 and Asn686) and three hydrophobic regions in HMGR binding site. The calculation results show that atorvastatin analogues obtained by introduction of F atoms or gem-difluoro groups could obviously improve the inhibitory activity. The new HMGR inhibitor analogues design in this paper had been submitted which is being currently synthesized by our laboratories.
Keywords: HMGR statins 3D-QSAR gem-difluoro molecular dynamics simulations
2014 Elsevier Ltd. All rights reserved.
∗ Corresponding author. Tel.: +86 21 60873250; fax: +86 21 60 877231; E-mail: [email protected] (L.P. Cheng) §
These two authors (Zhi Wang and Liping Cheng) contributed equally to this work and should be considered as co-first authors.
1
The diseases caused by atherosclerosis (e.g., Coronary heart disease (CHD) and Cerebral infarction) are one of the major health problems in the contemporary world. The high cholesterol level is the primary reason for causing coronary artery disease.1, 2 HMGR is the rate-controlling enzyme which catalyzes the conversion of HMG-CoA to mevalonate. This reaction is a fourelectron oxidoreduction and it is the committed step to endogenous cholesterol biosynthesis and isoprenoids.3 The HMGR is a tight tetramers protein.4 It has two isoforms, isoform1 and isoform2, which share high sequence homology (87% identity).5 The isoform2 plays an important role in mevalonate path way along with isoform1. HMGR inhibitors (statins) have been considered to be the best drug to cure the CHD and other diseases. Over the past decades, a number of HMGR inhibitors had been synthesized and listed, including Lovastatin, Pravastatin, Fluvastatin, Atorvastatin, Rosuvastatin and Pitavastatin.6 The atorvastatin is the most efficient inhibitor against both the two isoforms of HMGR.5 Novel statins should be not only with the ability of reducing the risk of cardiovascular disease and the LDL cholesterol level, but also is safe to health. Furthermore, the latter is the most important. Although statins are considered to be generally safe since the first statin (lovastatin) listed, skeletal muscle-related toxicity remains an associated adverse effect.7-9 Therefore, it is necessary to develop more and new efficient and safe HMGR inhibitors.
Figure 1 the structure of 9f
In recent years, many papers have reported some progresses in HMGR or its inhibitors in theoretical calculations. For example, da Silva et al.10 used virtual screening, flexible docking and molecular interaction fields to study and design novel HMGR inhibitors. Thilagavathi et al. 11 had developed a 3D-QSAR model for HMGR inhibitors based on a training set of 29 compounds by COMSIA method.12 The built model can be extrapolated to predict novel and more potent molecules. Besides the 29 compounds used by Thilagavathi, Zhang et al.13 added another 17 compounds from the MDDR database to build a new 3D-QSAR model. In their paper, different chemometric tools had been used including comparative molecular field analysis (CoMFA),14 molecular docking and pharmacophore model. However, on one hand, the total number of samples in their data set is less than 50. On the other hand, they did not explore the interactions between the inhibitors and receptor proteins. To obtain more understanding to the structural and chemical properties required for HMGR inhibitory activity, more 3D-QSAR and other computational explorations analyses for comparatively large number of newly synthesized and tested HMGR inhibitors are necessary.In the present work, 120 atorvastatin analogues reported by the same research group15-20 as potent and selective HMGR inhibitors were collected as dataset. A series of studies were performed using some computational methods include 3DQSAR tools like CoMFA and CoMSIA, molecular docking and MD simulation. The results of CoMFA and CoMSIA studies will not only explain the conformation or spatial orientation of atorvastatin analogues but also provide useful information for the design of potent and selective HMGR inhibitor. The introduction of F atom often improves the biological activity of target
compounds21 and it is a common and effective method to structural modifications. 22 Fluorinated compounds play an increasing important role in medicinal chemistry. Our laboratories23-25 had successfully synthesized some compounds by introducing the F atom or fluoride group into the proper position. For example, Wang26 had designed and synthesized a series of gem-difluoro derivatives of statins, in which the compound 9f (Fig. 1) has a good inhibitory activity to HMGR. These results encouraged us to design and synthesize more high inhibitory activity containing F derivatives. Moreover, the docking and MD analysis, performed using HMGR crystal structures, might expound the probable binding modes of theses inhibitors at the HMG bonding site. On the basis of molecular field information of 3D-QSAR tools, molecular docking and MD simulations, a few proposals were offered to design novel molecules with improved activity. In this study, we used the ligand-based alignment rule. The alignment result is shown in Fig. 2. According to the partial least squares (PLS) method,27,28 some statistical parameters were used to evaluate and analysis the stand or fall of these models, including q2, r2, standard error of estimate (SEE) and F-statistic values. Table 1 lists the statistical parameter results of the CoMFA and CoMSIA analyses. From Table 1, the optimal CoMFA and CoMSIA models show good predictive power with the cross-validated q2 of 0.558 and 0.582. The correlation coefficient r2, SEE and F-statistic values are 0.977, 0.087, and 354.857 for CoMFA model and 0.919, 0.161, and 121.009 for CoMSIA model. For the CoMFA model, the contributions of the steric and electrostatic fields are 81.8 and 18.2%. The main contribution in the model was steric field. For the optimal CoMSIA model, the steric, electrostatic, hydrophobic, H-Bond donor and acceptor field contribution was 16.3, 12.7, 27.7, 24.0, and 19.3%, respectively. The H-bond donor and hydrophobic fields were found to be the important contributions in the optimal CoMSIA model.
Figure 2 Alignment of the training set and compound 20 used as a template for alignment, with the common substructure shown in bold, and some substituents including the important R2 R3, R4 and Y substituents shown in dashed red border. Molecules are colored in white for common C, blue for N, red for O, yellow for S, cyan for H, green for F, respectively. Table 1 Statistical results for the CoMFA and CoMSIA models CoMFA q2 0.558 noc 10 2 r 0.977 SEE 0.087 F 354.857 fraction Steric 0.818 Electrostatic 0.182 H-acceptor H-donor Hydrophobic -
CoMSIA 0.582 8 0.919 0.161 121.009 0.163 0.127 0.193 0.240 0.277
Before the validation of the 3D-QSAR models, an initial inspection of the inhibitor molecules is very necessary. The
2
residual values of compounds between the predicted and experimented pIC50 exceeded one logarithm unit29were considered as outliers. According to the above rules, outliers of the optimal CoMFA and CoMSIA models are compound 05, 08 and 96. These outliers are different from the template compound 20 at some substituent position. For example, compared with compound 20, compound 05 is only with the Bn at position R3, but compound 20 with the –CN at Bn substituent, the –CN substituent is very important to inhibitory activity. 30 In compound 09, the R3 and R4 substituent is a ring, which is different from the compound 20. Similar to compound 08, compound 96 has not any substituent at R3 position. The calculation results based on the optimal CoMFA and CoMSIA models are shown in Supplementary data (Table A2). The linear fit between predicted and actual activity of the optimal CoMFA and CoMSIA model is described in Fig. 3. According to the optimal CoMFA and CoMSIA models, the StDev*Coeff contour maps were built. To aid in simplification, compound 20 is used to illustrate all contour maps of the optimal CoMFA and CoMSIA models. From Figs. 4-6, we can insight into the five fields and explain the key structural features required for inhibitory activity.
Figure 3 Plot of experimental activity [log1/IC50] against activity predicted by the best CoMFA (A) and CoMSIA (B) models
Figs. 4 and 5 show the steric and electrostatic contour maps of CoMFA and CoMSIA. From Fig. 4, the red (electronegative groups favored) and blue (electropositive groups) contours indicate the default 20% and 80% level contributions, respectively. Fig. 4A depicts the CoMFA electrostatic contours. A big and medium-sized red contour near the X substituent (pharmacophore) indicates electronegative groups would be favorable. A medium-sized and small red contour is also near the R1 position. Some large blue contours around the ring A (R2R3 substituent) show electropositive groups are favored for inhibitory activity. This is in agreement with the experimental data. For example, the order of some compounds inhibitory activity is: 30 (H) >32 (Me) >33 (Et). Fig. 4B depicts the CoMSIA electrostatic contour maps. Compared with the CoMFA contour map, only one big red contour is near the –COO- of the pharmacophore. There is also a big blue contour near the –COO-. Another big blue contour around the R2R3 substituent position shows the electropositive groups are favor to bioactivity.
Figure 4 Electrostatic contours of the CoMFA (A) and CoMSIA (B) models. The red color shows the favored negative electrostatic area and the blue color shows the favored positive electrostatic area
Figure 5 Steric contours of the CoMFA (A) and CoMSIA (B) models. The green color shows the favored steric area and the yellow color show steric area
From Fig. 5, the green (sterically favorable) and yellow (sterically unfavorable) contours also represent default level contributions. Fig. 5A shows the CoMFA steric contour maps of the best model with compound 20 surround. The contour maps mainly focus on the part of hydrophobic substituents. The green contour region around ring A (R2R3 substituent), B (R4 substituent) and beside the R1 substituent shows that the steric field give a positive effect to improve bioactivity. Some high bioactivity compounds such as 08, 13, 16 and 20 have large functional groups at the R2R3 region around the big green contour. Nevertheless, due to only the small substituent at the R2R3 region, the inhibitory capacity of some compounds (32, 33) is lower than that of compound 20 by 2 logarithm units. In the database set, some compounds are absent from ring B, such as 35-73 and 93111, the inhibitory activity is lower than that of the most potent inhibitor. Every green contour has a yellow contour region at hand, which indicates that the substituent of the above three positions (ring A and B, R1 substituent) with too large group would lead to a decrease of bioactivity, such as compounds 20 and 21. In short, the appropriate bulky or large groups are benefit to inhibitory activity. The contour maps of CoMSIA are described in Fig. 5B. As the Fig. 5B described, a large green (sterically favorable) contour map is in the center of the ring A, and a bulky yellow (sterically unfavorable) contour map surround the R2R3 region of compound 20. Since the CoMSIA steric contour map is similar to that of the CoMFA model, and the analysis of the CoMSIA steric contour map is neglected herein. The hydrophobic contour map of CoMSIA is shown in Fig. 6A. From the Fig, one large white (hydrophilic favorable) contour map is around the ring pyrrole, suggesting that introduction of hydrophilic substituents into the ring pyrrole will be benefit for inhibitory activity. Besides the X substituent (pharmacophore), other positions were replaced by the hydrophobic groups. One large yellow (hydrophobic favorable) contour map crossed the R2R3 substituent, indicating this position side with hydrophilic groups. The –NH2 is the hydrophilic group, compound 30 has this substituent with high bioactivity, but compound 31-35 own hydrophobic groups at the same position with low bioactivity. Compound 20 is used as template molecule to analysis the influence of H-bond donor and acceptor fields on inhibitory activity. As shown in Fig. 6B, a big cyan (H-bond donor favorable) contour map above the ring B suggests that the electropositive H-Bond donors in these regions are favorable for the inhibitory activity. A big purple contour map was observed farther from compound 20, indicating the H-bond acceptor has little effect on inhibitory activity. A big red contour map appears near the 5-OH and 7-COO- of the X substituent implies that Hbond acceptor groups could improve the inhibitory activity. A big magenta contour map near the R1 substituent indicates the Hbond acceptor groups could have a positive influence on the inhibitory activity, and a medium sized red contour map is near the 4-position of ring A.
3
Figure 6 Hydrophobic (A) and Hydrogen bonding (B) contours of the CoMSIA models. The yellow color shows the favored hydrophobic area, the white color shows the disfavored hydrophobic area, the magenta color shows the favored H-acceptor area, the red color shows the disfavored H-acceptor area, the cyan color shows the favored H-donor area, and the purple color represents the disfavored H-donor area
Based on the above theoretical calculation results and the known synthesized gem-difluoro derivatives of statins. We designed and predicted some HMGR inhibitor analogues by introduction of F atoms into proper position of pharmacophore. Compound 20, with the highest inhibitory activity was taken as a template to design new compounds. The predicted values are shown in Table 2. Table A3 shows that the pIC50 values of all designed new analogues are higher than that of 20, indicating their high inhibitory activity. These new designed compounds are being currently synthesized by our laboratories. The corresponding results will be published in further publications. The pIC50 value of the above mentioned 9f was also predicted by our 3D-QSAR model. The predicted pIC50 value is 8.24 (CoMFA) and 8.567 (CoMSIA), which is close to its experimental value (8.051).26 Therefore, the theoretical results are consistent with our experimental results and the model built in the present study is reliable.
Figure 7 Docking result of the compounds in the study. The binding site (2Q1L) surrounding the compound 20. Hydrogen bonds are shown as yellow dashed lines, with distance unit of Å. The inhibitor is also shown as purple stick model. Molecular surface is shown as cyan.
20_2
9.894
9.434
20_3
9.766
9.420
20_4
9.763
9.414
20_5
9.764
9.414
20_6
9.746
9.431
20_7
9.766
9.422
9.767
9.434
8.240
8.567
20_8
OH OH COF
9f
Docking protocols are widely used to explore the binding affinities of a series of ligands.31 In this paper, to study the binding environment in which the inhibitor interacts within the HMGR, molecular docking studies were performed on all 120 compounds. The docking results show that the highest total-score of compound 22 was 16.56, the total-scores of docking studies are shown in Supplementary data (Table A2), from the 120 listed total-scores, the pIC50 of most inhibitors keep in touch with the docking total-scores. Fig.7 describes the dock result of compound 20 and HMGR crystal structure. As shown in Fig.7, it is easily seen that some key amino acids (Arg568, Ser565, Lys 735, Ser684, Lys692, Asp690 and Lys691) interact with the HMGR inhibitors by hydrogen bonds at the binding site. The hydrogen bond distances observed are 2.52 Å (Arg568-HN-H…NC-), 2.01 Å (Ser565-OH…O-), 2.29 Å (Lys735-HN-H… O-), 1.89 Å (Ser684-OH…O-), 2.70 Å (Lys692-HN-H… O-), 1.93 Å (Asp690O…HO-), 1.86 Å (Lys691-HN-H… O-), respectively.
Table 2 Designed molecules and predicted inhibit activity values of HMGR No. R
20 20_1
Skeleton type Predicted activity value CoMFA CoMSIA 9.765 9.420 9.891
9.518
4
Figure 8 (A) Plot of the Total-energy docked complex versus the MD simulation time in the MD-simulated structures.(B) View of superimposed backbone atoms of the lowest energy structure of the MD simulation (green) and the initial structure (pink) for compound 20-2Q1L complex. Compound 20 is represented as carbon-chain in pink for the initial complex and carbonchain in green for the lowest energy complex.
Figure 9 MD conformation derived for compound 20 with HMG the binding site of HMGR (PDB entry: 2Q1L). Hydrogen bonds are shown as yellow dashed lines, with distance unit of Å. The inhibitor is also shown as purple stick model. Molecular surface is shown as cyan.
MD simulation is used to explore the interaction of the receptor protein and ligand. As the highest bioactivity inhibitor, compound 20 was used as the template molecule to explain the MD simulations. In our study, a 5 ns simulation of the ligandreceptor complex was run to energy balance at 2 ns to obtain a stable conformation. The total-energy of ligand-receptor compound ranging from 10200 to 9750 KJ/mol is shown in Fig. 8A. After 2 ns, the total-energy of the complex decreases to 9750 KJ/mol and maintains similar energy value at last 3 ns simulation. The overlay of the stable structure in the whole 5 ns MD simulations and the original docked structure is depicted in Fig. 8B. From the Fig. 8B, the pink and green ribbon represents the original docked structure and the lowest energy structure, respectively. The carbon-chain of compound 20 from two conformation structures also has the same color with the ribbon of the protein. Last but not least, the original docked structure and the overlay of the stable structure are docked into the same binding site, and their pharmacophore structure and other structures are basically similar. After 5 ns MD simulation studies, we found some residues including the Cys561, Leu857, Met657, Arg590 et al. are still essential to the interaction between inhibitor and HMGR. However, compared with the initial molecular docking results, the number of amino acid residues and hydrogen bonds at the binding site decrease. As shown in Fig.9, it has only four amino acid residues, Lys735, Asp690, Asn686 and Arg590 to
form 5 hydrogen bonds with the compound 20. In addition, from the MD results shown in Fig. 8B, the amino acid residues of HMGR binding pocket can form not only hydrogen bonds but also the electrostatic and van der Waals force. Met, Leu and Cys are the hydrophobic amino acids,32-34 they can form powerful interaction with the ring A, B and C. Here the rings A, B, and C are hydrophobic substituents and they may be regarded as three hydrophobic regions. Therefore, we predict that the interaction between the hydrophobic amino acids and the three hydrophobic regions perhaps play key roles in the inhibitory activity. Based on the results of QSAR, docking and MD simulation, we suppose that the inhibitor binds to the HMGR with a “scorpion” conformation depicted in Fig. 10B, where the pyrrole ring form the main brain, the three hydrophobic regions act as the two big pincers and mouth, and the X substituent is the key tail of the scorpion, respectively. Combined with the Fig 10A, firstly, three corners of the main brain of “scorpion” tightly interact with three hydrophobic regions of the ligand with the residues of Leu857, Cys561 and Met657, respectively. The hydrophobic groups are wonderful to inhibitors that form hydrophobic interaction with protein receptor, which correlating well with the previous QSAR and docking studies. The tail of the “scorpion” is the pharmacophore structure interaction with the HMGR by Lys735, Asp690, Asn686 and Arg590, respectively.
Figure 10 (A) MD conformation derived for compound 20 with HMG the binding site of HMGR (PDB entry code: 2Q1L). (B) The “scorpion” conformation of compound 20 at the binding sit of MHGR.
In this study, a series of computer-aided drug design processes, such as 3D-QSAR studies, molecular docking, and MD simulations, were used to explore the feasibility of introduction of F atoms into atorvastatin analogue as HMGR inhibitors. The built 3D-QSAR models show good internal and external predictive ability and can be extrapolated to predict novel and more potent molecules. The contour maps obtained from the CoMFA and CoMSIA analysis could guide to design new chemical entities with high HMGR inhibitory activity. To study the binding modes of inhibitors at the active site of HMGR protein, molecular docking studies of representative compounds was performed. Some key residues such as Lys735, Arg590, Asp690 and Asn686 as well as three hydrophobic regions in HMGR binding site were found. To further confirm the reliability of docking results, MD simulations were carried out for representative compounds. Based on the built QSAR models, several F-containg novel HMGR inhibitors were designed. The results show that the atorvastatin analogues obtained by introduction of F atoms or fluoride substituent groups could improve the inhibitory activity. Acknowledgments Thanks for financial support given by the National Natural Science Foundation of China (No. 21172148), the Scientific Research Innovation Project of Shanghai (No. 14YZ148) and
5
“College Students' (PE2014044).
Innovative
Activities”
of
Shanghai
References and notes 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.
16.
17.
18.
19.
Grundy, S. M. J. Am. Med. Assoc.1986, 256, 2849 Thelle, D.S. Drug Invest. 1990, 2,1 Tabernero, L.; Bochar, D. A.; Rodwell, V. W.; Stauffacher, C. V. Proc. Natl. Acad. Sci. U.S.A.1999, 13, 7167 Istvan, E. S.; Palnitkar, M.; Buchanan, S. K.; Deisenhofer, J. J. EMBO. 2000, 19, 819 Karthik, M. V. K.; Satya Deepak, M. V. K. N.; Shukla, P. Comput. Biol. Med. 2012, 42, 156 Istvan, E.S.; Deisenhofer, J. Science. 2001, 292, 1160 Thompson, P.D.; Clarkson, P.; Karas, R.H. JAMA. 2003, 289, 1681 Ballantyne, C. M.; Corsini, A.; Davidson, M. H.; Holdaas, H.; Jacobson, T. A.; Leitersdorf, E.; Marz, W.; Reckless, J. P.; Stein, E. A. Arch. Intern. Med. 2003,163, 553 Omar, M.A.; Wilson, J. P.; Cox, T.S. Ann. Pharmacother. 2001, 35, 1096 da Silva, V. B.; Taft, C.A.; Silva, C. H. T. P. J. Phys. Chem. A. 2008, 112, 2007 Thilagavathi, R.; Kumar, R.; Aparna; Sobhia, M. E.; Gopalakrishnan, B.; Chakraborti, A. K. Bioorg. Med. Chem. Lett. 2005, 15, 1027 Klebe, G.; Abraham, U.; Mietzner, T. J. Med. Chem. 1994, 37, 4130 Zhang, Q. Y.; Wan, J.; Xu, X.; Yang, G. F.; Ren, Y. L.; Liu, J. J.; Wang, H.; Guo, Y. J. Comb. Chem. 2007, 9, 131 Cramer, R. D. III.; Patterson, D. E.; Bunce, J. D. J. Am. Chem. Soc. 1988, 110, 5959 Pfefferkorn, J. A.; Choi, C.; Song, Y. T.; Trivedi, B. K.; Larsen, S. D.; Askew, V.; Dillon, L.; Hanselman, J. C.; Lin, Z. W.; Lu, G. H.; Robertson, A.; Sekerke, C.; Auerbach, R.; Pavlovsky, A.; Harris, M. S.; Bainbridge, G.; Caspers, N. Bioorg. Med. Chem. Lett. 2007, 17, 4531 Pfefferkorn, J. A.; Song, Y.; Sun, K. L.; Miller, S. R.; Trivedi, B. K.; Choi, C.; Sorenson, R. J.; Bratton, L. D.; Unangst, P. C.; Larsen, S.D.; Poel, T. J.; Cheng, X. M.; Lee, C.; Erasga, N.; Auerbach, B.; Askew, V.; Dillon, L.; Hanselman, J. C.; Lin, Z. W.; Lu, G.; Robertson, A.; Olsen, K.; Mertz, T.; Sekerke, C.; Pavlovsky, A.; Harris, M. S.; Bainbridge, G.; Caspers, N.; Chen, H. F.; Eberstadt, M. Bioorg. Med. Chem. Lett. 2007, 17, 4538 Larsen, S. D.; Poel, T. J.; Filipski, K. J.; Kohrt, J. T.; Pfefferkorn, J.A.; Sorenson, R. J.; Tait, B. D.; Askew, V.; Dillon, L.; Hanselman, J. C.; Lu, G. H.; Robertson, A.; Sekerke, C.; Kowala, M. C.; Auerbach, B. J. Bioorg. Med. Chem. Lett. 2007, 17, 5567 Park, W. K. C.; Kennedy, R.M.; Larsen, S. D.; Miller, S.; Roth, B. D.; Song, Y. T.; Steinbaugh, B. A.; Sun, K.; Tait, B. D.; Kowala, M.C.; Trivedi, B. K.; Auerbach, B.; Askew, V.; Dillon, L.; Hanselman, J. C.; Lin, Z. W.; Lu, G. H.; Robertson, A.; Sekerke, C. Bioorg. Med. Chem. Lett. 2008, 18, 1151 Pfefferkorn, J.A.; Litchfield, J.; Hutchings, R.; Cheng, X. M.; Larsen, S. D.; Auerbach, B.; Bush, M. R.; Lee, C.; Erasga, N.; Bowles, D. M.; Boyles, D. C.; Lu, G. H.; Sekerke, C.; Askew, V.; Hanselman, J.C.; Dillon, L.; Lin ,Z. W.; Robertson, A.; Olsen, K.; Boustany, C.; Atkinson, K.; Goosen, T. C.; Sahasrabudhe, V.; Chupka, J.; Duignan, D. B.; Feng, B.; Scialis, R.; Kimoto, E.; Bi, Y. A.; Lai, Y. R.; El-Kattan, A.; Bakker-Arkema, R.; Barclay, P.; Kindt, E.; Le, V.; Mandema, J. W.; Milad, M.; Tait, B. D.; Kennedy, R.; Trivedi, B. K.; Kowala, M. Bioorg. Med. Chem. Lett.
2011, 21, 2725 20. Bratton, L. D.; Auerbach, B.; Choi, C.; Dillon, L.; Hanselman, J. C.; Larsen, S. D.; Lu, G.; Olsen, K.; Pfefferkorn, J. A.; Robertson, A.; Sekerke, C.; Trivedi, B. K.; Unangst, P. C. Bioorg. Med. Chem. 2007, 15, 5576 21. Kirk, K. L. J. Fluorine Chem. 2006, 127, 1013 22. Mogi, R.; Morisaki, K.; Hu, J.B.; Prakash, G. K.S.; Olah, G. A. J. Fluorine Chem. 2007, 128, 1098 23. Wang, X. G.; Fang, X.; Yang, X. Y.; Ni, M.; Wu, F. H. Chin. J. Chem. 2012, 30, 2767 24. Wang, X. G.; Fang, X.; Xiao, H. Y.; Yin, Y.; Xia, H. M.; Wu, F. H. J. Fluorine Chem. 2012, 133, 178 25. Yang, B.; Shi, L.Y.; Wu, J. J.; Fang, X.; Yang, X. Y.; Wu, F. H. Tetrahedron. 2013, 69, 3331 26. Wang, X.G. East China University of Science and Technology. 2013 27. Stǻhle, L.; Wold, S. J. Chemom. 1987, 1, 185 28. Wold, S.; Albano, C.; Dunn III, W. J.; Edlund, U.; Esbenson, K.; Geladi, P.; Hellberg, S.; Lindburg, W.; Sjostrom, M. Chemometrics. 1984, 138, 17 29. Patel, P.D.; Patel, M. R.; Kaushik-Basu, N.; Talele, T. T. J. Chem. Inf. Model. 2008, 48, 42 30. Bonnichon, F.; Grabner, G.; Guyot, G.; Richard, C. J. Chem. Soc., Perkin. Trans. 2, 1999, 6, 1203 31. Politi, A.; Durdagi, S.; Moutevelis-Minakakis, P.; Kokotos, G.; Mavromoustakos. T. J. Mol. Graph. Model. 2010, 29, 425 32. Lee, B.; Richards, F. M. J. Mol. Biol. 1971, 55, 379 33. Manavalan, P.; Ponnuswamy, P. K. Nature 1978, 275, 673 34. Krigbaum, W. R.; Komoriya, A. Biochim. Biophys. 1979, 576, 204
6
Figure 1 the structure of 9f
Figure 2 Alignment of the training set and compound 20 used as a template for alignment, with the common substructure shown in bold, and some substituents including the important R2R3, R4 and Y substituents shown in dashed red border. Molecules are colored in white for common C, blue for N, red for O, yellow for S, cyan for H, green for F, respectively.
7
Figure 3 Plot of experimental activity [log1/IC50] against activity predicted by the best CoMFA (A) and CoMSIA (B) models
Figure 4 Electrostatic contour of the CoMFA (A) and CoMSIA (B) models. The red color shows the favored negative electrostatic area and the blue color shows the favored positive electrostatic area
Fig. 5 Steric contour of the CoMFA (A) and CoMSIA (B) models. The green color shows the favored steric area and the yellow color show steric area
8
Figure 6 Hydrophobic (A) and Hydrogen bonding (B) contours of the CoMSIA models. The yellow color shows the favored hydrophobic area, the white color shows the disfavored hydrophobic area, the magenta color shows the favored H-acceptor area, the red color shows the disfavored H-acceptor area, the cyan color shows the favored H-donor area, and the purple color represents the disfavored H-donor area
Figure 7 Docking result of the compounds in the study. The binding site (2Q1L) surrounding the compound 20. Hydrogen bonds are shown as yellow dashed lines, with distance unit of Å. The inhibitor is also shown as purple stick model. Molecular surface is shown as cyan.
9
Figure 8 (A) Plot of the Total-energy docked complex versus the MD simulation time in the MD-simulated structures.(B) View of superimposed backbone atoms of the lowest energy structure of the MD simulation (green) and the initial structure (pink) for compound 20-2Q1L complex. Compound 20 is represented as carbon-chain in pink for the initial complex and carbon-chain in green for the lowest energy complex.
10
Figure 9 MD conformation derived for compound 20 with HMG the binding site of HMGR (PDB entry code: 2Q1L). Hydrogen bonds are shown as yellow dashed lines, with distance unit of Å. The inhibitor is also shown as purple stick model. Molecular surface is shown as cyan.
Figure 10 (A) MD conformation derived for compound 20 with HMG the binding site of HMGR (PDB
entry code: 2Q1L). (B) The “scorpion” conformation of compound 20 at the binding sit of MHGR.
11
Table 1 Statistical results for the CoMFA and CoMSIA models CoMFA 0.558 q2 10 noc 0.977 r2 0.087 SEE 354.857 F fraction 0.818 Steric 0.182 Electrostatic H-acceptor H-donor Hydrophobic
CoMSIA 0.582 8 0.919 0.161 121.009
0.163 0.127 0.193 0.240 0.277
Table 2 Designed molecules and predicted inhibit activity values of HMGR No. R
20
Skeleton type Predicted activity value CoMFA CoMSIA 9.765 9.420
20_1
9.891
9.518
20_2
9.894
9.434
20_3
9.766
9.420
20_4
9.763
9.414
20_5
9.764
9.414
20_6
9.746
9.431
20_7
9.766
9.422
12
20_8
9.767
9.434
9f
8.240
8.567
13
(A) MD conformation derived for compound 20 with HMG the binding site of HMGR (PDB entry code:
2Q1L). (B) The “scorpion” conformation of compound 20 at the binding sit of MHGR.
14