3D-QSAR and molecular docking analysis of (4-piperidinyl)-piperazines as acetyl-CoA carboxylases inhibitors

Udghosh Singh; Rahul Prakashchand Gangwal; Gaurao V. Dhoke; Rameshwar Prajapati; Mangesh Damre; Abhay T. Sangamwar

doi:10.1016/j.arabjc.2012.10.023

View/Download PDF

Buy Reprints

PDF

Translate this page into:

Original article

10 (

1_suppl

); S617-S626

doi:

10.1016/j.arabjc.2012.10.023

3D-QSAR and molecular docking analysis of (4-piperidinyl)-piperazines as acetyl-CoA carboxylases inhibitors

Udghosh Singh¹, Rahul Prakashchand Gangwal¹, Gaurao V. Dhoke, Rameshwar Prajapati, Mangesh Damre, Abhay T. Sangamwar^⁎

Department of Pharmacoinformatics, National Institute of Pharmaceutical Education and Research (NIPER), Sect-67, S.A.S. Nagar, Punjab 160062, India

⁎Tel.: +91 0172 221 4682. abhays@niper.ac.in (Abhay T. Sangamwar)

Received: 2012-7-12, Accepted: 2012-10-31, Published: 2012-02-01

Disclaimer:
This article was originally published by Elsevier and was migrated to Scientific Scholar after the change of Publisher.

1 Authors with equal contribution.

Abstract

Acetyl-CoA carboxylase (ACC) is a crucial metabolic enzyme, which plays a vital role in fatty acid metabolism and obesity induced type 2 diabetes. Herein, we have performed 3D-QSAR and molecular docking analysis on a novel series of (4-piperidinyl)-piperazines to design potent ACC inhibitors. This study correlates the ACC inhibitory activities of 68 (4-piperidinyl)-piperazine derivatives with several stereo-chemical parameters representing steric, electrostatic, hydrophobic, hydrogen bond donor and acceptor fields. The CoMFA and CoMSIA models exhibited excellent $r_{ncv}^{2}$ values of 0.974 and 0.985, and $r_{cv}^{2}$ values of 0.671 and 0.693, respectively. CoMFA predicted $r_{pred}^{2}$ of 0.910 and CoMSIA predicted $r_{pred}^{2}$ of 0.963 showed that the predicted values were in good agreement with experimental values. Glide5.5 program was used to explore the binding mode of inhibitors inside the active site of ACC. We have accordingly designed novel ACC inhibitors by utilising the LeapFrog and predicted with excellent inhibitory activity in the developed models.

Keywords

Acetyl-CoA carboxylases

(4-Piperidinyl)-piperazines

CoMFA

CoMSIA

Molecular docking

LeapFrog

Show Related Articles from PubMed

1 Introduction

Obesity-induced type 2 diabetes is a global health scourge, which accounts for 5% global deaths (Bengtsson et al., 2011). It is mainly attributed to sedentary lifestyle, high-fat diet and the demographic shift to a more aged population. Worldwide obesity has more than doubled since 1980. According to WHO 2008 global estimate, more than 1.4 billion adults (20 and older), were overweight. 65% of the world’s population live in countries where overweight and obesity kill more people than underweight. More than 40 million children under the age of five were overweight in 2010 (Turkoglu et al., 2003 and Steyn et al., 2004). In obesity-induced diabetes, the ability of adipose tissue to store excess nutrients exceeds its capacity due to increased energy intake. This leads to an increase in ectopic fat accumulation in non-adipose tissue at a cellular level, such as liver, pancreas and skeletal muscles. It has been shown that the level of intracellular lipids found in the skeletal muscle, and the liver is a strong predictor of the development of insulin resistance and type 2 diabetes (Hulver et al., 2003 and Sinha et al., 2002). Therefore, there has been an increase in research interest in the paradigm of altering lipid disorders by targeting enzymes that are involved in lipid synthesis and oxidation (Kusunoki et al., 2006).

Acetyl-CoA carboxylases (ACC) catalyse the biotin-dependent carboxylation of acetyl-CoA to form malonyl-CoA, a key intermediate metabolite that regulates the rate of fatty acid metabolism (McGarry and Foster, 1980). It is a homo oligomeric protein composed of a carboxyltransferase (CT), a biotin carboxyl carrier protein and biotin carboxylase (BC) domains. The conversion of acetyl-CoA to malonyl-CoA by ACC is an ATP-dependent process. Malonyl-CoA, thus formed serves two functions, firstly, it acts as a substrate for de novo fatty acid synthesis, and secondly, it acts as an allosteric inhibitor of carnitine palmitoyl transferase (CPT-1), a key enzyme that positively regulates mitochondrial β-oxidation. Thus, the inhibition of ACC will block de novo fatty acid synthesis and enhance β-oxidation of fatty acid; this could be utilised for the treatment of metabolic disorders such as obesity and type 2 diabetes (Harwood et al., 2003; Harwood and James, 2005; Corbett, 2009 ). Mammalian ACC exists in two isoforms with distinct subcellular localizations. ACC1, a 265 kDa cytosolic protein, is primarily expressed in lipogenic tissues, while ACC2, a 280 kDa protein, is predominantly expressed in oxidative tissues such as liver, skeletal muscle and heart (Mao et al., 2003; Abu-Elheiga et al., 1997, 2001, 2005 ). It has been shown that the inhibitor of ACC2 particularly, reduces intracellular fat by the subsequent switch to fat utilisation for energy production; however, the suppression of both ACC1 and ACC2 expression using antisense oligonucleotides stimulates fat oxidation in hepatocytes more potently than that of either ACC1 or ACC2 alone (Savage et al., 2006). Thus, non-specific inhibition of ACC is more efficacious in stimulating fat oxidation than an ACC2-selective inhibitor.

Several small-molecule inhibitors have been reported as the blocker of ACC. The (4-piperidinyl)-piperazines are a recent series of ACC inhibitors discovered through a structure based focused screening. These derivatives have shown improved properties, including excellent cellular potency and oral bioavailability, which makes them promising candidates for further studies. In the present study, 3D-QSAR approaches including comparative molecular field analysis (CoMFA) and comparative molecular similarity analysis (CoMSIA) and molecular docking offer further insight into the understanding of the structure–activity relationship of (4-piperidinyl)-piperazine derivatives as ACC inhibitors. Based on the structure–activity relationship revealed in the present study, novel ACC inhibitors were accordingly designed using the LeapFrog. The stepwise description of the methodology used for 3D-QSAR analysis, molecular docking and new inhibitor designing of ACC is as shown in Fig. 1.

Stepwise description of methodology used for 3D-QSAR analysis, molecular docking and new inhibitors designing of ACC.

2 Materials and methods

2.1

2.1 Data set

Two series of (4-piperidinyl)-piperazine derivatives and their biological activities were collected from the literature reported by the same research group (Chonan et al., 2010, 2011). The inhibitory activities of the ACC inhibitors were expressed in IC₅₀ and converted to pIC₅₀ using the formula pIC₅₀ = −logIC₅₀ These (4-piperidinyl)-piperazine derivatives were divided into a training set of 50 compounds, and a test set of 18 compounds. Chemical structures and corresponding biological data are shown in Table 1.

Table 1 The structure and pIC₅₀ values (experimental and predicted) of training and test set molecules.

Sr.No.	R	pIC₅₀	CoMFA	Residual	CoMSIA	Residual
1	phenyl	6.90	6.93	0.03	6.89	0.01
2	4-HO-phenyl	7.33	7.30	0.03	7.38	0.05
3	3-HOCH₂-phenyl	7.33	7.35	0.02	7.38	0.05
4	4-HOCH₂-phenyl	7.38	7.38	0.00	7.37	0.01
5^a	4-H₂NCO-phenyl	7.10	7.10	0.00	7.11	0.01
6^a	4-HOCH₂CH₂-phenyl	7.57	7.50	0.08	7.65	0.08
7	4-HOOCCH₂-phenyl	7.12	7.12	0.00	7.13	0.01
8	4-HOCH₂CH₂O-phenyl	7.33	7.27	0.06	7.31	0.02
9	4-HOCH₂CH₂CH₂-phenyl	7.15	7.21	0.06	7.13	0.02
10^a	1H-indole-5-yl	7.26	7.27	0.01	7.24	0.02
11	1H-indole-6-yl	7.33	7.23	0.10	7.30	0.03
12	1H-indazole-5-yl	7.38	7.30	0.08	7.31	0.07
13	1H-indazole-6-yl	7.24	7.27	0.03	7.32	0.08
14	Indolin-2-one-5-yl	7.27	7.27	0.00	7.28	0.01

Sr.No.	R₁	R₂	R₃	pIC₅₀	CoMFA	Residual	CoMSIA	Residual
15	Me	H	H	7.16	7.04	0.12	7.07	0.09
16	H	Me	H	7.23	7.27	0.04	7.28	0.05
17	H	H	Me	7.44	7.49	0.05	7.47	0.03
18	Me	Me	H	6.93	6.94	0.01	6.96	0.03
19^a	Me	H	Me	6.90	6.91	0.01	6.90	0.00
20	H	Me	Me	7.48	7.40	0.08	7.35	0.13
21	Me	Me	Me	6.98	7.09	0.11	7.04	0.06
22	H	Me	Et	6.77	6.81	0.04	6.76	0.01
23^a	H	Me	CN	8.01	7.83	0.18	8.10	0.09
24	H	CH₂NMe₂	Me	8.08	8.08	0.00	8.08	0.00

Sr.No.	R₁	R₂	pIC₅₀	CoMFA	Residual	CoMSIA	Residual
25	F	H	7.48	7.48	0.00	7.52	0.04
26	H	F	7.32	7.37	0.05	7.34	0.02
27^a	Me	H	7.43	7.42	0.01	7.38	0.05
28^a	H	Me	7.52	7.52	0.00	7.51	0.01
29^a	MeO	H	7.49	7.49	0.00	7.50	0.01
30^a	H	MeO	7.68	7.66	0.04	7.67	0.02
31	CF3	H	7.00	7.00	0.00	7.00	0.00
32^a	H	CF3	7.14	7.16	0.02	7.14	0.00
33	F	Me	7.31	7.32	0.01	7.37	0.06
34	F	MeO	7.21	7.33	0.12	7.21	0.00
35	MeO	MeO	7.52	7.45	0.07	7.48	0.04

Sr.No.	R	pIC₅₀	CoMFA	Residual	CoMSIA	Residual
36	1-Acetylpiperidin-4-yl	7.13	7.14	0.01	7.13	0.00
37	Me	5.73	5.76	0.03	5.76	0.03
38^a	Et	6.22	6.53	0.01	6.20	0.31
39	n-Pr	6.62	6.54	0.08	6.62	0.00
40^a	n-Bu	6.68	6.68	0.00	6.62	0.06
41	iso-Bu	6.84	6.85	0.01	6.87	0.03
42	iso-Pr	6.88	6.88	0.01	6.89	0.01
43	c-Pentyl	7.31	7.38	0.07	7.29	0.02
44	c-Hex	7.10	7.07	0.03	7.10	0.00
45^a	tert-Bu	6.96	6.97	0.01	6.98	0.02
46	MeO	6.76	6.75	0.01	6.71	0.05
47^a	EtO	7.26	7.25	0.01	7.21	0.05
48	iso-PrO	7.55	7.61	0.06	7.57	0.02
49	c-PeO	7.28	7.25	0.03	7.21	0.07
50	c-HexO	7.12	7.14	0.02	7.19	0.07
51	tert-BuO	7.62	7.62	0.00	7.62	0.00
52	tert-BuNH	6.92	6.90	0.02	6.93	0.01

Sr.No.	R	R₁	pIC₅₀	CoMFA	Residual	CoMSIA	Residual
53	CH₂F	H	7.77	7.67	0.10	7.77	0.00
54	CF₃	H	7.51	7.46	0.05	7.51	0.00
55	H	5-Me	7.49	7.55	0.06	7.43	0.06
56^a	H	6-Me	7.22	7.22	0.00	7.22	0.00
57	CH₂F	6-Me	7.49	7.53	0.04	7.57	0.08
58	CF₃	6-Me	7.24	7.23	0.01	7.21	0.03
59	CF₃	6-CF₃	7	7	0.00	6.98	0.02
60	CF₃	6-F	7.21	7.35	0.14	7.29	0.08
61	CF₃	6-Cl	7.42	7.33	0.09	7.42	0.00
62	CF₃	6-MeO	7.57	7.56	0.01	7.53	0.04
63^a	CF₃	4-Aza	7.92	7.42	0.50	7.91	0.01
64	CF₃	5-Aza	6.93	6.90	0.03	6.93	0.00
65	CF₃	6-Aza	7.39	7.39	0.00	7.40	0.01
66^a	CF₃	7-Aza	7.35	7.34	0.01	7.36	0.01
67^a	–	7.12	7.18	0.06	7.11	0.01
68	–	7.48	7.43	0.05	7.48	0.00

a Test set compounds.

2.2

2.2 Molecular modelling

All the molecular modelling studies were performed using SYBYL 7.1 molecular modelling package installed on a Silicon Graphics Fuel Work station running IRIX 6.5. 3D molecular structures of all the ACC inhibitors were constructed using the Sketch Molecule module. Compound 24 was selected as a template molecule for the sketching and alignment of all the compounds. The fragment of the template molecule used as the common structure is as shown in Fig. 2. Further, the molecules were minimised by applying tripos molecular mechanics force field with conjugate gradient method (Powell, 1977). The minimisation was terminated when the energy gradient convergence criterion of 0.05 kcal/mol was reached or when the 10,000 steps minimisation cycle was exceeded. Gasteiger–Hückel charges (Gasteiger and Marsili, 1980) were applied to all the training and test set molecules, used for 3D-QSAR studies.

Common structure used during alignment of all (4-piperidinyl)-piperazines derivatives.

2.3

2.3 Molecular alignment for 3D-QSAR analysis

Molecular alignment is the most important part of the 3D-QSAR studies. A common substructure-based alignment method was used to align all inhibitors with a template molecule. The aligned training set molecules are as shown in Fig. 3.

Alignment of all molecules used for CoMFA and CoMSIA molecular field generation.

2.4

2.4 Calculation of CoMFA and CoMSIA descriptors

For all the training set molecules, the steric and electrostatic CoMFA fields at each lattice intersection of a regularly spaced grid of 2.0 Å were calculated using the default probe, a sp³ carbon atom with a charge of +1 and a van der Waals radius of 1.52 Å. The Lennard–Jones potential terms represent steric field and Coulombic terms represents electrostatic fields, which were calculated using tripos force field, where their energy values were truncated at 30 kcal/mol. The minimum column filtering of 2.0 kcal/mol was used during model generation to improve the signal-to-noise ratio by omitting those lattice points whose energy variation was below this threshold. The leave-one-out (LOO), a cross-validation method was implemented during regression analysis. The non-cross-validated conventional analysis was produced with the optimal number of components equal to that yielding the highest $r_{cv}^{2}$ , and the corresponding conventional correlation coefficient $r_{ncv}^{2}$ , F value and standard error of estimate (SEE), were also calculated (Singh and Sobhia, 2011).

The CoMSIA method, incorporating steric, electrostatic, hydrogen bond donor, hydrogen bond acceptor and hydrophobic fields, was carried out using a probe atom with radius 2.0 Å, +1.0 charge, and hydrophobic and hydrogen bond properties of +1. The attenuation factor and column filtering were set to 0.3 and 2.0 kcal/mol, respectively. The nine different CoMSIA models were statistically evaluated in a similar way as described in CoMFA.The predictive abilities of developed models were determined from a test set of 18 compounds. pIC₅₀ values for test molecules were predicted by using developed CoMFA and CoMSIA models. The predictive correlation coefficient ( $r_{pred}^{2}$ ), based on the molecules of test set, was calculated using the following equation. $r_{pred}^{2} = \frac{(SD - PRESS)}{SD}$ where SD represents the sum of squared deviations between the inhibitory activities of the test set and mean activities of the training molecules and PRESS represents the sum of squared deviations between predicted and actual activity values for each molecule in the test set (Nandi and Bagchi, 2010 and Gupta et al., 2011).

Y randomisation test is a widely used approach to establish the robustness of developed QSAR models. In this test, we have developed new QSAR models by shuffling the inhibitory activity randomly and keeping the original CoMFA and CoMSIA fields as such. The new QSAR models are expected to have low $r_{ncv}^{2}$ and $r_{cv}^{2}$ values for several trials, which confirm the robustness of the developed QSAR models.

2.5

2.5 Molecular docking

To investigate the detailed intermolecular interactions between the (4-piperidinyl)-piperazine derivatives and ACC, an automated docking program Glide5.5 was used. Three-dimensional structure information of ACC was taken from the protein data bank (PDB ID: 3FF6). Co-crystallised ligand (CP-640186) was docked into ACC to validate docking protocol. The protocol followed during the docking studies of (4-piperidinyl)-piperazine derivative included ligand preparation, processing of the protein, grid preparation and molecular docking. The ligand molecules were deleted, and hydrogen atoms were added during protein preparation. Solvent molecules were deleted, and bond orders for crystal protein were adjusted and minimised up to 0.30 Å RMSD (Singh and Elizabeth Sobhia, 2011 and Dhoke et al., 2012). An active site of 10 Å was created around the co-crystallised ligand. Extra precision (XP) mode and other default parameters of Glide software were used for the docking studies of known and designed ACC inhibitors.

2.6

2.6 Designing of potent analogues using LeapFrog

LeapFrog is a de novo tool to design a series of potentially active ligand molecules even when the target protein structure is unknown (Rasmussen et al., 2002). LeapFrog calculates binding energy by considering three major components viz., direct steric, electrostatic, and implicit hydrogen bonding enthalpies of ligand–cavity binding using the tripos force field. The method of calculating binding energies in LeapFrog is similar to that of Goodford’s GRID program (Goodford, 1985). LeapFrog makes use of the CoMFA contours and crystal structure of the target receptor for the generation of its cavity. The cavity thus obtained is used to generate the site points. The charge of a site point atom is positive, negative, or lipophilic. Its value is compared with 1.0. If the atom charge is smaller in magnitude than 1.0 the site point is lipophilic, and it seeks a positive atom in the approaching fragment; and if greater than 1.0, the site point seeks a negative atom. The JOIN, FUSE and OPTIMIZE moves in LeapFrog were utilised to optimise the compounds (Ambure et al., 2011).

3 Results and discussion

3.1

3.1 CoMFA and CoMSIA analyses

The statistical parameters obtained from the CoMFA and CoMSIA analyses of (4-piperidinyl)-piperazine derivatives are listed in Table 2. The analysis of the resulting 3D-QSAR models showed that the best CoMFA model was obtained with combined steric and electrostatic fields, that yields a cross-validated $r_{cv}^{2}$ of 0.671 with 9 optimum number of components, a non-cross-validated $r_{ncv}^{2}$ of 0.974, an estimated F value of 167.45, and low standard error of estimation (SEE) of 0.062. In this model, the steric and electrostatic contributions were found to be 59.0% and 41.0%, respectively.

Table 2 Summary of results from CoMFA and CoMSIA analysis.

	$r_{cv}^{2}$ ^j	$r_{ncv}^{2}$ ^k	SEE^l	ONC^m	Fⁿ	$r_{pred}^{2}$ ^o	Relative % contributions
	$r_{cv}^{2}$ ^j	$r_{ncv}^{2}$ ^k	SEE^l	ONC^m	Fⁿ	$r_{pred}^{2}$ ^o	S	E	H	D	A
CoMFA	0.671	0.974	0.062	9	167.45	0.910	59.0	41.0	–	–	–
CoMSIA^a	0.552	0.975	0.062	10	151.42	0.943	56.4	43.6	–	–	–
CoMSIA^b	0.693	0.985	0.048	10	252.766	0.963	31.8	33.1	35.1	–	–
CoMSIA^c	0.452	0.889	0.126	7	46.876	0.828	34.8	37.6	–	27.7	–
CoMSIA^d	0.509	0.974	0.064	7	143.841	0.840	26.2	25.0	28.2	20.6	–
CoMSIA^e	0.400	0.938	0.093	7	88.83	0.897	28.9	28.4	29.5	–	13.1
CoMSIA^f	0.441	0.874	0.134	7	41.769	0.804	30.8	33.0	–	24.3	11.8
CoMSIA^g	0.343	0.935	0.100	7	56.36	0.774	–	29.8	45.3	25.0	–
CoMSIA^h	0.282	0.860	0.142	7	36.939	0.781	–	27.0	34.5	26.2	12.4
CoMSIAⁱ	0.487	0.968	0.070	10	118.922	0.864	24.5	21.0	26.2	19.1	8.5

S, Steric; E, electrostatic; H, hydrophobic; D, hydrogen donor; A, acceptor.

a SE.

b SEH.

c SED.

d SEHD.

e SEHA.

f SEDA.

g EHD.

h EHDA.

i SEHDA.

j Cross validated correlation coefficient.

k No validation correlation coefficient.

l Standard Error of Estimate.

m Optimal number of components.

n F-test value.

o Predictive correlation coefficient.

Using various combinations of CoMSIA descriptor fields, nine different CoMSIA models were developed. A model consisting of steric, electrostatic and hydrophobic CoMSIA fields with a cross-validated $r_{cv}^{2}$ of 0.693 with 10 optimum number of components and conventional $r_{ncv}^{2}$ of 0.985 was selected for further analysis. F value and SEE of this model were 252.76 and 0.048, respectively. The relative contributions of steric, electrostatic, and hydrophobic were found to be 31.8%, 33.1%, and 35.1%, respectively. The scattered plots of experimental pIC₅₀ against the CoMFA and CoMSIA predicted pIC₅₀ of both training set and test set are also shown in Fig. 4. These results indicate that the model thus developed was consistent and was supported by bootstrapping results for 100 runs, with $r_{bs}^{2}$ of 0.989 and 0.991 for CoMFA and CoMSIA, respectively.

Graph of experimental versus predicted pIC50 of the training set and the test set using (a) CoMFA and (b) CoMSIA.

The final non cross-validated partial least square (PLS) analysis model of CoMFA and CoMSIA was further selected to generate the 3D contour maps. The contour maps illustrate lattice points and the variation in the molecular field values at lattice points, which is strongly associated with the variation in the receptor binding affinity. Molecular field analysis aids in finding the favourable or unfavourable interaction energies of aligned molecules with the help of a probe atom, surrounding the molecules. These 3D colour contour maps provide hints for the modification required to design new molecules with improved activity.Fig. 5a and b shows the distribution of steric and electrostatic fields generated by using the validated CoMFA model. The green and yellow colour (80% and 20% contributions) contour maps show the favourable and unfavourable steric interactions, respectively, while blue and red (80% and 20% contributions) contour maps indicate the favourable and unfavourable electrostatic interaction with the molecules, respectively. The green plot was found around the carboxy group attached to piperidinyl ring of (4-piperidinyl)-piperazine derivatives indicating the favourable region for the presence of bulky groups for the activity. (4-piperidinyl)-piperazine derivatives with small substituents like methyl (compound 37) on carboxy group attached to piperidinyl were weakly active. Many compounds with an increased bulk at carboxy group attached to piperidinyl ring (compounds 37–44 and 53–66), particularly aliphatic rings, gave a greater activity (compounds 43, 44, 49 and 50). Another green plot located on 3,4 substituted phenyl ring attached to the pyridine indicates that moderate bulky groups favour the activity (compound 2–14), while compound not bearing any substitution on phenyl ring was least active (compound 1). Yellow region is sighted in steric contour plots at the 3rd position of the indole ring attached to the pyridine suggesting that bulky groups in these regions disfavour the ACC inhibitory activity and SAR data also suggested that the substitution at this position results in a decrease in the activity as evidenced by the compounds bearing substitution like methyl and ethyl groups (compound 19, 21 and 22), which were less active.

CoMFA contour maps for ACC inhibitors (a) the steric field distribution and (b) the electrostatic field distribution for highly active compound 24.

The electrostatic effects of the substituents were analysed by the presence of blue and red colour electrostatic contour maps. One blue contour was present in the proximity of the carboxy group of 1-acetylpiperidin-4-yl substituted (4-piperidinyl)-piperazine derivatives (1–36). Another blue contour map was found around the substitution at the 4th position of the phenyl ring with polar (hydroxyl and carbamoyl) functional group (compound 2–9). The blue contour indicates that electropositive potential enhances ACC inhibitory activity. The red contour plot covering the oxygen of the carboxylic group indicates that the electronegative potential favours the activity (compounds 53–66).

The CoMSIA contour map consists of steric, electrostatic and hydrophobic fields. CoMSIA model generated steric and electrostatic contour maps are comparatively similar to the steric and electrostatic contour maps of CoMFA model. The hydrophobic field effect is shown in Fig. 6. Influence of the hydrophobic group is explained by the presence of an orange and white colour map; orange contour explains the favourable hydrophobic fields while the white explains the unfavourable hydrophobic fields. Orange colour contour plot located around the 1-acetylpiperidin-4-yl indicates that the hydrophobic group exhibits the positive effect on the biological activity. Another orange colour contour plot at the 6th position of benzothiophene derivatives shows that the hydrophobic group substitution such as methyl, fluorine and chlorine favours the activity (compound 56–62). A big white colour contour in the proximity of hydrophobic substitution at the carboxylic group attached to the (4-piperidinyl)-piperazine derivatives indicates that the hydrophobic substitution disfavours the ACC inhibitory activity (compound 43 and 44). The white colour contour around the 3rd position of the phenyl ring indicates that the hydrophobic group disfavours the activity as observed in the case of compounds 10 and 31. Another white region is sighted in hydrophobic contour plots at the 3rd position of the indole ring attached to the pyridine suggesting that hydrophobic groups in these regions disfavour the ACC inhibitory activity (compound 19), which was less active as compared to non-substituted compounds at the 3rd position of the indole ring.

The distribution of hydrophobic field (CoMSIA contour maps) around the highly active compound (24) of the training set.

The best feature of 3D-QSAR was used to predict the activities of test set molecules, which are not included in the training set. The derived CoMFA and CoMSIA models were validated to get the significant utility by predicting activity of external test set compounds. All the test set molecules were constructed, minimised and aligned with a template, in a like manner. The predicted activities of test set molecules were in good agreement with the observed activities within an acceptable error range and verified by the CoMFA $r_{pred}^{2}$ value of 0.910 and CoMSIA $r_{pred}^{2}$ value of 0.963.

To evaluate robustness of the developed QSAR models, we have performed the Y randomisation test. Several random shuffles of the pIC₅₀ values were performed and the results are shown in Table 3. The low $r_{ncv}^{2}$ and $r_{cv}^{2}$ values show that the good results in our original model are not due to a chance correlation. These results show that both the CoMFA and CoMSIA models are reliable and can be useful in designing new potent ACC inhibitors.

Table 3 Ten random correlation coefficients (

r_{ncv}^{2}

and

r_{cv}^{2}

) from the activity (Y) randomizsation test study.

Sr. No.	CoMFA		CoMSIA
Sr. No.	$r_{ncv}^{2}$	$r_{cv}^{2}$	$r_{ncv}^{2}$	$r_{ncv}^{2}$
1	0.081	−0.031	0.052	−0.024
2	0.077	−0.014	0.215	−0.000
3	0.275	−0.001	0.143	−0.034
4	0.069	−0.050	0.277	−0.009
5	0.536	0.112	0.820	0.108
6	0.032	−0.076	0.020	−0.066
7	0.116	−0.070	0.062	−0.050
8	0.085	−0.064	0.064	−0.044
9	0.129	−0.044	0.061	−0.056
10	0.235	0.068	0.032	−0.039

3.2

3.2 Docking studies

The docking studies were carried out to explore the interaction mechanism between inhibitors and the receptor. To validate the docking protocol, co-crystallised ligand (CP-640186) was re-docked into the active site of acetyl-CoA carboxylases 2 (ACC2) with a root mean squared deviation of 0.714 Å. It binds at the dimer interface and interacts with residues from both chains of the ACC2 protein. The carbonyl group adjacent to the anthracene of CP-640186 forms the hydrogen bond with the main-chain amide N atom of Glu2230. Another weak hydrogen bond was observed between the carbonyl group adjacent to morpholine and the amide N atom of Gly2162. All (4-piperidinyl)-piperazine derivatives were docked into the active site of ACC, and they have shown a similar binding pattern as that of the co-crystallised ligand. Fig. 7 shows the docking of highly active molecule 24 into the active site of ACC. One hydrogen-bond interaction between the backbone NH of Glu2230 and the carbonyl group of all inhibitors was conserved. Some inhibitors have also shown the additional hydrogen bond interactions with Lys1967 and Tyr1974. The binding interaction patterns observed during docking studies were complementary to those of CoMFA and CoMSIA contour maps.

The binding pocket of ACC (PDB ID: 3FF6) with docked conformation of highly active compound 24.

3.3

3.3 Designing of potent analogues

The region file of the contour generated during partial least square (PLS) analysis of the CoMFA model was used for cavity generation calculations for our study. Most potent inhibitors were subjected for LeapFrog optimisation. The novel ligands were built using different moves such as JOIN, FUSE and OPTIMIZE. The designed molecules were minimised similarly as that of training set molecules. All designed molecules were aligned, and their activities were predicted using CoMFA and CoMSIA models. Interestingly two molecules showed an increase in potency as the substituents present on these molecules were well correlated with the contours of the CoMFA and CoMSIA models. These two molecules were also docked into the active site of ACC, and they showed a similar binding interaction pattern with a high docking score in comparison to compound 24. Table 4 shows the structure, CoMFA predicted activity and Glide docking score of these molecules.

Table 4 Chemical structures, predicted pIC₅₀ activities and Glide score of designed ACC inhibitors.

Compound name	Structure	CoMFA Predicted pIC₅₀ value	Glide score
Compound A		8.31	−11.09
Compound B		8.23	−10.95

4 Conclusions

3D-QSAR and molecular docking analysis of 68 (4-piperidinyl)-piperazine inhibitors were performed to determine the structural requirements for potency against ACC. Stable and statistically reliable CoMFA and CoMSIA models were developed, which have shown significant $r_{cv}^{2}, r_{ncv}^{2}, r_{bs}^{2}, r_{pred}^{2}$ and a small standard deviation indicating a better statistical relationship between the activity and descriptors. The putative binding conformations and interactions of the known and designed ACC inhibitors in the active site were studied by molecular docking. Both 3D-QSAR models and docking studies rendered complementary information, and the steric, electrostatic and hydrophobic contour maps derived from the CoMFA and CoMSIA models provided crucial clues that can be used in the successful designing of highly active analogues of (4-piperidinyl)-piperazine derivatives against ACC.

Acknowledgement

The authors acknowledge the financial support from Department of Science and Technology (DST), New Delhi.

References

Abu-Elheiga L., Almarza-Ortega D.B., Baldini A., Wakil S.J., . Human acetyl-CoA carboxylase 2 molecular cloning, characterization, chromosomal mapping, and evidence for two isoforms. J. Biol. Chem.. 1997;272:10669-10677.
[Google Scholar]
Abu-Elheiga L., Matzuk M.M., Abo-Hashema K.A.H., Wakil S.J., . Continuous fatty acid oxidation and reduced fat storage in mice lacking acetyl-CoA carboxylase 2. Science. 2001;291:2613-2616.
[Google Scholar]
Abu-Elheiga L., Matzuk M.M., Kordari P., Oh W.K., Shaikenov T., Gu Z., Wakil S.J., . Mutant mice lacking acetyl-CoA carboxylase 1 are embryonically lethal. Proc. Natl. Acad. Sci. USA. 2005;102:12011.
[Google Scholar]
Ambure P.S., Gangwal R.P., Sangamwar A.T., . 3D-QSAR and molecular docking analysis of biphenyl amide derivatives as p38α mitogen activated protein kinase inhibitors. Mol. Diversity. 2011;16:377-388.
[Google Scholar]
Bengtsson C., Blaho S., Saitton D.B., Brickmann K., Broddefalk J., Davidsson Ö., Drmota T., Folmer R., Hallberg K., Hallén S., . Design of small molecule inhibitors of acetyl-CoA carboxylase 1 and 2 showing reduction of hepatic malonyl-CoA levels in vivo in obese Zucker rats. Bioorg. Med. Chem.. 2011;19:3039-3053.
[Google Scholar]
Chonan T., Tanaka H., Yamamoto D., Yashiro M., Oi T., Wakasugi D., Ohoka-Sugita A., Io F., Koretsune H., Hiratate A., . Design and synthesis of disubstituted (4-piperidinyl)-piperazine derivatives as potent acetyl-CoA carboxylase inhibitors. Bioorg. Med. Chem. Lett.. 2010;20:3965-3968.
[Google Scholar]
Chonan T., Wakasugi D., Yamamoto D., Yashiro M., Oi T., Tanaka H., Ohoka-Sugita A., Io F., Koretsune H., Hiratate A., . Discovery of novel (4-piperidinyl)-piperazines as potent and orally active acetyl-CoA carboxylase 1/2 non-selective inhibitors: F-Boc and triF-Boc groups are acid-stable bioisosteres for the Boc group. Bioorg. Med. Chem.. 2011;19:1580-1593.
[Google Scholar]
Corbett J.W., . Review of recent acetyl-CoA carboxylase inhibitor patents: mid-2007–2008. Expert Opin. Ther. Pat.. 2009;19:943-956.
[Google Scholar]
Dhoke G.V., Gangwal R.P., Sangamwar A.T., . A combined ligand and structure based approach to design potent PPAR-alpha agonists. J. Mol. Struct.. 2012;1028:22-30.
[Google Scholar]
Gasteiger J., Marsili M., . Iterative partial equalization of orbital electronegativity – a rapid access to atomic charges. Tetrahedron. 1980;36:3219-3228.
[Google Scholar]
Goodford P.J., . A computational procedure for determining energetically favorable binding sites on biologically important macromolecules. J. Med. Chem.. 1985;28:849-857.
[Google Scholar]
Gupta P., Garg P., Roy N., . Comparative docking and CoMFA analysis of curcumine derivatives as HIV-1 integrase inhibitors. Mol. Diversity. 2011;15:1-18.
[Google Scholar]
Harwood H.J., Petras S.F., Shelly L.D., Zaccaro L.M., Perry D.A., Makowski M.R., Hargrove D.M., Martin K.A., Tracey W.R., Chapman J.G., . Isozyme-nonselective N-substituted bipiperidylcarboxamide acetyl-CoA carboxylase inhibitors reduce tissue malonyl-CoA concentrations, inhibit fatty acid synthesis, and increase fatty acid oxidation in cultured cells and in experimental animals. J. Biol. Chem.. 2003;278:37099-37111.
[Google Scholar]
Harwood J., James H., . Treating the metabolic syndrome: acetyl-CoA carboxylase inhibition. Expert Opin. Ther. Targets. 2005;9:267-281.
[Google Scholar]
Hulver M.W., Berggren J.R., Cortright R.N., Dudek R.W., Thompson R.P., Pories W.J., MacDonald K.G., Cline G.W., Shulman G.I., Dohm G.L., . Skeletal muscle lipid metabolism with obesity. Am. J. Physiol. Endocrinol. Metabol.. 2003;284:E741-E747.
[Google Scholar]
Kusunoki J., Kanatani A., Moller D.E., . Modulation of fatty acid metabolism as a potential approach to the treatment of obesity and the metabolic syndrome. Endocrine. 2006;29:91-100.
[Google Scholar]
Mao J., Chirala S.S., Wakil S.J., . Human acetyl-CoA carboxylase 1 gene: presence of three promoters and heterogeneity at the 5â€²-untranslated mRNA region. Proc. Natl. Acad. Sci. USA. 2003;100:7515-7520.
[Google Scholar]
McGarry J.D., Foster D.W., . Regulation of hepatic fatty acid oxidation and ketone body production. Annu. Rev. Biochem.. 1980;49:395-420.
[Google Scholar]
Nandi S., Bagchi M.C., . 3D-QSAR and molecular docking studies of 4-anilinoquinazoline derivatives: a rational approach to anticancer drug design. Mol. Diversity. 2010;14:27-38.
[Google Scholar]
Powell M.J.D., . Restart procedures for the conjugate gradient method. Math. Prog.. 1977;12:241-254.
[Google Scholar]
Rasmussen B.B., Holmback U.C., Volpi E., Morio-Liondore B., Paddon-Jones D., Wolfe R.R., . Malonyl coenzyme A and the regulation of functional carnitine palmitoyltransferase-1 activity and fat oxidation in human skeletal muscle. J. Clin. Invest.. 2002;110:1687-1694.
[Google Scholar]
Savage D.B., Choi C.S., Samuel V.T., Liu Z., Zhang D., Wang A., Zhang X., Cline G.W., Yu X.X., Geisler J.G., . Reversal of diet-induced hepatic steatosis and hepatic insulin resistance by antisense oligonucleotide inhibitors of acetyl-CoA carboxylases 1 and 2. J. Clin. Invest.. 2006;116:817.
[Google Scholar]
Singh R., Elizabeth Sobhia M., . Homology modeling of human CCR2 receptor. Med. Chem. Res.. 2011;20:1704-1712.
[Google Scholar]
Singh R., Sobhia M.E., . Synergistic application of target structure-based alignment and 3D-QSAR study of protein tyrosine phosphatase 1B (PTP1B) inhibitors. Med. Chem. Res.. 2011;20:714-725.
[Google Scholar]
Sinha R., Dufour S., Petersen K.F., LeBon V., Enoksson S., Ma Y.Z., Savoye M., Rothman D.L., Shulman G.I., Caprio S., . Assessment of skeletal muscle triglyceride content by 1H nuclear magnetic resonance spectroscopy in lean and obese adolescents. Diabetes. 2002;51:1022-1027.
[Google Scholar]
Steyn N.P., Mann J., Bennett P.H., Temple N., Zimmet P., Tuomilehto J., Lindstrom J., Louheranta A., . Diet, nutrition and the prevention of type 2 diabetes. Public Health Nutr.. 2004;7:147-166.
[Google Scholar]
Turkoglu C., Duman B.S., Gunay D., Cagatay P., Ozcan R., Buyukdevrim A.S., . Effect of abdominal obesity on insulin resistance and the components of the metabolic syndrome: evidence supporting obesity as the central feature. Obes. Surg.. 2003;13:699-705.
[Google Scholar]

Introduction

Materials and methods

Data set
Molecular modelling
Molecular alignment for 3D-QSAR analysis
Calculation of CoMFA and CoMSIA descriptors
Molecular docking
Designing of potent analogues using LeapFrog

Results and discussion

CoMFA and CoMSIA analyses
Docking studies
Designing of potent analogues

Conclusions

Fulltext Views
702

PDF downloads
339

View/Download PDF
Download Citations

BibTeX
RIS

Show Sections

[1] Abu-Elheiga L., Almarza-Ortega D.B., Baldini A., Wakil S.J., . Human acetyl-CoA carboxylase 2 molecular cloning, characterization, chromosomal mapping, and evidence for two isoforms. J. Biol. Chem.. 1997;272:10669-10677.
[Google Scholar]

[2] Abu-Elheiga L., Matzuk M.M., Abo-Hashema K.A.H., Wakil S.J., . Continuous fatty acid oxidation and reduced fat storage in mice lacking acetyl-CoA carboxylase 2. Science. 2001;291:2613-2616.
[Google Scholar]

[3] Abu-Elheiga L., Matzuk M.M., Kordari P., Oh W.K., Shaikenov T., Gu Z., Wakil S.J., . Mutant mice lacking acetyl-CoA carboxylase 1 are embryonically lethal. Proc. Natl. Acad. Sci. USA. 2005;102:12011.
[Google Scholar]

[4] Ambure P.S., Gangwal R.P., Sangamwar A.T., . 3D-QSAR and molecular docking analysis of biphenyl amide derivatives as p38α mitogen activated protein kinase inhibitors. Mol. Diversity. 2011;16:377-388.
[Google Scholar]

[5] Bengtsson C., Blaho S., Saitton D.B., Brickmann K., Broddefalk J., Davidsson Ö., Drmota T., Folmer R., Hallberg K., Hallén S., . Design of small molecule inhibitors of acetyl-CoA carboxylase 1 and 2 showing reduction of hepatic malonyl-CoA levels in vivo in obese Zucker rats. Bioorg. Med. Chem.. 2011;19:3039-3053.
[Google Scholar]

[6] Chonan T., Tanaka H., Yamamoto D., Yashiro M., Oi T., Wakasugi D., Ohoka-Sugita A., Io F., Koretsune H., Hiratate A., . Design and synthesis of disubstituted (4-piperidinyl)-piperazine derivatives as potent acetyl-CoA carboxylase inhibitors. Bioorg. Med. Chem. Lett.. 2010;20:3965-3968.
[Google Scholar]

[7] Chonan T., Wakasugi D., Yamamoto D., Yashiro M., Oi T., Tanaka H., Ohoka-Sugita A., Io F., Koretsune H., Hiratate A., . Discovery of novel (4-piperidinyl)-piperazines as potent and orally active acetyl-CoA carboxylase 1/2 non-selective inhibitors: F-Boc and triF-Boc groups are acid-stable bioisosteres for the Boc group. Bioorg. Med. Chem.. 2011;19:1580-1593.
[Google Scholar]

[8] Corbett J.W., . Review of recent acetyl-CoA carboxylase inhibitor patents: mid-2007–2008. Expert Opin. Ther. Pat.. 2009;19:943-956.
[Google Scholar]

[9] Dhoke G.V., Gangwal R.P., Sangamwar A.T., . A combined ligand and structure based approach to design potent PPAR-alpha agonists. J. Mol. Struct.. 2012;1028:22-30.
[Google Scholar]

[10] Gasteiger J., Marsili M., . Iterative partial equalization of orbital electronegativity – a rapid access to atomic charges. Tetrahedron. 1980;36:3219-3228.
[Google Scholar]

[11] Goodford P.J., . A computational procedure for determining energetically favorable binding sites on biologically important macromolecules. J. Med. Chem.. 1985;28:849-857.
[Google Scholar]

[12] Gupta P., Garg P., Roy N., . Comparative docking and CoMFA analysis of curcumine derivatives as HIV-1 integrase inhibitors. Mol. Diversity. 2011;15:1-18.
[Google Scholar]

[13] Harwood H.J., Petras S.F., Shelly L.D., Zaccaro L.M., Perry D.A., Makowski M.R., Hargrove D.M., Martin K.A., Tracey W.R., Chapman J.G., . Isozyme-nonselective N-substituted bipiperidylcarboxamide acetyl-CoA carboxylase inhibitors reduce tissue malonyl-CoA concentrations, inhibit fatty acid synthesis, and increase fatty acid oxidation in cultured cells and in experimental animals. J. Biol. Chem.. 2003;278:37099-37111.
[Google Scholar]

[14] Harwood J., James H., . Treating the metabolic syndrome: acetyl-CoA carboxylase inhibition. Expert Opin. Ther. Targets. 2005;9:267-281.
[Google Scholar]

[15] Hulver M.W., Berggren J.R., Cortright R.N., Dudek R.W., Thompson R.P., Pories W.J., MacDonald K.G., Cline G.W., Shulman G.I., Dohm G.L., . Skeletal muscle lipid metabolism with obesity. Am. J. Physiol. Endocrinol. Metabol.. 2003;284:E741-E747.
[Google Scholar]

[16] Kusunoki J., Kanatani A., Moller D.E., . Modulation of fatty acid metabolism as a potential approach to the treatment of obesity and the metabolic syndrome. Endocrine. 2006;29:91-100.
[Google Scholar]

[17] Mao J., Chirala S.S., Wakil S.J., . Human acetyl-CoA carboxylase 1 gene: presence of three promoters and heterogeneity at the 5â€²-untranslated mRNA region. Proc. Natl. Acad. Sci. USA. 2003;100:7515-7520.
[Google Scholar]

[18] McGarry J.D., Foster D.W., . Regulation of hepatic fatty acid oxidation and ketone body production. Annu. Rev. Biochem.. 1980;49:395-420.
[Google Scholar]

[19] Nandi S., Bagchi M.C., . 3D-QSAR and molecular docking studies of 4-anilinoquinazoline derivatives: a rational approach to anticancer drug design. Mol. Diversity. 2010;14:27-38.
[Google Scholar]

[20] Powell M.J.D., . Restart procedures for the conjugate gradient method. Math. Prog.. 1977;12:241-254.
[Google Scholar]

[21] Rasmussen B.B., Holmback U.C., Volpi E., Morio-Liondore B., Paddon-Jones D., Wolfe R.R., . Malonyl coenzyme A and the regulation of functional carnitine palmitoyltransferase-1 activity and fat oxidation in human skeletal muscle. J. Clin. Invest.. 2002;110:1687-1694.
[Google Scholar]

[22] Savage D.B., Choi C.S., Samuel V.T., Liu Z., Zhang D., Wang A., Zhang X., Cline G.W., Yu X.X., Geisler J.G., . Reversal of diet-induced hepatic steatosis and hepatic insulin resistance by antisense oligonucleotide inhibitors of acetyl-CoA carboxylases 1 and 2. J. Clin. Invest.. 2006;116:817.
[Google Scholar]

[23] Singh R., Elizabeth Sobhia M., . Homology modeling of human CCR2 receptor. Med. Chem. Res.. 2011;20:1704-1712.
[Google Scholar]

[24] Singh R., Sobhia M.E., . Synergistic application of target structure-based alignment and 3D-QSAR study of protein tyrosine phosphatase 1B (PTP1B) inhibitors. Med. Chem. Res.. 2011;20:714-725.
[Google Scholar]

[25] Sinha R., Dufour S., Petersen K.F., LeBon V., Enoksson S., Ma Y.Z., Savoye M., Rothman D.L., Shulman G.I., Caprio S., . Assessment of skeletal muscle triglyceride content by 1H nuclear magnetic resonance spectroscopy in lean and obese adolescents. Diabetes. 2002;51:1022-1027.
[Google Scholar]

[26] Steyn N.P., Mann J., Bennett P.H., Temple N., Zimmet P., Tuomilehto J., Lindstrom J., Louheranta A., . Diet, nutrition and the prevention of type 2 diabetes. Public Health Nutr.. 2004;7:147-166.
[Google Scholar]

[27] Turkoglu C., Duman B.S., Gunay D., Cagatay P., Ozcan R., Buyukdevrim A.S., . Effect of abdominal obesity on insulin resistance and the components of the metabolic syndrome: evidence supporting obesity as the central feature. Obes. Surg.. 2003;13:699-705.
[Google Scholar]

3D-QSAR and molecular docking analysis of (4-piperidinyl)-piperazines as acetyl-CoA carboxylases inhibitors

Abstract

Keywords

Acetyl-CoA carboxylases

(4-Piperidinyl)-piperazines

CoMFA

CoMSIA

Molecular docking

LeapFrog

1 Introduction

2 Materials and methods

2.1 Data set

2.2 Molecular modelling

2.3 Molecular alignment for 3D-QSAR analysis

2.4 Calculation of CoMFA and CoMSIA descriptors

2.5 Molecular docking

2.6 Designing of potent analogues using LeapFrog

3 Results and discussion

3.1 CoMFA and CoMSIA analyses

3.2 Docking studies

3.3 Designing of potent analogues

4 Conclusions

Acknowledgement

References

Suggested read for related articles:

We value your privacy