2D-QSAR, 3D-QSAR, molecular docking and ADMET prediction studies of some novel 2-((1H-indol-3-yl)thio)-N-phenyl-acetamide derivatives as anti-influenza A virus

Figures & data

Table 1. Chemical structures of the 2-((1 H-indol-3-yl)thio)-N-phenyl-acetamide derivatives along with their anti-IAV values.

Sr No.	R1	R2	Y	R3	EC50 (μM)	pEC50
1	CH3	H	S	2-Cl	27.94	4.5538
2	CH3	H	S	3-Cl	9.98	5.0009
3	CH3	H	S	4-Cl	9.82	5.0079
4	CH3	H	S	2,3-di-Cl	13.96	4.8551
5	CH3	H	S	2,4-di-Cl	11.93	4.9234
6	CH3	H	S	2,5-di-Cl	15.75	4.8027
7	CH3	H	S	2,6-di-Cl	3.5	5.4559
8	CH3	H	S	3,4-di-Cl	0.64	6.1938
9	CH3	H	S = O	2-Cl	29.78	4.5261
10	CH3	H	S = O	3-Cl	39.27	4.4059
11	CH3	H	S = O	4-Cl	9.93	5.0031
12	CH3	H	SO2	2,3-di-Cl	36.34	4.4396
13	CH3	H	SO2	2,4-di-Cl	33.71	4.4722
14	CH3	H	SO2	2,5-di-Cl	18.03	4.7440
15	H	6-CH3O	S	2- CH3CH2O	4.48	5.3487
16	H	5-CH3O	S	2- CH3CH2O	1.27	5.8962
17	H	4-CH3O	S	2- CH3CH2O	11.21	4.9504
18	H	7-Cl	S	2- CH3CH2O	1.75	5.7570
19	H	6-Cl	S	2- CH3CH2O	1.53	5.8153
20	H	5-Cl	S	2- CH3CH2O	4.19	5.3778
21	H	4-Cl, 6-CH3O	S	2- CH3CH2O	0.58	6.2366
22	-	-	-	2- CH3CH2O	31.42	4.5028
23	H	7-CH3O	S	3,4-di-Cl	13.81	4.8598
24	H	7-CH3O	S	2- CH3CH2O	12.2	4.9136
25	H	6-CH3O	S	3,4-di-Cl	15.8	4.8013
26	H	5-CH3O	S	3,4-di-Cl	15.45	4.8111
27	H	4-CH3O	S	3,4-di-Cl	15.08	4.8216
28	H	6-Cl	S	3,4-di-Cl	7.36	5.1331
29	H	5-Cl	S	3,4-di-Cl	7.62	5.1180
30	H	4-Cl	S	3,4-di-Cl	5.72	5.2426
31	-	-	-	3,4-di-Cl	31.56	4.5009
32	-	-	-	-	25.08	4.6007

Figure 1. Optimized structures (a) structure of compound 21, (b) alignment and superposition of the dataset compounds (capped sticks model).

Table 2. Internal validation of the 2D-QSAR models.

Internal validation	GFA- MLR Model	GFA-ANN Model (5-4-1)	Threshold	Comment	Reference
Lack of fit (LOF)	0.12354	-
Pearson Correlation (r)	0.9410	0.9570	R > 0.6	Passed	[32]
Pearson Correlation squared (${\mathrm{r}}_{\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{i}\mathrm{n}}^2$)	0.8861	0.9163	R^2train > 0.6	Passed	[32]
Adjusted r^2	0.8525	-	R^2adj > 0.6	Passed	[32]
Spearman Rank Correlation (ρ)	0.8754	0.9199	ρ > 0.6	Passed
Root Mean Squared Deviation (RMSE)	0.1350	0.1420	Low	Passed
Cross Validated Squared (q^2)	0.7864	0.8739	Q^2 > 0.6	Passed	[32]
Y-randomization ($c{R}_p^2)$	0.7818	-	$c{R}_p^2$> 0.6	Passed	[17]

Table 3. External validation parameters of the 2D-QSAR models.

External validation metrics	GFA-MLR Model	GFA-ANN Model (5-4-1)	Threshold	Comment	Reference
Pearson Correlation squared (${\mathrm{r}}_{\mathrm{t}\mathrm{e}\mathrm{s}\mathrm{t}}^2$)	0.7230	0.7368	R^2test > 0.6	Passed	[32]
$r_{pred}^2$	0.6872	0.6257	$R_{pred}^2$> 0.5	Passed	[32]
Δ${\bar{R}}_m^2$ (test)	0.4881	-	<0.5	Passed	[32]
$r_0^2$	0.6615	-	>0.5	Passed	[32]
RMSEP	0.4042	0.4421	-	-
r’^20	0.7195	-	>0.5	Passed	[32]
(r^2-r0^2)/r^2	0.0850	-	(r^2-r0^2)/r^2 < 0.1	Passed	[17]
k	1.0486	-	0.85 < k < 1.15	Passed	[17]
(r^2- r’^20)/r^2	0.0047	-	(r^2- r’^20)/r^2 < 0.1	Passed	[17]
k’	0.9503	-	0.85 < k’ < 1.15	Passed	[17]
$\left|r_0^2-\mathrm{r}{\prime }_0\right|$	0.0580	-	$\left|r_0^2-\mathrm{r}{\prime }_0\right|$< 0.3	Passed	[17]

Figure 2. RMSE values of training set against the number of epochs and number of neurons in the hidden layer (1 to 5).

Figure 3. Schematic representation of the 5-4-1 neural architecture for the GFA-ANN model.

Figure 4. Relevance plot of the feature selected descriptors in the GFA-MLR and GFA-ANN models.

Figure 5. Scatter plot of the leverage scores against standardized residuals of the GFA-ANN model (Williams plot).

Table 4. Statistical validation results of probable CoMFA models.

Descriptors	q^2	r^2	SEE	N
Steric (S)	0.606	0.886	0.1848	3
Electrostatic (E)	0.842	0.995	0.0432	7
S + E	0.59	0.925	0.1538	4

q²: Leave one out cross-validated correlation coefficient

r²: Non-cross validated correlation coefficient

SEE: Standard error of estimation

N: number of optimum components

Table 5. Statistical validation results of the best 3D-QSAR models.

					Field fractions
Model	q^2	r^2train	F	r^2test	S	E	H	D	A
CoMFA_E	0.842	0.995	456.82	0.8317	-	1.00	-	-	-
CoMFA_ES	0.59	0.925	58.77	0.7034	0.644	0.356	-	-	-
CoMSIA_EA	0.759	0.925	82.73	0.6269	-	0.786	-	-	0.214
CoMSIA_SEA	0.751	0.923	79.38	0.6243	0.397	0.391	-	-	0.212
CoMSIA_EAD	0.767	0.929	86.96	0.6162	-	0.759	-	0.074	0.167
CoMSIA_SEDA	0.752	0.924	81.17	0.6311	0.372	0.384	-	0.074	0.170
CoMSIA_EHDA	0.760	0.939	102.96	0.585	-	0.667	0.160	0.051	0.121
CoMSIA_SEHDA	0.747	0.939	102.86	0.5841	0.301	0.368	0.153	0.053	0.126

Figure 6. 3D fields of the CoMFA model for the most active compound (21), (a) green area depict desirable steric bulk while yellow area disfavor steric bulk, (b) blue region favors positive charge and red region favors negative charge.

Figure 7. 3D fields contribution of the CoMSIA_EAD model for the most active compound (compound 21), (a) electrostatic contour map where blue region favor positive charge and red region favor negative charge, (b) magenta contour represent the region for desirable hydrogen bond acceptor while red area represents undesirable acceptor, (c) cyan contours represent the area for desirable hydrogen bond donor while purple area represents the undesirable donor.

Figure 8. General description of the 3D QSAR analysis.

Table 6. Molecular docking scores of the 2-((1 H-indol-3-yl)thio) acetamide derivatives as anti-IAV agents.

S/N	Score	rmsd_refine	E_conf	E_place	E_score1	E_refine	E_score2
1	−6.4704	1.7910	20.5151	−63.1710	−9.5719	−28.6277	−6.4704
2	−6.3978	1.9279	−2.0012	−83.1593	−9.8690	−28.2970	−6.3978
3	−6.3507	1.3791	5.0251	−45.0998	−11.0389	−25.4566	−6.3507
4	−6.4224	1.9460	21.1921	−47.7252	−9.9613	−32.6838	−6.4224
5	−6.5399	1.7959	10.1379	−56.9896	−10.3130	−32.6339	−6.5399
6	−6.6172	1.2467	14.0751	−77.7281	−10.8902	−28.9187	−6.6172
7	−6.5942	1.5289	32.1700	−68.0969	−9.6906	−30.2713	−6.5942
8	−6.3784	1.5814	3.9788	−64.3388	−10.0639	−28.1994	−6.3784
9	−6.6945	1.3668	7.6522	−108.4647	−11.7388	−35.2875	−6.6945
10	−6.7401	1.7653	−15.3426	−93.2008	−11.2134	−36.8298	−6.7401
11	−6.3637	0.9860	−20.5806	−68.7072	−9.9775	−23.3522	−6.3637
12	−6.6639	2.1425	16.7122	−70.4718	−10.8073	−36.9502	−6.6639
13	−6.7937	1.4824	−2.5416	−56.2266	−12.1327	−40.0038	−6.7937
14	−6.8329	1.9367	2.7722	−77.2653	−10.4397	−37.1532	−6.8329
15	−6.7451	1.6896	17.9035	−66.6872	−11.6219	−33.5884	−6.7451
16	−7.1303	2.2249	12.2309	−45.2351	−10.0616	−36.5056	−7.1303
17	−6.7349	2.2951	32.1663	−59.6583	−10.3008	−38.0070	−6.7349
18	−6.4453	3.3581	21.6193	−57.5597	−10.1230	−36.6166	−6.4453
19	−6.7273	1.1573	19.5224	−69.4917	−10.7190	−32.7859	−6.7273
20	−6.7150	1.6514	17.4370	−66.1153	−10.0576	−33.0545	−6.7150
21	−7.0431	1.7231	18.5511	−82.7573	−10.4805	−37.9177	−7.0431
22	−6.6093	1.0685	26.5938	−87.7603	−11.6759	−39.3687	−6.6093
23	−6.5676	2.0696	6.8814	−61.3086	−10.5411	−33.6283	−6.5676
24	−6.3978	1.3737	7.3658	−52.5790	−9.6361	−28.8639	−6.3978
25	−6.4427	2.2170	8.9476	−40.3052	−10.1557	−33.6551	−6.4427
26	−6.5794	1.4865	6.7760	−65.6818	−10.3202	−38.2883	−6.5794
27	−6.5971	1.9896	23.7852	−61.3695	−10.3674	−33.6521	−6.5971
28	−6.2924	4.1064	11.7588	−39.7686	−9.8481	−30.3737	−6.2924
29	−6.5729	1.5330	8.3900	−94.8744	−9.5519	−32.2424	−6.5729
30	−6.3875	1.7441	22.8514	−46.2033	−9.9025	−26.7708	−6.3875
31	−6.2991	1.7688	19.4041	−84.8582	−10.1595	−34.2715	−6.2991
32	−5.8682	1.4007	14.4801	−80.4247	−9.5620	−36.4932	−5.8682
ZMR	−9.1687	1.4566	−359.6887	−118.9620	−17.9872	−69.7331	−9.1687

Figure 9. 3D docking view of compound 16 with the H1N1 neuraminidase (PDB: 4B7Q): (a) the best pose of compound 16, (b) 3D hydrogen bond surfaces around the ligand, (c) 2D residual interaction of compound 21-complex.

Figure 10. 3D docking view of compound 21 with the H1N1 neuraminidase (PDB: 4B7Q): (a) the best pose of compound 21, (b) 3D hydrogen bond surfaces around the ligand 21, (c) 2D residual interaction of compound 21-complex.

Table 7. Binding interaction of the H1N1 neuraminidase receptor with compound 16.

Distance (Å)	Interaction type	From	Chemistry	To	Chemistry
2.81596	Hydrogen Bond	A: ARG118	H-Donor	16	H-Acceptor
2.16458	Hydrogen Bond	A: ARG293	H-Donor	16	H-Acceptor
2.17962	Hydrogen Bond	A: ARG368	H-Donor	16	H-Acceptor
2.83677	Carbon Hydrogen Bond	16	H-Donor	A:ASP151	H-Acceptor
2.66168	Carbon Hydrogen Bond	16	H-Donor	A: TRP179	H-Acceptor
2.74365	Carbon Hydrogen Bond	16	H-Donor	A: GLU119	H-Acceptor
2.99412	Carbon Hydrogen Bond	16	H-Donor	A:ASP151	H-Acceptor
3.90682	Electrostatic	A: ARG152	Positive	16	Pi-Orbitals
4.71855	Electrostatic	A: ARG152	Positive	16	Pi-Orbitals
3.06108	Electrostatic	A: ARG368	Positive	16	Pi-Orbitals
4.17481	Electrostatic	A:ASP151	Negative	16	Pi-Orbitals
4.35022	Electrostatic	A:ASP151	Negative	16	Pi-Orbitals
4.69859	Electrostatic	A: GLU228	Negative	16	Pi-Orbitals
3.48848	Electrostatic	A: GLU278	Negative	16	Pi-Orbitals
3.04774	Electrostatic	A: GLU278	Negative	16	Pi-Orbitals
5.02341	Hydrophobic	16	Alkyl	A: ILE149	Alkyl
5.46962	Hydrophobic	16	Pi-Orbitals	A: ARG225	Alkyl
4.42977	Hydrophobic	16	Pi-Orbitals	A:PRO431	Alkyl

Table 8. Binding residual interaction of the H1N1 neuraminidase receptor with compound 21.

Distance (Å)	Interaction type	From	Chemistry	To	Chemistry
3.04455	Hydrogen Bond	A: ARG118	H-Donor	21	H-Acceptor
2.49282	Hydrogen Bond	A: ARG293	H-Donor	21	H-Acceptor
2.58207	Hydrogen Bond	21	H-Donor	A: ASP151	H-Acceptor
4.55051	Electrostatic	A: ARG152	Positive	21	Pi-Orbitals
3.42318	Electrostatic	A: ARG368	Positive	21	Pi-Orbitals
3.44829	Electrostatic	A: ARG368	Positive	21	Pi-Orbitals
4.16643	Electrostatic	A: ASP151	Negative	21	Pi-Orbitals
3.11298	Electrostatic	A: GLU278	Negative	21	Pi-Orbitals
4.37383	Hydrophobic	21	Alkyl	A: ILE427	Alkyl
4.37138	Hydrophobic	21	Alkyl	A:PRO431	Alkyl
4.99098	Hydrophobic	21	Pi-Orbitals	A: ARG368	Alkyl
3.95830	Hydrophobic	21	Pi-Orbitals	A:PRO431	Alkyl
5.22865	Hydrophobic	21	Pi-Orbitals	A: PRO431	Alkyl

Figure 11. Physicochemical radar chart of the lead compounds.

Table 9. Lipinski’s rule parameters of compounds 16 and 27.

Molecule	MW	LogP	nHA	nHD	TPSA (Å^2)	nLV
16	356.12	3.558	5	2	63.35	0
21	390.08	4.172	5	2	63.35	0
Oseltamivir	330.15	−1.317	10	9	187.56	0
Zanamivir	312.2	1.013	6	3	90.65	0
Rule	≤ 500	≤ 5	≤ 10	≤ 5	< 140	≤1

Table 10. Drug-likeness evaluation of compounds 16 and 21 based on Ghose, Veber, Egan, Muegge, QED and SA scores.

Molecule	QED	SA	Ghose	Veber	Egan	Muegge
16	0.619	2.08	Yes	Yes	Yes	Yes
21	0.56	2.325	Yes	Yes	Yes	Yes
Oseltamivir	0.213	4.392	Yes	Yes	Yes	Yes
Zanamivir	0.691	3.758	Yes	Yes	Yes	Yes

Table 11. ADMET properties of the lead molecules in the dataset.

Category	Properties	Prediction probability values (symbols)
	Human Intestinal Absorption (HIA)	0.004(–)	0.004(–)
	MDCK Permeability	9.22 × 10^−06	1.06 × 10^−05
Absorption	Caco-2 Permeability	−4.864	−4.994
	Pgp-inhibitor	0.921 (+++)	0.892 (++)
	Pgp-substrate	0.305 (-)	0.27 (–)
	Plasma Protein Binding (PPB)	98.17%	99.79%
Distribution	Volume of Distribution (VD)	0.749	0.91
	Blood Brain Barrier (BBB)	0.165(–)	0.088(–)
	CYP1A2 inhibitor	0.898 (++)	0.94(+++)
	CYP1A2 substrate	0.898 (++)	0.845(++)
	CYP2C19 inhibitor	0.968 (+++)	0.97(+++)
	CYP2C19 substrate	0.422 (-)	0.275(–)
Metabolism	CYP2C9 inhibitor	0.936 (+++)	0.95(+++)
	CYP2C9 substrate	0.954(+++)	0.952(+++)
	CYP2D6 inhibitor	0.756(++)	0.741(++)
	CYP2D6 substrate	0.902(+++)	0.887(++)
	CYP3A4 inhibitor	0.937(+++)	0.908(+++)
	CYP3A4 substrate	0.553(+)	0.305(-)
Excretion	CL	8.746	7.328
	T1/2	0.65	0.47
	AMES Toxicity	0.746(++)	0.698(+)
Toxicity	Carcinogenicity	0.177(–)	0.191(–)
	Eye Irritation	0.633(+)	0.608(+)
	Respiratory Toxicity	0.602(+)	0.718(++)

Key: The prediction probability values are classified into six symbols: 0–0.1(–), 0.1–0.3(–), 0.3–0.5(-), 0.5–0.7(+), 0.7–0.9(++), and 0.9–1.0(+++). Generally, ‘+++’ or ‘++’ represents the molecule is more likely to be toxic or defective, while ‘−−’or‘−’ represents nontoxic or appropriate.

Tropsha A. Best practices for QSAR model development, validation, and exploitation. Mol Inform. 2010;29(6‐7):476–488.

 PubMed Web of Science ®Google Scholar

Roy K, Das RN, Ambure P, et al. Be aware of error measures. Further studies on validation of predictive QSAR models. Chemometr Intell Lab Syst. 2016;152:18–33.

 Web of Science ®Google Scholar