Identification of 4-aminoquinoline core for the design of new cholinesterase inhibitors

In vitro cholinesterase Inhibition Assay

The assay followed the method of Ellman et al. (1961) using a Thermo Scientific Varioskan Flash. AChE (C3389, Type VI-S, from Sigma) and BuChE (C0663, from human erythrocytes), 5,5′-dithiobis (2-nitrobenzoic acid) (Sigma reagent, DTNB, D218200), acetylthiocholine (ATC), and butyrylthiocholine (BTC) iodides were purchased from Sigma-Aldrich (Shanghai, China). AChE/BuChE stock solution was prepared by adjusting 500 units of the enzyme and 1 mL of gelatin solution (1% in water) to 100 mL with water. This enzyme solution was further diluted before use to give 2.5 units/mL. ATC/BTC iodide solution (0.075 M) was prepared in water. DTNB solution (0.01 M) was prepared in water containing 0.15% (w∕v) sodium bicarbonate. For buffer preparation, potassium dihydrogen phosphate (1.36 g, 10 mmol) was dissolved in 100 mL of water and adjusted with KOH to pH = 8.0 ± 0.1. Stock solutions of the test compounds were prepared in ethanol, 100 µL of which gave a final concentration of 10⁻⁴ M when diluted to the final volume of 132 µL. For each compound, a dilution series of at least five different concentrations (normally 10⁻⁴ ∼ 10⁻⁹ M) was prepared.

For measurement, a cuvette containing 100 µL of phosphate buffer, 10 µL of the respective enzyme, and 10 µL of the test compound solution was allowed to stand for 5 min before 10 µL of DTNB were added. The reaction was started by addition of 2 µL of the substrate solution (ATC/BTC). The solution was mixed immediately, and exactly 2 min after substrate addition the absorption was measured at 25 °C at 412 nm. For the reference value, 10 µL of water replaced the test compound solution. For determining the blank value, additionally 10 µL of water replaced the enzyme solution. Each concentration was measured in triplicate at 25 °C. The inhibition curve was obtained by plotting percentage enzyme activity (100% for the reference) versus logarithm of test compound concentration. Calculation of the IC₅₀ values was performed with Graph Pad Prism 5.0.

Kinetic study

Kinetic measurements were performed in the same manner, while the substrate (ATC/BTC) was used in concentrations of 25, 50, 90, 150, 226, 452 and 678 µM for each test compound concentration and the reaction was extended to 4 min before measurement of the absorption. Vmax and Km values of the Michaelis–Menten kinetics were calculated by nonlinear regression from substrate-velocity curves using Graph Pad Prism 5.0. Linear regression was used for calculating the Lineweaver–Burk plots.

Molecular docking

Computational methods are useful tools for drug discovery and evaluation that has been widely applied in drug discovery campaign (Zheng et al., 2013; Ford & Ho, 2016). The docking study was performed by CDOCKER module implemented in Discovery Studio 3.0. The principle of CDOCKER can be breifly summarized as follow: CDOCKER generates ligand “seeds” to populate the binding pocket. Each seed is then subjected to high temperature molecular dynamics (MD) using a modified version of CHARMm force field (Wu et al., 2003). The structure after MD run is then fully minimized under the forcefield. The solutions are then clustered according to position and conformation and ranked by energy. The cocrystal structure of Torpedo Californica AChE bound with bis(7)-tacrine (TcAChE, PDB id: 2CKM) was used for molecular docking. The binding sites were defined by residues around the CAS and PAS of AChE (in 6 Å radius). The heating step, cooling steps, and cooling temperature were set to 5,000, 5,000, and 310, respectively. Other parameters were kept as default.

Compound information

All compounds except 03, 10, 15, 17 and 22 were purchased from Sigma Aldrich (Sigma Aldrich, Shanghai, China: http://www.sigmaaldrich.com/china-mainland.html), with purity >95.0%. Compounds 03, 10, 15, 17 and 22 were bought from J&K Scientific (J&K Scientific, Shanghai, China: http://www.jkchemical.com/). The detailed information is listed in Table S1.

Identification of 4-aminoquinoline as the potent core targeting AChE

Multiple aminoquinolines, with amino group substituted at different position of quinolone ring (Table 1, compound 01∼07), were firstly collected. Preliminary evaluations of these compound were performed by determine their AChE (eeAChE) % inhibition at 10 µM following Ellman’s method (Ellman et al., 1961). It was previously reported that T5369186 (23) and tacrine were used as positive controls. Results showed that only 4-aminoquinoline core had strong AChE inhibition (% inhibition of 03 and 07 68.29 ± 1.83% and 90.59 ± 0.28%, respectively). Compound 07 was further determined for the AChE inhibitory curve and IC₅₀ (23 used as positive control). The compound showed dose-dependent manner and well fitted inhibitory curve (Fig. 1), with IC₅₀ 0.72 ± 0.06 µM.

Kinetics study

To gain information on the mechanism of inhibition, compound 07 was selected for kinetic studies of AChE inhibition by using Lineweaver–Burk plots, which were reciprocal rates versus reciprocal substrate concentrations for the different inhibitor concentrations resulting from the substrate-velocity curves for AChE (Fig. 2). The compound exhibited a mixed-type inhibition of AChE, for the plot showed both increased slopes (decreased Vmax) and intercepts (higher Km) when the concentration of the inhibitors were increased, indicating that the compounds may bind to both CAS and PAS (Chen et al., 2015). The detailed Km and Vmax values from the non-linear regression fitting in Lineweaver–Burke is shown in Table S2.

Table 1: The preliminary assay of collected compounds with aminoquinoline or other similar cores.

Cpd.	Structure	AChE % inhibitiona
01		4.88 ± 1.46
02		18.89 ± 2.77
03		68.29 ± 1.83
04		13.14 ± 3.48
05		4.02 ± 2.96
06		3.21 ± 2.32
07		90.59 ± 0.28
08		16.04 ± 2.72
09		13.55 ± 2.0
10		37.51 ± 1.52
11		31.74 ± 2.17
12		17.36 ± 1.98
13		3.32 ± 1.79
14		4.26 ± 4.43
15		13.64 ± 3.29
16		33.15 ± 3.87
17		35.88 ± 4.01
18		26.02 ± 6.82
19		7.56 ± 2.11
20		16.63 ± 3.01
21		-8.59 ± 1.23
22		5.48 ± 0.77
23		98.53 ± 1.81
Tacrine		97.36 ± 2.54; 0.08 ± 0.01b

DOI: 10.7717/peerj.2140/table-1

Notes:

a% inhibition at 10 µM.

bIC₅₀ of tacrine (µM).

Structure–activity relationship and binding mode analysis by molecular docking

To deeply understand the binding mode of between AChE and the potent compounds, molecular docking was applied to further analyze compound 07 and 23 (Fig. 3). The binding conformation suggested that the two compounds bound to catalytic site (CAS) of AChE in a very similar manner. In detail, 07 formed strong π–π interactions with Trp84, Phe330 and Tyr334 of the CAS of AChE. The amino group formed a H-bond with His440, which was considered as a critical member of the catalytic triad of AChE. The methyl group inserted into a small sub-pocket surrounded by Asp72 and Ser81 and contacted the backbone of them through hydrophobic interactions, which enhanced the activity of the inhibitor. Compound lacked this group (03, 68.29 ± 1.83% inhibition at 10 µM) showed reduced activity, and this further confirmed that methyl was a pharmocophoric group for 4-aminoquinoline core. The acetyl group of 23 inserted into a hydrophobic groove formed by the aromatic side chains of Trp84, Phe330 and Tyr334, which contributed to the binding affinity of 23 (region in the red dot line in Fig. 3). However, this group was missed in 07, leaving the pocket unoccupied. This could be the reason for the decreased activity of compound 07 compared to 23.

Ligand efficiency (LE) is an important parameter when evaluate the advantage of lead compounds or active fragments (Reynolds, Tounge & Bembenek, 2008). It is an attempt to normalize the activity of a compound by its molecular size (Congreve et al., 2008). To further recognize the importance of the acetyl group, ligand efficiency was calculated for 07 and 23 based on their −pIC₅₀ and molecular weight (Table 2). Although 23 was slightly more potent than 07, the LE of 23 was lower than 07, indicating that the acetyl group was useful but not important to the inhibitory activity. Considering the hydrophobic character of the AChE sub-pocket around this group, proper optimization, especially those groups easily to form hydrophobic contacts, may help to further enhance the activity as well as LE. The CLogP value of 07 and 23 was also predicted (1.43 and 1.64, respectively). Considering that the further design of MTDLs based on the core will enhance the CLogP because of the introduction of hydrophobic groups, the initial CLogP value of the two compounds is acceptable.

Amino group at other position of quinoline (01, 02, 04 ∼ 06), however, exhibited remarkably reduced activity, with % inhibition ranging from 3.21 ± 2.32 to 33.51 ± 1.52% at 10 µM. The results indicated that spatial location of the amino group on quinolone ring was one of the determinants to the activity of the compound. According to the previously reported results, the catalytic triad including Ser200, Glu327 and His440 played a critical role in hydrolyzing acetylcholine by AChE. Inhibitors directly interacting or closing to this triad can impede the catalytic function of AChE. According to the binding mode (Fig. 3), the amino group at 4-position of quinolone pointed to Glu327, and interacted with His440. This mode can inhibit the approach of acetylcholine to the catalytic site, thus exerted AChE inhibition. Oppositely, 04 (Fig. 4, blue) bound to AChE with a completely different mode from 07 (Fig. 4, green). The 5-amino group pointed to Glu199 and Tyr130 and formed two H-bonds to the residues, leading to a different location of the quinolone ring compared to that of compound 07. Although it interacted with Trp84 through π–π stacking, it was far from Phe330 and Tyr334, and did not form any interaction with these residues, which were important when 07 bound to AChE. Under such binding mode, 04 was moved away from the catalytic triad, and could not inhibit the approach of acetylcholine to the CAS of AChE; therefore, 04 exhibited very poor inhibitory activity. This can also be the reason for the loss of the activity of other aminoquinolines. Additionally, the results indicated that the CAS site was large enough to endure structural modification of 4-aminoquinoline ring, especially on the benzene ring. Proper optimization at this site could improve the binding affinity of the compound through forming polar recognitions or hydrophobic contacts towards the sub-pocket around Tyr130 and Glu199.

To further confirm the importance of 4-amino group, we replaced it with other substituents including halogen, hydroxyl, carboxyl and nitro groups (08∼12). Only 4-hydroxyl and 4-carboxyl exhibited moderate inhibition, while 4-chloro, 4-bromo and 4-nitro compounds loss the inhibitory activity. The results further confirmed that polar contact near the catalytic triad of CAS was a determinant for AChE inhibition.

Next we changed the core to other rings including isoquinoline (13 ∼ 15), naphthalene (16 ∼ 17), 5,6,7,8-tetrahydronaphthalene (18) and 1H-indole (19 ∼ 22) to verify the function of quinolone ring. The isoquinoline and 1H-indole, which had different arrangement of the polar group or atom to that of quinolone, led to completely loss of inhibitory activity, indicating that electrostatic character of the bicyclic core was another determinant for AChE inhibition.