Imidazopyranotacrines as Non-Hepatotoxic, Selective Acetylcholinesterase Inhibitors, and Antioxidant Agents for Alzheimer′s Disease Therapy

Herein we describe the synthesis and in vitro biological evaluation of thirteen new, racemic, diversely functionalized imidazo pyranotacrines as non-hepatotoxic, multipotent tacrine analogues. Among these compounds, 1-(5-amino-2-methyl-4-(1-methyl-1H-imidazol-2-yl)-6,7,8,9-tetrahydro-4H-pyrano[2,3-b]quinolin-3-yl)ethan-1-one (4) is non-hepatotoxic (cell viability assay on HepG2 cells), a selective but moderately potent EeAChE inhibitor (IC50 = 38.7 ± 1.7 μM), and a very potent antioxidant agent on the basis of the ORAC test (2.31 ± 0.29 μmol·Trolox/μmol compound).


Computational Chemistry: Molecular Modeling of imidazopyranotacrine 8
A modeling study was carried out through docking simulations to shed light on the nature and spatial location of the key interactions of the (R)-and (S)-enantiomers of 8, selected as the most potent ChEI, on the AChE binding, by using AutoDockVina [1].
The kinetic data provide evidence that compound 8 displays a non-competitive type inhibition and argue in favor of interactions of compound 8 with the peripheral anionic site (PAS) of AChE. Molecular modeling studies have been carried in order to validate this assumption. Figure S1. Binding mode of inhibitor (R)-8 at the active site of EeAChE. Ligand is rendered as balls and sticks and illustrated in blue. The side chains conformations of the mobile residues are illustrated in the same color light as the ligand. Different subsites of the active site were colored: catalytic triad (CT) in green, oxyanion hole (OH) in pink, anionic sub-site (AS) in orange, except Trp86, acyl binding pocket (ABP) in yellow, and PAS in blue. Red dashed lines are drawn among atoms involved in hydrogen bond interactions.
The most energetically favorable binding mode of compound (R)-8 at the active site of EeAChE is shown in figure S1. This ligand shows a binding energy of −12.2 kcal/mol, it is located in the PAS and no interactions with the catalytic active site (CAS) were found The examination of the first shell of residues surrounding (R)-8 reveals that the pyranotacrine moiety was well fitted in the hydrophobic pocket composed by Tyr72, Tyr124, Trp286, Ile294, Phe295 and Tyr341. It is stacked against the indole and the benzene rings of Trp286 and Tyr341, respectively, through π-π interactions. The naphthalene moiety showed T-shape interaction with Tyr72. Moreover, three hydrogen bonds involving the amino group were observed. Asp74-O is engaged in a bifurcated hydrogen bond with both N-H protons of the amino group. The third one is established with Tyr124.
The binding pose of ligand (S)-8 based on the docking results is presented in Figure S2. This compound binds effectively to the PAS through hydrogen bonding interactions and π-π stacking interactions (predicted binding energy of −11.8 kcal/mol). It is able to bind in the PAS by face-to-face π-π interactions between the pyranotacrine moiety of the ligand and Tyr341 phenyl ring and the Trp286 indole ring, and edge-to-face π-π interactions between naphthalene moiety of the ligand and Tyr72 phenyl ring. A hydrogen bond between the nitrogen atom of the pyridine ring and Tyr124-OH is established. The amino group of the ligand forms hydrogen bonding to the oxygen of the carbonyl Tyr341. Consequently, it can be postulated that racemic compound 8, due to its large size, is unable to enter completely into the narrow active site gorge of the AChE receptor and hence acts only as a PAS binding site, as a non-competitive inhibitor. Lastly, the binding modes overlay of compound 8 was examined. Both enantiomers share the position of cyclohexane and phenyl rings from which the molecules are arranged as mirror images ( Figure S3). Next, we carried out the docking analysis of compound 8 on BuChE, in order to explain why this compound was a poor BuChE inhibitor.
In figure S4 we show the position of the top-scored poses of both enantiomers. The docking simulation of (R)-8 and (S)-8 shows that these compounds could not be accommodated efficiently inside the active site gorge, as the orientation of the rings does not allow penetration of molecules deep into the gorge, thereby, hindering their interactions with the amino acids in the active site.

Molecular Modeling Methods
As the inhibitors were tested as racemic mixtures in the assay, both enantiomeric forms were built up and used for docking. (R)-8 and (S)-8 were assembled within Discovery Studio, version 2.1, software package, using standard bond lengths and bond angles. With the CHARMm force field [2], and partial atomic charges, the molecular geometries of (R)-8 and (S)-8 were energy-minimized using the adopted-based Newton-Raphson algorithm. Structures were considered fully optimized when the energy changes between iterations were less than 0.01 kcal/mol [3].
The coordinates of EeAChE (PDB ID: 1C2B), were obtained from the Protein Data Bank (PDB). For docking studies, initial protein was prepared by removing all water molecules, heteroatoms, any co-crystallized solvent and the ligand. Proper bonds, bond orders, hybridization and charges were assigned using protein model tool in Discovery Studio, version 2.1, software package. CHARMm force field was applied using the receptor-ligand interactions tool in Discovery Studio, version 2.1, software package. Docking calculations were performed with the program Autodock Vina [1]. AutoDockTools (ADT; version 1.5.4) was used to add hydrogens and partial charges for proteins and ligands using Gasteiger charges. Flexible torsions in the ligands were assigned with the AutoTors module, and the acyclic dihedral angles were allowed to rotate freely. Trp286, Tyr124, Tyr337, Tyr72, Asp74, Thr75, Trp86, and Tyr341 receptor residues were selected to keep flexible during docking simulation using the AutoTors module. Because VINA uses rectangular boxes for the binding site, the box center was defined and the docking box was displayed using ADT. For Electrophorus electricus AChE (PDB ID: 1C2B) the docking procedure was applied to whole protein target, without imposing the binding site ("blind docking"). The search space was defined as a box of 60 × 60 × 72 with grid points separated 1 Å, which centered at the middle of the protein (x = 21.5911; y = 87.752; z = 23.591). The num_modes were set to 40 and the other parameters were left as default values. Finally, the most favorable conformations based on the free energy binding were selected for analyzing the interactions between the AChE and inhibitor. All the 3D models are depicted using Discovery Studio, version 2.1. The AutoDock Vina docking procedure used was previously validated [4,5].
The eqBuChE model has been retrieved from the SWISS-MODEL Repository [6][7][8]. This is a database of annotated three-dimensional comparative protein structure models generated by the fully automated homology-modeling pipeline SWISS-MODEL. A putative three-dimensional structure of eqBuChE has been created based on the crystal structure of hBuChE (PDB ID: 2PM8), these two enzymes exhibited 89% sequence identity. Initial protein was prepared and docking calculations were performed following the same protocol described before for EeAChE. All dockings were performed as blinds dockings where a cube of 75 Å with grid points separated 1 Å, was positioned at the middle of the protein (x = 29.885; y = −54.992; z = 58.141). Default parameters were used except num_modes, which was set to 40. The lowest docking-energy conformation was considered as the most stable orientation. Finally, the docking results generated were directly loaded into Discovery Studio, version 2.1.

ADMET of Imidazopyranotacrines 2, 4, 8 and 10
The Absorption Distribution Metabolism and Elimination (ADME) properties were calculated using the QikProp module of Schrodinger suite (QikProp, version 3.8, Schrodinger, LLC, New York, NY, 2013) running in normal mode, for assessing the druggability, and are shown in Tables S1 and S2.