Discovery of new butyrylcholinesterase inhibitors via structure-based virtual screening

Abstract Butyrylcholinesterase (BChE) plays an important role in the progression of the Alzheimer’s disease. In this study, we used a structure-based virtual screening (VS) approach to discover new BChE inhibitors. A ligand database was filtered and docked to the BChE protein using Glide program. The outcome from VS was filtered and the top ranked hits were thoroughly examined for their fitting into the protein active site. Consequently, the best 38 hits were selected for in vitro testing using Ellman’s method, and six of which showed inhibition activity for BChE. Interestingly, the most potent hit (Compound 4) exhibited inhibitory activity against the BChE enzyme in the low micromolar level with an IC50 value of 8.3 µM. Hits obtained from this work can act as a starting point for future SAR studies to discover new BChE inhibitors as anti-Alzheimer agents.


Introduction
Alzheimer's disease (AD) is one of the most prevalent neurodegenerative disorders that affect humans especially in older stage of their lives. It is estimated that the disease affected 50 million people worldwide according to World Alzheimer Report 2018 and the number is expected to grow to more than quadruple in the next 30 years 1 .
AD which is characterised by a gradual decline in cognitive function combined with behavioural and psychiatric symptoms 2,3 . Being multifactorial brain disorder, the exact pathophysiology of AD is not yet entirely known 4 . However, several pathogeneses of AD have been suggested: b-amyloid oligomerisation, s-protein aggregation, cholinergic dysfunction, oxidative stress, and inflammation are implicated in the development of AD 5,6 . Most of the current treatments are based on the cholinergic dysfunction hypothesis, which states that degeneration of cholinergic neurons and a deficiency of the neurotransmitter acetylcholine (ACh) are the ultimate reasons for the loss of memory and decline of cognitive function [7][8][9] .
Cholinesterase is the enzyme which catalyses the hydrolysis of ACh and thus inhibition of cholinesterase can be an effective way for the treatment of AD 10 . It is well-known that two different cholinesterase (ChE) enzymes, acetylcholinesterase (AChE), and butyrylcholinesterase (BChE), are responsible for the hydrolysis of ACh within the brain 11 . However, AChE has stronger acetylcholine hydrolytic activity than BChE does under the same condition 12 , as former enzyme can cleave more than 10,000 ACh molecules per second 13 . AChE is substrate specific in nature and is found in high concentrations in the brain, while BChE is nonspecific and is distributed throughout the body 14 . In particular, it is associated with glial and endothelial cells in the brain 15 . In AD, the AChE level in the brain decreases progressively, but BChE level remains the same or increases up to 165% of the normal level 16 . This altered enzymes levels contribute to their different activity as the severity of AD advances 17 . ratio of BChE: AChE shifting from 0.6 to as high as 1.1 contributes to the formation of cholinergic deficits in these regions, leading to the behavioural, and cognitive dysfunction [18][19][20][21] .
Anti-cholinesterase drugs which are used in the treatment of AD are reversible protein inhibitors 22 . They used to postpone the onset of symptoms that are associated with patient lifestyle such as memory ability to think and speak, rationale judgment, and other thought processes 23 . Selective BChE inhibition is potentially advantageous for the treatment of AD. It circumvents the classical cholinergic toxicity that is a common side effect of AChE inhibition 24 .
Current Food and Drug Administration (FDA) approved cholinesterase inhibitors namely: donepezil 25 , rivastigmine, 26 and galantamine 27 , help only in postponing the onset of the symptoms of AD and do not treat the underlying disease 28 .
In this study, we used a structure-based drug design approach to discover new BChE inhibitors which may serve as a valuable inspiration in the search for new treatment options of patients with advanced AD. First, an extensive literature survey was conducted to investigate the critical chemical features that contribute to the inhibition of BChE function and to better understand the binding fashion. Subsequently, we used structure-based virtual screening (VS) followed by Ellman's method to determine the experimental activity of selected hits.

Protein and ligand library preparation
The crystal structure of human BChE protein was obtained from the protein data bank (http://www.rcsb.org, PDB: 4TPK) 29 . All solvent molecules were eliminated then the MOE protein preparation module 30 was used to check out the protein crystal structure for any missing atoms or residues and correct them accordingly 31 . The protein was then prepared by Protein Preparation Wizard of the Schr€ odinger modeling suite, 32 that included adding hydrogen atoms to the protein structure and assigning partial charges to each atom.
Ligand database was downloaded from the NCI website (https://cactus.nci.nih.gov/download/nci/). The database was prepared by processing the molecules through LigPrep, the ligand preparation module in the Schr€ odinger software package 33 . The ligands were filtered based on drug-like rules; the Veber's rules 34 and Lipinski's rule of five 35 . These filters included: molecular weight 500, hydrogen bond donor (HBD) 5, hydrogen bond acceptor (HBA) 10, logP 5, polar surface area (PSA) 140, and rotatable bonds 10.

Virtual screening
The targeted BChE protein was subjected to a docking-based VS workflow using the docking module of the Schr€ odinger suite, Glide (Grid-based ligand docking with energetics) 33 . The prepared drug-like ligand library was docked into the active site of the protein. The Glide virtual screening workflow (VSW) is a three-step docking protocol involving three levels of increasing docking precisions: high-throughput virtual screening (HTVS); standard precision (SP); and extra-precision (XP). The 2000 ligands docked and ranked by GLIDE-XP were then clustered based on the MACCS algorithm 36 and were then visually inspected. Finally, 38 ligands belonging to various structural scaffolds and showing convenient binding modes in the BChE active site were selected for biological assessment.

Quantification of BChE inhibitory activity in a spectrophotometric assay
Selected hits were obtained from NCI and they were provided as dry powders in variable quantities (5-10 mg). Compounds were initially dissolved in DMSO to give stock solutions of 100 mM. Subsequently, they were diluted to the required concentrations with Tris buffer pH 8.0 for the assay. Enzymatic inhibition assays were performed on BChE from equine serum (Sigma), according to the spectrophotometric Ellman's method 37 . The experiment was performed in 48-well plates in a final volume of 100 mL. Each well contained 0.22 U/mL eqBChE dissolved in Tris-HCl buffer, pH 8.0. They were preincubated for 20 min at different compound concentrations at 37 C. Then 0.5 mM butyrylthiocholine iodide (Sigma) and 0.35 mM 5,5 0 -dithiobis -2-nitrobenzoico (DTNB; Sigma) were added. Colour development was measured spectrophotometrically at 412 nm using microplate reader (BioTek ELx800) at a rate of one measurement per minute over 15 min period. Positive (Eserine, sigma 100 uM) and negative (no inhibitors) controls were tested. All samples were assayed in at least duplicate measurements. In general, the amount of DMSO was kept below 1% in the assay. IC50 values were determined graphically from inhibition curves using Graph Pad prism version 6, Graph Pad Software, Inc.. (602697) (11052)

Results and discussion
Computer-aided molecular modeling protocol should involve sufficient study on the active site of the target enzyme before conducting the main VS 38 . Hence, the catalytic pocket of BChE protein was investigated and found to be consisting of a catalytic site at which the hydrolysis reaction takes place, its located at the bottom of a deep and narrow gorge, which is composed of conserved aromatic amino acids 39 . This most important site, also called the esteratic site, contains the three essential amino acids, Ser198, His438 and Glu325, which create the catalytic triad. They are involved in the transfer of the acetyl group from ACh to Ser198 39 . An essential role in the hydrolysis process is also played by amino acids in the BChE anionic site, which is responsible for binding the substrate quaternary ammonium group with cation-p interactions. Due to the interactions with the anionic site, the proper orientation of ACh in the gorge is provided.
One of the serine hydrolase features is stabilisation of the transition state by amino acids of the oxyanion hole through highly conserved N-H dipoles, derived from amino acids of the main chain: Gly116, Gly117, and Ala199 40 . During the enzymatic reaction, the transition complex is created and stabilised by those amino acids. The acyl pocket is responsible for substrate specificity. Comparison of human butyrylcholinesterase (HuBChE) and human acetylcholinesterase (HuAChE) shows differences in size, especially of the acyl binding pocket. The active gorge is larger in HuBChE than in HuAChE (500 Å 3 versus 300 Å 3 ). In HuBChE, the shape of the acyl pocket is determined by two residues: these are the aliphatic residues Leu286 and Val288, the presence of the smaller amino acids, valine and leucine, makes the hollow in the acyl pocket larger in BChE and enables larger molecules to get in 41 .
To sum up, the main residues in the active site were classified into three groups: catalytic residues (Ser198, His438, and Glu325), acyl binding pocket (Gly116, Gly117, Trp231, Leu286, and Val288), and choline binding pocket (Trp82) are shown below in Figure 1. After reviewing the literature, we can identify the following interactions in the active site as important for binding BChE inhibitors: interactions with the aromatic rings of Trp231, Trp82, and Phe329, to lesser extent H-bonding interaction with His438 and the tendency of the hit structure to enter the acyl pocket of BChE protein.
With regards to the BChE virtual screening, a schematic representation of the used protocol is shown in Figure 2. The VS started with filtering the NCI-ligand library according to drug-like characteristics The obtained drug-like ligand library where screened through three subsequent docking steps: HTVS, GLIDE-SP then GLIDE-XP so that docking precision increases gradually as ligands advance from one stage to another. The final outcome from VS process was filtered and preference was given to molecules containing features that are important for aforementioned interactions with active site residues.
To evaluate the inhibitory action of the top ranked hits, these compounds were ordered from NCI and were tested against the BChE protein using Ellman's method 37 . Eserine inhibitor was used as positive control. Table 1 shows the inhibition pattern of the top ranked compounds again the BChE enzyme. The IC 50 values of these six hits were then identified as shown in Table 1. These active compounds exhibited variable inhibitory concentrations ranging from low to high micromolar range. Four out of these hits showed IC 50 values below 100 mM (hit compounds 4, 5, 12, and 26). Interestingly, compound 4 exhibited the maximum inhibitor activity against BChE enzyme with an IC 50 value of 8.3 mM.    The regression coefficient of the dose-response curve as calculated by Graph pad prism 6.0. Inhibitory activity of the tested compounds was measured by half maximal inhibitory concentration (IC 50 ). a The regression coefficient of the dose-response curve as calculated by Graph pad prism 6.0. b Reference inhibitor.
With regard to the in silico data, docking results for the six active hits are shown in Table 2. All active compounds exhibited low binding energies ranging from (À10.3 to À16.3 kcal/mol). These scores seem to be even more interesting if they got related to the size of the molecule since these hits scored ligand efficiency scores of around À0.5 kcal/mol. Ligand efficiency scores are calculated by dividing the docking score over the molecule weight of the compound and it indicates for suitability of the ligand to act as lead compound. Hence, these compounds look to have interesting binding energies with relatively small size when compared with the co-crystallised ligand.
Docked compounds showed convenient-binding modes in the BChE active site. In particular, to the best two hits (compounds 4 and 12), the main interactions attributed to their activity relied on p-p and p-hydrogen interactions with Trp231, Trp82, and Phe329 along with hydrogen bonding with Gly117 and Ser198 (as shown in Figures 3 and 4). Additionally, the entry of the molecule within the acyl binding pocket of the active site along with the aforementioned bonding are believed to be substantially important for inhibitory activity. 1 H-NMR spectroscopy was used to confirm the correct structure and the purity of the compounds. See spectra in supporting information. The structures of the compounds 4 were confirmed on the basis of 1 H-NMR spectroscopy analysis. The 1 H-NMR spectrum showed five characteristic signals resonates at d ¼ 1.62, 2.45, 2.59, 2.71, and 3.88 ppm attributed to the CH 2 groups with an integration value of 16 protons. Another signal appeared at d ¼ 3.56 as broad band corresponding to CH proton. The eight aromatic protons resonated as a multiplet at d ¼ 6.89, 6.96, 7.27, 7.41, and 7. 50 ppm. In addition, signal appeared at d ¼ 7.36 as broad band corresponding to NH-proton exchanged in D 2 O. The 1 H-NMR spectrum of 5 showed a characteristic of two signals at d ¼ 0.91 and 2. 17 ppm corresponding to the three methyl groups with an integration value of nine protons. The four signals attributed to the five CH 2 groups resonates at d ¼ 1.84, 2.03, 2.75, and 3.62 ppm attributed to the CH 2 proton with an integration value of 10   with an integration value of 4 protons. Hence, all required integral values matches the number of protons without any other peak in the NMR spectra which confirm the purity of the compounds.

Conclusion
Computer-aided drug design was employed to study and discover new BChE inhibitors that could be a starting point for a promising drug candidate in the treatment of AD. NCI database was filtered, treated, and subsequently screened against the BChE protein.
Furthermore, the top docked compounds were tested against BChE activity which showed a direct correlation between the biochemical and drug design approaches by revealing four hits with IC 50 values below 100 mM. The four hits' identity was checked and their purity was confirmed. Compound 4 (IC 50 value of 8.3 mM) and compound 5 (IC 50 value of 32.6 mM) showed strong inhibitory activity that could become a starting point for future development and structure activity relationship studies for the discovery of new BChE inhibitors.