Anticholinesterase activity of β-carboline-1,3,5-triazine hybrids

The β-carboline-1,3,5-triazine hydrochlorides 8-13 were evaluated in vitro against acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE). The analysed compounds were selective to BuChE, with IC 50 values in the range from 1.0–18.8 µM being obtained. The N -{2-[(4,6-dihydrazinyl-1,3,5-triazin-2-yl)amino]ethyl}-1-phenyl-β-carboline-3-carboxamide (12) was the most potent compound and kinetic studies indicate that it acts as a competitive inhibitor of BuChE. Molecular docking studies show that 12 strongly interacts with the residues of His438 (residue of the catalytic triad) and Trp82 (residue of catalytic anionic site), confirming that this compound competes with the same binding site of the butyrylthiocholine.


INTRODUCTION
Alzheimer's disease (AD) is a multifactorial neurodegenerative disorder (Larson, Kukull, Katzman, 1992), characterised by cognitive impairment, which is associated by means of cholinergic hypothesis, to the loss cholinergic of neurons and a decrease in levels of cholinergic neurotransmission (Francis et al., 1999;Pinto, Lanctôt, Herrmann, 2011). The hydrolysis of the neurotransmitter acetylcholine (ACh) into choline and acetic acid, a reaction catalysed by enzymes of the cholinesterase family, is necessary to allow a cholinergic neuron to return to its resting state after activation (Čolović et al., 2013). The acetylcholinesterase (AChE) predominates in the healthy brain, while the butyrylcholinesterase (BuChE) is considered to play a minor role in the regulation of synaptic ACh levels (Silva et al., 2014). In fact, most of the current drugs used for AD treatment are based on this hypothesis, acting mainly by the inhibition of AChE (Silva et al., 2014;Čolović et al., 2013). However, studies have shown that as AD progresses, the BuChE activity progressively increases, while AChE activity remains unchanged or gradually decreases. When this occurs, the BuChE assumes function of metabolize the ACh in the synapse (Darvesh, Hopkins, Geula, 2003;Anand, Singh, 2013;Li et al., 2017). Thus, also inhibiting BuChE, cognitive improvements associated with the current cholinesterase inhibitors can be obtained. As a result, both AChE and BuChE can be considered as therapeutic targets in Alzheimer's disease treatment (Darvesh, Hopkins, Geula, 2003;Anand, Singh, 2013;Li et al., 2017).
The treatment of AD usually is performed with cholinesterase inhibitors (donepezil, rivastigmine and galantamine), which enhance cholinergic signalling in the central nervous system and cognitive symptoms. However, these drugs do not prevent the AD progression, and unfortunately there is still no cure for this disease (Silva et al., 2014;Pinto, Lanctôt, Herrmann, 2011;Čolović et al., 2013).
highly active and selective for BuChE, and it was found that over 60% of the studied compounds showed a better inhibitory activity of BuChE than the drug galantamine (Horton et al., 2017).
In our previous work, we investigated the properties of β-carbolines related to neurodegenerative diseases, which led us to identify compound VI (Figure 1) as a potent DYRK1A and MAO-A inhibitor (Drung et al., 2014). By continuing our research, and taking in account the related proprieties of β-carbolines and 1,3,5-triazines, in this work we evaluated the anticholinesterase activity of β-carboline-1,3,5-triazine hybrids VII (Figure 1) against AChE and BuChE. Additionally, kinetic and molecular docking studies were carried out for the most potent compound, aiming to evaluate its inhibition mode against BuChE.
Anticholinesterase activities of β-carboline-1,3,5triazine hydrochlorides against AChE and BuChE were evaluated according to Ellman's modified method (Ellman et al., 1961). Stock solutions of the tested compounds 8-13 were prepared in milli-Q water. The tests were performed in polystyrene 96-well plate, with 125 µL of DTNB (0.5 mM), 50 µL of buffer solution of phosphate (pH 8), 25 µL of sample solution at different concentrations and without the inhibitor (control), and 25 µL of enzyme solution of AChE (0.23 U mL -1 prepared in buffer solution of phosphate) or BuChE (0.23 U mL -1 prepared in buffer solution of phosphate) were added to each well. The plate was incubated at 30°C for 15 minutes under stirring, and then absorbance was measured at a wavelength of 412 nm. After this time, 25 µL of substrate (acetylthiocholine or butyrylthiocholine, 5 mM, prepared in milli-Q water) was added to each well. The plate was kept at 30°C, under stirring, and the absorbance was measured again at the same wavelength after 4 minutes. The tests were performed in triplicate.
The rates of reactions were calculated using appropriate software (Origin 6.1). The inhibition percentages were calculated by comparing of control reaction rate with the sample reaction rate using Eq.1: %inhibition = ((control reaction rate-sample reaction rate)/control reaction rate) x 100 (1) The inhibition curve was obtained by plotting an inhibition percentage graph versus the logarithm of the inhibitor concentration.
Re-docking simulations were performed to validate the parameters that had been chosen and were repeated four times which gave a RMSD values below 0.5 Å. The best results were submitted to energy minimisation with the NAMD2 program (Phillips et al., 2005). The force field adopted for proteins was CHARMM C35b2-C36a2, and for the ligands, they were generated in the same format as the SwissParam server (Zoete et al., 2011). The results were shown with the CCP4 Molecular Graphics software (McNicholas et al., 2011).

Chemistry
The β-carboline-1,3,5-triazine hybrids 8-13 (Scheme 1) were synthesised as described for Baréa et al. (2018). Briefly, the β-carboline intermediate 1, obtained from L-tryptophan commercial, was subjected to reaction with cyanuric chloride in the absence or presence of different amines, under basic medium. The hydrochloride salts were prepared from the treatment of compounds 8-13 with hydrochloric acid in methanol (Scheme 1).

Anticholinesterase activity
The anticholinesterase activities of β-carboline-1,3,5-triazine hydrochlorides 8-13 and donepezil (reference compound) were evaluated according to the Ellman method (Ellman et al., 1961). Firstly, the compounds 8-13 were evaluated in vitro at concentrations of 10 µM and 100 µM against AChE and BuChE and their percentages of inhibition were determined ( Table  I). All compounds showed less than 50% or no inhibition for AChE, and their IC 50 (50% Inhibitory Concentration) values were not determined for this enzyme. On the other hand, all hybrids inhibited the BuChE at a concentration of 100 µM, showing inhibition percentage in the range from 67.3-91.9%, and most of them also inhibited this enzyme at concentration of 10 µM (40.5-82.7%). Thus, the IC 50 values were determined for compounds that inhibited more than 40% of this enzyme, at a concentration of 10 µM ( Table I). The obtained IC 50 values ranged from 1.0-18.8 µM, with derivative 12, containing the hydrazinyl group at 6-and 4-positions of 1,3,5-triazine ring, the most active among the tested compounds for BuChE.
In summary, our results show that the analysed β-carboline derivatives were selective to BuChE. This selectivity can be explained based on the volume of the BuChE active site gorge, which is approximately 200 Å 3 larger than the AChE gorge (Saxena et al., 1997;Johnson, Moore, 2012;Masson, Carletti, Nachon, 2009), allowing the accommodation of tested β-carboline-1,3,5triazine hybrids. Moreover, other compounds of the β-carboline class with a flexible linker also exhibited selectivity for BuChE (Zhao et al., 2018). The authors explain that folded molecules are suitable for the relatively spherical and large cavity of BuChE, but not stretched and slender enough to fit the narrow gorge of AChE (Zhao et al., 2018). Therefore, the presence of the flexible N-aminoethyl-carboxamide group between the β-carboline and 1,3,5-triazine moieties in 8-13 probably corroborated the obtained selectivity.
Graphical analysis of Lineweaver-Burk plots ( Figure  2) and the kinetic parameters of BuChE activity showed a practically unchanged V max value (V max = 0.11, 0.10 and 0.11 µMs -1 in the absence and presence of 0.3 and 3.0 µM of 12, respectively) and an increasing K m value (K m = 0.28, 0.30 and 1.58 mM in the absence and presence of 0.3 and 3.0 µM of 12, respectively) with increasing inhibitor concentrations, i.e. increasing slopes and the same intercepts on the y-axis (-1/V max ). This pattern indicates a competitive type of inhibition (Copeland, 2000). It is shown that compound 12 and substrate (butyrylthiocholine) compete for the same active site, i.e. the inhibitor interacts with the same binding site as the substrate.
For hybrid 12, the dissociation constant (K i ) value obtained was 0.55 µM while the K m value of butyrylthiocholine iodide for BuChE was 0.28 mM, which shows that the binding capacity of 12 with BuChE is approximately 509-fold that of the substrate. In addition, the hybrid 12 exhibited a K i value similar to derivative Vb (Figure 1, K i = 0.64 µM for BuChE) (Torres et al., 2012) and showed a binding capacity that was approximately 164-fold greater than that of harmane (IV, Figure 1, K i = 90 µM for BuChE) (Torres et al., 2012).

Molecular modelling studies
The molecular docking calculations were performed using AutoDockVina program (Trott, Olson, 2010) implemented at the interface PyRx 0.9 (Wolf, 2009

Enzyme kinetics
Due to the potent activity observed for compound N-{2-[(4,6-dihydrazinyl-1,3,5-triazin-2-yl)amino] ethyl}-1-phenyl-β-carboline-3-carboxamide (12), this compound was submitted to kinetics studies to investigate its type of BuChE inhibition. The kinetics studies were performed using the modified Ellman's method (Ellman et al., 1961). To assess the kinetic parameters, we measured the initial rate of enzyme activity at different concentrations of substrate butyrylthiocholine (0.05 to 2.0 mM) in the absence and presence of the compound 12. Lineweaver-Burk plots (Figure 2) were generated by plotting the reciprocal of the initial rate (1/V 0 ) against the reciprocal of substrate concentrations (1/[S]) for the different concentrations of 12, resulting from the substrate-velocity curves for BuChE. significant activity and selectivity for BuChE. The kinetics and molecular docking studies for the most active hybrid 12 indicate that this compound inhibited BuChE via a competitive type of inhibition.
complex of the enzymes with butyrylcholinesterase obtained from the Protein Data Bank (PDB). The best docked poses, i.e. the lowest energy conformer in the most populated cluster of conformers, were subjected to energy minimisation by NAMD (Phillips et al., 2005) program and analysed to explain interactions between ligands and the target enzyme. Figure 3 shows that compound 12 is oriented in active site gorge of BuChE. The hydrogen atom of protonated triazine moieties interacts with the carboxylate oxygen atom of Gly115 via an H-bond. The oxygen atom of the carbonyl group forms an H-bond with OH group of Thr120 and NH group also forms an H-bond with the carboxylate oxygen atom of Gly116. The protonated nitrogen of hydrazine moieties interacts with Tyr440 and Trp82 by π-cation interaction. Both hydrogen atoms of the protonated hydrazine moieties are therefore likely to form an H-bond with the carboxylate oxygen atoms of His438 and Gly439 and the other protonated hydrazine moieties also interact with Glu197 by H-bond interactions. This strong interaction with His438 (residue of the catalytic triad) (Nicolet et al., 2003) and the π-cation interaction with Trp82 (residue of catalytic anionic site) (Nicolet et al., 2003) confirm that the compound competes with the same binding site of the butyrylthiocholine.

CONCLUSION
In conclusion, we evaluated the anticholinesterase activity of the β-carboline-1,3,5-triazine hybrids against AChE and BuChE. All of the compounds showed