Dual BACE1 and Cholinesterase Inhibitory Effects of Phlorotannins from Ecklonia cava—An In Vitro and in Silico Study

Alzheimer′s disease (AD) is one of the most common neurodegenerative diseases with a multifactorial nature. β-Secretase (BACE1) and acetylcholinesterase (AChE), which are required for the production of neurotoxic β-amyloid (Aβ) and the promotion of Aβ fibril formation, respectively, are considered as prime therapeutic targets for AD. In our efforts towards the development of potent multi-target, directed agents for AD treatment, major phlorotannins such as eckol, dieckol, and 8,8′-bieckol from Ecklonia cava (E. cava) were evaluated. Based on the in vitro study, all tested compounds showed potent inhibitory effects on BACE1 and AChE. In particular, 8,8′-bieckol demonstrated the best inhibitory effect against BACE1 and AChE, with IC50 values of 1.62 ± 0.14 and 4.59 ± 0.32 µM, respectively. Overall, kinetic studies demonstrated that all the tested compounds acted as dual BACE1 and AChE inhibitors in a non-competitive or competitive fashion, respectively. In silico docking analysis exhibited that the lowest binding energies of all compounds were negative, and specifically different residues of each target enzyme interacted with hydroxyl groups of phlorotannins. The present study suggested that major phlorotannins derived from E. cava possess significant potential as drug candidates for therapeutic agents against AD.


Introduction
Alzheimer s disease (AD) is a progressive and irreversible neurodegenerative disorder with characteristic features of cognitive dysfunction, memory impairment, and behavior disturbances. The neuropathological hallmarks of AD patients are the presence of extracellular deposits of amyloid plaques and intracellular filamentous neurofibrillary tangles in the brain [1]. Amyloid plaques and neurofibrillary tangles are aggregates of amyloid-β peptide (Aβ) and hyperphosphorylated tau protein, respectively. In recent years, the "amyloid hypothesis" has arisen as the major pathological mechanism in AD, and the evidence from transgenic mice models revealed that Aβ triggered tau phosphorylation and neurofibrillary tangles formation [2].
Aβ is generated by the sequential proteolytic cleavage of two aspartic proteases, βand γ-secretase, in the amyloidogenic pathway. β-Secretase (BACE1) initially cleaves amyloid precursor γ-secretase expression and inhibiting Aβ-induced neurotoxicity [19,33]. In our previous study, three major phlorotannins of E. cava-eckol, dieckol, and 8,8 -bieckol-exhibited anti-apoptotic and anti-neuroinflammatory properties against Aβ-induced cellular damage, which led to our interest in the study of phlorotannins-mediated suppression of related enzymes in Aβ production and aggregation [34]. Therefore, the purpose of the present study is to evaluate the inhibitory effects of these compounds against both BACE1 and AChE through in vitro and in silico approaches.
Mar. Drugs 2019, 17, x FOR PEER REVIEW 3 of 14 the study of phlorotannins-mediated suppression of related enzymes in Aβ production and aggregation [34]. Therefore, the purpose of the present study is to evaluate the inhibitory effects of these compounds against both BACE1 and AChE through in vitro and in silico approaches.
The IC 50 values (µM) were calculated from a log dose inhibition curve and expressed as the mean ± standard deviation (SD). All assays were performed in three independent experiments. DMSO was used as a negative control in the BACE1 and AChE assays. b Resveratrol and c galantamine were used as positive controls in the BACE1 and AChE assays, respectively. d Inhibition constant (K i ) and e inhibition mode were determined using Dixon plot and Lineweaver-Burk plot, respectively.  Dimethyl sulfoxide (DMSO) was used as negative controls in the BACE1 and AChE assays. Three phlorotannins displayed high potencies as AChE inhibitors (Table 1 and Figure 2). Both dieckol and 8,8 -bieckol exhibited potent AChE inhibitory activity (IC 50 values of 5.69 ± 0.42 and 4.59 ± 0.32 µM, respectively) and about twofold greater than the eckol (IC 50 , 10.03 ± 0.94 µM).
To demonstrate the specificity of the targeted enzymes, all compounds were tested against tumor necrosis-converting enzyme (TACE), which is a candidate for α-secretase, and other serine proteases, including trypsin, chymotrypsin, and elastase ( Table 2). With serine proteases being found in nearly all body tissues and involved in various physiological functions, including digestion, reproduction and immune response, off-target activity causing severe side effects is possible and likely if an inhibitor is not specific to BACE1 and AChE [35]. Therefore, to determine whether phlorotannins inhibited only targeted enzymes without affecting the normal pathway, all compounds were tested against TACE and serine proteases. Up to 100 µM, our tested compounds did not show significant inhibition against the above enzymes, indicating that three phlorotannins appeared to be relatively specific inhibitors of BACE1 and AChE. The inhibitory activity (%) was expressed as the mean ± SD of three independent experiments. DMSO was used as a negative control in TACE and serine proteases assays. b Comparison of concentration level in each sample is not significantly different.

Kinetic Study of BACE1 and AChE Inhibition
As shown in Table 1 and Figure 3, Lineweaver-Burk plots for the inhibition of BACE1 by eckol, dieckol, and 8,8 -bieckol were fitted well to the noncompetitive inhibition mode in visual inspection, and the K i values of eckol, dieckol, and 8,8 -bieckol were 31.2, 20.1, and 13.9 µM, respectively. On the other hand, our tested compounds were competitive inhibitors of AChE, where the Lineweaver-Burk plots intersected a common point on the y-axis (Table 1 and Figure 4). The K i values of eckol, dieckol, and 8,8 -bieckol were 37.3, 12.3, and 11.4 µM, respectively, and were obtained from the Dixon plot.

In Silico Docking Study of the Inhibition of BACE1 and AChE by Phlorotannins
According to the in silico docking simulation results, BACE1 and phlorotannins complexes had an allosteric inhibition mode (Table 3 and Figure 5). GLY34 and SER36 of BACE1 formed two hydrogen bonds with the hydroxyl group of eckol with bonding distances of 3.277 and 3.239 Å , respectively. In the dieckol-BACE1 complex, TRP76, THR232, and LYS321 participated in three hydrogen bonds (bonding distance: 2.960, 3.149, and 3.488 Å , respectively). 8,8′-Bieckol-BACE1 complex had four hydrogen bonding interactions with residues LYS107, GLY230, THR231, and

In Silico Docking Study of the Inhibition of BACE1 and AChE by Phlorotannins
According to the in silico docking simulation results, BACE1 and phlorotannins complexes had an allosteric inhibition mode (Table 3 and Figure 5). GLY34 and SER36 of BACE1 formed two hydrogen bonds with the hydroxyl group of eckol with bonding distances of 3.277 and 3.239 Å, respectively. In the dieckol-BACE1 complex, TRP76, THR232, and LYS321 participated in three hydrogen bonds   As indicated in Table 4 and Figure 6, the docking results for eckol, dieckol, and 8,8′-bieckol indicated negative binding energies of −8.8, −9.5, and −9.2 kcal/mol, respectively. Hydrogen bonding  As indicated in Table 4 and Figure 6, the docking results for eckol, dieckol, and 8,8 -bieckol indicated negative binding energies of −8.8, −9.5, and −9.2 kcal/mol, respectively. Hydrogen bonding interactions between eckol and THR83, TRP86, TYR124, and SER125 of AChE were observed by five hydrogen bonds (

Discussion
With respect to the development of anti-AD agents, enzyme inhibition is one of the most promising potential therapeutic strategies. Since BACE1 is the initiating and rate-limiting enzyme in Aβ formation, it is considered as a key target for lowering cerebral Aβ levels [7][8][9]. Additionally, AChE plays a critical role in cholinergic neurotransmission and participates in non-cholinergic mechanisms such as accelerating Aβ fibril formation through conformational change of Aβ and increasing Aβ toxicity by Aβ-AChE complexes [14][15][16]. Thus, multi-enzyme target inhibition against BACE1 and AChE may provide a promising therapeutic approach for AD. In search of a candidate for AD prevention and/or treatment, numerous researchers over the past few decades have focused on discovering natural enzyme inhibitors. Several natural inhibitors of BACE1 and AChE such as coumarins, citrus flavanones, triterpenoids, and alkaloids have been reported [36][37][38][39]. However, efforts to explore bioactive constituents from marine organisms against BACE1 and AChE have been relatively limited.
Interestingly, the difference in inhibitory properties among phlorotannins is related to the number of hydroxyl groups present. In our new findings, it was shown that 8,8 -bieckol containing 11 OH groups had the highest inhibitory efficacy against BACE1 when compared to dieckol (10 OH groups) and eckol (6 OH groups). When phlorotannins from Eisenia bicyclis, one of the brown algae, were investigated for their BACE1 inhibitory effects, the result that dieckol was stronger than eckol was similar to that of our present study [40]. Consistent with our result, Ahn and colleagues have reported that the inhibitory effect on HIV-reverse transcriptase of 8,8 -bieckol containing a biaryl linkage was tenfold higher than that of 8,4 -dieckol with a diphenyl ether linkage [41]. This observation indicated that the steric hindrance of the hydroxyl and aryl groups near the biaryl linkage of 8,8 -bieckol noticeably enhanced its inhibitory potency.
Among three phlorotannins, 8,8 -bieckol showed the most potent inhibitory activity against AChE. Similar results regarding the correlation between the molecular size of phlorotannins and enzyme inhibitory efficacy was revealed in a previous study. 8,8 -bieckol showed more potent activity against hyaluronidase, with an IC 50 of 40 µM, than dieckol and eckol (IC 50 , 120 and >800 µM, respectively) [42]. In addition, the present study first demonstrated the specific molecular docking interaction as well as biological properties of eckol, dieckol, and 8,8 -bieckol against AChE.
The BACE1 inhibition kinetics indicated that the tested compounds act as non-competitive inhibitors, which means that these compounds can bind either another regulatory site or to the subsite of BACE1. The inhibition level is dependent on the concentration of the inhibitor but is not reduced by increasing concentrations of substrate. Because of this, V max is reduced, but K m is unaffected. In BACE1 inhibitory activity, phlorotannins decreased the V max values without affecting the affinity of BACE1 toward the K m , which demonstrated that these compounds exhibited non-competitive inhibition against BACE1. However, AChE kinetics results exhibited that our tested compounds are competitive inhibitors with unchanged V max and increased K m . In other words, these compounds interacted directly with the catalytic site of AChE instead of with other allosteric pockets.
In silico docking analysis is a valuable drug discovery tool and can be used to discover prospective, biologically active molecules from natural product databases. The results of the molecular docking score were provided to evaluate the capacity of different protein-ligand complex interactions and to compare the biological activities and the inhibition mode. In the BACE1 docking simulation, multiple hydrogen interactions were observed in the BACE1-phlorotannins complexes. Eckol interacted with both GLY34 and SER36 of BACE1, and dieckol bounded to TRP76, THR232, and LYS321. In addition, 8,8 -bieckol formed four hydrogen bonds with BACE1 residues, including LYS107, GLY230, THR231, and SER325. These docking results showed that hydrogen bonds between phlorotannins and allosteric residues of BACE1 play an important role in enzyme inhibition.
AChE docking analysis provides insight into the mechanism underlying active site binding interaction. The hydroxyl group of eckol formed five hydrogen bonds with THR83, TRP86, TYR124, and SER125 of AChE. In particular, the choline-binding site residue (TRP86) of AChE was involved in hydrogen bond interaction with eckol. Dieckol showed four hydrogen-bond interactions with ASN233, THR238, ARG296, and HIS405, whereas 8,8 -bieckol made one hydrogen bond with ARG296 located in the active site of AChE. These docking results from the in silico study were in agreement with our in vitro experimental data.
Nagayama and coworkers demonstrated no significant toxic effects in the oral administration of up to 1,500 mg/kg of phlorotannins for 14 days in male and female Institute of Cancer Research (ICR) mice [45]. In a human study, E. Cava extract was shown to be safe for use in food supplements at a maximum daily intake level of 263 mg/day for adults [46]. Collectively, phlorotannins are toxicologically very safe, explaining their traditional and present consumption as foods and medicinal products.
Bioavailability parameters such as biotransformation and conjugation during absorption from the GI tract are principle factors influencing in vivo biological activity. Lipinski s rule of five is a widespread strategy to define bioavailability predictions of drug molecules. According to this predictive model, a compound needs to exhibit optimum GI absorption with a molecular weight of < 500 Da, no more than five hydrogen bond donors, no more than ten hydrogen bond acceptors, and a calculated partition coefficient (LogP) that is no more than five [47]. Fortunately, eckol meets Lipinski s requirements for acceptable oral bioavailability, while dieckol and 8,8 -bieckol have limitations on bioavailability [48]. However, the compounds absorbed by specific transporters are an exception to this rule, and a recent study demonstrated that dieckol successfully penetrated into the brain via crossing the blood-brain barrier (BBB), suggesting that the compound may be transported through an unknown mechanism [49]. A study of the permeability of eckol and 8,8 -bieckol was limited, but it is likely that similar results might also be predictable as that of dieckol. Overall, our marine compounds from E. Cava are safe, potent, and selective natural dual inhibitors against BACE1 and AChE that can be used for the multi-target, directed agents of AD.

Enzyme inhibition Studies
Fluorometric assays with a recombinant human BACE1 or TACE were conducted according to manufacturer instructions. Briefly, reaction mixtures containing human recombinant BACE1 (1.0 U/mL), the substrate (75 µM in 50 mM ammonium bicarbonate), and phlorotannins dissolved in an assay buffer (50 mM sodium acetate, pH 4.5) were incubated in darkness for 60 min at 25 • C in well plates. The increase in fluorescence intensity produced by substrate hydrolysis was observed on a fluorescence microplate reader with excitation and emission wavelengths of 545 and 590 nm, respectively. The inhibition ratio was obtained using the following equation: where C was the fluorescence of control (enzyme, assay buffer, and substrate) after 60 min of incubation, C 0 was the fluorescence of control at time 0, S was the fluorescence of tested samples (enzyme, sample solution, and substrate) after 60 min of incubation, and S 0 was the fluorescence of the tested samples at time 0.
A human recombinant TACE (0.1 ppm in 25 mM Tris buffer), the substrate (APP peptide YEVHHQKLV using EDANS/DABCYL), and phlorotannins were dissolved in an assay buffer, which were then combined and incubated for 60 min in the dark at 25 • C. The increase in fluorescence intensity produced by substrate hydrolysis was observed on a fluorescence microplate reader with excitation and emission wavelengths of 320 and 405 nm, respectively.
The colorimetric assays, including AChE, trypsin, chymotrypsin, and elastase were assayed according to previously described methods [36]. The hydrolysis of AChE was monitored according to the formation of yellow 5-thio-2-nitrobenzoate anions at 405 nm for 15 min, which were produced by the reaction of DTNB with thiocholine released from ACh. All reactions were performed in 96-well plates in triplicate and recorded using a microplate spectrophotometer.
N-benzoyl-L-Arg-pNA, N-benzoyl-L-Tyr-pNA, and N-succinyl-Ala-Ala-Ala-pNA were used as substrates to assay the inhibition of trypsin, chymotrypsin, and elastase, respectively. Enzyme, Tris-HCl buffer (0.05 M, in 0.02 M CaCl 2 , pH 8.2), and phlorotannins were incubated for 10 min at 25 • C; then, substrate was added for 30 min at 37 • C. The absorbance was recorded at 410 nm. The inhibition ratio was obtained using the following equation: where A was the absorbance of the control (enzyme, assay buffer, and substrate) after 60 min of incubation, and B was the absorbance of tested sample (assay buffer and sample solution) after 60 min of incubation.

Enzyme Kinetic Study
To evaluate the kinetic mechanisms of phlorotannins towards BACE1 and AChE, Dixon and Lineweaver-Burk plots were conducted by various concentrations of substrate and inhibitors. Kinetic parameters such as K i , V max , and K m values were calculated by Sigma Plot 12.3 (Systat Software, Inc., San Jose, CA, USA)

Molecular Docking Study
X-ray crystal structures of human BACE1 (PDB code: 2WJO) and AChE (PDB code: 4PQE) were retrieved from the Protein Data Bank (PDB, http://www.rcsb.org/). Three-dimensional (3D) structures of eckol, dieckol, and 8,8 -bieckol were obtained from PubChem with compound identification number (CID) of 145937, 3008868, and 3008867, respectively. The Autodock Vina software version 1.1.2 (The Scripps Research Institute, La Jolla, CA, USA,) was used to conduct molecular docking analysis. The dimensions of the grid were 30 × 30 × 30 Å, the cluster radius was 1 Å, and the Cα coordinates in each selected backbone binding residue of the protein receptor was used for the center of docking space. Other options for docking simulations were used as defaults. The atomic coordinates of the ligands were drawn and displayed using Marvin sketch (5.11.4, 2012, ChemAxon, One Broadway Cambridge, MA, USA).

Statistical Analysis
All results were presented as the mean ± SD of three independent experiments. Statistical significance was assessed by Duncan s multiple range tests using Statistical Analysis System (SAS) version 9.3 (SAS Institute, Cary, NC, USA).

Conclusions
The integration of enzyme activity, kinetics, and in silico docking studies provided principle insights into the molecular basis underlying ligand binding affinity and BACE1 and AChE inhibition. Accordingly, these results suggested that phlorotannins from E. cava, especially 8,8 -bieckol, have noteworthy potential for the possible development as treatments and/or preventative agents against AD.