Introduction

Parkinson’s disease (PD) is an exceedingly prevalent neurodegenerative disorder in the aging population, particularly in developed countries (Abdullah et al. 2015), second only to Alzheimer’s disease. The basis of PD appears to be the reduction of the dopamine level in neuronal synapses. Therefore, dopaminergic agonists are thought to play a crucial role in the therapy of PD.

Monoamine oxidases (MAOs) inhibitors also are important in the treatment of PD. MAOs are flavoproteins that catalyze the oxidative deamination of a mixture of neurotransmitters, such as norepinephrine, dopamine, and serotonin, as well as different exogenous and endogenous amines (i.e., tyramine, benzylamine, etc.) to their corresponding aldehydes (Lin et al. 2003). MAO exists as two isoforms, MAO-A and MAO-B, which are encoded by distinct genes sharing 70% sequence identity and the dependence on the covalently linked flavin adenine dinucleotide (FAD) cofactor in their amino acid active sites (Bach et al. 1988). The active sites of MAOs can accommodate different substrates and can be targeted by specific inhibitors. Several studies have implicated inhibitors of MAO A and B in the attenuation of oxidative stress (Sturza et al. 2013; Friedman et al. 2011). Generally, MAO-A inhibitors are used as antidepressant agents in treatment of depressive illness featuring a deficiency of serotonin and, to a lesser extent, norepinephrine (Lum and Stahl 2012; Schwartz 2013). MAO-B inhibitors are used in the therapy of PD where the key deficiency of dopamine is responsible for the characteristic motor deficits (Fernandez and Chen 2007). MAO-B is more abundant and more dynamic in the human basal ganglia (Youdim and Riederer 2007).

Monoamine oxidase inhibitors (MAOIs) have been widely employed clinically, but the use of nonselective and selective MAO-A inhibitors decreased after the recognition that these drugs potentiate the sympathomimetic effects of the dietary amine, tyramine (Da Prada et al. 1988; Flockhart 2012). Due to this interaction, blood pressure increases abnormally (the “cheese effect”). Selective MAO-B inhibitors do not cause tyramine-induced high blood pressure and the development of reversible MAO-A inhibitors that have low risk of the cheese effect have facilitated the continued role of MAOIs in the clinic (Finberg 2014). The older, irreversible MAO-A inhibitors are still used to treat certain types of depression. But from a drug design and drug safety point of view, reversible inhibition of MAO-A is a more desirable property; such drugs would not require dietary restrictions (Bonnet 2003; Provost et al. 1992). Both reversible and irreversible MAO-B inhibitors have excellent safety profiles and are not associated with changes in blood pressure (Finberg and Gillman 2011; Pae et al. 2012). The discovery of novel classes of compounds with potent inhibition activity is of interest.

Eisenia bicyclis (Kjellman) Setchell is a perennial brown alga (Phaeophyta), which belongs to the Laminariaceae family and is distributed along the mid-Pacific coastlines of Korea and Japan. It is consumed as a raw material for sodium alginate and phlorotannin-rich raw materials (Okada et al. 2004). Phlorotannins are secondary metabolites of phloroglucinol (1,3,5-trihydroxybenzene) that polymerize through ether, phenyl, or 1,4-dibenzodioxin linkages (Okada et al. 2004). Eckol and dieckol are well-known phlorotannins isolated from E. bicyclis (Jung et al. 2013). These phlorotannin compounds have a variety of biological activities including relief of anti-diabetic complications (Okada et al. 2004; Jung et al. 2008), anti-tumor (Noda et al. 1989), hepatoprotection (Kim et al. 2005), plasmin inhibition (Fukuyama et al. 1989), algicidal (Nagayama et al. 2003), tyrosinase inhibition (Kang et al. 2004), anti-inflammatory (Ryu et al. 2009), slowing of skin aging (Joe et al. 2006), anti-oxidant (Li et al. 2009), cholinesterase inhibition (Myung et al. 2005), anti-Alzheimer (Jung et al. 2010), anti-hyperlipidemic (Yoon et al. 2008), anti-diabetic (Lee et al. 2010, anti-allergic (Sugiura et al. 2006), and angiotensin converting enzyme-I inhibition (Jung et al. 2006) activities.

MAO inhibitory activity by eckol and dieckol has not been reported. This study examined the MAOs inhibitory activity of eckol, dieckol and methanol (MeOH) extract from E. bicyclis through an in vitro enzyme assay and molecular docking simulation.

Materials and methods

Chemicals

hMAO-A and hMAO-B were derived separately from BTI-TN-5B14 insect cells infected with recombinant baculovirus cDNA. The hMAO isozymes were purchased from Sigma-Aldrich (St. Louis, MO, USA) and a final concentration of 10 μg/mL of each isozyme was used. DEP (selegiline-irreversible MAO-BI) was purchased from Sigma-Aldrich.

Plant material

The leafy thalli of E. bicyclis were purchased from a local telemarketing company (www.ulleumgdomall.com) in Gangwon Province, Korea, in April 2014. The alga was authenticated by J. S. Choi of Pukyong National University. Voucher specimens (no. 20140428) have been deposited in the author’s laboratory (J. S. Choi).

Preparation of MeOH extract of E. bicyclis and its major phlorotannins, eckol and dieckol

Lyophilized powder (100 g) of E. bicyclis leafy thallus was refluxed with MeOH (3 × 1 L) for 3 h and each filtrate was concentrated to dryness in vacuo at 40 °C to render the MeOH extract (34.0 g). This extract was suspended in distilled H2O and then successively partitioned with n-hexane, chloroform (CH2Cl2), ethanol acetate (EtOAc), and butanol (n-BuOH) to yield the respective fractions (2.6, 2.4, 10.9, and 7.8 g), as well as H2O residue (11.1 g). Eckol and dieckol were isolated from the EtOAc fraction as previously described (Jung et al. 2013), and identified by direct comparison with authentic samples (1H and 13C–NMR). The structures of eckol and dieckol are shown in Fig. 1.

Fig. 1
figure 1

Structures of eckol and dieckol

Full size image

hMAO-A and hMAO-B activity assay

A chemiluminescent assay was used to confirm MAO-A and MAO-B inhibitory effects of eckol, dieckol, and the MeOH extract of E. bicyclis using a MAO-Glo kit (Promega, Madison, WI, USA). Arbitrary light units (ALU) of the enzymes were measured in the presence of eckol, dieckol, MeOH extract, and deprenyl (DEP) as a positive control. Briefly, the hMAO-A and hMAO-B isozymes were diluted to 2× with reaction buffer (pH 7.4) and preincubated with 4× eckol, dieckol, MeOH extract, or DEP working solutions at room temperature for 1 h in white opaque 96-well plates. Final concentrations (all 5 μg/mL) of eckol, dieckol, and MeOH extract, and 2 μg/mL were used to determine MAO-B inhibition activity. Eckol, dieckol, MeOH extract (4×), and DEP working solutions were serially diluted in reaction buffer (pH 7.4) to determine the IC50. The sample was substituted with reaction buffer for controls, and only reaction buffer was used as the blank. Based on our preliminary optimizations and as previously described (Valley et al. 2006), the reaction was initiated by adding 4× luciferin derivative substrate (LDS) to produce a final concentration of 40 and 4 μM for the hMAO-A and hMAO-B reactions, respectively. The final volume per well of each reaction was 50 μL. Reconstituted luciferin detection reagent (RLDR, 50 μL) was added to all wells and incubated for 20 min to stop the reaction and produce the luminescence signal. The ALU produced were detected using a SpectraMax Multi-Mode Reader (Molecular Devices, Sunnyvale, CA, USA).

Enzyme kinetics

Michaelis–Menten hyperbolic regression curves and parameters (K m and V max ) were determined using the MAO-Glo assay after determining the best incubation time for initial velocity of activity. The assay was carried out under the same conditions in white opaque 96-well plates. Seven LDS data points differing logarithmically were acquired by preparing a serial dilution of triplicate concentrations. The substrate preparation was carried out at gradual 4× final LDS 40, 20, and 10 μM and no substrate for hMAO-A, and 4, 2, and 1 μM and no substrate for hMAO-B. Sample concentrations of 10, 50, and 100 μM were prepared in triplicate for hMAO-A and hMAO-B. Controls without samples were prepared simultaneously. A 2× enzyme concentration for a constant final concentration of 20 μg/μL was used for a 20 min incubation. The reactions were started by mixing the enzyme and inhibitor (or buffer) with 4× LDS in the wells. After a 1 h (for both hMAO-A and hMAO-B) incubation at room temperature, the initial reaction rate was inhibited by doubling the volume of each well with RLDR. The reactions were measured after 20 min with a SpectraMax Multi-Mode illuminometer (Molecular Devices). The Michaelis–Menten equation data were plotted as a nonlinear curve, which was transformed and presented as a linear curve in double reciprocal Lineweaver–Burk plots. K m and V max values were computed and presented as relative fold of change after the Michaelis–Menten equation analysis using GraphPad Prism 6.01 (GraphPad Prism Software, Inc., La Jolla, CA, USA). The dissociation constant of each isozyme-inhibitor complex (K i ) was also calculated as a global share value for each group of data, according to the competitive mode of the inhibition equation for the samples model with hMAO-A and hMAO-B.

Molecular docking simulation

Docking studies were performed with eckol and dieckol to examine qualified binding poses against MAO-A and MAO-B. The crystalline structure of hMAO-A with harmin (HRM, PDB ID:2z5x) and hMAO-B (PDB ID:1gos) with N-[(E)-methyl](phenyl)-N-[(E)-2-propenylidene]methanaminium (NYP) were obtained from the RCSB Protein Data Bank with resolutions of 2.2 and 1.6 Å, respectively. HRM and NYP heteroatom were removed for the docking simulations, the protein was regarded as ligand free, and water molecules were removed from the protein structure using Accelrys Discovery Studio 4.1 (DS 4.1) (DS, http://www.accelrys.com; Accelrys, Inc. San Diego, CA, USA). All hydrogen atoms were added into the protein using the automated docking tool AutoDock 4.2.6 (Goodsell et al. 1996; Jones et al. 1997; Rarey et al. 1996). The HRM and NYP binding areas of the protein were considered the most likely regions for docking simulation to obtain the best ligand binding results. The three-dimensional structure data file (sdf) structure of eckol and dieckol were obtained from PubChem (CID: 145937 and CID: 3008868, respectively) and converted into pdb format using DS 4.1. The automated docking simulation was performed using Autodock tools (ADT) to assess the appropriate binding orientations and conformations of the ligand molecule with the protein inhibitor. A Lamarkian genetic algorithm method implemented in the Autodock 4.2 program was employed. Gasteiger charges were added by default, the rotatable bonds were set by the Autodock tools, and all torsions were allowed to rotate for the docking calculations. Grid maps were generated by the Autogrid program where a grid box size of 126 × 126 × 126 points with a default spacing of 0.375 Å between the grid points was used to cover almost the entire favorable protein binding site. The X, Y, and Z centers were 30.868, 28.77, and −14.856 for MAO-A (2z5x) and 53.502, 147,805, and 24.371 for MAO-B (1gos), respectively. The docking protocol for rigid and flexible ligand docking consisted of 10 independent genetic algorithms (GA) and other parameters used the ADT. Binding aspects of the MAO-A and MAO-B residues and their corresponding binding affinity scores were regarded as the best molecular interaction. The results were visualized and analyzed using UCSF Chimera (http://www.cgl.ucsf.edu/chimera/) and Ligplot.

Statistical analyses

Data analyses were performed using GraphPad Prism 6.01 software. Data are presented as mean ± standard error (SE) with triplicate replications, representing at least two independent experiments. Inhibitory potency was expressed as mean ± SE of at least two independent experiments. The IC50 was obtained by interpolating the best fit logarithmic concentration-inhibition curves for R 2. The selective index was determined by the ratio of IC50 of hMAO-A/IC50 of hMAO-B and K i of hMAO-A/K i of hMAO-B. One-way analysis of variance was used to detect differences. A p value < 0.05 was considered significant.

Results

Effect of E. bicyclis MeOH extract, eckol, and dieckol on hMAO-A and hMAO-B

The inhibitory effect of the MeOH extract, eckol and dieckol obtained from E. bicyclis as well as standard DEP were measured by a luminescence assay. The inhibitory potency of all on the MAO-A and -B isozymes was compared with that of DEP (Table 1). As demonstrated in Table 1 and Fig. 2, all displayed good inhibitory activity on both hMAO-A and hMAO-B, with more effectiveness inhibiting hMAO-A. The MeOH extract exerted good inhibitory activities with IC50 values of 16.82 ± 0.70 and 38.38 ± 1.78 μM, respectively, on hMAO-A and hMAO-B in dose-dependent manners. The luminescence assay revealed a DEP IC50 value for hMAO-A and hMAO-B of 10.54 ± 1.25 and 0.128 ± 0.01 μM, respectively (Table 1). Eckol significantly inhibited hMAO-A and hMAO-B, with IC50 values of 7.20 ± 0.71 and 83.44 ± 1.48 μM, respectively. Dieckol also showed good hMAO-A and hMAO-B inhibitory activities, with IC50 of 11.43 ± 1.06 and 43.42 ± 0.73 μM, respectively (Table 1).

Table 1 Inhibitory activities of MeOH extract, eckol, and dieckol from Eisenia bicyclis, and deprenyl on monoamine oxidase (MAO)-A and MAO-B with selective index (SI)
Full size table
Fig. 2
figure 2

Eckol (ECK) and dieckol (DEK) inhibitory efficacy on recombinant human monoamine oxidases (hMAOs) compared to Eisenia bicyclis methanolic extract (ME) and standard MAO inhibitor selegiline (DEP) at 5 μg/ml for MAO-A and 20 μg/ml for MAO-B (2 μg/ml for MAO-B DEP). Data are mean ± standard error of at least three determinations. ***P < 0.05

Full size image

Inhibition mode of hMAO-A and hMAO-B by eckol and dieckol

Incubation time was optimized to 1 h to inhibit hMAO-A and hMAO-B activities. The incubation period was within the range of initial velocities at the recommended substrate concentrations. The Michaelis–Menten kinetics curves of the MAOs with and without eckol/or dieckol at an initial rate of velocity (V) versus the change in luciferin derivative substrate (LDS) concentration are illustrated in Lineweaver–Burk plots (Figs. 3a, b, 4a, b). In the case of hMAO-A inhibition mode for eckol, all regression lines crossed in different point (V max ) on the y-axis, whereas the regression lines crossed at one points on the x-axis revealing the various K m values. These regression lines meet in one point outside of x-axes (Figs. 3a, b). All eckol regression lines crossed at different points (V max ) in y-axes, while that regression lines crossed at one point on the x-axis at variable K m values towards the hMAO-B inhibition. Figure 3c, d depict the maximum velocity (V max ) with folds of change for eckol. Figure 3e, f depict the K m values for eckol in the inhibition of hMAO-A and hMAO-B. In the case of hMAO-A and hMAO-B inhibition mode for dieckol, all regression lines crossed at different point (V max ) on the y-axes, while the regression lines crossed at one point on the x-axes at variable K m values. Figure 4c, d depict the maximum velocity (V max ) with the fold-changes for dieckol. Likewise, Fig. 4e, f depict K m values for dieckol for the inhibition of hMAO-A and hMAO-B.

Fig. 3
figure 3

Effects of eckol on human monoamine oxidases A and B (hMAO-A and hMAO-B) kinetics. Initial velocity (V) with substrate concentrations ([S]) with or without eckol as Lineweaver–Burk plots of hMAO-A (a) hMAO-B (b). Michaelis–Menten kinetic parameters are presented as fold-change of eckol. Parameters are maximum velocity (Vmax) and the Michaelis constant (Km) for hMAO-A (c, e) and hMAO-B (d, f), respectively. Data are mean ± standard error, n = 3, from three separate eckol experiments. ***P < 0.05

Full size image
Fig. 4
figure 4

Effects of dieckol on human monoamine oxidases A and B (hMAO-A and hMAO-B) kinetics. Initial velocity (V) with substrate concentrations ([S]) with or without dieckol as Lineweaver–Burk plots of hMAO-A (a) hMAO-B (b). Michaelis–Menten kinetic parameters are presented as fold change of dieckol. Parameters are maximum velocity (Vmax) and the Michaelis constant (Km) for hMAO-A (c, e) and hMAO-B (d, f), respectively. Data are mean ± standard error, n = 3, from three separate dieckol experiments. ***P < 0.05

Full size image

hMAO-A (Figs. 5a, 6a) and hMAO-B (Figs. 5b, 6b) Michaelis–Menten curve with eckol and dieckol had the best fit to the mixed and non-competitive mode of inhibition using GraphPad Prism 6.01. Since K i determination is independent of substrate, K i was determined and calculated according to the mixed and non-competitive mode of inhibition. As shown in Table 2, K i values of eckol were 20.26 ± 2.95 and 162.8 ± 3.45 μM for the inhibition of hMAO-A and hMAO-B, respectively. K i values of dieckol were 20.28 ± 1.08 and 18.50 ± 1.63 μM for the inhibition of hMAO-A and hMAO-B, respectively.

Fig. 5
figure 5

Inhibition of human monoamine oxidases A (hMAO-A) (a) and B (hMAO-B) (b) by eckol best fit mixed inhibition mode for hMAO-A and best fit non-competitive inhibition mode for hMAO-B. Global share fit values of K i  ± standard error, n = 3, were calculated by GraphPad Prism 6. K i values were calculated using the mixed for MAO-A and non-competitive for MAO-B mode of inhibition

Full size image
Fig. 6
figure 6

Inhibition of human monoamine oxidases A (hMAO-A) (a) and B (hMAO-B) (b) by dieckol best fit the non-competitive inhibition mode. Global share fit values of K i  ± standard error, n = 3, were calculated by GraphPad Prism 6. All K i values were calculated using the non-competitive mode of inhibition

Full size image
Table 2 Ki value of eckol and dieckol on monoamine oxidase (MAO)-A and MAO-B with the selective index (SI) and inhibition type
Full size table

Docking analysis of eckol and dieckol into hMAO-A and hMAO-B active sites

The docking scores and orientation of eckol and dieckol at the active sites of the hMAO-A and hMAO-B crystal structures are shown in Tables 3 and 4, and Figs. 7 and 8, respectively. The crystallographic poses are visualized for both analogs in the active site of each isozyme compared to their cognate ligands (2z5x and 2v5z) and the cofactor FAD (Figs. 7a, 8a). The hydrogen bond distance, amino acid interactions, binding energy, and orientation of the docked compound within the active site are the basic criteria for successful docking of eckol and dieckol and hMAOs (2z5x and 2v5z) to analyze the receptor-ligand complex model. Eckol showed lowest binding energy (−9.40 kcal/mol) with highest binding affinity as compared with reported inhibitor HRM (−6.95 kcal/mol) and dieckol (−7.48 kcal/mol) toward the inhibition of hMAO-A (Table 3). Eckol showed four hydrogen bond interactions with hMAO-A active site residues (Met445, Tyr69, and Gln215) (Fig. 7c). Some hydrophobic interactions were also observed between Gly443, Gly67, Tyr444, Ile180, Phe352, Tyr407, Cys406, and Arg51 with eckol towards hMAO-A inhibition. Dieckol also showed good binding affinity (−7.48 kcal/mol) in hMAO-A inhibition by forming eight hydrogen bond interactions. Dieckol displayed hydrophobic interaction with Leu259, Gln296, Asn292, Glu286, Ala279, Lys280, and Pro243 residues of the hMAO-A active site.

Table 3 Docking affinity scores and possible H-bond formation to human monoamine oxidase A (hMAO-A) (2z5x) active sites by eckol and dieckol compared with that of the inhibitor harmine (HRM)
Full size table
Table 4 Docking affinity scores and possible H-bond formation to human monoamine oxidase B (hMAO-B) (1gos) active sites by eckol and dieckol compared with that of the inhibitor N-[(E)-methyl](phenyl)-N-[(E)-2-propenylidene]methanaminium (NYP)
Full size table
Fig. 7
figure 7

Binding poses between human monoamine oxidase A (hMAO-A) and eckol, dieckol along with cofactor FAD and inhibitor harmine (HRM) (a). Binding mode of harmine, eckol and dieckol to the hMAO-A (2z5x) active sites (b–d)

Full size image
Fig. 8
figure 8

Binding poses between human monoamine oxidase B (hMAO-B) and eckol, dieckol along with cofactor FAD and inhibitor N-[(E)-methyl](phenyl)-N-[(E)-2-propenylidene]methanaminium (NYP) (a). Binding mode of NYP, eckol, and dieckol to the hMAO-B (1gos) active sites (b–d)

Full size image

Dieckol exhibited lowest binding energy (−7.89 kcal/mol) with a highest binding affinity in hMAO-B inhibition compared to the reported inhibitor NYP (−5.65 kcal/mol) and eckol (−6.77 kcal/mol). Dieckol had eight hydrogen bond interactions in the inhibition of hMAO-B (Table 4). Along with hydrogen bonds, dieckol also showed several hydrophobic interactions with the active site residues of hMAO-B (Fig. 8). In the case of hMAO-B inhibition, eckol showed good inhibitory potency with binding energy (−6.77 kcal/mol). Trp119, His115, Val106, Asn116, and Arg120 residues were involved with eckol in hydrogen bond interactions. Likewise, Asp123, Phe103, Glu483, Pro104, and Tyr112 residues interacted with eckol in hydrophobic interactions (Fig. 8).

Discussion

Brown algae are sea vegetables that are a popular food and health food in China, Japan, and South Korea (Nisizawa et al. 1987). The biological and pharmacological activities of marine brown algae have been clarified, which has revealed brown algae to be an enriched source of highly bioactive secondary metabolites that may be exploited in the development of novel pharmaceutical agents (Ali 2010). Biologically active compounds identified from brown algae include phlorotannins, diterpenes, polysaccharides, phytosterols, and phytopigments.

Of these compounds, phlorotannins—especially eckol and dieckol—have the greatest potential as a functional ingredient in food products, pharmaceuticals, and cosmeceuticals due to bioactivities that include antioxidant, bactericidal, anti-HIV, anticancer, radioprotective, antiallergic, and inhibition of α-glucosidase, α-amylase, angiotensin-I converting enzyme, matrix metalloproteinases, hyaluronidase, and tyrosinase (Li et al. 2011), neuroprotective effect on amyloid beta peptide-induced toxicity in PC12 cells (Ahn et al. 2012), and inhibition of acetylcholinesterase, butyrylcholinesterase and BACE1 in the treatment of Alzheimer’s disease (AD) (Li et al. 2011; Jung et al. 2010).

The present study was prompted by our idea that methanol (MeOH) extract of E. bicyclis, and the isolated eckol and dieckol constituents might have activity against hMAO isozymes for the treatment of PD. An in vitro enzyme assay and molecular docking simulation were used to explore this idea.

The two-step highly sensitive functional luminescence assay was used to investigate hMAO inhibitory activities of the MeOH extract, eckol, and dieckol (Fig. 2). The luminescence assay was more useful than a spectrophotometric assay to determine inhibitory efficacy, potency, selectivity, and mode of inhibition of MeOH extract, eckol, and dieckol. The effects of MeOH extract, eckol, and dieckol on hMAO-A and hMAO-B activities were very clear compared to that of DEP positive control (Fig. 2). Eckol and dieckol showed potent hMAO-A inhibitory IC50 values of 7.20 ± 0.71 and 11.43 ± 1.06 μM, respectively, whereas DEP showed inhibitory IC50 of 10.54 ± 1.25 μM for hMAO-A (Table 1). MeOH extract also showed good inhibitory IC50 of 16.82 ± 0.70 μM for hMAO-A. On the other hand, both eckol and dieckol showed significant inhibitory IC50 values of 83.44 ± 1.48 and 43.42 ± 0.73 μM for hMAO-B, respectively (Table 1). Positive control DEP showed inhibitory IC50 of 0.128 ± 0.01 μM for hMAO-B (Table 1). In the case of hMAO-B inhibition, dieckol was more potent than eckol. Interestingly, MeOH extract exhibited higher hMAO-B inhibitory activity with lower IC50 of 38.38 ± 1.78 μM compared with the inhibitory activity of eckol and dieckol inhibitory. This means that MeOH extract contains other compounds besides eckol and dieckol. In fact, MeOH extract derived brown alga contains different kinds of phlorotannins. This is the first work demonstrating the anti-PD activity of marine edible brown algae and its phlorotannins via inhibiting MAOs.

To explore the mechanism of enzyme inhibition by eckol and dieckol, kinetic analyses were performed at different concentrations of substrate and using various concentrations of eckol and dieckol. Mixed and non-competitive inhibitions were shown by eckol and dieckol after preparing double-reciprocal and Lineweaver–Burk plots for hMAO-A and hMAO-B (Figs. 3, 4). Eckol showed a mixed and non-competitive mode of inhibition for hMAO-A and hMAO-B (Fig. 3), whereas dieckol showed a non-competitive mode of inhibition for both MAO isozymes (Fig. 4). Both eckol and dieckol showed reversible inhibitory activity on MAOs, which is important for the development of enzyme inhibitors. The Ki value for eckol to inhibit hMAO-A and hMAO-B isozymes was 20.26 ± 2.95 and 162.8 ± 3.45 μM, respectively (Fig. 5). The respective Ki values for dieckol were 20.28 ± 1.08 and 18.50 ± 1.63 μM. The lower K i values of eckol and dieckol to inhibit hMAO-A compared to hMAO-B implied tighter binding of eckol and dieckol to hMAO-A, and presumably more effective inhibition.

We also carried out molecular docking simulations to determine the mechanism of the interactions between eckol/or dieckol, and the active site of MAOs isozymes. Molecular docking is a computational technique that has been validated as a means of exploring possible binding modes of a ligand to a given receptor, or another binding site (Goodsell et al. 1996). Eckol showed higher binding affinity with lower binding energy (−9.40 kcal/mol) to hMAO-A than hMAO-B (−6.77 kcal/mol). The NYP inhibitor showed binding affinity of −5.65 kcal/mol for hMAO-B and HRM had a −6.95 kcal/mol binding affinity of hMAO-A. The Met445, Tyr69, and Gln215 hMAO-A active site residues formed strong hydrogen bonds with eckol hydroxyl groups (O2, O4, and O5). The eckol-hMAO-A complex was further stabilized by hydrophobic interactions with the −Gly443, Gly67, Tyr444, Ile180, Phe352, Tyr407, Cys406, and Arg51 residues. On the other hand, dieckol showed higher binding affinity with lower binding energy (−7.89 kcal/mol) to hMAO-B than hMAO-A (−7.48 kcal/mol). The Gly101 and Asn116 hMAO-B active sites residues formed hydrogen bonds with OH groups (O10 and O11) of dieckol. Similarly, the -Thr201, Glu84, and Thr202 residues interacted with the dieckol O16 hydroxyl group through hydrogen bonds, whereas the -Thr195, Thr196, and Asp123 residues formed hydrogen bonds with dieckol hydroxyl group (O17). The dieckol–hMAO-B complex was further strengthened by hydrophobic interactions with -Asn203, Pro102, Arg100, Ser200, Ile477, Thr478, Glu483, Trp119, Arg120, and Phe103 residues. Eckol exhibited lower binding energy than dieckol to inhibit hMAO-A isozyme, but dieckol also showed good inhibitory binding energy (−7.48 kcal/mol) compared to HRM (−6.95 kcal/mol). Likewise, eckol showed significant binding energy (−6.77 kcal/mol) towards the inhibition of hMAO-B isozyme compared to NYP (−5.65 kcal/mol).

The molecular docking analysis of eckol and dieckol on hMAO-A and hMAO-B crystalline structures predicted very different binding behaviors, which were illustrated in a LigPlot + analysis (Laskowski and Swindells 2011) (Figs. 7b–d, 8b–d). During this docking simulation, the presence of hydroxyl group in the compound structure was very important and increased the tendency of hydrogen bond interaction. There are more hydroxyl groups in dieckol than in eckol structure, which was reflected in the greater hydrogen bond interaction of dieckol compared to eckol in the inhibition of the MAO-A and MAO-B isozymes. No specific residues were responsible for the inhibition of the MAO-A and MAO-B isozymes. Concerning the inhibition of hMAO-A, the inhibitory activity of eckol differed from that of dieckol, likely reflecting their different binding sites on hMAO-A (Fig. 7a). Similarly, the difference in the inhibitory activity of eckol and dieckol for hMAO-B was reflected in the crystallographic structures (Fig. 8a).

Eckol and dieckol have not been used to treat PD, However, both possess anti-inflammatory activities (Jung et al. 2013) and a neuroprotective effect on amyloid beta peptide-induced toxicity in PC12 cells (Ahn et al. 2012). PD is related with oxidative stress (Hwang 2013). Eckol and dieckol have antioxidant activity (Kang et al. 2003; 2013). Eckol and dieckol also potently inhibit peroxisome proliferator-activated receptor γ (Jung et al. 2014), which protect against PD neurodegeneration (Carta 2013). The capability of oral absorption and the ability of both eckol and dieckol to cross the blood–brain barrier could be of potential value in the treatment of PD, if these attributes can be proven in future studies. The present in vitro data studies suggest that eckol and dieckol inhibit MAOs; in vivo studies are needed for confirmation. The anti-inflammatory properties and neuroprotective activities of eckol and dieckol in addition to their reversible inhibition of the MAO enzymes implicate them as useful compounds for PD therapy.

In conclusion, the marine phlorotannins, eckol and dieckol can reversibly inhibit MAOs. Biochemical and molecular docking examinations revealed mixed and non-competitive catalytic site related mechanisms. Hence, eckol and dieckol could be safer substitutes for currently used classical MAOs inhibitors. Eckol and dieckol could be a novel class of natural reversible MAO inhibitor. The properties of these phlorotannins may be of value to therapeutically manage PD and other neurological disorders.