In Vitro and In Silico Characterization of G-Protein Coupled Receptor (GPCR) Targets of Phlorofucofuroeckol-A and Dieckol

Phlorotannins are polyphenolic compounds in marine alga, especially the brown algae. Among numerous phlorotannins, dieckol and phlorofucofuroeckol-A (PFF-A) are the major ones and despite a wider biological activity profile, knowledge of the G protein-coupled receptor (GPCR) targets of these phlorotannins is lacking. This study explores prime GPCR targets of the two phlorotannins. In silico proteocheminformatics modeling predicted twenty major protein targets and in vitro functional assays showed a good agonist effect at the α2C adrenergic receptor (α2CAR) and an antagonist effect at the adenosine 2A receptor (A2AR), δ-opioid receptor (δ-OPR), glucagon-like peptide-1 receptor (GLP-1R), and 5-hydroxytryptamine 1A receptor (5-TH1AR) of both phlorotannins. Besides, dieckol showed an antagonist effect at the vasopressin 1A receptor (V1AR) and PFF-A showed a promising agonist effect at the cannabinoid 1 receptor and an antagonist effect at V1AR. In silico molecular docking simulation enabled us to investigate and identify distinct binding features of these phlorotannins to the target proteins. The docking results suggested that dieckol and PFF-A bind to the crystal structures of the proteins with good affinity involving key interacting amino acid residues comparable to reference ligands. Overall, the present study suggests α2CAR, A2AR, δ-OPR, GLP-1R, 5-TH1AR, CB1R, and V1AR as prime receptor targets of dieckol and PFF-A.


Introduction
G protein-coupled receptors (GPCRs) are a family of membrane receptors that regulate human pathophysiology and are the leading target class for pharmaceuticals. At present, GPCRs mediate the effect of approximately one-third of the FDA-approved drugs [1][2][3]. However, these drugs target mainly biogenic amine receptors, which comprise around 30 members of the GPCR family [3]. There is, therefore, an immense potential within pharmaceuticals/natural products to exploit, considering the remaining family members for which no existing ligands have been identified.
In the traditional drug development process, the high-throughput screening (HTS) approach against drug targets of choice is the very first step to uncover new drugs, which has now been augmented by the in silico method to maximize the probability of novel leads discovery. Traditional Chinese medicine (TCM) is an important research object of network

In Silico Target Prediction
Proteocheminformatics (PCM) modeling is a quantitative bio-modeling technique that can predict the affinity and potency of a ligand against multiple different protein targets simultaneously by combining chemical and biological information from the ligand and related targets into a single machine learning model [23]. From in silico PCM modeling, the highest-ranked twenty potential protein targets were predicted for the phlorotannins. Table 1 presents a list of the target proteins with an average score value. As shown in the Table 1, the V 1A receptor was predicted as a top target for dieckol and PFFA. For PFF-A, 5-hydroxytryptophan 1A (5-HT 1A R), 5-hydroxytryptophan 1B (5-HT 1B R), and cannabinoid 1 (CB 1 R) receptors were among the predicted top twenty protein targets. Based on this prediction and reported biological activities of the phlorotannins in the literature, we proceeded to validate adenosine A 2A receptor (A 2A R), alpha-2A adrenergic receptor (α 2A AR), alpha-2C adrenergic receptor (α 2C AR), δ-opioid receptor (δ-OPR), CB 1 R, free fatty acid receptor 1 (FFA 1 R or GPR40), glucagon-like peptide-1 receptor (GLP-1), V 1A R, 5-HT 1A R, and 5-HT 1B R cell-based functional assays.
Based on the functional effect above 50% at 100 µM, the concentration-dependent effect was further tested and compared with the reference agonists and antagonists (Figures 2 and 3 and Tables 3 and 4) followed by molecular docking simulation. Molecular docking simulation of test ligands to the crystal structures of target proteins and comparison with the reference ligands results revealed the mechanism of ligand-targetprotein interaction. a Value was extracted from our previous study [22]. b The test compound induces at least a 25% agonist effect at this concentration, which results in an apparent inhibition.
In the docking simulation, dieckol formed two H-bond interactions with Ile80 and Asp170 ( Figure 4B) while four H-bond interactions (His278, Ala59, Ala81, Ser67) were observed for PFF-A ( Figure 3C). The binding of reference ligands to the A 2A R crystal structure showed the involvement of residues Phe168, Leu249, Asn253, and Met270. The total number of hydrophobic and electrostatic interactions involved in dieckol binding was greater than that of PFF-A binding (Table S2). Interestingly, only one interacting residue (Leu249) was in common with the reference ligand. However, PFF-A had two common interacting residues (Leu249 and Phe168) with reference antagonist ZM241385 (Table S1).

Dieckol and PFF-A as α 2C AR Agonists
Evaluation of the concentration-dependent agonist effect of phlorotannins (Table 3 and Figure 2A) at α 2C AR depicted dieckol as a moderate agonist (EC 50 : 98.80 ± 7.71 µM) and PFF-A as a good agonist (EC 50 : 23.67 ± 3.32 µM). Even at a 25-µM concentration, PFF-A stimulated the effect of 1 µM epinephrine by 55%. The reference agonist epinephrine had an EC 50 value of 0.86 nM. To further support the functional effect and delineate the difference in activity between the two phlorotannins, a molecular docking simulation of test ligands and target protein was performed.
The reference antagonist naltrindole showed an H-bond interaction with aspartic acid residue (Asp128) and numerous hydrophobic interactions with tryptophan residues -Trp284 (π-π-T-shaped), Trp284 (π-alkyl), and Trp274 (π-alkyl). Only Asp128 was a common interacting residue among the test and reference ligands while Tyr308 was observed for PFF-A and reference ligand binding, but not for dieckol (Tables S1 and S2).

Dieckol and PFF-A as GLP-1R Antagonists
Results from the functional assay on mouse GLP-1 receptor-expressed βTC6 cells demonstrated dieckol and PFF-A as full antagonists of the GLP-1 receptor. At a concentration of 100 µM, both the compounds inhibited the effect of 0.3 nM GLP-1(7-37) by 100%. However, at the 25-µM concentration, PFF-A inhibited the reference agonist-response by 57.37% and dieckol by 21.23%. Additionally, a dose-dependent response curve yielded IC 50 values of 47.19 ± 2.46 and 21.56 ± 2.16 µM for dieckol and PFF-A, respectively (Table 4 and Figure 3C). The potency of PFF-A was two-fold higher than that of dieckol. The reference antagonist exendin-3(9-39) had an IC 50 value of 4.6 nM. From the molecular docking study, hydrogen-bond interactions with Ser352 and Thr355 (Table S3) and hydrophobic interactions with Leu354, Lys351, and Val405 (Table S4) were common observations in test ligands and reference antagonist NNC0640 binding with an inactive-state GLP-1R (5vex) ( Figure 5C-E) in our molecular docking simulation. An unfavorable contact between dieckol and GLP-1R receptor was observed via the Asn407 residue.

Dieckol and PFF-A as 5-HT 1A R Antagonists
An antagonist effect was observed for dieckol and PFF-A in a cell-based functional assay. At 100-µM concentration, dieckol and PFF-A inhibited the response of 30 nM serotonin by 91.0 ± 3.11% and 77.00 ± 11.03%, respectively (Table 2). A concentration-dependent dose-response showed that dieckol and PFF-A inhibited the 50% response of 30 nM serotonin at 43.31 ± 3.22 and 17.75 ± 3.42 µM, respectively (Table 4 and Figure 3E). However, the agonist effect at 5-HT 1A R was negligible for both the compounds when tested at the 100-µM concentration. As a result, the EC 50 value was not determined. Reference drug serotonin had an EC 50 value of 0.72 nM and antagonist GR55562 had an IC 50 value of 4.4 nM.
Docking of test and reference ligands to the active site of 5-HT 1A R demonstrated that aspartic acid residue Asp116 is one of the important binding residues ( Figure 6D). Dieckol formed an H-bond interaction with Asp116, Thr200, Ser190, Asn386, and Tyr96 while PFF-A did with Thr188, Glu372, Tyr96, and Asn386 ( Figure 6E,F). Reference ligands serotonin and WAY 100635 formed an H-bond interaction with Asp116 via a salt-bridge. Interactions with Thr200, Phe361, and Val117 were observed for test ligands and serotonin binding (Tables S5 and S6).

Discussion
Dieckol and PFF-A are phloroglucinol (1,3,5-trihydroxybenzene)-based polyphenols with a varied number of phloroglucinol units attached via dibenzofuran and dibenzodioxin linkages. Dieckol is a phloroglucinol hexamer and PFF-A is a phloroglucinol pentamer. A structure-activity relationship between phloroglucinol and its oligomers in our recent study [22] showed that more than three repeating phloroglucinol units are necessary for hMAOs inhibition and D 3 /D 4 receptor agonist effect. Likewise, oligomerization of phloroglucinol with more than five repeating units is essential for the antagonist effect at D 1 , NK 1 , and 5-HT 1A receptors. Here, although the monomer phloroglucinol is not included in the study, the pentamer (PFF-A) showed better activity than a hexamer (dieckol). An interesting observation in this study is that regardless of the receptors at which these two phlorotannins showed functional effects (except the hV 1A R), PFF-A was two-fold more potent than dieckol. In contrary to the findings that the phenolic -OH groups attached to the benzene ring of polyphenols play a vital role in the antioxidant effect [24][25][26] and that an increase in the number of hydroxyl groups increases antioxidant activity, the functional effect of PFF-A at tested GPCRs was higher than that of dieckol despite having a lower number of hydroxyl groups. The possible reason underlying this might be the structure or orientation of PFF-A that enables it to reach the core active site cavity of receptors where it binds to conserved interacting residues leading to conformational change.
Adenosine is an endogenous autacoid that regulates cellular physiology via adenosine A 1 , A 2A , A 2B , and A 3 receptors. These receptors are expressed in several cells and tissues throughout the body and play a crucial role in regulating the pathophysiology of the human body, suggesting a potential drug target. Of different adenosine receptor subtypes, A 2A R is the main receptor subtype in the striatum colocalized with dopamine D 2 receptor and it modulates motor function [27,28]. Activation of A 2A R decreases the binding affinity of D 2 R for agonists, implying A 2A R antagonists as novel therapeutics for Parkinson's disease [29]. At the synapse, A 2A R facilitates glutamate release and potentiates NMDA receptor effects. It also stimulates glutamate release in astrocytes by inhibiting glutamate transporter-1 (GLT-1), and the level of A 2A Rs in neurons and glia is significantly high in depression and Schizophrenia [30]. Hence, A 2A Rs antagonists are effective as antidepressants and anti-anxiety agents. Here, dieckol and PFF-A showed an antagonist effect at hA 2A R with IC 50 values of 87.18 ± 2.63 and <50 µM, respectively. Furthermore, molecular docking simulation showed that dieckol and phlorofucofuroeckol-A strongly interact with the Phe168 residue, which is known as one of the important residues for ligand binding, via pi-pi interaction [31]. Structurally, dieckol and PFF-A are powerful radical scavengers [32] and as such, dieckol, in a recent study [33], protected dopaminergic neuronal cells by preventing α-synuclein aggregation via antioxidant mechanism. In a previous study [34], dieckol suppressed LPS-induced excessive microglial activation and protected neuronal cells by downregulating extracellular signal-regulated kinases, protein kinase B (PKB/Akt), and nicotinamide adenine dinucleotide phosphate hydrogen (NADPH) oxidase-mediated pathways. Likewise, PFF-A inhibited glutamate-induced apoptotic PC12 cell death in a caspase-dependent manner [35].
Similarly, at α 2C AR, PFF-A showed a strong agonist effect and formed an H-bond with the Asp131 of α 2C AR, which is the conserved active site residue. Adrenergic receptors are targets for epinephrine and norepinephrine and are involved in maintaining homeostasis. Among several types of adrenergic receptors, highly expressed α 2 adrenoreceptors in astrocytes, and in glutamatergic and GABAergic neurons act by increasing intracellular Ca 2+ levels [36]. The α 2C AR subtype mediates cold-induced vasoconstriction, inhibits dopamine release in basal ganglia [37], and serotonin in the mouse hippocampus [38]. Therefore, α 2C AR selective ligands have a therapeutic role in neuropsychiatric disorders [38] and α 2C AR agonists are implicated in the treatment of neuropathic pain [39][40][41].
Serotonin (5-hydroxytryptamine, 5-HT) is a monoamine neurotransmitter that plays a crucial role in physiological functions, and of a total of 14 subtypes of 5-HT receptors, the 5-HT 1A receptor is a prominent target for the treatment of various neuropsychiatric and neurological disorders, prominently depression [42]. In the functional assay, dieckol and PFF-A showed a good antagonist effect at 5-HT 1A R. Furthermore, they interacted with conserved aspartate residue (Asp116) of 5-HT 1A R via H-bond and pi-anion binding, respectively.
Vasopressin is an antidiuretic hormone that plays a vital role in the central nervous system (CNS) and peripheral nervous system (PNS). The vasopressin receptor is one of the promising targets for CNS drugs, and vasopressin antagonists represent a novel approach for the treatment of stress, mood, and behavioral disorders [43]. Likewise, as a peripheral role, V 1A R is responsible for vasoconstriction, myocardial contractility, platelet aggregation, and uterine contraction [44]. Similarly, in a recent study [45], upregulated vasopressin 1 receptor (V 1 R) expression in hepatocytes of ischemia-reperfusion injury mouse model was identified and the V 1 R/Wnt/β-catenin/FoxO3a/Akt pathway was highlighted as vital for hepatoprotection.
Cannabinoid CB 1 receptors are among the most abundant GPCRs in the brain and they modulate CNS activity [46]. Cannabinoid CB 1 receptor agonist activation of the CB 1 receptor leads to decreased levels in cellular cAMP via inhibition of adenylyl cyclase. Moreover, CB 1 activation inhibits voltage-gated Ca 2+ channels and activates K + channels, and these overall intracellular signaling activities reduce cellular excitability [47]. Likewise, studies also indicate high expression levels of CB 1 R in various types of cancer [48,49]. Interestingly, a new study demonstrated a higher orexigenic effect of the CB 1 R agonist AM11101 than tetrahydrocannabinol [50]. This shows that CB 1 R agonists could be used as an appetite stimulant in underweight patients. In the present study, only PFF-A showed a promising agonist effect at CB 1 R with an EC 50 of 13.42 ± 2.03 µM. Several reports on PFF-A show neuroprotective effects mainly via antioxidant mechanisms [14,35,51]. Likewise, a recent study suggested the ATF3-mediated pathway as a possible mechanism of PFF-A-induced apoptosis in human colorectal cancer cells [52]. However, it remains unclear whether the neuroprotective and anticancer effect of PFF-A is via CB 1 R agonist activity.
The human CB 1 receptor is an important therapeutic target for obesity and obsessive disorders and the mechanism of its transition state (either active or inactive) is vital for understanding the regulatory action of the receptor [53]. A salt bridge between conserved Asp-Arg-Tyr (DRY) motif in the C-terminal region of transmembrane 3 (TM3) and transmembrane 6 (TM6) characterizes the active or inactive conformation of the rhodopsin-like GPCRs [54]. In an inactive conformation of CB 1 R, TM6 packs against TM3 and transmembrane 5 (TM5) and G protein-interacting residues-Phe200 (helix III) and Trp356 (helix VI) are obstructed [55]. The reference inverse agonist (taranabant) is bound to the inactive state crystal structure by forming an H-bond interaction between the NH of taranabant and the hydroxyl of Ser383 and the −CF 3 group with Ser173, Phe189, and Lys192. This result corroborates the findings of a previous study [56] which concluded that a strong H-bond between the -NH group of taranabant and the hydroxyl of Ser383 was vital for superior affinity to CB 1 R. Likewise, the agonist CP55940 formed π-π interactions with Phe170 and Phe268, and two H-bond interactions with Ser173 and Ser383 in a similar fashion, as reported earlier [56]. PFFA also formed a stable pi-pi interaction with Phe268 and an H-bond interaction with Ser173 of the active state crystal structure (6kqi), which could explain the agonist potency of PFFA in vitro.
Among the tested protein targets, CB 1 R, GLP-1, and GPR40 are obesity/T2DM related GPCRs and in the functional assays, PFF-A showed a good agonist effect at CB 1 R, while both the dieckol and PFF-A showed an antagonist effect at the GLP-1 receptor. Their effect at GPR40 was mild agonist. A gut-derived incretin hormone GLP-1 stimulates insulin and suppresses glucagon secretion, inhibits gastric emptying, and reduces appetite and food intake. In a previous study, intracerebroventricular injection of exendin (9-39), a specific GLP-1 antagonist, blocked the inhibitory effect of GLP-1 on food intake [57]. Hence, GLP-1 agonists represent a new class of antidiabetic agents [58]. In a recent study on the anti-diabetic effect in the zebrafish model [59], dieckol treatment reduced liver glucose-6-phosphate and phosphoenolpyruvate carboxykinase, and enhanced glucose transport and insulin sensitivity via protein kinase B (Akt) phosphorylation. It is of note that dieckol and PFF-A showed a good antagonist effect at GLP-1. Thus, the in vivo effects of these phlorotannins in GLP-1-mediated signaling are urgent.
In conclusion, the present study characterizes the receptors hA 2A R, hα 2C AR, hδ-OP, CB 1 R, GLP-1, hV 1A R, and h5-HT 1A R as prime protein targets of dieckol and PFF-A. Moreover, the binding mechanism of test ligands with the target proteins strengthens the study and warrants further in vivo studies.

Isolation of Phlorotannins
Phlorotannins-dieckol and PFF-A were isolated from the ethyl acetate fraction of E. stolonifera ethanolic extract, as described previously [11,22].

In Silico Prediction of Targets
To predict potential protein targets for the phlorotannins, a proteocheminformatics modeling (PCM) in silico target prediction method was employed, as described recently [60]. For full information on the model, readers are further directed to a previous report [61].
The in-house assay protocol and experimental conditions are reported in our previous reports [15,22,62]. The functional effect of dieckol and PFF-A was characterized based on their modulation effect on cytosolic Ca 2+ ion mobilization using a fluorimetric detection method or by measuring their effect on cAMP modulation using homogeneous timeresolved fluorescence (HTRF) detection.

Measurement of cAMP Level
Functional activity of phlorotannins over hA 2A R, hα 2C AR, hCB 1 R, GLP-1R, and h5-HT 1B R was determined by measuring their effects on cAMP production by the HTRF detection method using transected cells expressing human cloned receptors.

Functional Activity over hA 2A R
In brief, the PC12 cells were suspended in HBSS buffer (Invitrogen) complemented with 20 mM HEPES (pH 7.4), 0.2 U/mL ADA, and 100 µM rolipram, then distributed in microplates at a density of 2.10 3 cells/well and preincubated for 5 min at room temperature (RT) in the presence of HBSS (basal control), the test compound, or the reference agonist or antagonist. For stimulated control measurement, separate assay wells contained 3 µM NECA. Following 10 min incubation at RT, the cells were lysed and the fluorescence acceptor (D2-labeled cAMP) and fluorescence donor (anti-cAMP antibody labeled with europium cryptate) were added. After 60 min at RT, the fluorescence transfer was measured at λex = 337 nm and λem = 620 and 665 nm using an EnVision microplate reader EnSpire (PerkinElmer, Waltham, MA, USA). The cAMP concentration was determined by dividing the signal measured at 665 nm by that measured at 620 nm (ratio). Agonist result was expressed as a percent of the control response to 3 µM NECA while the antagonist effect as percent inhibition of the control response to 100 nM NECA. The standard reference agonist was NECA and the antagonist was ZM 241385, which were tested in each experiment at several concentrations to generate a concentration-response curve from which their EC 50 and IC 50 values were calculated.

Functional Activity over hα 2C AR
Briefly, the transfected CHO cells suspended in HBSS buffer (Invitrogen) complemented with 20 mM HEPES (pH 7.4) and 500 µM IBMX were distributed in microplates at a density of 10 4 cells/well in the presence of either of the following: For agonist assay-HBSS (basal control), epinephrine 1 µM (stimulated control) or various concentrations (EC 50 determination), or the test compounds. For antagonist assay-HBSS (stimulated controls), rauwolscine 10 µM (basal control) or various concentrations (IC 50 determination), or the test compounds. The reference agonist epinephrine and the adenylyl cyclase activator NKH 477 were added at respective final concentrations of 100 nM and 5 µM. For basal control measurements, epinephrine was omitted from the wells containing 3 µM rauwolscine. After 10 min at 37 • C, the cells were lysed and the fluorescence acceptor (D2-labeled cAMP) and fluorescence donor (anti-cAMP antibody labeled with europium cryptate) were added. After 60 min at RT, the fluorescence transfer was measured at λex = 337 nm and λem = 620 and 665 nm using a microplate reader (Envision, Perkin Elmer). The concentration of cAMP was determined by dividing the measured signal at 665 nm by that measured at 620 nm (ratio). The agonist result are shown as a percent of the control response to 1 µM epinephrine and the antagonist result are expressed as a percent inhibition of the control response to 30 nM epinephrine. Epinephrine and rauwolscine were the standard reference drugs used in each experiment at different concentrations.

Functional Activity over hCB 1 R
The transfected CHO cells were suspended in HBSS buffer (Invitrogen) complemented with 20 mM HEPES (pH 7.4). Then, the cells were distributed in microplates at a density of 5.10 3 cells/well in the presence of either of the following: For agonist assay-HBSS (basal control), 30 nM CP 55940 (stimulated control) or various concentrations (EC 50 determination), or the test compounds. For antagonist assay-HBSS (stimulated controls), 10 µM AM 281 (basal control) or various concentrations for IC 50 determination, or the test compounds. Thereafter, the reference agonist CP 55940 and the adenylyl cyclase activator forskolin were added at respective final concentrations of 1 nM and 25 µM. For basal control measurements, CP 55940 was excluded from the wells containing 10 µM AM 281. After 30 min of incubation at 37 • C, the cells were lysed and the fluorescence acceptor (D2-labeled cAMP) and fluorescence donor (anti-cAMP antibody labeled with europium cryptate) were added. The fluorescence transfer was measured at λex = 337 nm and λem = 620 and 665 nm using an Envision microplate reader (PerkinElmer, Waltham, MA, USA) after 60 min at RT. The agonist results are expressed as a percent of the control response to 10 nM CP 55940 and the antagonist results are expressed as percent inhibition of the control response to 1 nM CP 55940. CP 55940 and AM 281 were standard reference drugs that were tested in each experiment.

Functional Activity over GLP-1R
The HBSS buffer (Invitrogen) complemented with 20 mM HEPES (pH 7.4) and 500 µM IBMX was used to suspend and distribute the βTC6 cells at a density of 1.5x10 4 cells/well. The plate was then incubated for 10 min at RT in the presence of HBSS (basal and stimulated control), the test compound, or the reference agonist and antagonist. In the agonist assay, separate assay wells containing 100 nM GLP-1  were prepared for the stimulated control measurement, while in the antagonist assay, the reference agonist GLP-1(7-37) was added at a final concentration of 0.3 nM, and separate assay wells contained HBSS for basal control measurements. Following incubation, the cells were lysed and the fluorescence acceptor (D2-labeled cAMP) and fluorescence donor (anti-cAMP antibody labeled with europium cryptate) were added. After 60 min at room temperature, the fluorescence transfer was measured at λex = 337 nm and λem = 620 nm and 665 nm using an Envision microplate reader (PerkinElmer, Waltham, MA, USA). The results are expressed as either a percent of the control response to 100 nM GLP-1(7-37) or a percent inhibition of the control response to 0.3 nM GLP-1 . The standard reference agonist was GLP-1(7-37) and the antagonist was exendin-3(9-39). 4.5.5. Functional Activity over 5-HT 1B R Concisely, a plasmid containing the GPCR gene of interest (5-HT 1B ) was transfected into Hela cells. The resulting stable transfectants were suspended in HBSS buffer (Invitrogen, Carlsbad, CA, USA) containing 20 mM HEPES (pH 7.4), 400 mM NaCl, 1 mg/mL glucose, and 500 µM IBMX and distributed in microplates at a density of 2 × 10 4 cells/well. The plates were then incubated for 20 min at RT in the presence of either of the following: HBSS and 0.1% BSA (basal control), serotonin at 10 µM (stimulated control) or various concentrations for EC 50 determination, or the test phlorotannins. Thereafter, the adenylyl cyclase activator NKH 477 (5 µM) was added and the plates were incubated at 37 • C for 20 min. Then, the cells were lysed and a fluorescence acceptor (D2-labeled cAMP) and fluorescence donor (anti-cAMP antibody with europium cryptate) were added following 60 min incubation at RT. After incubation, the fluorescence transfer was measured using an Envision microplate reader (PerkinElmer, Waltham, MA, USA) and the results are expressed as a percentage of the control response to 10 µM serotonin for the agonist effect and as percent inhibition of the control response to 100 nM serotonin.

Functional Activity over hα 2A AR
The rat basophil leukemia cells were distributed in microplates at a density of 1.1 × 10 4 cells/well after suspending in a HBSS buffer (Invitrogen) containing 20 mM HEPES. Then, the fluorescent probe (Fluo8, AAT Bioquest) mixed with probenecid in HBSS buffer (Invitrogen) complemented with 20 mM Hepes (Millipore) (pH 7.4) was added into each well incubated for 60 min at 30 • C. Thereafter, the assay plates were positioned in a microplate reader (FlipR Tetra, Molecular Device) and we added test compounds, reference agonist/antagonist or HBSS buffer (basal control). Change in fluorescence intensity which varies proportionally to the free cytosolic Ca 2+ ion concentration was measured. For stimulated control measurements, separate assay wells containing 0.1 µM epinephrine bitartrate were prepared. The agonist effect was calculated as a % of control response to epinephrine bitartrate at 0.1 µM. Similarly, for the antagonist effect, % inhibition of the control response to epinephrine bitartrate at 3 nM was evaluated. Epinephrine bitartrate and RX-821002 were used as reference agonists and antagonists, respectively.

Functional Activity over hδ-OPR
At first, rat basophil leukemia cells were suspended in HBSS buffer (Invitrogen) complemented with 20 mM HEPES, and distributed in microplates at a density of 2.768 × 10 4 cells/well. Thereafter, a mixture of fluorescent probe (Fluo8, AAT Bioquest) and probenecid in HBSS buffer (Invitrogen) complemented with 20 mM Hepes (Millipore) (pH 7.4) was added and plates were incubated for 60 min at 30 • C. Then, the assay plates were positioned in a FlipR Tetra microplate reader (Molecular Device, San Jose, CA, USA) for the addition of the test compound, reference agonist/antagonist, or HBSS buffer (basal control). Change in fluorescence intensity that varies proportionally to the free cytosolic Ca 2+ ion concentration was measured.
For stimulated control measurements, 1 µM DPDPE was added in separate assay wells. The results are expressed as a percent of the control response to DPDPE at 1 µM or a percent inhibition of the control response to DPDPE at 25 nM. The standard reference agonist and antagonist were DPDPE and naltriben mesylate, respectively. 4.6.3. Functional Activity over FFA 1 R/GPR40 In general, transfected HEK-293 cells suspended in DMEM buffer (Invitrogen) containing 1% FCSd were distributed in microplates at a density of 2.10 4 cells/well. Then, the mixture of fluorescent probe (Fluo4 Direct, Invitrogen) and probenecid in HBSS buffer (Invitrogen) complemented with 20 mM Hepes (Invitrogen) (pH 7.4) was added into each well and incubated for 60 min at 37 • C, followed by 15 min incubation at 22 • C. Thereafter, the assay plates were positioned in a CellLux microplate reader (PerkinElmer, Waltham, MA, USA) which was used for the addition of the following: For agonist assay-test compound, reference agonist, or HBSS buffer (basal control). Linoleic acid at 100 µM was added in separate assay wells for stimulated control measurement. For antagonist assay-test compound or HBSS buffer (basal and stimulated control), then, 5 min later, 20 µM linoleic acid. Agonist results are expressed as a percent of the control response to 100 µM linoleic acid while antagonist results are expressed as percent inhibition of the control response to 20 µM linoleic acid. 4.6.4. Functional Activity over hV 1A R Briefly, CHO-V 1A R cells were separately suspended in DMEM buffer (Invitrogen, Carlsbad, CA, USA) complemented with 0.1% FCSd and distributed into microplates (4.5× 10 4 cells/well). Then, fluorescent probe (Fluo4, Invitrogen) mixed with probenecid in HBSS buffer (Invitrogen, Carlsbad, CA, USA) supplemented with 20 mM HEPES, pH 7.4 (Invitrogen) was added to each well, allowing to equilibrate with the cells for 60 min at 37 • C, then 15 min at 22 • C. Thereafter, the assay plates were positioned in a CellLux microplate reader (PerkinElmer, Waltham, MA, USA) and dieckol and PFF-A (12.5, 25, 50, 100, and/or 150 µM), reference agonist, or HBSS buffer (basal control) was added. For stimulated control measurements, AVP at 1 µM was added in separate assay wells. The agonist effect on V 1A R was calculated as a % of control response to 1 µM AVP. Similarly, for the antagonist effect, % inhibition of the control response to 10 nM AVP was evaluated. AVP and [d(CH 2 ) 5 1 , Tyr (Me) 2 ]-AVP were used as reference agonist and antagonist, respectively. 4.6.5. Functional Activity over h5-HT 1A R In brief, Ba/F3-5HT 1A R cells were first suspended in HBSS buffer (Invitrogen, Carlsbad, CA, USA) complemented with 20 mM HEPES buffer (pH 7.4). Then, the cells were distributed into microplates at a density of 1 × 10 6 cells/well. Subsequently, fluorescent probe (Fluo8, AAT Bioquest) mixed with probenecid in HBSS buffer (Invitrogen, Carlsbad, CA, USA) supplemented with 20 mM HEPES (Invitrogen) (pH 7.4) was added to each well, and the plates were incubated for 60 min at 37 • C. Thereafter, plates were fixed in a FlipR Tetra microplate reader (Molecular Device, San Jose, CA, USA) and dieckol and PFF-A (12.5, 25, 50, 100 and/or 150 µM), reference agonist, or HBSS buffer (basal control) was added. Fluorescence intensity was measured which varied in proportion to the free cytosolic Ca 2+ ion concentration. Agonist effect on 5-HT 1A R was calculated as a % of control response to 2.5 µM serotonin. Similarly, the percentage inhibition of the control response to 30 nM serotonin was calculated for the antagonist effect. Serotonin and (S)-WAY-100635 were used as reference agonists and antagonists, respectively.

Homology Modeling and Molecular Docking
The primary sequence of the human 5-HT 1A R and human V 1A R was obtained from UniProt (ID: P08908 and P37288, respectively). Based on the SWISS-MODEL, the 5-HT 1B receptor (PDB: 5V54) was selected as a template for homology modeling of human 5-HT 1A because it showed a good sequence similarity (0.42), sequence identity (42.97), and quaternary structure quality estimate (QSQE) (0.32) to this receptor. Similarly, µ-opioid receptor (PDB: 4DKL) was selected as a template for homology modeling of human V 1A R, because it showed a good sequence similarity (0.32), sequence identity (24.54), and QSQE (0.19) to this receptor. The constructed model was refined using the ModRefiner server. Automated docking simulations were carried out with the AutoDock 4.2. program [63]. The structures of dieckol and PFF-A were generated and converted into 3D structures using Marvin Sketch (v17,1,30, ChemAxon, Budapest, Hungary). Structures of dieckol and PFF-A were energy-minimized using a molecular mechanics 2 (MM2) force field. X-ray crystallographic structures of GPCRs were obtained from the RCSB protein data bank (PDB) with respective PDB IDs-hA 2A R (3eml) [31], hα 2C AR (6kuw), hδ-OP (4ej4) [64], hCB 1 R (6kqi) [65], and hGLP-1 [66]. The structures of reported agonists (5 -N-ethylcarboxamidoadenosine (NECA), epinephrine, DPI-287, CP 55940, PF-06882961, AVP, and serotonin, and antagonists (ZM241385, RS-79948, naltrindole, taranabant, NNC0640, SR49059, and WAY 100635) were downloaded from PubChem or PDB. For each ligand-protein complex, 10 docking poses were generated using the same grid parameters (size and center) and docking parameters (genetic algorithm and run options). The pose for the lowest binding energy was chosen for the final docking result. When the root-mean-square deviation (RMSD) value between our docking result and the original crystallographic structures of the protein was less than 0.15 nm, we considered our docking protocol to be valid and performed the simulation. Results were analyzed and visualized using Discovery Studio (v17.2, Accelrys, San Diego, CA, USA).