==== Front PLoS One PLoS One plos PLOS ONE 1932-6203 Public Library of Science San Francisco, CA USA 10.1371/journal.pone.0277873 PONE-D-22-04470 Research Article Physical Sciences Chemistry Physical Chemistry Chemical Bonding Hydrogen Bonding Physical Sciences Physics Thermodynamics Free Energy Medicine and Health Sciences Pharmacology Drug Interactions Physical Sciences Chemistry Computational Chemistry Molecular Dynamics Biology and Life Sciences Physiology Physiological Parameters Body Weight Obesity Medicine and Health Sciences Pharmacology Pharmacokinetics Biology and Life Sciences Toxicology Toxicity Medicine and Health Sciences Pathology and Laboratory Medicine Toxicology Toxicity Medicine and Health Sciences Pharmacology Drug Screening Poria cocos compounds targeting neuropeptide Y1 receptor (Y1R) for weight management: A computational ligand- and structure-based study with molecular dynamics simulations identified beta-amyrin acetate as a putative Y1R inhibitor A computational molecular docking study of neuropeptide Y1 receptor Y1R inhibitors and allosteric modulators Wong Ann Rann Conceptualization Data curation Formal analysis Investigation Methodology Writing – original draft Writing – review & editing 1 Hung Andrew Conceptualization Data curation Formal analysis Investigation Methodology Supervision Validation Writing – review & editing 2 Yang Angela Wei Hong Conceptualization Investigation Methodology Supervision Validation Writing – review & editing 1 Gill Harsharn Conceptualization Methodology Supervision Validation Writing – review & editing 2 https://orcid.org/0000-0002-1315-5242 Lenon George Binh Conceptualization Funding acquisition Investigation Methodology Validation Writing – review & editing 1 * 1 School of Health and Biomedical Sciences, RMIT University, Bundoora, Victoria, Australia 2 School of Science, RMIT University, Melbourne, Victoria, Australia Selvaraj Chandrabose Editor Alagappa University, INDIA Competing Interests: The authors have declared that no competing interests exist. * E-mail: george.lenon@rmit.edu.au 30 6 2023 2023 18 6 e027787314 2 2022 9 10 2022 © 2023 Wong et al 2023 Wong et al https://creativecommons.org/licenses/by/4.0/ This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Poria cocos (PC) is a medicinal herb frequently used in weight-loss clinical trials, however the mechanisms by which its compounds target orexigenic receptors including the neuropeptide Y1 receptor (Y1R) remain largely unknown. This study aimed to screen PC compounds for favourable pharmacokinetics profiles and examine their molecular mechanisms targeting Y1R. Forty-three PC compounds were systematically sought from pharmacological databases and docked with Y1R (PDB: 5ZBQ). By comparing the relative binding affinities, pharmacokinetics and toxicity profiles, we hypothesised that compounds designated PC1 3,4-Dihydroxybenzoic acid, PC8 Vanillic acid, PC40 1-(alpha-L-Ribofuranosyl)uracil, could be potential antagonists as they contact major residues Asn283 and Asp287, similar to various potent Y1R antagonists. In addition, PC21 Poricoic acid B, PC22 Poricoic acid G and PC43 16alpha,25-Dihydroxy-24-methylene-3,4-secolanosta-4(28),7,9(11)-triene-3,21-dioic acid, contacting Asn299, Asp104 and Asp200 proximal to the extracellular surface could also interfere with agonist binding by stabilising the extracellular loop (ECL) 2 of Y1R in a closed position. Owing to their selective interaction with Phe302, an important residue in binding of selective Y1R antagonists, PC12 beta-Amyrin acetate, PC26 3-Epidehydrotumulosic acid and PC27 Cerevisterol were proposed as putative antagonists. Following the consensus approach, PC12 beta-Amyrin acetate, PC26 3-Epidehydrotumulosic acid and PC27 Cerevisterol were identified as candidate compounds due to their high affinities (-12.2, -11.0 and -10.8 kcal, respectively), high drug-likeness and low toxicity profiles. Trajectory analyses and energy contributions of PC12-Y1R complex further confirmed their structural stability and favourable binding free energies, highlighting the feasibility and possible development of PC12 beta-Amyrin acetate as a future Y1R inhibitor. National Computing Infrastructure (NCI) LIEF LE170100200 Wong Ann Rann Australian Government Research Training Program (RTP) Scholarship Wong Ann Rann We thank the National Computing Infrastructure (NCI), which is supported by the Australian Government, for providing resources and services to assist with computational work in this study. Molecular dynamics simulations were undertaken using the LIEF HPC-GPGPU facility hosted at the University of Melbourne. This Facility was established with the assistance of LIEF Grant LE170100200. AW acknowledges financial support through the Australian Government Research Training Program (RTP) Scholarship. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Data AvailabilityNo ethical or legal restrictions on data sharing is apparent in our study. Input parameters for molecular docking and dynamics simulations have been specified in the methods section and in the supplementary file as follows: • S1 Table. Autodock Vina configuration file inputs for blind and focused docking. Receptor and ligands were presented in the pdbqt format, while all co-ordinates and box size dimensions are presented in Angstroms. • S2 Table. Protein-ligand interaction profiler thresholds for each parameter. The mode of analysis was specified as “detection of macromolecule-ligand interactions. Data Availability No ethical or legal restrictions on data sharing is apparent in our study. Input parameters for molecular docking and dynamics simulations have been specified in the methods section and in the supplementary file as follows: • S1 Table. Autodock Vina configuration file inputs for blind and focused docking. Receptor and ligands were presented in the pdbqt format, while all co-ordinates and box size dimensions are presented in Angstroms. • S2 Table. Protein-ligand interaction profiler thresholds for each parameter. The mode of analysis was specified as “detection of macromolecule-ligand interactions. ==== Body pmcIntroduction Obesity, characterised by excessive fat accumulation that poses a risk to health, has been a growing health concern as it has been associated with serious and often life-threatening comorbidities. Globally, the prevalence of obesity in adults has increased significantly in the past 4–5 decades, with estimates of 281 and 309 million men and women identified with obesity in 2016 [1]. Compared to diabetes and other medications related to comorbidities of obesity, the number of anti-obesity pharmacotherapy is relatively limited. Five FDA approved medications for obesity are orlistat, phentermine-topiramate, naltrexone-bupropion, liraglutide and the recently approved semaglutide [2–6]. Except orlistat which works to reduce fat digestion and absorption, four of these medications target the central nervous system to control food intake and minimise binge-eating behaviour. With these approaches, it is not without a risk of adverse events associated with gastrointestinal discomfort, depressive behaviours, and surgical complications [7–9]. Considering the limitations of current approaches, many patients resort to complementary and alternative strategies for obesity management, including Chinese herbal medicine. Poria cocos (PC) (Schw.) Wolf. also known as Fu ling, has been a widely used herb among Chinese herbal formulations for weight management and lipid-modulation. PC is the dried sclerotium of Wolfiporia cocos, a saprophytic fungus in the family of Polyporaceae. It is known to be enriched with triterpenes and polysaccharides, with anti-inflammatory, anti-tumour and anti-oxidant properties [10]. Its extensive therapeutic effects have attracted considerable research attention, particularly for the management of metabolic diseases. As observed in multiple systematic reviews and meta-analyses, PC was reported to be one of ten most highly prescribed herbs in obesity clinical trials [11, 12]. Furthermore, for dyslipidaemia studies on humans [13] and ob/ob mice [14, 15], PC has also been a common focus. Although various clinical and experimental studies of PC-containing herbs and formulations have indicated significant improvements in clinical and biochemical outcomes, the molecular mechanisms of PC compounds and their potential as druggable candidates remain largely unknown. To elucidate the mechanistic action of PC compounds, protein targets were sought from the neuroactive ligand-receptor pathway, one of the most important targets to regulate feeding behaviour and energy homeostasis. Among the neuropeptide Y (NPY) receptors, recent studies have been focusing on the Y1 subtype (Y1R) as a therapeutic target for the management of obesity and eating disorders [16, 17]. Belonging to class A (rhodopsin-like) G-protein coupled receptor (GPCR), Y1R governs a range of signal transduction and is modulated by a broad range of stimuli included small peptides, lipid analogues, and endogenous and exogenous small molecules. The Y1Rs are generally expressed in the basal ganglia and the limbic system, with specific co-localisation in the cerebral cortex, caudate putamen and central amygdala [18]. They have also been found in the subcutaneous and visceral adipose tissues by which they could alter metabolic parameters in visceral obesity [19]. The role of Y1R in obesity pathophysiology is primarily induced by a polymorphism in the Y1R gene with a Cystosine to Thymidine nucleotide substitution. Clinical studies on 306 obese participants with C to T polymorphism showed a lower tolerance for fasting [19]. Furthermore, the central administrations of antagonistic small molecules such as BIBP3226, LY357897 or 1229U91 have also resulted in a significant reduction in feeding behaviour in rats [20]. Thus, these studies support the role of Y1R on obesity development and therefore Y1R is selected to be the primary target of this research. This study aimed to explore the mechanisms of actions of PC compounds targeting Y1R and examine their pharmacokinetic profiles as druggable candidates. Computational molecular docking using blind and focused approaches [21] was conducted to predict possible binding sites and to examine the binding affinity and poses of PC compounds on the inactive Y1R. The drug-likeness, absorption, distribution, metabolism, excretion, and toxicity (ADMET) profiles were examined using QSAR approaches. By comparing the binding affinities, ligand-residue interaction, pharmacokinetics profile, and molecular simulations of PC compounds against the native co-crystalised ligand, we proposed several putative inhibitors, which could be worthwhile for further in vivo and in vitro investigations as novel small molecules targeting Y1R. Materials and methods Identification and preparation of PC compounds and Y1R target PC compounds were comprehensively searched on the Traditional Chinese Medicine Information Database (TCM-ID) [22] and the Traditional Chinese Medicine Systems Pharmacology platform (TCMSP) [23] using keywords including ‘Poria cocos’, ‘Fu ling’, and ‘茯苓’ up to October 2021. The criteria for screening of duplicated compounds were based on several molecular descriptors, including the PubChem CID, compound name, molecular formula, and molecular weight. The 3D conformer structures of included PC compounds were downloaded in sdf or mol2 format, and were subsequently converted into pdbqt format using Open babel (v 3.1.1) [24] following on the requirements of AutoDock Vina. Duplicates were removed based on similarities in molecular descriptors. Missing bonds and hydrogens were repaired, and all 3D structures were converted into pdbqt format for docking. No limit was set for the number of torsions. The structure of Y1R was obtained from RCSB PDB (PDB: 5ZBQ) [25] and processed in Visual Molecular Dynamics (VMD, v1.9.3) [26]. It adopts an inactive conformation co-crystallised with a known antagonist UR-MK299 at its canonical binding site. It has been reported that UR-MK299 has a PPARγ IC50 of 80 nM and Ki of 28.67nM [27] using a LanthaScreen competitive binding assay and was therefore used as a positive control for this study. The antagonist contains three benzene rings and various carboxyl groups, forming seven hydrogen bonds with Gln219, Asn283 and Asp287 on helices V and VI, by which two of three residues (Asn283 and Asp287) were suggested to be crucial for Y1R inhibition [25]. The native ligand (UR-MK299) was isolated from the original PDB structures for re-docking as controls in both blind and focused docking conditions. In the latter case, only the region near the known ligand binding site of Y1R was probed. For the preparation of the original Y1R PDB, missing loops and side chains were fixed through homology modelling with subsequent energy minimisation on the SWISS-MODEL [28] server based on target sequence obtained from UniProt [25]. The PDB structure was converted into the pdbqt format via Open babel (v 3.1.1) [24] for docking in AutoDock Vina (v1.1.2) [29]. Physicochemical space examination, drug-likeness prediction, and pharmacokinetics profiling PC compounds were subjected to quantitative structural-activity analyses (QSAR) using Data Warrior, an open source cheminformatics tool [30] to examine their chemical space and drug-likeness potential. Known Y1R ligands were obtained from ChEMBL database [31] and structural variables were used as comparison. Pharmacokinetics profiling including absorption, metabolism, distribution, excretion and toxicity (ADMET) indexes were obtained from running PC compounds through a graph-based structural signature modelling server, pkCSM [32]. These preliminary ligand-based QSAR predictions offer complementary roles to support structural screening such as the molecular docking and molecular dynamics (MD) simulations performed in the next sections. Molecular docking and ligand-residue interaction analyses Computational docking between Y1R and PC compounds were first conducted by blind docking to predict available binding sites of Y1R. Both allosteric and canonical binding sites were examined. The final binding site selection for focused docking was guided by (a) blind docking outcomes and (b) experimentally known sites [25] to provide more accurate binding poses and energy scores with less computational power [21]. Docking was performed in Autodock Vina [29] with all ligands set to allow full torsional flexibility while proteins were kept in a fixed position. By referencing the co-crystalised ligand coordinates, the grid size and grid center for blind docking were set to 46.69 x 48.12 x 78.56 Å (center x = -44.33, y = -21.39, z = 83.53) and that of focused docking were set to 36.14 x 28.99 x 32.20 Å (center x = -48.98, y = -20.01, z = 68.21). The Intel Xeon Cascade Lake processor nodes of the high-performance computing cluster (‘Gadi’) housed at the National Computational Infrastructure (Canberra, Australia) was used to perform AutoDock Vina calculations. Subsequently, hydrogen and hydrophobic interactions on docked protein-ligand complexes were assessed by the protein-ligand interaction profiler (PLIP) [33] according to pre-specified bond angles and distances cut-offs (S1 Table), and visualised as networks in Gephi (0.9.3) [34]. Molecular dynamics simulations and free binding energy calculations To examine the stability of protein-ligand complexes, the candidate PC compounds with the most favourable drug-likeness, ADMET properties, docking energy scores and binding poses were selected for molecular dynamics simulation against the native ligand UR-MK299. Ligand topology and parameter files were generated using CGenFF [35, 36] while the membrane system was prepared using CHARMM36m force field [37]. The system was constructed from CHARMM-GUI membrane builder module using the replacement method [38]. Disulfide bonds between Cys113-Cys198 and Cys33-Cys296 were specified, and the protein orientation was set by outputs from the PPM 2.0 server along the Z-axis. The system consists of POPC components distributed in a 1:1 ratio in the upper (129 molecules) and lower (130 molecules) leaflet, which was solvated with TIP3P water model (25190 molecules) and neutralised with 0.15M of K+ and Cl- counterions. Water thickness of at least 22.5 Å on the top and bottom ends was applied to the system box. Energy minimisation was carried out using the steepest descent algorithm following the suggested equilibration scheme to reduce force constraints [38] and periodic boundary conditions were applied to produce constant temperature at 310K with the Nose-Hoover thermostat [39] and constant pressure set at 1 atm using the Parrinello-Rahman barostat using semi-isotropic coupling [40]. Long-range electrostatic forces were calculated from the Particle Mesh Ewald (PME) method [41], and all covalent bonds involving hydrogen atoms were constrained using the LINCS algorithm [42]. The final production runs were performed on GROMACS 2020.4 for 100 ns in triplicates [43]. To examine the stability of ligand binding, the root-mean-squared deviation (RMSD), root-mean-squared fluctuation (RMSF), number of hydrogen bonds and number of contacts were examined. Additionally, binding free energies were calculated by Molecular Mechanics-Poisson Boltzmann Surface Area (MM-PBSA) using the g_mmpbsa tool [44] from the final 1ns of the stabilised trajectory. For these MM-PBSA calculations, the grid spacing was set to 0.05 nm, the solvent dielectric constant was set to 80, and the solute dielectric constant was set to 2. The non-polar energy contribution was approximated based on the solvent-accessible surface area (SASA) model, with a probe radius set to 0.14 nm. The linear Poisson−Boltzmann equation (LPBE) implemented in the Adaptive Poisson-Boltzmann Solver (APBS) package was used. The list of input parameters for MM-PBSA calculations applied in this study was determined following the specification in the original publication of g_mmpbsa tool [44]. Results and discussion Virtual screening through ligand- and structure-based approaches Ligand-based screening and pharmacokinetics profiling revealed a high number of promising drug-like PC compounds occupying similar chemical space as known Y1R inhibitors The QSAR modelling of 43 PC compounds (S1 Fig) and 173 known Y1R inhibitors suggest that they share similar chemical space characterised by molecular weight, cLogP and topological surface area, particularly for PC compounds with higher molecular weight (Fig 1a). This is not surprising as most therapeutic drug-like molecules range between 350 to 500 g/mol, whereas a molecular weight of < 350 g/mol appears to be more lead-like [45]. Fig 1b shows the drug-likeness scores of PC compounds, and 62.8% of compounds approached a positive score. Interestingly, compounds with a lower rotatable bond count, between 0 to 7, exhibited higher drug-likeness scores. This could be due to their relatively higher oral bioavailability, as previous studies have demonstrated a correlation between the number of rotatable bonds and oral bioavailability, irrespective of molecular weight [46, 47]. The absorption, metabolism, distribution, and excretion (ADMET) indexes were illustrated with detailed cut-off points in S2 Fig. Following pkCSM prediction, most PC compounds exhibit favourable ADMET ranges which suggest that they could be readily bioavailable. According to open molecules and pkCSM analyses, most PC compounds appear to be within safe limits of AMES toxicity, and are generally non mutagenic, carcinogenic or irritative (Fig 1c–1f). Although several compounds had low drug-likeness or unfavourable ADMET properties, possible compound re-design guided by these parameters may help synthesise better molecules for Y1R. For example, compounds predicted to have low GI absorption rates could be alternatively administered, via nasal, pulmonary, subcutaneous, or transdermal routes or assisted with micro and nanotechnologies, similar to insulin delivery methods [48]. In addition, nanotechnology such as using nano gold particles have been reported to improve absorption rates of PC compounds [49]. Hence these sub-optimal pharmacokinetics profiles may not necessarily be a major concern for the present study. However, it is important to note that the variability and limitations of machine-learning based predictions exist [50]. Notably, non-important fragments may be more heavily weighted over relevant bioactive fragments involved in binding, and many factors including the quality of input molecules, the choice of descriptors and the selection of statistical modelling and validation methods could significantly alter the accuracy of predictive models. Considering the limitations from ligand-based screening, all PC compounds were subjected to structure-based molecular docking for binding affinity scoring, and a candidate compound is proposed for molecular dynamics simulation based on consensus from ligand-based and structure-based methods. 10.1371/journal.pone.0277873.g001 Fig 1 Ligand- and structure-based screening outcomes. (a) Chemical space comparison of PC compounds (red) and known Y1R inhibitors (blue). (b) Drug-likeness prediction of PC compounds based on substructure fragments. Toxicity indexes include AMES (c), mutagenicity (d), tumorgenicity (e), and irritant (f) following QSAR analysis from pkCSM and open molecules. 3D ligand-residue interactions in blind (g) and focused (h) docking, with an example compound PC12 at the orthosteric (i) and allosteric (j) sites. The docking protocol was validated with a known antagonist (UR-MK299) and PC compounds were observed to bind at the orthosteric binding site of Y1R with favourable binding energies Docking protocol validations demonstrated that UR-MK299 bound favourably to Y1R at the canonical binding site (blind docking affinity: -11.2 kcal/mol, focused docking affinity: -11.7 kcal/mol). As observed in S3 Fig, the hydrogen bonds (with Asn283 and Asp287), pi-stacking contacts (with Phe282 and Phe286), and disulphide bridges (with Cys121) were reflected in both the re-docked experiments, consistent with previously reported PPARγ co-crystal structures (S3 Fig). Applying the validated protocols, 44 compounds, identified from TCM-ID and TCMSP databases, were docked with Y1R (PDB: 5ZBQ). A total of 396 docking poses, and binding energies were obtained, and an allosteric site within the transmembrane domain was identified through blind docking (Fig 1g–1j). Strikingly, all PC compounds in both blind and focused docking were scored with a negative (favourable) binding affinity, as shown in Table 1. Consistent with other docking studies [21] and the positive control UR-MK299, the median energy scores from focused docking (-10.1 kcal/mol) within the pre-defined box size were significantly more superior compared to blind docking (-7.95kcal/mol), and hence PC compounds in focused docking were selected for further analyses in the next few sections. 10.1371/journal.pone.0277873.t001 Table 1 Binding affinity results, calculated Ki values, and residue contacts from focused docking (kcal/mol) of PC ligands with Y1R (PDB: 5ZBQ). CID Index Compound name Formula ΔG (kcal/mol) Ki (μM) Weight (g/mol) Hydrogen bonds Hydrophobic contacts 15391339 PC37 3beta-p-Hydroxybenzoyldehydrotumulosic acid C38H52O6 -12.6 0.001 604.8 Asp104, Asp200, Asn283 Phe28, Asp104, Phe173, Val197, Phe199, Leu215, Phe282, Phe286, Ala294, His298 92156 PC12# beta-Amyrin acetate C32H52O2 -12.2 0.001 468.8 Gln219 Pro117, Gln120, Ile124, Phe173, Phe199, Trp276, Leu279, Phe302 9805290 PC25 Polyporenic acid C C31H46O4 -11.9 0.002 482.7 Asn299 Leu26, Phe28, Thr97, Tyr100, Asp104, Pro117, Val197, Phe199, Phe282, Ala294, Phe302 15226717 PC35 Dehydropachymic acid C33H50O5 -11.9 0.002 526.7 NA Leu26, Phe28, Thr97, Tyr100, Asp104, Pro117, Gln120, Val197, Phe199, Phe282, Ala294, Phe302 444679 PC16 Ergosterol C28H44O -11.8 0.002 396.6 NA Leu26, Phe28, Asp31, Asp104, Phe184, Val197, Phe199, Phe286, Ala294, Phe302 10743008 PC30 16alpha-Hydroxydehydrotrametenolic acid C30H46O4 -11.6 0.003 470.7 Asn116, Thr295 Leu26, Phe28, Thr97, Tyr100, Pro117, Val197, Phe199, Phe282, Ala294, Phe302 15391340 PC38 3-Dehydrotrametenolic acid C30H46O3 -11.6 0.003 454.7 Asn116, Asn299 Leu26, Phe28, Thr97, Tyr100, Pro117, Val197, Phe199, Phe282, Phe302 5283628 PC18 Stellasterol C28H46O -11.4 0.004 398.7 Asp104 Phe28, Val197, Phe199, Phe282, Asn283, Phe286, Ala294, Phe302 5484385 PC23 Pachymic acid C33H52O5 -11.1 0.007 528.8 Asp104, Asn299 Leu26, Phe28, Tyr100, Asp104, Phe173, Phe184, Val197, Phe282, Asn283, Phe302 21159065 PC39 See belowa C32H48O5 -11.1 0.007 512.7 Asp104, Asn299 Phe28, Thr97, Tyr100, Asp104, Phe184, Val197, Phe199, Phe282, His298, Phe302 10005581 PC26# 3-Epidehydrotumulosic acid C31H48O4 -11 0.008 484.7 Asp104, Asn299 Leu26, Phe28, Tyr100, Asp104, Pro117, Phe173, Phe184, Val197, Asn283, Phe302 12309443 PC32 Trametenolic acid C30H48O3 -11 0.008 456.7 Asp287, Asn299 Leu26, Phe28, Asp104, Phe184, Val197, Phe199, Asn283, Phe286, Phe302 10368709 PC29 25-Hydroxy-3-epidehydrotumulosic acid C31H48O5 -10.9 0.010 500.7 Asp287, Asn299 Leu26, Phe28, Asp104, Phe199, Asn283, Phe286, Phe302 12314446 PC33 Tumulosic acid C31H50O4 -10.9 0.010 486.7 Asp104, Asn299 Leu26, Phe28, Asp104, Val197, Phe199, Phe282, Asn283, Phe286, Phe302 73402 PC10 Eburicoic acid C31H50O3 -10.8 0.012 470.7 Asp104, Asn299 Leu26, Phe28, Asp104, Asn283, Phe286, Ala294, Phe302 10181133 PC27# Cerevisterol C28H46O3 -10.8 0.012 430.7 Pro117, Asn283 Phe28, Thr97, Tyr100, Thr101, Gln120, Val197, Phe199, Ala294, Phe302 5471852 PC21# Poricoic acid B C30H44O5 -10.7 0.014 484.7 Asp104, Gln177, Asp200, Asn299 Phe28, Asp104, Val197, Phe199, Phe282, Asn283, Phe302 15225964 PC34 Dehydrotumulosic acid C31H48O4 -10.7 0.014 484.7 His298, Asn299 Leu26, Phe28, Asp104, Val197, Phe199, Phe282, Asn283, Phe286, Phe302 5471966 PC22# Poricoic acid G C30H46O5 -10.6 0.017 486.7 Asp104, Gln177, Asp200, Asn299 Leu26, Phe28, Asp104, Val197, Phe199, Phe282, Asn283, Phe286, Phe302 15250826 PC36 Dehydroeburicoic acid C31H48O3 -10.6 0.017 468.7 Asn299 Leu26, Phe28, Asp104, Phe184, Val197, Phe199, Asn283, Phe286, Phe302 5471851 PC20 Poricoic acid A C31H46O5 -10.5 0.020 498.7 Gln177, Asp200, His298, Asn299 Leu26, Phe28, Phe184, Val197, Phe199, Phe282, Asn283, Phe286, Ala294, Phe302 72202422 PC42 See belowb C34H54O5 -10.5 0.020 542.8 Asp287, Asn299 Leu26, Tyr100, Pro117, Ile124, Phe173, Phe199, Phe282, Asn283, Phe286, Ala294, His298, Phe302 102378062 PC43 16alpha,25-Dihydroxy-24-methylene-3,4-secolanosta-4(28),7,9(11)-triene-3,21-dioic acid C31H46O6 -10.3 0.028 514.7 Asp104, Gln177, Asp200, Asn299 Phe28, Phe282, Asn283, Phe286, Ala294, Phe302 10918099 PC31 poricoic acid H C31H48O5 -10.2 0.033 500.7 Gln177, Asp200 Leu26, Phe28, Phe282, Asn283, Phe286, Phe302 56668247 PC41 Poricoic acid C C31H46O4 -10.1 0.039 482.7 Gln177, Asp200, Asn299 Leu26, Phe28, Asp104, Phe184, Val197, Phe199, Phe282, Asn283, Phe286, Phe302 14697 PC9 Methyl dehydroabietate C21H30O2 -10 0.046 314.5 NA Leu26, Phe28, Phe199, Phe282, Phe286, Ala294, Phe302 1203 PC5 (+/-)-Catechin C15H14O6 -9 0.249 290.27 Asp104, His298, Asn299 Leu26, Phe28, Asp104, Val197, Phe199, Ala294 182232 PC15 (+)-Epicatechin C15H14O6 -8.9 0.294 290.27 Asn283, Asn299 Leu26, Phe28, Phe199, Phe282, Phe286, Ala294, His298, Asn299 6441913 PC24 Coniferyl ferulate C20H20O6 -8.9 0.294 356.4 Gln177, Cys198, Asp200 Leu26, Phe28, Asp104, Phe184, Val197, Phe199 5319022 PC19 Ligustilide C12H14O2 -8 1.347 190.24 NA Phe28, Phe199, Phe282, Phe286, Ala294, His298 40428662 PC40# 1-(alpha-L-Ribofuranosyl)uracil C9H12N2O6 -6.6 14.348 244.2 Thr212, Gln219, Leu279, Asn283, Asp287 NA 121667 PC13 ethyl beta-D-glucopyranoside C8H16O6 -6.4 20.117 208.21 Gln219, Thr280, Asn283 Phe173 72 PC1# 3,4-Dihydroxybenzoic acid C7H6O4 -6.3 23.820 154.12 Thr212, Leu216, Asn283, Asp287 Asn283 985 PC4 Palmitic acid C16H32O2 -6.3 23.820 256.42 Glu29, Asn30 Phe28, Tyr100, Val197, Phe199, Phe282, Phe286, Ala294 8468 PC8# Vanillic acid C8H8O4 -6.2 28.205 168.15 Thr212, Asn283, Asp287 Leu216, Tyr220, Thr280 5282729 PC17 2-Dodecenoic acid C12H22O2 -6.2 28.205 198.3 Glu29, Asn30, Ala294 Phe28, Phe199, Phe282, Phe286 3893 PC6 Lauric acid C12H24O2 -5.9 46.824 200.32 Asp104 Phe28, Val197, Phe282, Phe286, Ala294, His298, Phe302 74112 PC11 Trimethyl citrate C9H14O7 -5.6 77.734 234.2 Asp104, Asn299 Phe28, Phe199, Ala294 8180 PC7 Undecanoic acid C11H22O2 -5.4 108.988 186.29 Thr295 Leu26, Phe199, Phe282, Phe286, Ala294, His298, Phe302 125207 PC14 11-Dodecenoic acid C12H22O2 -5.4 108.988 198.3 Asn299 Leu26, Phe199, Phe282, Phe286, Ala294, His298, Phe302 190 PC2 Adenine C5H5N5 -5.3 129.050 135.13 Gln219, Leu279, Thr280, Asn283 NA 379 PC3 Octanoic acid C8H16O2 -5.1 180.935 144.21 Glu29, Asn30, Ala294 Phe28, Asp104, Val197, Phe199, Ala294 10285815 PC28 (S)-Dimethyl 2-hydroxysuccinate C6H10O5 -5.1 180.935 162.14 Gln219, Asp287 Thr280 NA indicates no hydrogen bonds or hydrophobic residues. Each PC index represents a Poria cocos compound. Marked IDs# are PC compounds with putative antagonist roles, as discussed in text. Binding affinity scores were generated by AutoDock Vina (v1.1.2). Inhibition constant (Ki) = exp(ΔG×1000/RT), where ΔG is the docking energy, R is 1.985 x 10−3 kcal/mol/K and T is 298K. a(2R)-2-[(3S,5R,10S,13R,14R,16R,17R)-16-acetyloxy-3-hydroxy-4,4,10,13,14-pentamethyl-2,3,5,6,12,15,16,17-octahydro-1H-cyclopenta[a]phenanthren-17-yl]-6-methylhept-5-enoic acid. bmethyl (2R)-2-[(3S,5R,10S,13R,14R,16R,17R)-3-acetyloxy-16-hydroxy-4,4,10,13,14-pentamethyl-2,3,5,6,7,11,12,15,16,17-decahydro-1H-cyclopenta[a]phenanthren-17-yl]-6-methyl-5-methylideneheptanoate. Distances of hydrogen and hydrophobic contacts are listed in S2 and S3 Tables, respectively. Hydrogen bonds between PC compounds and key residues of Y1R binding site within the orthosteric ligand-binding domain and the deep V-shaped sub-pocket A high frequency (≥ 5) of hydrogen bonds between PC compounds and key polar or charged residues were observed (Fig 2a), including Asn299 (n = 31), Asp104 (n = 17), and Asp200 (n = 10), located near the extracellular domain, and Asn283 (n = 8), Asp287 (n = 8) found deep in the Y1R sub-pocket. Distance information on hydrogen bond interactions is provided in S2 Table. 10.1371/journal.pone.0277873.g002 Fig 2 3D ligand-residue interactions outlining hydrogen bonds (a) and hydrophobic interactions (b) in focused docking of Poria cocos compounds and Y1R. The node sizes of PC compounds (grey) and Y1R residues are based on the number of degrees, i.e., the higher the frequency of connections between residues and PC compounds, the larger the node size. Orthosteric ligand-binding domain As a residue with a high propensity to donate and accept hydrogen due to its carboxylate and amide side chains, it is not surprising to observe a large majority (48.8%) of PC compounds form polar contacts with Asn299. Although the role of Asn299 in Y1R is not yet well known, studies on Asn located in a similar location (near extracellular loop of helix VII) of other class A GPCRs, the M2 muscarinic acetylcholine receptor, suggest that a mutation from Asn to Cys (as well as Val to Cys) significantly reduced binding affinity of the orthosteric antagonist [51]. Similarly, Asp104 located in helix II could also play important roles in stabilising extracellular loop (ECL) 2 and affect ligand-binding activities, as the substitution of Asp104 to Ala completely abolished the binding of endogenous peptide, NPY [52]. In addition, the importance of Asp200 in endogenous ligand binding has been demonstrated in docking studies between peptide YY (PYY) and Y1R, as a hydrogen bond was indicated with 1.8Å between the peptide and receptor residue Asp200 [53]. A study on Asp200 conducted by Walker et al. (1994) [52] also revealed that a mutation of Asp200 to Ala completely abolished 1251-NPY binding activities at 50pM concentration. However, these findings contrasted with a later study conducted by Sautel et al. (1996) [54], which they found Asp200 to Ala mutation exhibited similar Ki values as the wild-type receptor for either the endogenous ligand (NPY) or non-peptide antagonist (BIBP 3226). Asp200 may therefore serve an indirect role to support binding events, pending confirmation with more recent investigations. Nevertheless, PC compounds (PC21, PC22 and PC43) interacting with Asn299, Asp104 and Asp200 by hydrogen bonds, could interfere with agonist binding and stabilising ECL2 in a closed position. Deep V-shaped sub-pocket Within the deep sub-pocket, two highly important residues for maintaining inactive conformation of Y1R are Asn283 on helix V and Asp287 on helix VI. These residues have been well studied for their ligand-recognition effects on Y1Rs. PC compounds forming hydrogen bonds with both Asn283 and Asp287 were PC1, PC8 and PC40. Mutagenesis studies indicate that alanine mutations of Asn283 and Asp287 both showed no or significantly lower binding of both endogenous agonist (NPY) and non-peptide antagonist (BIBP3226) [52, 54]. This observation is consistent with recent structural studies [25], by which the binding affinity scores of the antagonist UR-MK299 (positive control in this study) were reduced by over 2,000-fold when Asn283 and Asp287 were mutated to Ala and Asp residues, respectively. Regardless of peptide or non-peptide ligands, these studies strongly support the roles of Asn283 and Asp287 in substrate binding. The individual and synergistic roles of these residues on receptor function and the selectivity of peptide and non-peptide molecules on Y1R could be further explored. Extensive contacts with conserved and non-conserved Phe residues within the hydrophobic cluster and may interfere with binding of endogenous ligands near the extracellular surface of Y1R Aside from hydrogen bonding residues, PC compounds were stabilised by hydrophobic contacts with Y1R residues (Fig 2b), their distances were provided in S3 Table. The most targeted hydrophobic contacts within 4Å were Phe residues on extracellular loops (Phe28 and Phe199) and on transmembrane helices V and VI (Phe282, Phe286 and Phe302). As we aimed to identify possible inhibitors of Y1R, only PC compounds contacting the hydrophobic cluster indicated for Y1R antagonism were discussed. Among these hydrophobic clusters, it has been revealed that alanine mutation of the non-conserved Phe286 resulted in significantly lowered binding affinity for both agonist (NPY) and antagonist (BIBP3266) [54]. More specifically, Phe286 mutations to Ala resulted in a 3-fold increase in Kd values. Furthermore, Phe282 and Phe286 have also been implicated for G protein coupling, as substitutions of these residues aborted the activity of PYY without significant changes to the binding affinity of antagonists 1229U91 and J-104870 [55]. In contrast, alanine mutation of Phe302 has resulted in the complete abolishment of activities from known antagonists BMS-193885, BIBP3226, BIBO3304, UR-MK136 and UR-MK289 [25]. These findings suggest that Phe302 makes a profound impact on antagonistic activities of Y1R, and those compounds that solely interact with Phe302 were PC12, PC26 and PC27 (S4 Fig). A consensus approach identified PC12, PC26 and PC27 as putative Y1R inhibitors Following ligand- and structure-based screening, close examination of protein-ligand contacts at the important binding site residues identified several PC compounds for further analysis, as listed in Table 2. Considering the binding energies, drug-likeness profiles, and toxicity indexes, PC12, PC26 and PC27 could be proposed as candidate compounds targeting Y1R and is worth further investigation. PC12, beta-amyrin acetate was selected as the compound for molecular dynamics simulations as it produced the highest binding score out of the three. 10.1371/journal.pone.0277873.t002 Table 2 Putative Y1R inhibitors identified from ligand-based and structure-based screening approaches. Putative Y1R inhibitors ΔG (kcal/mol) Drug-likeness score AMES toxicity (yes/no) Mutagenic (high/none/low) Tumour-genic (high/none/low) Irritant (high/none/low) PC1 -6.3 -1.8 No high none none PC8 -6.2 -1.6 No high none none PC12* -12.2 -2.7 No none none none PC21 -10.7 -21.0 No none none none PC22 -10.6 -21.3 No none none none PC26* -11.0 -6.5 No none none none PC27* -10.8 -0.3 No none none none PC40 -6.6 -4.6 No none none none PC43 -10.3 -14.6 No none none high *Candidate compounds as putative Y1R inhibitors. Structural analyses from 100 ns of MDS trajectories suggest reasonable stability of PC12-Y1R complex The conformational dynamics on stability and fluctuations were examined with RMSD, RMSF, number of hydrogen bonds, number of contacts, minimum protein-ligand distance, and minimum residue-ligand distance. In all simulations, the RMSD increased steadily and reach a plateau by approximately 35ns. After reaching equilibrium, the average root mean-squared deviations (RMSD) of backbone atoms fitted to the starting structure (t = 0) for the apo-, PC12- and UR-MK299-Y1R complexes were 0.331nm, 0.339nm and 0.319nm, respectively (Fig 3a). Interestingly, PC12-Y1R appears to have marginally higher RMSD compared to UR-MK299- Y1R, this is attributed to the drift in the intracellular loop 3 of the second and third MD runs. As expected, most movements occurred in the extracellular and intracellular loop regions, while the transmembrane helices were relatively stable. Furthermore, the magnitude of RMSD were comparable to other rhodopsin-like GPCRs suspended in a POPC membrane system, with an average of approximately 0.35nm [56]. Per-residue fluctuation across the trajectory (Fig 3b) revealed several major peaks which corresponds to the loop regions, consistent with the RMSD trends. The highest peaks with the greatest flexibility among all simulations were between Lys239 and Thr258, the longest intracellular loop connecting helices 5 and 6. No significant differences between the RMSF peaks of Y1R were apparent, although a marginally lower average RMSF could be seen with the PC12-bound Y1R (red) compared to the UR-MK299-bound Y1R (blue), suggestive of a stronger loop-stabilising effect of PC12. 10.1371/journal.pone.0277873.g003 Fig 3 Structural analyses, snapshots of trajectories and binding free energy calculations from molecular dynamics simulation comparing PC12 (red) with its positive control UR-MK299 (blue) on Y1R (green). Structural analyses include RMSD (a), RMSF (b), number of hydrogen bonds (c), number of contacts (d), minimum protein-ligand distances (e), and minimum distance between residues (Pro117, Phe302) on either side of the ligand (PC12) (f). Snapshots of Y1R (pale green cartoon) and PC12 (red CPK) visualised in VMD throughout 100 ns (g-k). Per-residue energy decomposition for PC12 and UR-MK-299 (l), with a 3D ligand-residue inset of PC12 (red line and sticks). For both UR-MK299 and PC12 complexes after equilibrium, the number of hydrogen bonds remained relatively constant throughout the trajectory (Fig 3c), despite a slight decline in numbers for PC12 shortly after system relaxation. As the number of hydrogen bonds for PC12 were comparatively low, further assessments on number of contacts and protein-ligand distances were performed to examine whether protein-ligand interactions were present throughout the 100 ns trajectory. Intriguingly, the differences in number of contacts (Fig 3d) and the minimum protein-ligand distances (Fig 3e) for PC12 (2988 and 0.202nm, respectively), appear comparable to that of UR-MK299 (3443 and 0.187nm, respectively). Similarly, the distances between Pro117 (on helix 3) and Phe302 (on helix 7), two residues situated at either side of PC12 in Fig 3f appeared consistent at approximately 0.23 to 0.24nm across the smoothed trajectory. As such, the proximity of PC12 with Y1R residues indicates that the ligand remained bound and fitted well within the orthosteric ligand-binding domain. The stable and sustained interactions may be governed by non-polar contacts such as pi-stacking interactions, in stark contrast with the binding patterns of the positive control UR-MK299. Snapshots of the trajectories in Fig 3g–3k shows PC12-bound Y1R at different time series (ns), aside from a small shift in ligand orientation, PC12 remains bound to the orthosteric binding domain. Binding free energy estimated by MMPBSA indicate that PC12 could bind with relatively strong free energies to Y1R The stability of PC12 and its positive control (UR-MK299) was further assessed by the energy contributions from electrostatic, van der Waals, polar solvation and SASA terms using MMPBSA. Table 3 lists the components of binding energies (kcal/mol) averaged over three parallel simulations for binding of PC12 and UR-MK299 to the inactive Y1R. Interestingly, PC12 revealed stronger binding energies (ΔGbinding = -36.1 kcal/mol) compared to the positive control UR-MK299 (ΔGbinding = -32.4 kcal/mol), where most interactions were driven by van der Waals forces in both systems. The major difference in binding energies was attributed to the lower positive polar solvation energy arising from binding with PC12 (14.8 kcal/mol), compared to UR-MK299 (56.3 kcal/mol). Consistently, using the same binding free energy calculation approach, the ΔGbinding of PC12 beta-amyrin acetate appears in agreement to its pentacyclic triterpenol variant, alpha-amyrin acetate, bound to cannabinoid type 1 receptor (-36.6 kcal/mol) [57]. Whether PC12 or UR-MK299 were able to compete with the endogenous ligands NPY or PYY to inactive Y1R remains to be researched, as the ΔGbinding of PYY was found to be -50.6 kcal/mol, a significantly more favourable affinity compared to small molecules [53]. Nevertheless, the results suggest that PC12 could bind to the orthosteric ligand binding domain of Y1R with relatively strong binding free energies. 10.1371/journal.pone.0277873.t003 Table 3 Components of binding energies (kcal/mol) averaged over three parallel simulations. Components of binding energies (kcal/mol) PC12 UR-MK299 ΔGvdw -45.1 -65.5 ΔGelec -0.3 -15.7 ΔGpsolv 14.8 56.3 ΔGSASA -5.5 -7.6 ΔGbinding -36.1 -32.4 vdw, van der Waals interactions; elec, electrostatic energy; psolv, polar solvation energy; SASA, solvent-accessible surface area. To examine the relative importance of residues for ligand binding at the orthosteric site, the total energy contribution of each system was decomposed into each corresponding residue. The average per-residue binding free energies for the final 1ns of three trajectories is shown in Fig 3l. Trends in per-residue binding energies were mostly consistent between PC12 and UR-MK299. The most energetically favourable residues tended to be non-polar and aggregate around helix III (Pro117, Cys121), helix VI (Phe173), helix VI (Phe282, Phe286) and helix VII (Phe302). Interestingly, two charged asparagine residues on ECL2 and helix IV were involved in significantly stronger binding free energies for PC12 (Asp200: -1.98 kcal/mol, Asp287: -1.48 kcal/mol) compared to UR-MK299 (Asp200: -0.71 kcal/mol, Asp287: 0.05 kcal/mol). The contrasting contributions from Asp residues may be important for the selectivity of PC12 or PC compounds on Y1R. Notably, the hydrophobic cluster governed by Phe282, Phe286 and Phe302 is also apparent in previous experimental studies on known Y1R antagonists UR-MK299 and BMS-193885 [25, 58]. The most favourable contribution from Phe302 among the cluster (PC12: -1.67 kcal/mol, UR-MK299: -1.74 kcal/mol) is consistent with findings from the protein-ligand interaction analyses earlier, and warrants mutagenesis studies to confirm its role on receptor inactivation. As PC12 demonstrated strong hydrophobic interactions with residues in the hydrophobic cluster, this could serve as a pathway for stabilising the Y1R conformation in the inactive state. In contrast to the residues above, Arg208 and Gln219 were found to contribute to positive energies at the orthosteric binding site. The binding free energies of the basic residue, Arg208, appeared to be highly positive for UR-MK299 (3.40 kcal/mol) and to a slightly lesser extent for PC12 (1.39 kcal/mol). The presence of the basic residue may contribute to conformational changes at the apical surface of Y1R, which could be supported by charged-assisted hydrogen bonds when interacting with endogenous peptides or small molecules. This mechanism has also been discussed in an early study by Sylte et al. [59], demonstrating the role of Arg208 on the stability of ECL2 through intramolecular electrostatic interactions with Asp182 and Asp205. In addition, the binding free energies of Gln219 were also positive when bound to UR-MK299 (0.63 kcal/mol) and PC12 (0.71 kcal/mol). Consistent with UR-MK299 and PC12 in our study, polar interactions with Gln219 were not found to interact with antagonist BIBP3226, but was found to be within the binding region with Tyr36 of NPY, suggesting that this residue could be important for receptor activation [60]. Although Arg208 (helix V) and Gln219 (ECL 2), were not found to be within close contact with the ligands at the stabilised conformations throughout the 100 ns simulation, these conserved residues among neuropeptide Y receptor subtypes (Y4R, Y5R) could be important for ligand recognition as they were located the extracellular surface, pending further confirmation with future mutagenesis studies. Conclusion In summary, this study was conducted to critically examined the pharmacokinetics profiles of PC compounds and to explore their molecular mechanisms targeting Y1R. We discovered several putative Y1R inhibitors using ligand- and structure-based approaches, which could be used as starting points for further development, optimisation, and refinement as anti-obesity agents. Compounds designated PC1 3,4-Dihydroxybenzoic acid, PC8 Vanillic acid, PC40 1-(alpha-L-Ribofuranosyl)uracil were proposed as potential antagonists as they contact major residues Asn283 and Asp287, similar to various potent Y1R antagonists. PC21 Poricoic acid B, PC22 Poricoic acid G and PC43 16alpha,25-Dihydroxy-24-methylene-3,4-secolanosta-4(28),7,9(11)-triene-3,21-dioic acid, contacting Asn299, Asp104 and Asp200 proximal to the extracellular surface could also interfere with agonist binding by stabilising the extracellular loop (ECL) 2 of Y1R in a closed position. Owing to their selective interaction with Phe302, an important residue in binding of selective Y1R antagonists, PC12 beta-Amyrin acetate, PC26 3-Epidehydrotumulosic acid and PC27 Cerevisterol were also suggested to be putative antagonists. Following the consensus approach from ligand- and structure-based screening, candidate compounds PC12, PC26 and PC27 were identified due to their high affinities (-12.2, -11.0 and -10.8 kcal/mol, respectively), high drug-likeness and low toxicity profiles. In combination with molecular dynamics simulations and binding free energy calculations, we achieved successful validation of the binding feasibility of PC12, as shown by its structural stability and favourable binding free energies at the orthosteric Y1R transmembrane domain. As relatively few studies adopted multiple virtual screening strategies, the consensus framework from ligand- and structure-based approaches with subsequent validation of feasibility in molecular dynamics simulations makes this combinatorial approach a strength of this study. Aside from binding affinity values, binding poses and interactions, the pharmacokinetics properties were also examined, offering room for lead optimisation and refinement to assist with a more successful design of PC compounds as more potent Y1R inhibitors. Furthermore, the application of high-throughput screening and analyses methods not only enables more efficient identification of candidate PC compounds, but it also provides a comprehensive understanding of ligand-residue interactions, binding characteristics and pharmacological regulation globally, at the herb level, in addition to many studies solely reporting actions of selected compounds from Poria cocos. Although findings from molecular dynamics simulations corroborate the accuracy of results from docking analyses, they were only performed for a relatively short duration and in triplicate. More computationally intensive simulations up to and beyond microsecond timescales with higher number of parallel runs may reveal additional behaviours of ligand-protein complexes, including ligand-entry and possible concerted motions between other functionally relevant regions and the allosteric sites, not sampled by the 100 ns simulations in this study. Finally, as a limitation among computational studies, the predictive binding affinities, inhibitory constants, and pharmacokinetic properties may not be reflective of experimental values. Therefore, further investigations from in vitro and in vivo studies are desired to validate, characterise, and expand on the actions induced by PC compounds on Y1R for future obesity management. Supporting information S1 Table Protein-ligand interaction profiler thresholds for each parameter. The mode of analysis was specified as “detection of macromolecule-ligand interactions”. (PDF) Click here for additional data file. S2 Table Distance of hydrogen contacts between PC compounds and Y1R residues. H-A, Distance between H-Bond hydrogen and acceptor atom (Å); D-A, Distance between H-Bond donor and acceptor atoms (Å). (PDF) Click here for additional data file. S3 Table Distance of hydrophobic contacts between PC compounds and Y1R residues. Distance provided in Angstrom (Å). (PDF) Click here for additional data file. S1 Fig 2D conformer structures of all 43 PC compounds. Within each compound box, the following information were provided: PC index (top left), chemical formula (top right), PubChem CID (bottom center). (PDF) Click here for additional data file. S2 Fig Overview of ADMET properties of PC compounds based on pkCSM outputs. For each continuous parameter, molecular weight was plotted in the x-axis against the parameter. Red line shows suggested thresholds for continuous parameters, where applicable. (PDF) Click here for additional data file. S3 Fig Best binding poses of three molecules of co-crystalised known antagonist UR-MK299 on Y1R (PDB:5ZBQ). Ligand residue interactions in 3D and 2D views. (a) 3D diagram of 9AO superimposed in Y1R (green ribbons). (b) 2D diagram of 9AO (grey sticks) in its original co-crystallised form. (c) 2D diagram of 9AO (orange sticks) in blind docking. (d) 2D diagram of 9AO (yellow sticks) in focused docking. (PDF) Click here for additional data file. S4 Fig Interaction distance between Poria cocos compounds and residues of the hydrophobic phenylalanine (Phe) cluster in focused docking. The colour of tiles represents an interaction detected within 4Å (green), and more than 4Å (white). The colour gradient represents the distance between the Phe residues and Poria cocos compounds, the closer the distance the darker the colour. (PDF) Click here for additional data file. ==== Refs References 1 Bentham J , Di Cesare M , Bilano V , Bixby H , Zhou B , Stevens GA , et al . Worldwide trends in body-mass index, underweight, overweight, and obesity from 1975 to 2016: a pooled analysis of 2416 population-based measurement studies in 128·9 million children, adolescents, and adults. Lancet. 2017;390 : 2627–2642. doi: 10.1016/S0140-6736(17)32129-3 29029897 2 Wilding JPH , Batterham RL , Calanna S , Davies M , Van Gaal LF , Lingvay I , et al . Once-Weekly Semaglutide in Adults with Overweight or Obesity. N Engl J Med. 2021;384 : 989–1002. doi: 10.1056/NEJMoa2032183 33567185 3 Pi-Sunyer X , Astrup A , Fujioka K , Greenway F , Halpern A , Krempf M , et al . A Randomized, Controlled Trial of 3.0 mg of Liraglutide in Weight Management. N Engl J Med. 2015;373 : 11–22. doi: 10.1056/NEJMoa1411892 26132939 4 Wadden TA , Foreyt JP , Foster GD , Hill JO , Klein S , O’Neil PM , et al . Weight loss with naltrexone SR/bupropion SR combination therapy as an adjunct to behavior modification: The COR-BMOD trial. Obesity. 2011;19 : 110–120. doi: 10.1038/oby.2010.147 20559296 5 Allison DB , Gadde KM , Garvey WT , Peterson CA , Schwiers ML , Najarian T , et al . Controlled-release phentermine/topiramate in severely obese adults: A randomized controlled trial (EQUIP). Obesity. 2012;20 : 330–342. doi: 10.1038/oby.2011.330 22051941 6 Sjöström L , Rissanen A , Andersen T , Boldrin M , Golay A , Koppeschaar HPF , et al . Randomised placebo-controlled trial of orlistat for weight loss and prevention of weight regain in obese patients. Lancet. 1998;352 : 167–172. doi: 10.1016/S0140-6736(97)11509-4 9683204 7 Saunders KH , Umashanker D , Igel LI , Kumar RB , Aronne LJ . Obesity Pharmacotherapy. Med Clin North Am. 2018;102 : 135–148. doi: 10.1016/j.mcna.2017.08.010 29156182 8 Dorman RB , Miller CJ , Leslie DB , Serrot FJ , Slusarek B , Buchwald H , et al . Risk for hospital readmission following bariatric surgery. PLoS One. 2012;7 . doi: 10.1371/journal.pone.0032506 22412881 9 Srivastava G , Apovian CM . Current pharmacotherapy for obesity. Nat Rev Endocrinol. 2018;14 : 12–24. doi: 10.1038/nrendo.2017.122 29027993 10 Nie A , Chao Y , Zhang X , Jia W , Zhou Z , Zhu C . Phytochemistry and Pharmacological Activities of Wolfiporia cocos (F.A. Wolf) Ryvarden & Gilb. Front Pharmacol. 2020;11 : 1–17. doi: 10.3389/fphar.2020.505249 32116689 11 Wong AR , Yang AWH , Li K , Gill H , Li M , Lenon GB . Chinese herbal medicine for weight management: a systematic review and meta-analyses of randomised controlled trials. J Obes. 2021. doi: 10.1155/2021/3250723 33763238 12 Sui Y , Zhao HL , Wong VCW , Brown N , Li XL , Kwan AKL , et al . A systematic review on use of chinese medicine and acupuncture for treatment of obesity. Obes Rev. 2012;13 : 409–430. doi: 10.1111/j.1467-789X.2011.00979.x 22292480 13 Wong AR , Yang AWH , Li M , Hung A , Gill H , Lenon GB . The Effects and Safety of Chinese Herbal Medicine on Blood Lipid Profiles in Placebo-Controlled Weight-Loss Trials: A Systematic Review and Meta-Analysis. Chin K-Y , editor. Evidence-Based Complement Altern Med. 2022;2022 : 1368576. doi: 10.1155/2022/1368576 35082903 14 SUN S-S , WANG K , MA K , BAO L , LIU H-W . An insoluble polysaccharide from the sclerotium of Poria cocos improves hyperglycemia, hyperlipidemia and hepatic steatosis in ob/ob mice via modulation of gut microbiota. Chin J Nat Med. 2019;17 : 3–14. doi: 10.1016/S1875-5364(19)30003-2 30704621 15 Kim J-H , Sim HA , Jung DY , Lim EY , Kim YT , Kim BJ , et al . Poria cocus wolf extract ameliorates hepatic steatosis through regulation of lipid metabolism, inhibition of ER stress, and activation of autophagy via AMPK activation. Int J Mol Sci. 2019;20 : 4801. doi: 10.3390/ijms20194801 31569635 16 Yi M , Li H , Wu Z , Yan J , Liu Q , Ou C , et al . A Promising Therapeutic Target for Metabolic Diseases: Neuropeptide y Receptors in Humans. Cell Physiol Biochem. 2018;45 : 88–107. doi: 10.1159/000486225 29310113 17 Yan C , Zeng T , Lee K , Nobis M , Loh K , Gou L , et al . Peripheral-specific Y1 receptor antagonism increases thermogenesis and protects against diet-induced obesity. Nat Commun. 2021;12 : 1–20. doi: 10.1038/s41467-021-22925-3 33397941 18 Eva C , Serra M , Mele P , Panzica G , Oberto A . Physiology and gene regulation of the brain NPY Y1 receptor. Front Neuroendocrinol. 2006;27 : 308–339. doi: 10.1016/j.yfrne.2006.07.002 16989896 19 Hohmann S , Buchmann AF , Witt SH , Rietschel M , Jennen-Steinmetz C , Schmidt MH , et al . Increasing association between a neuropeptide Y promoter polymorphism and body mass index during the course of development. Pediatr Obes. 2012;7 : 453–460. doi: 10.1111/j.2047-6310.2012.00069.x 22941950 20 Yulyaningsih E , Zhang L , Herzog H , Sainsbury A . NPY receptors as potential targets for anti-obesity drug development. Br J Pharmacol. 2011;163 : 1170–1202. doi: 10.1111/j.1476-5381.2011.01363.x 21545413 21 Ghersi D , Sanchez R . Improving accuracy and efficiency of blind protein-ligand docking by focusing on predicted binding sites. Proteins Struct Funct Bioinforma. 2009;74 : 417–424. doi: 10.1002/prot.22154 18636505 22 Chen X , Zhou H , Liu YB , Wang JF , Li H , Ung CY , et al . Database of traditional Chinese medicine and its application to studies of mechanism and to prescription validation. Br J Pharmacol. 2006;149 : 1092–1103. doi: 10.1038/sj.bjp.0706945 17088869 23 Ru J , Li P , Wang J , Zhou W , Li B , Huang C , et al . TCMSP: A database of systems pharmacology for drug discovery from herbal medicines. J Cheminform. 2014;6 : 1–6. doi: 10.1186/1758-2946-6-13 24397863 24 O’Boyle NM , Banck M , James CA , Morley C , Vandermeersch T , Hutchison GR . Open Babel: An open chemical toolbox. J Cheminform. 2011;3 : 1–14.21214931 25 Yang Z , Han S , Keller M , Kaiser A , Bender BJ , Bosse M , et al . Structural basis of ligand binding modes at the neuropeptide y Y1 receptor. Nature. 2018;556 : 520–524. doi: 10.1038/s41586-018-0046-x 29670288 26 Humphrey W , Dalke A , Schulten K . VMD: visual molecular dynamics. J Mol Graph. 1996;14 : 33–38. doi: 10.1016/0263-7855(96)00018-5 8744570 27 Choi JH , Banks AS , Kamenecka TM , Busby SA , Chalmers MJ , Kumar N , et al . Antidiabetic actions of a non-agonist PPARγ ligand blocking Cdk5-mediated phosphorylation. Nature. 2011;477 : 477–481. doi: 10.1038/nature10383 21892191 28 Waterhouse A , Bertoni M , Bienert S , Studer G , Tauriello G , Gumienny R , et al . SWISS-MODEL: homology modelling of protein structures and complexes. Nucleic Acids Res. 2018;46 : W296–W303. doi: 10.1093/nar/gky427 29788355 29 Trott O , Olson AJ . AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem. 2010;31 : 455–461. doi: 10.1002/jcc.21334 19499576 30 Sander T , Freyss J , Von Korff M , Rufener C . DataWarrior: An open-source program for chemistry aware data visualization and analysis. J Chem Inf Model. 2015;55 : 460–473. doi: 10.1021/ci500588j 25558886 31 Gaulton A , Hersey A , Nowotka M , Bento AP , Chambers J , Mendez D , et al . The ChEMBL database in 2017. Nucleic Acids Res. 2017;45 : D945–D954. doi: 10.1093/nar/gkw1074 27899562 32 Pires DEV , Blundell TL , Ascher DB . pkCSM: Predicting small-molecule pharmacokinetic and toxicity properties using graph-based signatures. J Med Chem. 2015;58 : 4066–4072. doi: 10.1021/acs.jmedchem.5b00104 25860834 33 Salentin S , Schreiber S , Haupt VJ , Adasme MF , Schroeder M . PLIP: fully automated protein–ligand interaction profiler. Nucleic Acids Res. 2015;43 : W443–W447. doi: 10.1093/nar/gkv315 25873628 34 Bastian M, Heymann S, Jacomy M. Gephi: an open source software for exploring and manipulating networks. Third international AAAI conference on weblogs and social media. 2009. 35 Vanommeslaeghe K , MacKerell AD Jr . Automation of the CHARMM General Force Field (CGenFF) I: bond perception and atom typing. J Chem Inf Model. 2012;52 : 3144–3154. doi: 10.1021/ci300363c 23146088 36 Vanommeslaeghe K , Raman EP , MacKerell AD Jr . Automation of the CHARMM General Force Field (CGenFF) II: assignment of bonded parameters and partial atomic charges. J Chem Inf Model. 2012;52 : 3155–3168. doi: 10.1021/ci3003649 23145473 37 Huang J , Rauscher S , Nawrocki G , Ran T , Feig M , De Groot BL , et al . CHARMM36m: an improved force field for folded and intrinsically disordered proteins. Nat Methods. 2017;14 : 71–73. doi: 10.1038/nmeth.4067 27819658 38 Wu EL , Cheng X , Jo S , Rui H , Song KC , Dávila-Contreras EM , et al . CHARMM-GUI membrane builder toward realistic biological membrane simulations. J Comput Chem. 2014;35 : 1997–2004. doi: 10.1002/jcc.23702 25130509 39 Evans DJ , Holian BL . The nose–hoover thermostat. J Chem Phys. 1985;83 : 4069–4074. 40 Parrinello M , Rahman A . Crystal structure and pair potentials: A molecular-dynamics study. Phys Rev Lett. 1980;45 : 1196. 41 Darden T , York D , Pedersen L . Particle mesh Ewald: An N⋅ log (N) method for Ewald sums in large systems. J Chem Phys. 1993;98 : 10089–10092. 42 Hess B , Bekker H , Berendsen HJC , Fraaije JGEM . LINCS: a linear constraint solver for molecular simulations. J Comput Chem. 1997;18 : 1463–1472. 43 Abraham MJ , Murtola T , Schulz R , Páll S , Smith JC , Hess B , et al . Gromacs: High performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX. 2015;1–2 : 19–25. doi: 10.1016/j.softx.2015.06.001 44 Kumari R , Kumar R , Consortium OSDD, Lynn A . g_mmpbsa—A GROMACS tool for high-throughput MM-PBSA calculations. J Chem Inf Model. 2014;54 : 1951–1962. doi: 10.1021/ci500020m 24850022 45 Teague SJ , Davis AM , Leeson PD , Oprea T . The design of leadlike combinatorial libraries. Angew Chemie Int Ed. 1999;38 : 3743–3748. doi: 10.1002/(SICI)1521-3773(19991216)38:24<3743::AID-ANIE3743>3.0.CO;2-U 10649345 46 Veber DF , Johnson SR , Cheng H-Y , Smith BR , Ward KW , Kopple KD . Molecular properties that influence the oral bioavailability of drug candidates. J Med Chem. 2002;45 : 2615–2623. doi: 10.1021/jm020017n 12036371 47 Hou T , Wang J , Zhang W , Xu X . ADME evaluation in drug discovery. 6. Can oral bioavailability in humans be effectively predicted by simple molecular property-based rules? J Chem Inf Model. 2007;47 : 460–463. doi: 10.1021/ci6003515 17381169 48 Mo R , Jiang T , Di J , Tai W , Gu Z . Emerging micro- and nanotechnology based synthetic approaches for insulin delivery. Chem Soc Rev. 2014;43 : 3595–3629. doi: 10.1039/c3cs60436e 24626293 49 Li W , Wan H , Yan S , Yan Z , Chen Y , Guo P , et al . Gold nanoparticles synthesized with Poria cocos modulates the anti-obesity parameters in high-fat diet and streptozotocin induced obese diabetes rat model. Arab J Chem. 2020;13 : 5966–5977. 50 Lo Y-C , Rensi SE , Torng W , Altman RB . Machine learning in chemoinformatics and drug discovery. Drug Discov Today. 2018;23 : 1538–1546. doi: 10.1016/j.drudis.2018.05.010 29750902 51 Avlani VA , Gregory KJ , Morton CJ , Parker MW , Sexton PM , Christopoulos A . Critical role for the second extracellular loop in the binding of both orthosteric and allosteric G protein-coupled receptor ligands. J Biol Chem. 2007;282 : 25677–25686. doi: 10.1074/jbc.M702311200 17591774 52 Walker P , Munoz M , Martinez R , Peitsch MC . Acidic residues in extracellular loops of the human Y1 neuropeptide Y receptor are essential for ligand binding. J Biol Chem. 1994;269 : 2863–2869. doi: 10.1016/s0021-9258(17)42022-9 8300621 53 Choong YS , Lim YY , Soong JX , Savoo N , Guida C , Rhyman L , et al . Theoretical study of the interactions between peptide tyrosine tyrosine [PYY (1–36)], a newly identified modulator in type 2 diabetes pathophysiology, with receptors NPY1R and NPY4R. Hormones. 2021;20 : 557–569. doi: 10.1007/s42000-021-00278-2 33782920 54 Sautel M , Rudolf K , Wittneben H , Herzog H , Martinez R , Munoz M , et al . Neuropeptide Y and the nonpeptide antagonist BIBP 3226 share an overlapping binding site at the human Y1 receptor. Mol Pharmacol. 1996;50 : 285–292. 8700135 55 Kannoa T , Kanatani A , Keen SLC , Arai-Otsuki S , Haga Y , Iwama T , et al . Different binding sites for the neuropeptide Y Y1 antagonists 1229U91 and J-104870 on human Y1 receptors. Peptides. 2001;22 : 405–413. doi: 10.1016/s0196-9781(01)00350-3 11287095 56 Filizola M , Wang SX , Weinstein H . Dynamic models of G-protein coupled receptor dimers: indications of asymmetry in the rhodopsin dimer from molecular dynamics simulations in a POPC bilayer. J Comput Aided Mol Des. 2006;20 : 405–416. doi: 10.1007/s10822-006-9053-3 17089205 57 Siraj MA , Rahman MS , Tan GT , Seidel V . Molecular docking and molecular dynamics simulation studies of triterpenes from Vernonia patula with the cannabinoid type 1 receptor. Int J Mol Sci. 2021;22 : 3595. doi: 10.3390/ijms22073595 33808384 58 Park C , Kim J , Ko S-B , Choi YK , Jeong H , Woo H , et al . Structural basis of neuropeptide Y signaling through Y1 receptor. Nat Commun. 2022;13 : 1–12.34983933 59 Sylte I , Andrianjara CR , Calvet A , Pascal Y , Dahl SG . Molecular dynamics of NPY Y1 receptor activation. Bioorg Med Chem. 1999;7 : 2737–2748. doi: 10.1016/s0968-0896(99)00229-1 10658578 60 Du P , Salon JA , Tamm JA , Hou C , Cui W , Walker MW , et al . Modeling the G-protein-coupled neuropeptide Y Y1 receptor agonist and antagonist binding sites. Protein Eng. 1997;10 : 109–117. doi: 10.1093/protein/10.2.109 9089810