Cloud Hosting

Libraries preparation

A total of 265 medicinal plants were selected and introduced into the ChEBI database to extract the effective compounds. After removing duplicates, 87 potential compounds were obtained from the 265 native plant species. The PharmMapper website has introduced 423 potential target candidates for these compounds. In addition, the NCBI database has introduced 3312 colorectal cancer-related targets. Finally, 34 human protein molecules were selected, which could be targets of these compounds and play a role in colorectal cancer.

Of these compounds, 7 did not have common targets in their top 30 targets. The names and 2D structures of both the groups of compounds used in the screening and their targets are listed in Supplementary Tables S1S3.

A DT network with 87 compounds and 34 targets was constructed consisting of 114 nodes and 416 degrees. The NetworkAnalyser results calculate parameters such as the number of degrees, betweenness, and edge betweenness centrality values, which are used to select the best molecular targets and potential chemical compounds (Fig.1). Figure1A shows a bipartite network configured with node sizes based on the number of degrees; the larger the node size, the greater is the number of associations. Based on the network analysis, the compounds showed degrees ranging from 1 to 8. To screen the best compounds as ligands for molecular docking analysis against targets, compounds with six or more nodes were selected. Therefore, 31 compounds listed in Table 1 were selected for further study. Carbonic anhydrase 2, with a score of 0.44, showed the highest centrality among the targets, followed by mitogen-activated protein kinase 14 (0.14), estrogen receptor (0.07), and angiogenin (0.07). The remaining molecules had a score of less than 0.05. Moreover, Carbonic anhydrase 2 (71), mitogen-activated protein kinase 14 (38), bone morphogenetic protein 2 (35), and estrogen receptor (34) showed the most connections, based on the number of degrees associated with the nodes.

(A) The bipartite DT network. (B) The circle layout of DT network based on degree. (C) The circle layout of DT network based on betweenness centrality. (D) The betweenness centrality of nods. The green circular nodes represent the targets, and the brown triangular nodes represent the chemical compounds. The larger the size of the node, the greater the number of nodes and associations.

Finally, the compounds showed a number of degrees ranging from 1 to 8. To screen the best compounds as ligands for the molecular docking step against the targets, all compounds with six or more nodes were selected, in which there were 31 compounds with six, seven, or eight nodes.

After checking using the GEPIA database, the TOP targets that showed increased expression in colorectal cancer were selected. The target structures were downloaded from the PDB database (PDB IDs: 6AY2 for CTSB, 5T4E for DPP4, 6FVE for MIF, 7AUV for MAPK1, 4QTD for MAPK8, and 1SMO for TERM1). The Chimera 1.8.1 software was used for essential protein preparations, including removing water, ATP, ligands, and adding hydrogen charges. Critical ligand-binding sites were considered and docked to 31 compounds using the PyRx software. 31 molecules were selected after analyzing the pharmacokinetic properties and parameters of ADME. The centroid of the binding sites for the targets was calculated as the coordinates of the centroid of the ligand-binding sites using the UniProt database. The docked complexes were analyzed using PyMol and DiscoveryStudio software. Finally, molecules that presented the highest interaction energy against their target (above 7.5 for TERM1, 8 for CTSB and MIF, above 9 for DPP4 and MAPK1 and above 10 for MAPK8), respectively and had more binding with the amino acids of the binding site in the study with DiscoveryStudio. These findings led to the selection of Multiorthoquinone, Liquiritin, Hispaglabridin A, Isoliquiritin, Gibberellin A98, Cyclomulberrin, Cyclomorusin A, Cudraflavone B for simulation (MD), which produced a more stable complex with a lower energy level than TREM1, MAPK8, MAPK1, MAPK8, CTSB, MAPK1, MIF, and DPP4, respectively. The positions and amino acids involved in the binding are illustrated in Fig.2, 3, 4, 5, 6, 7, 8 and 9, Supplementary Tables S4S6. The RMSD score showed slight fluctuations and was approximately 0.2nm. These findings indicate the stability of the compounds in the target complex. The RMSF values for all protein structures were computed to accurately determine how the binding of the compounds affects flexibility.

Two-dimensional representations of the Multiorthoquinone against TREM1. (A) Root-mean square deviation of the complexes (RMSD). (B) Radius of gyration (Rg). (C) Hydrogen bond analysis from the simulation system. (D) Root-mean-square fluctuation (RMSF). (E) The binding conformation of 3D view. (F) Binding site interactions of 2D view.

Two-dimensional representations of the Liquiritin against MAPK8. (A) Root-mean square deviation of the complexes (RMSD). (B) Radius of gyration (Rg). (C) Hydrogen bond analysis from the simulation system. (D) Root-mean-square fluctuation (RMSF). (E) The binding conformation of 3D view. (F) Binding site interactions of 2D view.

Two-dimensional representations of the Isoliquiritin against MAPK8. (A) Root-mean square deviation of the complexes (RMSD). (B) Radius of gyration (Rg). (C) Hydrogen bond analysis from the simulation system. (D) Root-mean-square fluctuation (RMSF). (E) The binding conformation of 3D view. (F) Binding site interactions of 2D view.

Two-dimensional representations of the Hispaglabridin A against MAPK1. (A) Root-mean square deviation of the complexes (RMSD). (B) Radius of gyration (Rg). (C) Hydrogen bond analysis from the simulation system. (D) Root-mean-square fluctuation (RMSF). (E) The binding conformation of 3D view. (F) Binding site interactions of 2D view.

Two-dimensional representations of the Cyclomulberrin against MAPK1. (A) Root-mean square deviation of the complexes (RMSD). (B) Radius of gyration (Rg). (C) Hydrogen bond analysis from the simulation system. (D) Root-mean-square fluctuation (RMSF). (E) The binding conformation of 3D view. (F) Binding site interactions of 2D view.

Two-dimensional representations of the Cudraflavone B against DPP4. (A) Root-mean square deviation of the complexes (RMSD). (B) Radius of gyration (Rg). (C) Hydrogen bond analysis from the simulation system. (D) Root-mean-square fluctuation (RMSF). (E) The binding conformation of 3D view. (F) Binding site interactions of 2D view.

Two-dimensional representations of the Gibberellin A98 against CTSB. (A) Root-mean square deviation of the complexes (RMSD). (B) Radius of gyration (Rg). (C) Hydrogen bond analysis from the simulation system. (D) Root-mean-square fluctuation (RMSF). (E) The binding conformation of 3D view. (F) Binding site interactions of 2D view.

Two-dimensional representations of the Cyclomulberrin against MIF. (A) Root-mean square deviation of the complexes (RMSD). (B) Radius of gyration (Rg). (C) Hydrogen bond analysis from the simulation system. (D) Root-mean-square fluctuation (RMSF). (E) The binding conformation of 3D view. (F) Binding site interactions of 2D view.

Analysis of the liquiritin- and isoliquiritin-associated RMSF plots revealed that Liquiritin and Isoliquiritin flexibility were significantly different in the five regions of 35, 183, 199, 246, and 307 and two regions in 200 and 328 in MAPK8, respectively. In addition, multiorthoquinone flexibility was extremely high in the three regions 34, 76, and 107 in TREM1, which may be attributed to a lack of interaction between the three regions and multiorthoquinone. For Cyclomulberrin and Hispaglabridin A residues 94320, and 35, 300350 in MAPK1, the flexibility of amino acids was lower and higher, respectively. The RMSF plot associated with DPP4 indicated high interaction between most regions of DPP4 and Cudraflavone B. Additionally, Cyclomorusin A and Gibberellin A98 flexibility were extremely different in the two regions of 5070 and 150170 in the CTSB, and 80100 region in the MIF, respectively.

RG nature was constant for the individual domains during the entire simulation period associated with Gibberellin A98, Cyclomorusin A, and Cudraflavone B, indicating that the individual domains did not melt or unfold. These compounds did not affect the secondary structures of CTSB, MIF, or DPP4. However, the RG value associated with Liquiritin, Isoliquiritin and Hispaglabridin A, Cyclomulberrin throughout the MD simulation led to unfolding and activation of MAPK8 and MAPK1, respectively.

Supplementary Fig. S1 shows the projection of the trajectories on special eigenvectors (vectors 1 and 2) and time-dependent motions of the components in a particular vibration mode. The overall analysis of the eigenvector plots indicated that most vibrations occurred along eigenvector 1. According to the first two PCs, TREM1 and MIF proteins have almost the same trace values of the covariance matrix for the bound and unbound states with a slight shift. This indicates that the ligand is well equilibrated and stabilized with the protein, as reflected by theleast conformational changes due to reduced collective motions from unbound states. However, in other cases, 2D projection plots of the trajectories revealed that the ligand reduced the conformational diversity during the simulations, leading to a more compact cluster distribution. Sampling of different regions and showing different movement behaviors of proteinligand complexes compared to unbound proteins points to the binding of ligand effects on the rigidity of the structural conformation, which also affects the function of proteins.

The MM/PBSA binding free energy results are shown in Table 2 including the van der Waals energy (kJ/mol), electrostatic energy (kJ/mol), polar solvation energy (kJ/mol), SASA energy (kJ/mol), SAV energy (kJ/mol), WCA energy (kJ/mol), and binding energy (kJ/mol), are shown in Table 2. Compared to liquiritin, isoliquiritin had the lowest binding energy score of 183.04kJ/mol interacting with MAPK8 and formed a stronger binding. In addition, hypoglabrin A and cyclomulbrin interact with MAPK1 at almost the same binding energy (154kJ/mol). However, the number of hydrogen bonds in the interaction with hypoglabrin A is relatively high. MIF-Cyclomorusin A, with the most negative energy score (243.768kJ/mol), showed the strongest interaction among all docked complexes.

Originally posted here:

Investigation of the therapeutic role of native plant compounds ... - Nature.com