In-silico Screening and Identification of Antidiabetic Inhibitors Sourced from Phytochemicals of Philippine Plants against Four Protein Targets of Diabetes (PTP1B, DPP-4, SGLT-2, and FBPase)

The Supplementary Files contain the figures, tables, and videos relevant to the manuscript. All supplementary data are appropriately cited in the main text

Table S1. List of ligands with favorable ADMET profile. Results are obtained from ADMETlab 2.0 
Table S2. ADMET results of the control compounds used in this study.
Table S3. Docking scores of the 373 ligands that passed the ADMET filter. Also shown are the docking scores of the control compounds. Two docking software were used, Autodock 4.2 and Autodock Vina. Reference compounds are italicized. Compounds in bold are the top-performing ligands further analyzed in MD simulations. Values are in kJ/mol. 
Table S4. Average binding energies of the potential natural inhibitors against PTP1B using MM/PBSA. 
Table S5. Average binding energies of the potential natural inhibitors against DPP-4 using MM/PBSA. 
Table S6. Average binding energies of the potential natural inhibitors against SGLT-2 using MM/PBSA. 
Table S7. Average binding energies of the potential natural inhibitors against FBPase using MM/PBSA. 

Figure S1. Top ten highest-scoring ligands against PTP1B. Also shown are the structures of the three reference compounds used.
Figure S2. Top ten highest-scoring ligands against DPP-4. Also shown are the structures of the three reference compounds used.
Figure S3. Top ten highest-scoring ligands against SGLT-2. Also shown are the structures of the three reference compounds used.
Figure S4. Top ten highest-scoring ligands against FBPase. Also shown are the structures of the three reference compounds used.
Figure S5. The DPP-4 protein has a patternless loop formed by a group of amino acids from residues 240 to 260 exposed to the water solvent that caused the large peak observed in the RMSF plots. 
Figure S6. The noticeable high peaks in the RMSF plots of SGLT-2 complexes are from the partnerless loops connecting two alpha helices. These loops are exposed to the solvent and interact with the mobile water molecules.  
Figure S7. Docking poses of C-0690, C-1433, and C-1883 against the binding pocket of FBPase. Hydroxyl and methoxy groups are situated away from the binding pocket and towards the solvent. 
Figure S8. The FBPase protein has a large patternless loop formed by a group of amino acids from residues 50 to 70 exposed to the water solvent that caused the large peak observed in the RMSF plots.
Figure S9. Energetic breakdown of PTP1B residues with the strongest binding energy contribution. MM is the sum of the van der Waals and electrostatic energies.
Figure S10. Energetic breakdown of DPP-4 residues with the strongest binding energy contribution. MM is the sum of the van der Waals and electrostatic energies.
Figure S11. Energetic breakdown of SGLT-2 residues with the strongest binding energy contribution. MM is the sum of the van der Waals and electrostatic energies.
Figure S12. Energetic breakdown of FBPase residues with the strongest binding energy contribution. MM is the sum of the van der Waals and electrostatic energies.

Video S1 - Sample 100-ns simulation clip (C-0671) of a stable complex against PTP1B
Video S2 - 100-ns simulation clip of C-0888 which had an unstable behaviour against PTP1B
Video S3 - 100-ns simulation clip of C-1847 which had an unstable behaviour against PTP1B
Video S4 - Sample 100-ns simulation clip (C-2083) of a stable complex against DPP-4
Video S5 - Sample 100-ns simulation clip (C-0914) of a stable complex against SGLT-2
Video S6 - Sample 100-ns simulation clip (C-1254) of a stable complex against FBPase
Video S7 - 100-ns simulation clip of C-0690 which had an unstable behavior against FBPase
Video S8 - 100-ns simulation clip of C-1433 which had an unstable behavior against FBPase
Video S9 - 100-ns simulation clip of C-1883 which had an unstable behavior against FBPase