Screening glycosides containing disaccharide and trisaccharide moieties at P2X7. (A) Initial experiments used a fixed concentration of ATP (200 µM final) and glycoside (10 µM final) to screen selected glycosides at P2X7. Data are collated from two to four independent experiments. Ginsenoside-CK was used as the control PAM and is shown in blue. YO-PRO-1 uptake was measured as area under curve (50–300 seconds), and data are expressed as percentage of control, where the control is ATP + DMSO. One-way ANOVA with Dunnett’s multiple comparisons test was performed. *P < 0.05 compared with DMSO control. (B) Dose-response curve to ATP in the presence of vehicle (DMSO), ginsenoside-CK, or gypenoside XVII (10 µM). (C) Dose-response curve to ATP in the presence of vehicle (DMSO), ginsenoside-CK, or gypenoside XLIX (10 µM). (D) Dose-response curve to ATP in the presence of vehicle (DMSO), ginsenoside-CK, or saikosaponin A (10 µM). Data points are means ± S.D. The same curves from DMSO and ginsenoside-CK are shown in (B and D). (E) Summary of data from YO-PRO-1 uptake experiments in HEK-hP2X7 cells or parental nontransfected HEK-293 cells in response to drug alone (saikosaponin A or solanine). (F) Lack of effect of the P2X7-selective antagonist AZ10606120 (AZ106; 10 µM) on ATP-induced YO-PRO-1 uptake when solanine or saikosaponin were used. One-way ANOVA with Sidak’s multiple comparisons test was performed. *P < 0.05, ns denotes not significant.

Screening glycosides containing monosaccharide moieties at P2X7. Dose-response curves to ATP in the presence of vehicle (DMSO), ginsenoside-CK, and the following: daucosterol (10 µM) (A), stevioside (SV; 10 µM) (B), ginsenoside-F1 (F1; 10 µM) (C), ginsenoside-F2 (F2; 10 µM) (D), ouabain (10 µM) (E), or scilliroside (10 µM) (F). Ginsenoside-CK is demonstrated throughout as the reference compound (same data shown in each plot). Data are collated from three independent experiments each performed in triplicate. Error bars represent S.D.

Molecular docking of ginsenoside-CK and ginsenoside-F1 to the central vestibule pocket of hP2X7. Ginsenoside-CK (cyan) docked into the central vestibule in the ATP-bound homology model of human P2X7 and right, ginsenoside-F1 (green) docked into the same site. Side chains of amino acid residues of hP2X7 implicated in interactions are shown.

Diastereoisomers of ginsenosides have different activity at hP2X7. (A) Dose-response curves to ATP in the presence of vehicle (DMSO), 20(S)-ginsenoside-Rg3, or 20(R)-ginsenoside-Rg3 (10 µM). (B) Dose-response curves to ATP in the presence of vehicle (DMSO), ginsenoside-20(S)-Rh2, or ginsenoside-20(R)-Rh2 (10 µM). Data are collated from three independent experiments and are means ± S.D. The same DMSO data are shown in both plots. (C) Chemical structures of ginsenoside-Rg3 and ginsenoside-Rh2 are shown with the stereo-centers highlighted in red.

Induced-fit docking of 20S-Rg3 at human P2X7 ginsenoside-20(S)-Rg3 (green) docked into the central vestibule site in a homology model of ATP-bound human P2X7 (open state). The predicted orientation of the stereocentre C-20 is such that the –OH is pointing away from the hydrophobic face of the binding site, thus minimizing any repulsive interactions. Key side chains of residues D318, L320, and F322 are indicated.

The PAM effects of glycosides in human THP-1 monocyte cell line. (A) ATP-induced calcium responses were measured in fura-2 AM loaded THP-1 cells in suspension using a Flexstation 3 plate reader. Agonist (500 µM ATP) and PAM (ginsenoside-CK 10 µM) were coinjected after establishment of a baseline for 40 seconds. Cells were preincubated with various P2X7 antagonists for 10 minutes prior to start of plate recordings. (B) Summary of collated data from calcium measurements. Fura-2 responses were calculated as area under curve. Data were analyzed using one-way ANOVA with Tukey’s multiple comparisons test to assess the effect of antagonists. *P < 0.05. (C) Investigating the P2X7 dependence of glycoside effects on ATP-induced calcium responses. AZ10606120 (AZ106; 10 µM) was added to cells to block P2X7 receptors prior to measuring calcium responses. Data were analyzed using one-way ANOVA with Dunnett’s multiple comparison test comparing each column against the control (500 µM ATP + DMSO). *P < 0.05. (D) Cell viability experiments were performed over 24 hours using HEK-hP2X7 cell line. Alamar blue fluorescence was measured, and data were normalized to percentage of control (DMSO). Data were collated from five independent experiments. One-way ANOVA was used to analyze the data with Sidak’s multiple comparisons test to compare selected pairs of columns (DMSO + ATP vs. ginsenoside + ATP). *P < 0.05. (E) Cell viability experiments were performed over 24 hours using THP-1 cells. Alamar blue fluorescence was measured and data were normalized to percentage of control (DMSO). Data were collated from five independent experiments. One-way ANOVA was used to analyze the data with Sidak’s multiple comparisons test to compare selected pairs of columns (DMSO + ATP vs. ginsenoside + ATP). *P < 0.05.