Yoda1 increases PIEZO2’s sensitivity to membrane stretch.

a) Exemplar PIEZO2 single channel pressure-clamp current trace measured from PIEZO2-transfected HEK293TΔPZ1 cells in presence of 20 µM Yoda1 or 0.05% DMSO vehicle control in the patch pipette. b) Scatter plots showing mean open and shut times from idealized single channel current traces (top; DMSO: n = 17, Yoda1: n = 7 independent patches) and open probability calculated for each patch using mean open and shut times (bottom). c) Dwell time histograms obtained from single channel recordings (DMSO: 2125 shut/open events; Yoda1: 1457 shut/open events). Histograms were best fitted with three exponential components (red traces) using Maximum likelihood estimation. d) Graphs comparing the mean dwell time of each lifetime component. e) Exemplar macroscopic PIEZO2 current traces (left: concatenated view; right: sweep stack view) and evoked by brief suction pulses of varying amplitude in presence of vehicle control or 20 µM Yoda1 in the bath (Right: same traces shown as sweep stacks). f) Top: Graph showing relative PIEZO2 peak current (I/Imax) vs. patch pressure as idealized curves corresponding to Boltzmann fits to the data obtained from independent cells. Bottom: Scatter plots of the mid-point activation pressure (P1/2) determined from Boltzmann fits shown above (DMSO: n = 9, Yoda1: n = 12 independent cells). Numbers next to bottom plots in (b) and (f) are exact p-values from Mann-Whitney U tests.

Yoda1 elicits PIEZO2-dependent calcium influx.

a) Time-course of calcium-sensitive fluorescence obtained with confocal imaging from cells transfected with GC6 (control), with PIEZO1-IRES-GC6 (PIEZO1) or PIEZO2-IRES-GC6 (PIEZO2) plasmids in the presence of Hank’s buffer saline (HBSS) or calcium boosted (CBS) solution. Cells were first exposed to 0.25% DMSO or 30 µM Yoda1 followed by 3 µM ionomycin. Dots and shaded areas respectively represent mean and s.e.m. from n = 12 (control DMSO HBSS), 11 (PIEZO1 DMSO HBSS), 13 (PIEZO2 DMSO HBSS), 23 (control Yoda1 HBSS), 4 (PIEZO1 Yoda1 HBSS), 24 (PIEZO2 Yoda1 HBSS), 26 (control DMSO CBS), 22 (PIEZO1 DMSO CBS), 12 (PIEZO2 DMSO CBS), 183 (control Yoda1 CBS), 105 (PIEZO1 Yoda1 CBS) cells, 55 (PIEZO2 Yoda1 CBS). b) Scatter plots comparison of ionomycin-normalized maxΔF/F0 obtained from data shown in a). c) Calcium influx measured in cells transfected with PIEZO2-IRES-GC6 and elicited by acute perfusion with a HBSS solution at isotonic (1X) or hypertonic (2X and 3X) concentrations containing either 100 µM Yoda1 or DMSO vehicle (Yoda1/DMSO: 1X, n = 29/29; 2X, n = 23/18; 3X, n = 30/30 independent cells). Data are normalized by dividing maxΔF/F0 values obtained in each cell treated with Yoda1 by the mean maxΔF/F0 value of their corresponding DMSO control. Numbers above plots in b) and c) show p-values from non-parametric Kruskal-Wallis test with Dunn’s multiple comparison against control.

Modulation of PIEZO2 poking currents by Yoda1.

a) Exemplar poking current traces from cells expressing the indicated PIEZO2 constructs in presence of 30 µM Yoda1 or 0.075% DMSO. Vertical bars represent 200 pA. b) Scatter plots showing inactivation Tau obtained from fitting individual traces as shown in a) with a mono-exponential function. Numbers above plots are exact p-values from unpaired t-tests. c) Left: Whole-cell ionic currents evoked by a 30 ms poke stimulus in HEK293TΔPZ1 cells transfected with a plasmid encoding WT PIEZO1 or the PIEZO1/2 1961-2063 chimera in the presence of 0.075% DMSO (black traces) and after addition of 30 µM Yoda1 (green traces). Right: Pairwise comparison of poking current deactivation kinetics (mono-exponential fits are shown as red traces on the left) in the WT (n = 5 independent cells) and chimeric (n = 4 independent cells) channels. Numbers are exact p-values from paired t-tests.

Modulation of PIEZO2 inactivation kinetics by other small molecules.

a) Exemplar PIEZO2 poking current traces obtained at -80 mV and evoked with a 5 µm poke in the presence of Yoda1 (grey) or Yoda2 (red) in the bath solution at indicated concentrations. Vertical bars represent 200 pA for all traces. b) Left: Dose-response curves from data shown in (a) depicting mono-exponential Tau inactivation of PIEZO2 poking currents as a function of the concentration of Yoda1 (grey, n = 10, 8, 11, 11, 15, 7, 46 independent cells from low to high dose) or Yoda2 (red, n = 2, 4, 8, 10, 13, 14 46 independent cells from low to high dose) in the bath solution. Lines are fit to the data using a standard dose-response equation. Right: Fit-idealized dose-response curves shown on the left are shown here normalized to the maximal response. c) Left: Exemplar PIEZO2 poking current traces measured in the presence of 30 µM Dooku1, Cmpd15, Cmpd64, or vehicle control (0.1 % DMSO) in the bath solution. Right: Scatter plots of Tau inactivation from traces shown in left (DMSO, n = 28; Dooku1, n = 27; Cmpd15, n = 8; Cmpd64, n = 8 independent cells). Numbers above plots show exact p-values from a Kruskal-Wallis test with Dunn’s correction for multiple comparisons.

Molecular basis for Yoda2’s increased potency.

a) Zoomed-in structural alignment of PIEZO1 (PDBID: 6b3R) and PIEZO2 (PDBID: 6kg7) showing the Yoda binding region colored as a function of sequence similarity. b) Left: chemical structures of Yoda1 and Yoda2. Right: localization and sequence conservation of PIEZO1 residues R1724 and R2098 (licorice). c) Left: dose-responses curves obtained for Yoda1 and Yoda2 against WT PIEZO1 (dark grey), R1724S (red) and R2098S (blue) using calcium imaging measured from n = 3 independent experiments (n = 4 for Yoda1/R1724S). Right: scatter plots showing WT-normalized EC50 for both Yoda molecules against each mutant. The numbers next to plots indicate exact p-values from two-tailed Mann-Whitney U-tests.

A conserved arginine electrostatically stabilizes Yoda2 in the Yoda binding site.

a) Structural overlap of indicated PIEZO1 PDBID structures showing rotameric flexibility of R1724. b) The side chain orientation of R1724 is calculated by measuring the angle between two planes: one formed by L1800-Cα R1724-Cα and the guanidinium group (R1724-Cζ), the other formed by Y1323-Cα, R1724-Cα and R1724-Cζ. Right: example of two R1724 orientations from our simulations with corresponding torsion angle. c) Torsion angle dynamics and histogram distribution of R1724 from previous simulations of the full-length PIEZO1 under 9.0 mN/m membrane tension (15). Vertical markers next to the Replica1 histograms indicate torsion angles reported for nine PIEZO1 PDBID structures indicated below. d) All-atom simulations of a truncated PIEZO1 blade (3 replicas of 20, 0.6 and 0.6 µs) showing the binding dynamics of Yoda2 (top), the dynamics of Yoda2-R1724 salt-bridge formation (distance between R1724-Cζ and the carboxylic carbon of Yoda2) (middle), and the time-evolution of the R1724 torsion angle (bottom). e) The dominant Yoda2 binding pose observed in >55% of the combined simulation frames from 3 replicas shows non-covalent interactions between Yoda2 and the indicated PIEZO1 residues, including the salt bridge with R1724.