Molecular and structural basis of the dual regulation of the polycystin-2 ion channel by small-molecule ligands

Abstract

Mutations in the PKD2 gene, which encodes the polycystin-2 (PC2, also called TRPP2) protein, lead to autosomal dominant polycystic kidney disease (ADPKD). As a member of the transient receptor potential (TRP) channel superfamily, PC2 functions as a non-selective cation channel. The activation and regulation of the PC2 channel are largely unknown, and direct binding of small-molecule ligands to this channel has not been reported. In this work, we found that most known small-molecule agonists of the mucolipin TRP (TRPML) channels inhibit the activity of the PC2_F604P, a gain-of-function mutant of the PC2 channel. However, two of them, ML-SA1 and SF-51, have dual regulatory effects, with low concentration further activating PC2_F604P, and high concentration leading to inactivation of the channel. With two cryo-electron microscopy (cryo-EM) structures, a molecular docking model, and mutagenesis results, we identified two distinct binding sites of ML-SA1 in PC2_F604P that are responsible for activation and inactivation, respectively. These results provide structural and functional insights into how ligands regulate PC2 channel function through unusual mechanisms and may help design compounds that are more efficient and specific in regulating the PC2 channel and potentially also for ADPKD treatment.

Keywords: cryo-EM structure; ion channel; ligand; polycystic kidney disease; transient receptor potential channel.

Fig. 1.

ML-SA1 further activates the GOF PC2_F604P channel. (A) Representative gap-free recording at +20 mV shows that ML-SA1 activates PC2_F604P in a bath solution containing 100 mM Na⁺ and 2 mM Ca²⁺. The structure of the ML-SA1 molecule is shown as an insertion. (B) Representative I–V curves (Left) and scatter plot and bar graph (Right) show currents of PC2_F604P before and after applying 20 µM ML-SA1 in a bath solution containing 100 mM Na⁺ and 2 mM Ca²⁺. In the scatter plot and bar graph, the current at +80 mV of each oocyte was normalized to its current before applying ML-SA1. Data are presented as mean ± SD in bar graphs here and those in the other figures (***P < 0.001, Student’s t test). (C) Concentration–response curve of the effect of ML-SA1 on increasing the current of PC2_F604P at +80 mV when recorded in 100 mM Na⁺/2 mM Ca²⁺ solution. Data were collected when currents reached the peak seconds after the ML-SA1 application. (D) Same as in B except that the currents were recorded in a bath solution without Ca²⁺. (E) Representative I–V curves of PC2_F604P, recorded before and after treatment of 20 µM ML-SA1 in 70 mM Ca-gluconate, and then switched to 70 mM CaCl₂ with 20 µM ML-SA1. (F) Representative I–V curves of PC2_F604P recorded in 70 mM CaCl₂ before and after treatment of 20 µM ML-SA1.

Fig. 2.

The activating binding site of ML-SA1 in PC2_F604P. (A) Sequence alignment shows conservation between PC2 and TRPML1 at the indicated regions. Gray boxes indicate the key residues at the ML-SA1 binding site in TRPML1 (23) and the aligned residues in PC2. Blue and pink boxes indicate the selectivity filter and lower gate, respectively. Asterisk: identical; double dots: conserved; single dot: similar. (B) Cryo-EM structure of the human TRPML1/ML-SA1 complex (PDB #5WJ9) (23). (C) Structural alignment between TRPML1/ML-SA1 and WT PC2 (PDB #5T4D) (21) (Left), and TRPML1/ML-SA1 and PC2_F604P (PDB #6D1W) (27) (Right). Side chains of key residues in the ML-SA1 binding site in TRPML1 and their aligned residues in PC2 are shown. (D and E) Representative I–V curves (D) and scatter plots and bar graphs (E) show the currents of PC2_F604P-F670A and PC2_F604P-F676A mutants before and after 20 µM ML-SA1 treatment. Currents at +80 mV are shown in bar graphs and were normalized to currents before ML-SA1 treatment (***P < 0.001, n.s.: no significance, Student’s t test). (F) A docking model showing a ML-SA1 molecule docked in a cavity of PC2_F604P that matches the ML-SA1 binding site in TRPML1. In the zoomed-in picture (Right), key residues in PC2_F604P that are aligned to those involved in ML-SA1 binding in TRPML1 are shown.

Fig. 3.

SF-33 and SF-71 may bind to the same site as ML-SA1 to inhibit PC2_F604P current. (A) Representative I–V curves (Left) and scatter plot and bar graph (Right) show currents of PC2_F604P before and after applying 20 µM SF-33 in a bath solution containing 100 mM Na⁺ and 2 mM Ca²⁺. Normalized currents at +80 mV are shown in the scatter plot and bar graph (***P < 0.001, Student’s t test). The structures of SF-33 is shown as insertion. (B) Concentration–response curves of the effect of SF-33 on inhibiting the current of PC2_F604P at +80 mV. (C and D) Same as in A and B except that 20 µM SF-71 was applied instead of SF-33. (E and F) Representative gap-free recording at +20 mV showing applying SF-33 or SF-71 removed the activation effect of ML-SA1 (E) and applying ML-SA1 also removed inhibition of SF-33 or SF-71 (F). (G) Scatter plots and bar graphs show the inhibitory effects of SF-33 and SF-71 at the indicated concentration on the PC2_F604P channel carrying either F670A (Left) or F676A (Right) mutation. Normalized currents at +80 mV are shown in bar graphs (*P < 0.05, **P < 0.01, ***P < 0.001, n.s.: no significance, Student’s t test). Representative I–V curves are shown in SI Appendix, Fig. S5.

Fig. 4.

High concentrations of ML-SA1 and SF-51 inhibit the activity of the PC2-F604P. (A and B) Representative gap-free recording at +20 mV (A) and I–V curve (B) of the PC2_F604P current, showing the activation after applying 20 µM ML-SA1 and the inhibition after applying 100 µM ML-SA1. (C) The plot of the activation index (ratio of current at +80 mV after ML-SA1 application vs. current before application, I_ML-SA1/I_bath) of ML-SA1 at 20 and 100 µM, showing the initial activation after applying 20 µM ML-SA1 and the inhibition after further applying 100 µM ML-SA1 for 2 min. Data from four individual oocytes are shown. (D) Gap-free recording of PC2_F604P at +20 mV showing the effects of applying 20 µM (Middle) or 100 µM (Right) ML-SA1 for a relatively long time. Applying 100 µM ML-SA1 led to a sharp increase followed by a gradual decrease in the current. The current at the initial activation peak and after 2 min application, which were used in generating curves in E, are indicated. DMSO, which was used to help dissolve ML-SA1, was tested as negative control (Left). (E) Plot of the activation index of ML-SA1 at the “initial” activation peak, 2 min, and 4 min after applying indicated concentrations of ML-SA1, showing that concentrations above 20 µM led to inhibition after the initial activation. (F) Activation indexes of 100 µM ML-SA1 or 100 µM SF-51, measured at around 5 s when the current reached the peaks and 1 min after application, showing that SF-51 caused stronger inhibition than ML-SA1. (G) Representative I–V curves showing the current of PC2_F604P before and after applying 100 µM SF-51 for 5 s or 1 min.

Fig. 5.

Cryo-EM structures show that ML-SA1 binds at the TOP/outer pore site of PC2 and may lead to channel closing. (A) Two orthogonal surface views of cryo-EM reconstruction of PC2-F604P in the presence of 200 µM ML-SA1. The corresponding EM densities for ML-SA1 are in orange. One of them is highlighted within the boxes and shown in transparency on top. Structural figures were prepared in UCSF ChimeraX (49) if not otherwise indicated. (B–D) Close-up views of the density maps (Left) and cartoon (Right) of 3D reconstructions of PC2-F604P in ML-SA1-bound (blue, B) and lipid-bound (purple, C) state, compared to PC2-F604P apo structure reported previously (27) (green, D, PDB code: 6D1W). The EM densities corresponding to the lipid (pink) and ML-SA1 (orange) are represented in transparency in the cartoon view. All the maps are contoured at 4σ. The EM maps shown in this manuscript are under C4 symmetry unless otherwise indicated. (E) Close-up views of the electrostatic potential surface of the ML-SA1 binding site, calculated by coulombic electrostatic potential in ChimeraX (49), coloring ranging from red for negative potential through white to blue for positive potential. ML-SA1 is indicated by a black arrow. (F) Cartoon shows that ML-SA1 binds to the TOP/outer pore site. The detailed interaction interface between ML-SA1 and PC2_F604P is shown in the Right panel. The key side chains that may be involved in binding are represented in sticks. R654 in the outer pore is also labeled to indicate its cation-pi interaction with W380 in the TOP domain. (G) TOP: Side views of the ion permeation pathway of two opposing subunits for the PC2 WT [orange, PDB code: 5Z1W (21)], PC2-F604P_apo [green, PDB code: 6D1W (27)], PC2-F604P_lipid (purple) and PC2-F604P_ML-SA1 (blue) structures. A π-helix in the middle of S6 in the WT PC2, which is transited to α-helix in other structures, is indicated. The key residues that form major constrictions in the pore are shown in sticks and the diagonal distances in Å between opposing residues are labeled. The corresponding EM densities for these residues are highlighted within the black boxes at the bottom. (H) Superimposing the pore domain of F604P_apo (green), F604P_lipid (purple), and F604P_mL-SA1 (blue) structures. The elongated intracellular gates in both F604P_lipid and F604P_ML-SA1 structures are indicated by red arrows. (I) Comparison of pore radii in structures shown in G, which are calculated using the HOLE program (50). The dotted line denotes a 1.0 Å-radius.

Fig. 6.

Mutations at the TOP/outer pore site reduce/abolish the high concentration ML-SA1-induced inhibition of PC2_F604P. (A) Detailed structure of PC2_F604P_ML-SA1 shows the key residues contribute to the ML-SA1 binding at the TOP/outer pore site. ML-SA1 is shown in orange. This structural image was prepared with PyMOL (The Pymol Molecular Graphics System). (B) Plots of the activation index of ML-SA1 at 20 and 100 µM from individual oocytes expressing the PC2_F604P, PC2_F604P/Q456W, or PC2_F604P/Q557A, showing the Q456W and Q557A mutations led to a reduced inhibition of 100 µM ML-SA1. Data of 100 µM were collected after 2 min ML-SA1 application. Results of the other tested mutants are included in SI Appendix, Fig. S11. (C) Scatter plot and bar graphs show the activation index of 20 µM ML-SA1 (Left) and the inactivation index of 100 µM ML-SA1 (ratio of current at +80 mV after 100 µM ML-SA1 was applied for 2 min vs. current after 20 µM ML-SA1 was applied for 10 to 15 s, I_{100 µM ML-SA1}/I_{20 µM ML-SA1}) (Right) of PC2_F604P and the indicated mutants. Oocyte numbers are indicated in parentheses. Data are presented as mean ± SD (n.s.: not significant, **P < 0.01, ***P < 0.001, Student’s t test). (D) Plots of the inhibition index of SF-71 (ratio of current after SF-71 application vs. current before application) at 20 and 100 µM when applied to PC2_F604P and the Q557A mutant, showing no change caused by Q557A mutation. Data of 100 µM were collected after 2 min SF-71 application.

Fig. 7.

A working model of the dual regulation of PC2_F604P by ML-SA1. (A) PC2_F604P is a GOF mutant that conducts Na⁺ (brown dots) influx but is not permeable to Ca²⁺ (red dot). (B) ML-SA1 (green triangle) at low concentration (<20 µM) binds to the S5-PH1-S6 site and further activates the PC2_F604P channel. The ML-SA1-activated channel gain weak permeability of Ca²⁺. (C) ML-SA1 at higher concentration (>20 µM) further binds to the TOP/outer pore site and causes inactivation of the PC2_F604P channel. The conformation changes in the inactivation may include the closing of the lower pore.