Mechanism of tethered agonist-mediated signaling by polycystin-1

Abstract

Polycystin-1 (PC1) is an important unusual G protein-coupled receptor (GPCR) with 11 transmembrane domains, and its mutations account for 85% of cases of autosomal dominant polycystic kidney disease (ADPKD). PC1 shares multiple characteristics with Adhesion GPCRs. These include a GPCR proteolysis site that autocatalytically divides these proteins into extracellular, N-terminal, and membrane-embedded, C-terminal fragments (CTF), and a tethered agonist (TA) within the N-terminal stalk of the CTF that is suggested to activate signaling. However, the mechanism by which a TA can activate PC1 is not known. Here, we have combined functional cellular signaling experiments of PC1 CTF expression constructs encoding wild type, stalkless, and three different ADPKD stalk variants with all-atom Gaussian accelerated molecular dynamics (GaMD) simulations to investigate TA-mediated signaling activation. Correlations of residue motions and free-energy profiles calculated from the GaMD simulations correlated with the differential signaling abilities of wild type and stalk variants of PC1 CTF. They suggested an allosteric mechanism involving residue interactions connecting the stalk, Tetragonal Opening for Polycystins (TOP) domain, and putative pore loop in TA-mediated activation of PC1 CTF. Key interacting residues such as N3074–S3585 and R3848–E4078 predicted from the GaMD simulations were validated by mutagenesis experiments. Together, these complementary analyses have provided insights into a TA-mediated activation mechanism of PC1 CTF signaling, which will be important for future rational drug design targeting PC1.

Keywords: Gaussian accelerated molecular dynamics; cellular signaling; mutagenesis; polycystin-1; tethered agonist.

Fig. 1.

Atomic structure of PC1 CTF and experimental effects of stalk region variants on signaling to NFAT reporter. (A) Atomic structure of the PC1 CTF as extracted from the cryo-EM structure of the PC1–PC2 complex (PDB: 6A70). The missing regions added through homology modeling, including Stalk, PL, and C-tail, are highlighted in red. The TOP and PLAT domains are shown in green and pink surfaces, respectively. (B) Sequence alignment of the stalk region in WT, stalkless (ΔStalk), and the G3052R, R3063C, and R3063P ADPKD missense mutants of human PC1 CTF and structural view of PC1 stalk variants with the mutated residues highlighted as sticks. The sequence alignment shows residue numbers (superscripted), GPS cleavage site (arrow), and ADPKD-associated missense mutations/polymorphisms (red). (C) Representative Western blot of WT and mutant CTF proteins from one of the experiments summarized in D. Blot was originally probed with C20 antibody and reprobed with antibody against actin. (D) Relative, average total expression level (+SD) for WT and mutant CTF constructs from signaling transfections represented in E. (E) Average fold NFAT reporter activation (+SD) relative to empty vector (ev) for WT and mutant CTF proteins from three to five separate transfection experiments with n = 3 wells per expression construct per experiment. *P < 0.005 and **P ≤ 0.0001 relative to WT CTF levels.

Fig. 2.

Reduced correlations in residue motions and disrupted domain interactions in the stalk variants of PC1 CTF compared with the WT. (A) The correlation matrix of residue motions averaged over three GaMD simulations of the WT PC1 CTF. High correlations between Stalk–TOP and TOP–PL domains are highlighted in red circles. (B–D) The average correlation matrix of residue motions (lower triangle) and corresponding differences relative to the WT (upper triangle) calculated from three GaMD simulations of the (B) ΔStalk, (C) G3052R, and (D) R3063C systems of PC1 CTF. Important regions with statistically significant differences (P < 0.05) are highlighted in blue circles. (E–H) Two-dimensional free energy profiles of the (E) WT, (F) ΔStalk, (G) G3052R, and (H) R3063C systems of PC1 CTF regarding the number of atom contacts between the Stalk and TOP domains and the R3848-E4078 distance (the CZ atom in R3848 and the CD atom in E4078) calculated from the GaMD simulations. Residue contacts were calculated for heavy atoms with a distance cutoff of 4 Å. Important low-energy conformational states are identified, including the “Closed,” “Intermediate I1,” “Intermediate I2,” “Intermediate I3,” “Open1,” and “Open2.”

Fig. 3.

Distinct low-energy conformations of the WT and stalk variants of PC1 CTF. (A–C) Residue interactions observed in “Closed” protein conformation between (A) Stalk–TOP, (B) Stalk–TM5, and (C) TOP–PL domains. (D–F) Residue interactions observed in “Open” conformation of the Stalk variants between (D) Stalk–TOP, (E) Stalk-TM5, and (F) TOP–PL domains. The Stalk (orange cartoon), TM helices (blue cartoon), TOP domain (green surface), and PL (magenta cartoon) are labeled in the PC1 CTF. Residue interactions are represented in ball and stick.

Fig. 4.

Intermediate conformational states of PC1 CTF observed from GaMD simulations of the R3063P mutant. (A) The average correlation matrix of residue motions (lower triangle) and corresponding differences relative to the WT (upper triangle) calculated from three GaMD simulations of the R3063P mutant. (B) Two-dimensional free energy profile of the R3063P mutant regarding the number of contacts between the Stalk and TOP domains and the R3848–E4078 distance calculated from the GaMD simulations. Important low-energy conformational states are identified, including the “Intermediate I1” and “Intermediate I3.” (C) Structural conformation of Stalk–TOP interaction for the R3063P mutant. (D) Residue interactions between TOP–PL as observed in the “Intermediate I1” conformational state. (E) Residue interactions between TOP–PL as observed in the “Intermediate I3” conformational state.

Fig. 5.

Experimental analyses of GaMD simulation-predicted residue interactions between TOP–PL domains important for WT stalk-mediated activation of PC1 CTF. (A) Average fold NFAT reporter activation (+SD) relative to empty vector (ev) for WT and GaMD mutant CTF proteins from three separate transfection experiments with n = 3 wells per expression construct per experiment. (B) Average relative total expression level (+SD) for WT and GaMD mutant CTF proteins from experiments in A. (C) Representative Western blots of WT and GaMD mutant CTF proteins from one of the experiments represented in B probed with C20 antibody. Images are from two separate Western blots. Removal of intervening lanes is indicated by solid line. Blots were originally probed with C20 antibody and reprobed for actin. (D) Average relative cell surface expression level (+SD) for WT and GaMD mutant CTF proteins from two to six separate surface biotinylation experiments. (E) Representative Western blots of the surface biotinylation analyses of WT and GaMD mutant proteins from one of the experiments represented in D. Supernatant (Sup) lane is from the pulldown with WT CTF. Images are from two separate blots. Removal of intervening lanes indicated by solid lines. Blots were originally probed with C20 or E8 antibody as indicated and reprobed for actin. *P < 0.05, **P ≤ 0.01, ***P < 0.005, ^#P < 0.0005, and ^##P < 0.0001 relative to WT CTF levels.

Fig. 6.

Model of the Stalk TA-mediated activation signaling of the PC1 CTF. (A) A “Closed” low-energy conformation is identified for the WT CTF, in which signal is transduced through an allosteric pathway connecting the Stalk–TOP–PL domains and ultimately to the C-tail for G protein activation. (B) Stalk variants (including ΔStalk, G3052R, and R3063C) of CTF adopt a distinct “Open” conformation, which exhibits reduced residue interactions between the Stalk–TOP domains and broken salt bridge interaction of residues R3848 and E4078 between the TOP–PL domains, leading to significantly reduced signaling activity of these PC1 mutants.