The evolution and mechanism of GPCR proton sensing

Abstract

Of the 800 G protein-coupled receptors (GPCRs) in humans, only three (GPR4, GPR65, and GPR68) regulate signaling in acidified microenvironments by sensing protons (H⁺). How these receptors have uniquely obtained this ability is unknown. Here, we show these receptors evolved the capability to sense H⁺ signals by acquiring buried acidic residues. Using our informatics platform pHinder, we identified a triad of buried acidic residues shared by all three receptors, a feature distinct from all other human GPCRs. Phylogenetic analysis shows the triad emerged in GPR65, the immediate ancestor of GPR4 and GPR68. To understand the evolutionary and mechanistic importance of these triad residues, we developed deep variant profiling, a yeast-based technology that utilizes high-throughput CRISPR to build and profile large libraries of GPCR variants. Using deep variant profiling and GPCR assays in HEK293 cells, we assessed the pH-sensing contributions of each triad residue in all three receptors. As predicted by our calculations, most triad mutations had profound effects consistent with direct regulation of receptor pH sensing. In addition, we found that an allosteric modulator of many class A GPCRs, Na⁺, synergistically regulated pH sensing by maintaining the pK_a values of triad residues within the physiologically relevant pH range. As such, we show that all three receptors function as coincidence detectors of H⁺ and Na⁺. Taken together, these findings elucidate the molecular evolution and long-sought mechanism of GPR4, GPR65, and GPR68 pH sensing and provide pH-insensitive variants that should be valuable for assessing the therapeutic potential and (patho)physiological importance of GPCR pH sensing.

Keywords: G protein–coupled receptor; allosteric modulator; coincidence detection; evolution; proton; proton sensing; sodium.

Figure 1

GPR4, GPR65, and GPR68 share a triad of buried acidic residues.A, homology model of GPR65 showing the triad of buried acidic residues that originate from transmembrane helix 4 (TM4) (a Glu residue we call the apEx site) and TM2 and TM7 (a pair of Asp residues we call the DyaD site). B–D, Homology models of GPR65 (B), GPR4 (C), and GPR68 (D) showing the location of the buried acidic triad and 26 reported sites (rSites) of all GPR4, GPR65, and GPR68 mutations available in the literature. An illustration of the pHinder algorithm for identifying networks of buried ionizable residues is available in Fig. S1.

Figure 2

Deep variant profiling of GPR4, GPR65, and GPR68 at pH 5.0 and 7.0.A, the DCyFIR strain platform and DVP technology for rapidly screening human GPCRs and their mutants in yeast (42, 43). Each DCyFIR strain consists of a human GPCR and a humanized Gα chimera (hGα). GPCR activation and subsequent hGα coupling result in expression of the fluorescent transcriptional reporter, mTurquoise2. DVP is used to interrogate the functional significance of GPCR residues. Data from GPR68 (WT), GPR68–H17F (rSite), and GPR68–D282N (triad) were used here to demonstrate the approach. B, DVP of 390 DCyFIR strains at pH 5.0. GPCR signaling in each mutant strain was quantified relative to the WT by calculating the log₂ fold change (FC) in mTq2 fluorescence (n = 2). Results were scored as an increase in signaling (cyan; log₂ FC > 0.5), no change in signaling (black; log₂ FC ± 0.5), decrease in signaling (gray; log₂ FC < −0.5), and no signaling or Gα coupling (white). C, left, DVP of the 195 functional DCyFIR-strain mutants at pH 7.0. Results are shown for the 9 triad sites and the only 3 His to Phe rSite mutations that exhibited increased signaling at pH 7.0. Coloring is based on log₂ FC in mTq2 fluorescence (n = 2) and scored as an increase in signaling (cyan; log₂ FC > 2.0 for GPR4 and GPR68; log₂ FC > 0.5 for GPR65), no change in signaling (black; log₂ FC ± 2.0 for GPR4 and GPR68; log₂ FC ± 0.5 for GPR65), decrease in signaling (gray; log₂ FC < −2.0 for GPR4 and GPR68; log₂ FC < −0.5 for GPR65), or no detectable signaling or Gα coupling (white). C, right, Summed log₂ FC values for the DCyFIR strains of each mutant in the left panel. Error bars represent SD (n = 2–12). D, top, pH profiles of WT GPR4 (open circles), GPR65 (open squares), and GPR68 (closed circles) measured using the Gα_i DCyFIR strain. Data are the mean ± SD (n = 4). D, bottom, Based on our predictions, mutation of Asp and Glu triad residues to permanently neutral Asn or Gln side chains should cause right-shifted midpoints of proton activation (pH₅₀ values) and greater signaling as indicated by increased mTq2 fluorescence at and above pH 7.0. Primary data, numerical details, and experimental errors for the data in panels B and C can be found in Figures S2 and S3. DCyFIR, Dynamic Cyan Induction by Functional Integrated Receptors; DVP, deep variant profiling.

Figure 3

Validating the mechanism of GPR4, GPR65, and GPR68 proton sensing in HEK293 cells.A, the BRET-based mini G protein (mGp) assay used in HEK293 cells. Coupling of a Venus-tagged mGp (V-mG) to a luciferase-tagged GPCR (GPCR–Rluc8) produces a BRET signal (left) that can be used as a direct readout of GPCR pH activation profiles, as shown for the representative pH profile of WT GPR4 (right). B, relative pH₅₀ changes (ΔpH₅₀) of 24 GPR4, GPR65, and GPR68 triad (pink) and rSite (gray) variants and His to Phe variants at sites shared by all 3 GPCRs (black, solid, or striped). Variants indicated by an asterisk correspond to both rSite and conserved His sites. White bars (solid or striped) indicate nonfunctional variants (e.g., GPR68–D67N and –H245F) or variants with pH₅₀ values below the measured pH range (e.g., GPR68–H20F and –H169F). Data are the mean ± SEM of the net BRET signal (netBRET) (n = 3–11). C, pH profiles of select triad variants (pink) that exhibited upshifted pH₅₀ values and nullified pH sensing within the physiologic pH range (gray shading). Black curves correspond to the pH profiles of WT receptors. Inset values are the pH₅₀ values (mean ± SEM) for each triad variant and its corresponding WT receptor. D, interpretive model of triad (de)protonation and GPCR activation as a function of pH. The net charge and sequence of titration steps is only intended to illustrate the combination of (de)protonation events that underlie the pH-sensing mechanism. E, overlaid pH profiles and structural mapping of pH-insensitive variants shown in panel C (GPR4–E145Q, squares, dotted line; GPR65–D286N, triangles, dashed line; and GPR68–E149Q, circles, solid line). Data were normalized by adjusting the smallest and largest netBRET values for each variant to 0 and 100, respectively. Gray shading indicates the physiologically relevant range for GPCR proton sensing. Data shown in panels A–C are the mean ± SEM (n = 5–9). Mini G protein Venus–mGsi (V-mGsi) was used for all data in panels A–C to be consistent with our yeast titration data in Figure 2D. Thirty three additional pH profiles related to panel B can be found in Figure S4, B–D. BRET, bioluminescence resonance energy transfer; GPCR, G protein–coupled receptor.

Figure 4

Na⁺is an allosteric modulator of GPR4, GPR65, and GPR68 proton sensing.A–C, pH profiles in the presence and absence of 150-mM K⁺ or Na⁺ for the 3 WT pH sensors (A), their DyaD site Asp to Asn double mutants (B), and Glu to Gln apEx site single mutants (C). Data were collected using mini G protein V-mGsi and are the mean ± SEM of the net BRET signal (n = 6–12). For these experiments, slight increases in pH₅₀ values were observed as compared with those in Figure 3C, which we attributed to the different buffers required for each experiment (see Experimental procedures for details). D, interpretive model of triad (de)protonation, GPCR activation, H⁺ and Na⁺ coincidence detection, and Na⁺ negative allosteric modulation as a function of extracellular pH. The net charge and sequence of titration steps and Na⁺ binding are only intended to illustrate the combination of (de)protonation and Na⁺ binding events that underlie the pH-sensing mechanism. BRET, bioluminescence resonance energy transfer; GPCR, G protein–coupled receptor.

Figure 5

The molecular evolution of GPR4, GPR65, and GPR68 proton sensing.A, ranked number of extracellular His residues in 416 human GPCR structures, represented by 70 nonredundant GPCR genes, and the 4 GPR4, GPR65, GPR68, and GPR132 homology models. B, pH profiles of GPR4, GPR65, GPR68, and GPR132 in HEK293 cells measured using the mGp assay. C, phylogenetic analysis of 373 nonolfactory GPCRs illustrating the emergence of the buried acidic triad in GPR65 and additional eHis residues in GPR4 and GPR68. The DyaD in GPR132 is an EyaD because the conserved D^2.50 is E^2.50. D, pH profiles of WT GPR65 (solid triangles) and mutations of the conserved eHis residues GPR65–H14F (H^1.28) and GPR65–H243F (H^6.52) (open triangles). Data shown in panels B and D were collected using mini G protein V-mGsi and are the mean ± SEM (n = 4–9). GPCR, G protein–coupled receptor. eHis, extracellular His; mGp, mini G protein; GPCR, G protein–coupled receptor.