Concerted conformational changes control metabotropic glutamate receptor activity

Abstract

Allosteric modulators bear great potential to fine-tune neurotransmitter action. Promising targets are metabotropic glutamate (mGlu) receptors, which are associated with numerous brain diseases. Orthosteric and allosteric ligands act in synergy to control the activity of these multidomain dimeric GPCRs. Here, we analyzed the effect of such molecules on the concerted conformational changes of full-length mGlu2 at the single-molecule level. We first established FRET sensors through genetic code expansion combined with click chemistry to monitor conformational changes on live cells. We then used single-molecule FRET and show that orthosteric agonist binding leads to the stabilization of most of the glutamate binding domains in their closed state, while the reorientation of the dimer into the active state remains partial. Allosteric modulators, interacting with the transmembrane domain, are required to stabilize the fully reoriented active dimer. These results illustrate how concerted conformational changes within multidomain proteins control their activity, and how these are modulated by allosteric ligands.

Fig. 1.. Site-specific fluorescence labeling of mGlu receptors by genetic code expansion and bioorthogonal click chemistry.

(A) Resting open-open (R_OO) and active closed-closed (A_CC) structures of mGlu2 [Protein Data Bank (PDB) 7EPA and 7E9G]. The major domains are annotated including the upper and lower lobes of the VFT, the CRD, and the 7TM. The upper and lower lobes are highlighted in gray and agonist binding to the orthosteric site is shown as black spheres. (B) Schematic representation of genetic code expansion and bioorthogonal click chemistry. Coexpression of the orthogonal synthetase together with its cognate amber suppressor tRNA_CUA (otRNA/RS) in the presence of the noncanonical amino acid propargyloxy-l-phenylalanine (PrF) in HEK293T cells allows the reassignment of a premature Amber codon TAG within the ORF of the receptor gene to code for PrF. This leads to incorporation of PrF at a predefined position in the translated receptor, which subsequently reacts with a picolyl azide (pAz) dye using copper(I)-catalyzed azide-alkyne cycloaddition (CuAAC). (C and D) Ligand dose–dependent IP-1 production of SNAP-mGlu2 and SNAP-mGlu2-PrF248, respectively. IP-1 accumulation was mediated by coexpression of chimeric G_iq9. (E) Correlation of log EC₅₀ values obtained for SNAP-mGlu2 (C) and SNAP-mGlu2-PrF248 (D). (F) Schematic representation of N-terminal SNAP-labeling followed by click labeling of the lower lobe (A248) to obtain the FRET sensor used in (G). (G) LRET imaging of mGlu2 labeled with the Lumi4Tb-BG donor at N-terminal SNAP-tags and at position 248 in the lower lobe of the VFT with the "green" pAz acceptor at the plasma membrane of live HEK293T cells. Data in (C) and (D) represent the mean of 3 to 12 measurements ± SD.

Fig. 2.. Establishment of a sensor to monitor VFT closure in a single protomer.

(A) Ligand dose–dependent LRET changes between the upper lobes, labeled via SNAP-tags with Lumi4-Tb donor, and the lower lobes, labeled through PrF248 with pAz green acceptor. Both protomers in the dimer are labeled, giving rise to a maximum of two donor and two acceptor dyes per receptor. (B) Correlation of log EC₅₀ values obtained from LRET reporting on VFT closure (A) and downstream signaling given by IP-1 production (see Fig. 1D). (C) Comparison of dose-response behavior on VFT closure measured by LRET for SNAP-mGlu2-PrF248 homodimers and specifically engineered heterodimers composed of one unlabeled (mGlu2-C2) and one labeled (SNAP-mGlu2-PrF248-C1) protomers (see the main text). (D) Ligand dose–dependent IP-1 production of SNAP-mGlu2-PrF248 + 358. IP-1 accumulation was mediated by coexpression of chimeric G_iq9. (E) Correlation of log EC₅₀ values obtained for SNAP-mGlu2-PrF248 + 358 (D) and SNAP-mGlu2 (see Fig. 1C). (F) Dose-dependent increase in LRET as a result of VFT closure in response to LY379268 for the engineered heterodimer, in which only one VFT protomer is labeled on PrF248 + 358. Data in (A) to (F) represent the mean of 3 to 12 measurements ± SD. WT, wild type.

Fig. 3.. Agonists stabilize the VFT in its closed state.

(A) Side view on structures of the VFT in the inactive resting open/open (PDB 7EPA) and active closed/closed (PDB 7E9G) conformation with residues R358 in the upper lobe and A248 in the lower lobe of one VFT replaced by PrF (red spheres). (B to G) FRET histograms of VFT closure sensor in the absence of ligand (B, Apo) and in the presence of saturating 100 μM LY341495 (C), 1 mM DCG-IV (D), 1 mM DCG-IV + 1 μM BINA (E), 10 mM glutamate (F), and 10 mM glutamate +1 μM BINA (G). FRET histograms in (B) to (G) show the accurate FRET efficiency as the mean ± SEM of three independent biological replicates each normalized to 2000 events in the DA population (S = 0.3 to 0.7). Histograms display the fitting with three gaussians (white = very low FRET, yellow = low FRET, brown = high FRET) together with the global fit (black). The mean FRET efficiency of the Apo condition (A) in comparison to the antagonist LY341495 (B) is indicated by a dashed line. (H) Measured fraction of molecules in the closed state. Shown are the means ± SD of three independent biological replicates obtained from the number of molecules found in the high FRET population (brown) over the sum of all molecules found in the low FRET (yellow) and high FRET (brown) populations. Black circles show the Apo condition and orthosteric ligands alone, while blue squares are given for agonists in the presence of 1 μM BINA. Statistical differences were determined using a one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test and are given as *P ≤ 0.01; not significant (ns) > 0.05.

Fig. 4.. The equilibrium between resting and active upper lobe conformation follows agonist efficacy and is further promoted by allosteric modulations.

(A) Side view on structures of the VFT in the inactive resting open/open (PDB ID 7EPA) and active closed/closed (PDB ID 7E9G) conformation with residues R358 replaced by PrF (red spheres). (B to G) FRET histograms of upper lobe sensor in the absence of ligand (B, Apo) and in the presence of saturating 100 μM LY341495 (C), 1 mM DCG-IV (D), 10 mM glutamate (E), 1 mM DCG-IV + 10 μM BINA (F), and 10 mM glutamate + 10 μM BINA (G). FRET histograms in (B) to (G) show the accurate FRET efficiency as the mean ± SEM of three independent biological replicates each normalized to 2000 events in the DA population (S = 0.3 to 0.7). Histograms display the fitting with 4 gaussians (white = very low FRET, yellow = low FRET, brown = high FRET, white = very high FRET) together with the global fit (black). The mean FRET efficiency of the Apo condition (A) in comparison to the antagonist LY341495 (B) is indicated by a dashed line. (H) Measured fraction of molecules in the fully reoriented state. Shown are the means ± SD of three independent biological replicates obtained from the number of molecules found in the high FRET population (brown) over the sum of all molecules found in the low FRET (yellow) and high FRET (brown) populations. Black circles show the Apo condition and orthosteric ligands alone, while blue squares are given for agonists in the presence of 10 μM BINA. Statistical differences were determined using a one-way ANOVA with Tukey’s multiple comparisons test and are given as ****P ≤ 0.0001, ***P ≤ 0.001, ns > 0.05.

Fig. 5.. The lower lobe sensor reveals the agonist-dependent equilibrium between two distinct resting and one active state.

(A) Bottom view on structures of the VFT in the inactive resting open/open (PDB ID 7EPA) and active closed/closed (PDB ID 7E9G) conformations with residues A248 replaced by PrF (red spheres). (B to I) FRET histograms of lower lobe sensor in the absence of ligand (Apo, B) and in the presence of saturating 100 μM LY341495 (C), 1 mM DCG-IV (D), 10 mM glutamate (E), 100 μM LY379268 (F), 1 mM DCG-IV + 10 μM BINA (G), 10 mM glutamate +10 μM BINA (H), and 100 μM LY379268 + 10 μM BINA (I). FRET histograms show the accurate FRET efficiency (mean ± SEM) of three independent biological replicates, normalized to 2000 events in the DA population (S = 0.3 to 0.7). Histograms display the fitting with four gaussians (white = very low FRET, yellow = low FRET, brown = high FRET, white = very high FRET) together with the global fit (black). (J) Measured fraction of molecules in the fully reoriented/active state, as the means ± SD of three independent biological replicates obtained from the number of molecules found in the high FRET population (brown) over the sum of all molecules. Black circles, the Apo condition and orthosteric ligands alone; blue squares, agonists in the presence of 10 μM BINA. Statistical differences were determined using a one-way ANOVA with Tukey’s multiple comparisons test, given as ***P ≤ 0.001, **P ≤ 0.01, *P ≤ 0.05, ns > 0.05. (K) Correlation of the fraction of molecules in the active state obtained from the upper lobe sensor (see Fig. 4H) and the lower lobe sensor (J). (L) Mean FRET efficiency of the low FRET population at different ligand conditions [yellow gaussian in (B) to (F)].

Fig. 6.. Three-state model of mGlu2 activation.

(A) Correlation of VFT closure and lower lobe reorientation illustrates that agonists binding leads to an equilibrium between open and closed VFT, with full agonists favoring the closed state (black circles). In contrast, a maximal stabilization of the reoriented state requires the synergistic action of the PAM (BINA, blue circles). Shown are the normalized means (VFT closure: min = apo, max = LY379268; reorientation: min = apo, max = LY379268 + BINA) ± SD of three independent biological replicates. (B) The initial step of mGlu2 activation can be described by a three-state equilibrium between the R_oo (resting/open-open), the R_cc (resting/closed-closed) and the A_cc (active/closed-closed) states. Agonists activate the receptor through establishment of the R_oo to R_cc equilibrium, where agonists populated the R_cc state in a manner reflecting their pharmacological efficacies. In the absence of allosteric modulators, the VFTs then readily visit the A_cc state. PAMs exclusively act on the R_cc to A_cc equilibrium by increasing the residence time of molecules in the A_cc state. Antagonists induce a slight closure of the VFTs together with a small increase in separation of the lower lobes (R_oo*). The arrows on the cartoons indicate the degree of FRET. The boxes indicating the different states are colored according to the respective FRET states shown in Figs. 3 to 5.