Prolonged signaling of backbone-modified glucagon-like peptide- 1 analogues with diverse receptor trafficking

Abstract

Signal duration and subcellular location are emerging as important facets of G protein-coupled receptor (GPCR) function. The glucagon-like peptide-1 receptor (GLP-1R), a clinically relevant class B1 GPCR, stimulates production of the second messenger cyclic adenosine monophosphate (cAMP) upon activation by the native hormone, GLP-1. cAMP production continues after the hormone-receptor complex has been internalized via endocytosis. Here, we report GLP-1 analogues that induce prolonged signaling relative to GLP-1. A single β-amino acid substitution at position 18, with the residue derived from (S,S)-trans-2-aminocyclopentanecarboxylic acid (ACPC), enhances signaling duration with retention of receptor endocytosis. Pairing ACPC at position 18 with a second substitution, α-aminoisobutyric acid (Aib) at position 16, abrogates endocytosis, but prolonged signaling is maintained. Prolonged signaling is sensitive to the structure of the β residue at position 18. Cryoelectron microscopy structures of two GLP-1 analogues bound to the GLP-1R:Gs complex suggest substantial alterations to bound peptide structure and dynamics compared to the GLP-1:GLP-1R:Gs complex. These structural findings strengthen an emerging view that agonist dynamics in the receptor-bound state influence signaling profiles. Our results advance understanding of the structural underpinnings of receptor activation and introduce tools for exploring the impact of spatiotemporal signaling profiles following GLP-1R activation.

Keywords: GLP-1; cryo-EM; dynamics; peptide; trafficking.

Fig. 1.

cAMP washout experiments with GLP-1 analogues. (A) Top: Sequences of GLP-1 and analogues. Select residue numbers are shown above. Bottom: Structures of non-native residues. β³-residues share the same sidechain (depicted as R) as the native residue indicated inside the blue circle. (B) Depiction of the washout assay protocol. HEK293 GS22 cells transiently expressing hGLP-1R were challenged with agonist for 15 min (prewash), washed, and then luminescence was monitored for 4 h (postwash). (C) Time-course postwash cAMP measurements for selected agonists (data for all agonists are shown in SI Appendix, Fig. S1). The time courses for GLP-1 are shown in black. Peptide 1 (Left) is shown in red and peptide 2 (Right) is shown in orange. (D) AUC values for 0.1 nM agonist prewash. (E) AUC values for 0.1 nM agonist postwash. n = 3. Washout assay results where data were not normalized to GLP-1 (100%), for which statistical tests were performed (* indicates P < 0.05, one-way ANOVA with Dunnett’s posttest), are available in SI Appendix, Fig. S1 and Table S1. Error bars are SEM.

Fig. 2.

Characterization of GLP-1 analogues. (A) Single time point (~15 min after ligand addition) maximal, concentration–response cAMP production in HEK293 GS22 cells transiently transfected with hGLP-1R. (B) Concentration–response curves showing recruitment of β-arrestin-1 to GLP-1RLuc8 in transiently transfected HEK239FT cells (~45 min postligand addition). (C) Competition binding measurements with HEK293A cells transiently transfected with nLuc-GLP-1R. 10 nM Ex4Rox was used as a tracer. n = 4. (D) Receptor internalization as measured by reduction in maximal luminescence in HEK293 cells expressing nLuc-GLP-1R. 0% internalization was determined by cells treated with vehicle. Error bars indicate SD. n = 3 except for peptides 2 and 3 which are n = 2. Internalization assay data that have not been normalized to GLP-1 (100%), for which statistical tests were performed, are available in SI Appendix, Table S1. (E) Sequence of 6-carboxytetramethylrhodamine (TMR)-labeled GLP-1 analogues. (F) Representative (n = 2) confocal microscopy images of HEK293 GS22 cells transiently expressing GLP-1R-eGFP. Cells were treated with 10 nM agonist for 30 min before washing and paraformaldehyde fixation. Top: Cells treated with 10 nM GLP-1-TMR Bottom: Cells treated with 10 nM 1-TMR. From Left to Right: Nuclear staining with Hoechst, GLP-1R-eGFP, TMR-labeled peptide, and overlay of the three channels. The scale bars indicate 10 µm. (G) Desensitization assay where HEK293 GS22 cells transiently transfected with hGLP-1R were treated with the indicated peptide for 15 min, a washout step was performed, the cells were incubated for 4 h, and then the cells were rechallenged with 100 nM GLP-1. Values indicate percent loss in cAMP signal compared to vehicle treatment. (H) Desensitization vs. Postwash AUC values (from Fig. 1E). The line indicates a linear regression, and the P value was calculated from an F test. Error bars represent SEM unless otherwise stated.

Fig. 3.

Cryo-EM analysis of GLP-1 complexes including peptides 1 and 2. (A) EM density maps of peptides 1 and 2 bound to GLP-1R/Gs complexes. The transparent map with a black silhouette is contoured at low threshold and colored maps are contoured at higher threshold. (B) A comparison of the models of the GLP-1R bound to GLP-1 (PDB: 6X18), peptide 1, peptide 2, (conformer 1) and peptide 2 (conformer 2) shown in gray, red, orange, and blue, respectively. Peptide ligands were removed for clarity. The extracellular domains for complexes containing peptide 1 and peptide 2 (conformer 2) were not modeled due to low local resolution. (C) The crystal structure of the GLP-1R extracellular domain (PDB: 3IOL) rigidly fit (using Chimera 1.14) to each receptor-focused cryo-EM density maps indicated by each color. The aligned EM map of peptide 2 conformer 1 is shown in silhouette for reference. (D) The structure of peptide 2 bound to GLP-1R in conformer 1.

Fig. 4.

Molecular dynamics simulations of peptide 1 and peptide 2 initiated from a GLP-1-like conformation. (A) Per-residue RMSF difference between the GLP-1R bound to peptide 2 or peptide 1, plotted on the structures of the GLP-1R: Residues in red were more flexible during MD simulations of peptide 1, while green residues were more flexible during MD simulations of peptide 2. (B) Position of the ECL3 centroid (cyan points) plotted onto the GLP-1R (white transparent ribbon) in complex with peptide 1 (purple ribbon) or peptide 2 (orange ribbon), every nanosecond of MD simulation. The distance between T11 and ECL3 is shown as a dashed red line.