Conformational biosensors reveal GPCR signalling from endosomes

Abstract

A long-held tenet of molecular pharmacology is that canonical signal transduction mediated by G-protein-coupled receptor (GPCR) coupling to heterotrimeric G proteins is confined to the plasma membrane. Evidence supporting this traditional view is based on analytical methods that provide limited or no subcellular resolution. It has been subsequently proposed that signalling by internalized GPCRs is restricted to G-protein-independent mechanisms such as scaffolding by arrestins, or GPCR activation elicits a discrete form of persistent G protein signalling, or that internalized GPCRs can indeed contribute to the acute G-protein-mediated response. Evidence supporting these various latter hypotheses is indirect or subject to alternative interpretation, and it remains unknown if endosome-localized GPCRs are even present in an active form. Here we describe the application of conformation-specific single-domain antibodies (nanobodies) to directly probe activation of the β2-adrenoceptor, a prototypical GPCR, and its cognate G protein, Gs (ref. 12), in living mammalian cells. We show that the adrenergic agonist isoprenaline promotes receptor and G protein activation in the plasma membrane as expected, but also in the early endosome membrane, and that internalized receptors contribute to the overall cellular cyclic AMP response within several minutes after agonist application. These findings provide direct support for the hypothesis that canonical GPCR signalling occurs from endosomes as well as the plasma membrane, and suggest a versatile strategy for probing dynamic conformational change in vivo.

Figure 1. Nb80–GFP detects activated β₂-ARs in the plasma membrane and endosomes

a, The main events in β₂-AR cAMP signalling include agonist binding (step 1), conformational activation of the receptor (step 2) that is coupled to conformational activation of G_s (step 3) that produces guanine nucleotide exchange on G_s and subsequent activation of adenylyl cyclase (AC) (step 4). b, Scheme for detecting conformational activation of β₂-AR with Nb80–GFP. c, Representative Nb80-GFP (green) and β₂-AR (red) localization at the indicated time (left) after 10 µM isoprenaline addition (>30 Nb80–GFP positive endosomes per cell observed at 20 min; n = 29 cells, 10 experiments). d, Representative individual Nb80–GFP line scans (shown at the same magnification as panel c). e, Representative Nb80–GFP (green) and β₂-AR-3S (red) localization after 20 min of isoprenaline treatment (top) followed by reversal with 50 µM CGP-12177 for the indicated times (6.4 Nb80–GFP positive endosomes per cell; n = 40 cells, 3 experiments). f, Representative individual Nb80–GFP line scans. g, Recovery of cytoplasmic Nb80–GFP fluorescence (mean ± s.e.m., n = 5 experiments). Scale bars, 10 µm.

Figure 2. Nb80–GFP accumulates on β₂-AR-containing endosomes after their formation

a, β₂-AR (red) and Nb80–GFP (green) at the indicated times after isoprenaline addition. Scale bar, 10 µm. b, Average Nb80–GFP fluorescence measured in the TIRF illumination field at the indicated times (mean ± s.e.m., n = 7 cells). c, TIRF image series showing β₂-AR (red) and Nb80–GFP (green) in sequential frames. d, Kymograph of an individual β₂-AR-containing endosome (red, Alexa555) showing Nb80–GFP (green) acquisition over 4 min. e, Kymograph of β₂-AR (green, Alexa488) and Nb80–mRuby (red) over 6 min.

Figure 3. Nb80-GFP does not accumulate in clathrin-coated pits

a, Representative TIRF microscopy frames showing Nb80–GFP (green) and β-arrestin-2–mCherry (red) before and after agonist addition. Fluorescence intensity profiles are shown below for the indicated region and path; representative of n = 39 cells, 5 experiments and 4,849 punctae. b, Equivalent analysis comparing Nb80–GFP (green) to clathrin light chain-dsRed (red); representative of n = 26 cells, 3 experiments and 3,965 punctae. Arrowheads indicate examples of Nb80-GFP labeled endosomes. Scale bars, 5 µm. c, Model for two phases of Nb80–GFP recruitment by the activated β₂-AR, first at the plasma membrane and then at endosomes.

Figure 4. Internalized β₂-ARs contribute to the acute cAMP response

a, Scheme for detecting conformational activation of G_s with Nb37–GFP. b, Confocal image frames showing Nb37–GFP (green) and β₂-AR (red) at the indicated time points after isoprenaline addition (representative of n = 14 cells; estimated Nb37–GFP recruitment delay ranged from 0.7 to 2.65 min). Scale bar, 5µm. Yellow arrowhead indicates Nb37-GFP recruitment to the PM. White arrowhead indicates an endosome containing recently internalized β₂-AR and not associated with Nb37-GFP, white arrow indicates Nb37–GFP recruitment. c, Representative confocal frames showing a discrete endosomal structure labelled with Nb37–GFP (green) and β₂-AR (red) at higher magnification. Scale bar, 2 µm. Arrowhead indicates non-uniform localization of Nb37-GFP to the endosome. d, Forskolin-normalized β₂-AR-mediated cAMP response in the absence or presence of 30 µM Dyngo-4a (mean ± s.e.m., n = 10 experiments, P values in grey). e, Isoprenaline (20 min)-induced β₂-AR and β₂-AR-3S internalization measured by flow cytometry (n = 4 experiments). f, Forskolin-normalized β₂-AR-3S-mediated cAMP response in the absence or presence of 30 µM Dyngo-4a (n = 8 experiments; P = 0.1192). g, Model for two phases of β₂-AR G_s activation, first at the plasma membrane and then at endosomes, separated by the endocytic event.