Ligand efficacy modulates conformational dynamics of the µ-opioid receptor

µOR expression and purification

The wild-type Mus musculus µOR (6-398) with an N-terminal HA signal sequence followed by a Flag tag and a C-terminal 8×His tag was cloned in the pFastBac1 vector. The minimal-cysteine construct (µOR∆7) was created by introducing the mutations⁵¹ C13S, C22S, C43S, C57S, C170T, C346A and C351L into the wild-type µOR. Double-cysteine mutation constructs (µOR∆7(R182C/R276C) for DEER, µOR∆7(T180C/R276C) and µOR∆7(R182C/R273C) for smFRET experiments) were generated based on the µOR∆7 construct. The µOR was expressed and purified following a previous protocol¹³ with some modifications. The µOR was expressed in Sf9 insect cells (Expression Systems, authenticated by supplier, not tested for mycoplasma) using Bac-to-Bac baculovirus systems with 10 µM naloxone. Cells were collected 48 h post infection and were lysed in a buffer of 10 mM Tris pH 7.5, 1 mM EDTA, 100 µM TCEP, 10 µM naloxone, 160 µg ml⁻¹ benzamidine and 2.5 µg ml⁻¹ leupeptin. The receptor was extracted from the Sf9 membrane using buffer of 20 mM HEPES pH 7.5, 500 mM NaCl, 0.7% N-dodecyl-β-d-maltoside (DDM), 0.3% CHAPS, 0.03% cholesteryl hemisuccinate (CHS), 30% (v/v) glycerol, 5 mM imidazole, 2 mM MgCl₂, 160 µg ml⁻¹ benzamidine, 2.5 µg ml⁻¹ leupeptin, 10 µM naloxone, 100 µM TCEP and 2 µl benzonase in the cold room for 1 h. After centrifugation, Ni-NTA resin was added to the supernatant in a 500-ml centrifuge tube (Corning) and rotated for 2 h at 4 °C. Ni-NTA resin was washed in batch with washing buffer of 20 mM HEPES pH 7.5, 500 mM NaCl, 0.1% DDM, 0.03% CHAPS, 0.03% CHS, 5 mM imidazole and 10 µM naloxone and protein was eluted in washing buffer supplemented with 250 mM imidazole. Ni-NTA eluate was supplemented with 2 mM CaCl₂ and loaded onto anti-Flag M1 resin (Millipore-Sigma) for further purification. The detergent was exchanged to LMNG on a Flag column by gradually increasing the proportion of the exchange buffer (20 mM HEPES pH 7.5, 100 mM NaCl, 0.5 LMNG, 0.05% CHS, 2 mM CaCl₂ and 10 µM naloxone) over the Ni-NTA washing buffer supplemented with 2 mM CaCl₂ at room temperature for 1 h. The µOR was finally eluted with buffer of 20 mM HEPES pH 7.5, 100 mM NaCl, 0.01% LMNG, 0.001% CHS, 5 mM EDTA, 0.2 mg ml⁻¹ Flag peptide and 10 µM naloxone. After concentrating with a 4-ml 100-kDa cutoff concentrator (Amicon Ultra), the µOR was further purified by size-exclusion chromatography (SEC) using an SD200 increase 10/300 column (GE Healthcare) equilibrated with SEC buffer of 20 mM HEPES pH 7.5, 100 mM NaCl, 0.01% LMNG, 0.001% CHS and 10 µM naloxone. Fractions containing monomeric µOR were collected and concentrated with a 500-µl 100-kDa cutoff concentrator (Amicon Ultra). The µOR was supplemented with 15% (v/v) glycerol and flash frozen in liquid nitrogen.

G_i heterotrimer expression and purification

DNA for the human Gα_i1 was cloned into the pFastBac1 vector. DNA of human Gβ₁ with an N-terminal 6×His tag and HRV 3C protease cleavage site (LEVLFQGP) and Gγ₂ were cloned into the vector of pFastBac Dual under the promoter of ph and p10, respectively. P2 viruses of Gα_i1 and Gβ₁γ₂ were generated following the same protocol for the µOR. G_i1 heterotrimer was expressed in Hi5 cells (Expression Systems, authenticated by supplier, not tested for mycoplasma) with 4 ml P2 of Gα_i1 and 10 ml P2 of Gβ₁γ₂ per liter cells when cells reached a density of 3 million per ml. Cells were collected 48 h post infection and kept in −80 °C freezer until use.

Cell pellets were thawed in lysis buffer (10 mM Tris pH 7.5, 1 mM MgCl₂, 5 mM β-mercaptoethanol (β-ME), 10 µM GDP, 160 µg ml⁻¹ benzamidine, 2.5 µg ml⁻¹ leupeptin). After centrifugation, pellets were solubilized in solubilization buffer (20 mM HEPES pH 7.5, 100 mM NaCl, 1% sodium cholate, 0.05% LMNG, 5 mM MgCl₂, 20 mM imidazole, 5 mM β-ME, 10 µM GDP, 160 µg ml⁻¹ benzamidine, 2.5 µg ml⁻¹ leupeptin) and were stirred in a cold room for 1 h. After centrifugation at 14,000 rpm for 20 min, the supernatant was mixed with Ni-NTA resin and rotated at 4 °C for 1 h. Ni-NTA resin was then washed four times in batch with solubilization buffer. Detergent was exchanged to LMNG on the Ni-NTA column by gradually increasing LMNG concentration at room temperature. Protein was eluted with elution buffer (20 mM HEPES pH 7.5, 50 mM NaCl, 0.01% LMNG, 2 mM MgCl₂, 5 mM β-ME, 10 µM GDP, 180 mM imidazole). The His tag was cleaved by 1:50 (w/w) HRC 3 C protease. G_i1 was treated with 5 µl of λ protein phosphatase and was dialysed against dialysis buffer (20 mM HEPES pH 7.5, 50 mM NaCl, 0.01% LMNG, 2 mM MgCl₂, 2 mM MnCl₂, 5 mM β-ME, 10 µM GDP) overnight at 4 °C to remove imidazole. The His tag and contaminates were removed by loading G_i1 onto 2-ml Ni-NTA resin. Flow-through of Ni-NTA resin was loaded onto a MonoQ column and G_i1 was further purified by anion exchange. The G_i1 heterotrimer peak was collected and concentrated. After being supplemented with 15% glycerol, G_i1 was flash froze and kept in −80 °C freezer. For DEER samples, ion-exchange purified G_i1 was further injected onto an SD200 increase 10/300 column (GE Healthcare) equilibrated with SEC buffer (20 mM HEPES pH 7.5, 100 mM NaCl, 0.01% LMNG, 2 mM MgCl₂ and 10 µM GDP). SEC fractions were pooled, concentrated to 336 µM and flash frozen.

GRK5 expression and purification

Human GRK5 DNA with a C-terminal 6×His tag was cloned into pFastBac1 vector. P2 virus was generated following the same protocol of the µOR. GRK5 was expressed in Sf9 insect cells with 25 ml of P2 virus and was collected 48 h after infection. Purification of GRK5 was performed on ice or at 4 °C. Cells were lysed in lysis buffer (20 mM HEPES pH 7.5, 150 mM NaCl, 20 mM imidazole, 5 mM β-ME, 160 µg ml⁻¹ benzamidine, 2.5 µg ml⁻¹ leupeptin) by sonication on ice. Cell debris was removed by centrifuge at 14,000 rpm for 20 min. GRK5 in supernatant was purified by Ni-NTA resin using wash buffer (20 mM HEPES pH 7.5, 150 mM NaCl, 20 mM imidazole, 5 mM β-ME). Protein was eluted in wash buffer supplemented with 160 mM imidazole. GRK5 was concentrated and injected in an SD200 increase 10/300 column equilibrated with cold SEC buffer (20 mM HEPES pH 7.5, 300 mM NaCl) in cold room. SEC fractions of GRK5 were pooled, concentrated and flash frozen.

β-Arrestin-1 expression and purification

To investigate the conformational changes of the µOR in the presence of β-arrestin-1, a C-terminal truncated β-arrestin-1 was used for smFRET and DEER measurements. The long splice variant of human, cysteine-free (C59V/C125S/C140L/C150V/C242V/C251V/C269S), truncated β-arrestin-1 (1-382) (βarr1(∆CT))⁵² with an N-terminal 6×His and HRV 3 C site was in vector of pET15b and was transformed into BL21 (DE3) competent cells. Escherichia coli cells were cultured in TB medium with 100 µg ml⁻¹ ampicillin until OD₆₀₀ reached 1.2 at 37 °C in a shaker at 220 rpm. The temperature was decreased to 18 °C and protein expression was induced with 200 µM IPTG for 16 h. Purification of βarr1(∆CT) was performed on ice or at 4 °C. Cells were collected and sonicated in buffer 1 (20 mM Tris 8.0 (25 °C), 300 mM NaCl, 20 mM imidazole) supplemented with 160 µg ml⁻¹ benzamidine and 2.5 µg ml⁻¹ leupeptin. After centrifugation, protein in the supernatant was incubated with Ni-NTA resin at 4 °C for 1 h. The Ni-NTA resin was extensively washed with buffer 1, then was further washed with 3 column volumes of buffer 2 (20 mM Tris 8.0 (25 °C), 50 mM NaCl and 20 mM imidazole). βarr1(∆CT) was eluted with buffer 2 supplemented with 160 mM imidazole. βarr1(∆CT) was loaded onto a Source 15Q 4.6/100 PE anion-exchange column (GE Healthcare). The column was washed with 2 column volumes of buffer A (20 mM Tris 8.0 (25 °C), 50 mM NaCl), and βarr1(∆CT) was eluted with 15 column volumes of a linear gradient from 0 to 30% buffer B (20 mM Tris 8.0 (25 °C), 1 M NaCl). The peak fractions were pooled and supplemented with NaCl to a final concentration of 300 mM, which prevented the protein from precipitating when concentrated to high concentration in the following step. The protein was concentrated and injected in an SD200 increase 10/300 column equilibrated with SEC buffer of 20 mM HEPES pH 7.5, 300 mM NaCl. For DEER samples, SEC buffer was made in D₂O, and βarr1(∆CT) was concentrated to 986 µM and flash frozen.

Phosphorylation of µOR

The µOR was purified following the standard µOR purification protocol except that the naloxone was replaced with 10 µM DAMGO on the anti-Flag M1 resin and SEC purification procedures. 4 µM of µOR∆7(R182C/R276C) purified in the presence of DAMGO was incubated in phosphorylation buffer of 20 mM HEPES pH 7.5, 35 mM NaCl, 5 mM MgCl₂, 100 µM TCEP, 20 µM 1,2-dioctanoyl-sn-glycero-3-phospho-(1′-myo-inositol-4′,5′-bisphosphate) (C8-PIP2), 0.01% LMNG, 0.001% CHS and 100 µM DAMGO at room temperature for 1 h. ATP and GRK5 were then added to the reaction to a final concentration of 1 mM and 0.8 µM, respectively, and incubated for 1 h before more GRK5 was added. GRK5 was added every 1 h four times in total and the reaction was kept at room temperature.

To evaluate the phosphorylation level and make sure it reaches completion using ion-exchange chromatography, 12 µl of the phosphorylation reaction containing about 50 picomoles of µOR at different time points was removed and diluted to 200 µl using the buffer of 20 mM Tris pH 8.0 (25 °C), 50 mM NaCl, 0.01% LMNG, 5 mM EDTA and 10 µM naloxone. The samples were then injected onto a MonoQ (5/50) anion-exchange column (GE Healthcare) equilibrated with buffer A of 20 mM Tris 8.0 (25 °C), 50 mM NaCl, 0.01% LMNG and 10 µM naloxone. The column was washed with 1 column volumes of buffer A, and then with 40 column volumes of a linear gradient from 0 to 40% buffer B of 20 mM Tris 8.0 (25 °C), 1 M NaCl, 0.01% LMNG and 10 µM naloxone at room temperature. Protein elution was monitored by a fluorescence detector (Shimadzu) with excitation at 280 nm and emission at 340 nm (Extended Data Fig. 12a).

After the 4-h incubation with GRK5, the reaction was diluted by tenfold with the wash buffer of 20 mM HEPES pH 7.5, 100 mM NaCl, 0.01% LMNG, 0.001% CHS, 2 mM CaCl₂ and 10 µM naloxone before loading onto 3 ml M1 resin. The M1 resin was washed with 30 ml of the wash buffer at room temperature for 30 min. The µOR was finally eluted using elution buffer of 20 mM HEPES pH 7.5, 100 mM NaCl, 10 µM naloxone, 5 mM EDTA and 0.2 mg ml⁻¹ Flag peptide. After concentration, the µOR was further injected onto an SD200 increase 10/300 column equilibrated with SEC buffer of 20 mM HEPES pH 7.5, 100 mM NaCl, 0.01% LMNG, 0.001% CHS and 10 µM naloxone. Fractions containing monomeric µOR were collected and concentrated with a 500-µl 100-kDa cutoff concentrator (Amicon Ultra). The µOR was supplemented with 15% (v/v) glycerol and flash frozen in liquid nitrogen.

Fluorophore synthesis

Iodoacetamide-conjugated Cy3 and Cy5 fluorophores were synthesized following a previous protocol³⁰. In brief, 1 µmol of sulfo-Cyanine3 NHS ester or sulfo-Cyanine5 NHS ester (Lumiprobe) was dissolved in 500 μl dry dimethyl sulfoxide (DMSO). It was then added dropwise to a solution of 50 μl cadaverine in 500 μl of dry DMSO at room temperature. The reaction solution was stirred at room temperature for 5 min, then poured into 15 ml of 5% formic acid in ethyl acetate. The precipitate was collected and purified by high-performance liquid chromatography using 10 mM triethylammonium acetate pH 7.0 aqueous buffer (solvent A) with 100% acetonitrile (solvent B) as the mobile phase. The product fraction was dried using a rotary evaporator. The resulting pure fluorophore–cadaverine compound was then dissolved in 1 ml dry DMSO. N,N-diisopropylethylamine (100 μl) was added to this solution, followed by 1 mg iodoacetic acid NHS ester. The reaction solution was stirred at room temperature for 15 min and then poured into 15 ml ethyl acetate. The precipitate was collected and purified by high-performance liquid chromatography.

Synthesis of HO-1427

The bromo derivative⁵³ (261 mg, 1.0 mmol) (HO-559) was dissolved in acetone (20 ml) and NaI (300 mg, 2 mmol) was added. The reaction mixture was refluxed for 1 h then evaporated. The residue was dissolved in ethyl acetate/diethyl ether (50:50, 20 ml) and washed with brine (2 × 10 ml). The organic phase was dried (MgSO₄), filtered, evaporated and purified with flash chormatography (hexane:diethyl ether) yielding yellow crystals 230 mg (74%); melting point: 132–134 °C; retention factor (R_f) = 0.4 (hexane:ethyl acetate 2:1); Elemental analysis calculated for C₁₀H₁₅INO₂ (Mw: 308.1) C: 38.98; H: 4.91; N: 4.55%; measured: C: 39.02; H: 4.78; N: 4.61%; IR (cm⁻¹): 1665, 1615; MS (EI, m/z,%): 308 (8), 294 (6), 278 (6), 151 (100), 136 (8), 109 (52), 43 (61).

The melting point was measured with a Boetius micro melting point apparatus. The infrared (IR) spectrum was obtained using a Bruker Alpha FT-IR instrument with an attenuated total reflectance support on a diamond plate. The mass spectrum was recorded on a Shimadzu GCMS-2020 spectrometer in electron ionization (EI) mode (70 eV). The elemental analysis was performed on a Fisons EA 1110 CHNS instrument. Flash column chromatography was performed on Merck Kieselgel 60 (0.040–0.063 mm) column. Qualitative thin layer chromatography (TLC) was carried out on commercially available plates (20 cm × 20 cm × 0.02 cm) coated with Merck Kieselgel.

µOR labelling with fluorophores

Minimal-cysteine µOR with cysteine mutations on TM4 and TM6, namely µOR∆7(T180C/R276C) and µOR∆7(R182C/R273C), was labelled by commercial maleimide-conjugated sulfo-Cy3 and sulfo-Cy7 (Lumiprobe) or by home-made iodoacetamide-conjugated Cy3 and Cy5, respectively. SEC purified µOR was diluted to 10 µM in 20 µl of labelling buffer (50 mM HEPES pH 7.5, 100 mM NaCl, 0.01% LMNG, 0.001% CHS, 10 µM naloxone). 30 µM of donor fluorophore and 60 µM of acceptor fluorophore were added into the reaction. After incubation at 20 °C for 30 min, free dyes were quenched with 10 mM l-cysteine. The reaction was then loaded onto a home-packed desalt column filled with 2-ml G50 resin (Sigma) equilibrated with the desalt buffer (20 mM HEPES pH 7.5, 100 mM NaCl, 0.01% LMNG, 0.001% CHS, 15% glycerol). Fractions containing µOR were pooled, aliquoted and flash frozen. The concentration of µOR was approximately 500 nM.

µOR labelling with nitroxide spin label

To make samples of the µOR alone or in complex with G protein for DEER studies, SEC purified µOR∆7(R182C/R276C) without phosphorylation was diluted to 20 µM in labelling buffer (20 mM HEPES pH 7.5, 100 mM NaCl, 0.01% LMNG, 0.001% CHS, 10 µM naloxone). Nitroxide spin label reagent HO-1427 was added to a final concentration of 400 µM. After incubation at room temperature for 3 h, the reaction was quenched with 5 mM l-cysteine and was injected into an SD200 increase 10/300 column equilibrated with SEC buffer (20 mM HEPES pH 7.5, 100 mM NaCl, 0.01% LMNG, 0.001% CHS, 2 mM CaCl₂ in D₂O). Fractions of the monodisperse peak were pooled and equally divided into ten 1.5-ml tubes. The protein was diluted fourfold with SEC buffer. Ligands were added to each tube at a final concentration of 1 mM for naloxone, TRV130, PZM21, MP, buprenorphine, and morphine, 400 µM for DAMGO, 200 µM for lofentanil, and 500 µM for BU72. One tube of protein was kept without ligand. The µOR and ligand were incubated at room temperature for 2 h. Protein in each individual tube was concentrated and split into two parts, one of which was mixed with 20% (v/v) D8-glycerol, transferred to a capillary, and flash frozen. The other part was mixed with a threefold molar excess of G_i1, which was purified in D₂O buffer, and incubated for 30 min at room temperature. 1:100 apyrase (v/v, NEB) was added to the G-protein samples to remove free GDP and incubated for 1 h at room temperature. The G-protein samples were then mixed with 20% (v/v) D8-glycerol, transferred to capillaries and flash frozen.

To make samples in complex with βarr1(∆CT) for DEER studies, µORp∆7(R182C/R276C) was labelled with HO-1427 following a similar protocol above. SEC fractions were pooled and equally divided into 10× 1.5-ml tubes. The protein was diluted fourfold with D₂O dilution buffer of 20 mM HEPES pH 7.5, 100 mM NaCl, 0.01% LMNG, 0.001% CHS, 5 µM C8-PIP2, and respective ligand at a final concentration as indicated above. The µOR was incubated with ligand for 2 h at room temperature. Protein was then concentrated, mixed with a fourfold molar excess of βarr1(∆CT) that was in D₂O buffer, and incubated at room temperature for 1 h. The samples were then mixed with 20% (v/v) D8-glycerol, transferred to capillaries and flash frozen.

Single-molecule FRET experiments and analysis

All smFRET experiments were performed at 25 °C following previous protocol with some modifications⁵⁴. In brief, single-molecule FRET studies were performed on a home-built objective-type TIRFM microscope, based on a Nikon Eclipse Ti-E with an EMCCD camera (Andor iXon Ultra 897), and solid-state 532 nm excitation lasers (Coherent Inc. OBIS Smart Lasers). Fluorescence emission from the probes was collected by the microscope and spectrally separated by interference dichroic (T635lpxr, Chroma) and bandpass filters, ET585/65 m (Chroma, Cy3) and ET700/75 m (Chroma, Cy5), in a Dual-View spectral splitter (Photometrics). No bandpass filter was used for Cy7 in the Dual-View spectral splitter. The hardware was controlled and smFRET movies were collected using Cell Vision software (Beijing Coolight Technology).

The µOR was immobilized on the cover slip via biotinylated M1 Fab and streptavidin. In brief, the assembled glass chamber, which had been cleaned and passivated with biotin-polyethylene glycol, was incubated with 0.05 mg ml⁻¹ streptavidin in 20 mM HEPES 7.5, 100 mM NaCl. One minute later, the unbound streptavidin was washed out by 25 nM biotinylated M1 Fab in incubation buffer (50 mM HEPES pH 7.5, 100 mM NaCl, 0.01% LMNG, 0.001% CHS, 2 mM CaCl₂, 5 mM MgCl₂ and 100 µM ligand). The biotinylated M1 Fab was incubated in the channel for one minute and the unbound M1 Fab was washed out by incubation buffer. The N-terminal Flag-tagged, fluorophore-labelled µOR was diluted to around 20 nM in incubation buffer and incubated on ice for 1 h before measurement. The µOR was diluted to about 1 nM and injected into the chamber. The unbound µOR was removed by imaging buffer (incubation buffer + 50 nM protocatechuate-3,4-dioxygenase (PCD), 2.5 mM protocatechuic acid (PCA), 1.5 mM aged Trolox, 1 mM 4-nitrobenzyl alcohol (NBA), 1 mM cyclooctatetraene (COT)). Movies were taken at a frame rate of 10 s⁻¹ using the Cell Vision software. For measurement in complex with GDP-free G_i1, 20 nM µOR in the presence of 100 µM ligand was incubated with 20 µM G_i1 for 30 min followed by addition of 1:100 (v/v, NEB) apyrase. After incubation on ice for 1 h, the complex was diluted and injected into the chamber and measured following the same protocol above. For measurement in the presence of G_i1 and GDP, the surface-immobilized µOR was incubated with imaging buffer, then 20 µM G_i1 and various concentrations of GDP in imaging buffer were injected into the chamber and imaged. For measurement in the presence of βarr1(∆CT), the phosphorylated, Cy3/Cy5-labelled µOR was diluted to about 20 nM in arrestin buffer (50 mM HEPES pH 7.5, 100 mM NaCl, 0.01% LMNG, 0.001% CHS, 2 mM CaCl₂, 5 mM MgCl₂ and 100 µM ligand, 20 µM C8-PIP2), and 90 µM βarr1(∆CT) was added. After incubation on ice for 1 h, the µOR was diluted to 1 nM in arrestin buffer with βarr1(∆CT) at a final concentration of 90 µM. After immobilization, unbound µOR was washed out with imaging buffer supplemented with 90 µM βarr1(∆CT) and movies were taken.

To extract the time trajectories of single-molecule fluorescence, collected movies were analysed by a custom-made software program developed as an ImageJ plugin (http://rsb.info.nih.gov/ij). Fluorescence spots were fitted by a 2D Gaussian function within a nine-pixel by nine-pixel area, matching the donor and acceptor spots using a variant of the Hough transform⁵⁵. The background subtracted total volume of the 2D Gaussian peak was used as raw fluorescence intensity I.

Actual FRET efficiency was calculated via equation \(E={\left(1+\frac{{I}_{{\rm{D}}}}{{I}_{{\rm{A}}}-\chi {I}_{{\rm{D}}}}\gamma \right)}^{-1}\), where I_D is raw fluorescence intensity of donor, I_A is raw fluorescence intensity of acceptor, and χ is the cross-talk of the donor emission into the acceptor channel. γ accounts for the differences in quantum yield and detection efficiency between the donor and the acceptor and is calculated as the ratio of change in the acceptor intensity (ΔI_A) to change in the donor intensity (ΔI_D) upon acceptor photobleaching⁵⁶ (γ = ΔI_A/ΔI_D). The χ was 0.05, and the γ was 1 and 0.2 for Cy3/Cy5 and Cy3/Cy7 dye pairs, respectively. FRET traces were picked by a custom-made Matlab script based on three criteria⁵⁷: (1) signal-to-nose ratio of trances, which is defined as the mean of total intensity before photobleaching divided by its standard deviation, was higher than 4 and 3 for Cy3/Cy5 and Cy3/Cy7 dye pairs, respectively; (2) donor traces have single-step photobleaching; (3) traces last for at least 2 s. To calculate the transition rate in the presence of G protein and GDP, only traces that showed at least one high/low-FRET transition were selected and analysed by a Hidden Markov model-based software (HaMMy)⁴⁴. Two FRET states were identified by HaMMy. The cumulative frequency count of high-FRET dwell times for each condition was fitted in Origin software to single exponential decay curves, generating high-FRET dwell time. The cumulative frequency count of low-FRET dwell times for each condition was fitted in Origin software to double exponential decay curves and the low-FRET dwell time was calculated as a weighted average accordingly.

DEER experiments and analysis

Setup

Four pulse, Q-band DEER data were collected at 50 K on a Bruker e580 equipped with a QT-II resonator and a 150 W TWT amplifier using the pulse sequence: π/2(ν_A) – τ₁ – π(ν_A) – (τ₁ + t) – π(ν_B) – (τ₂ − t) – π(ν_A) – τ₂ – echo, with τ₁ = 300 ns, τ₂ = 3.5 μs, Δt = 16 ns, 16-step phase cycling and a repetition time of 510 μs. The observer pulses (ν_A) were set to 18 ns and 36 ns for π/2 and π pulses, respectively, and applied 70 MHz below resonance. The 100 ns pump pulse (ν_B) was applied on resonance and consisted of a 50 MHz linear chirp pulse generated by an arbitrary waveform generator. We furthermore used an 8-step ESEEM suppression protocol. All experiments were implemented using Xepr v2.6b.163.

Analysis

DEER data were processed via Gaussian mixture models (GMM) implemented in Matlab (v.2019b) using the DEERlab toolbox (v.0.9.2)⁵⁸. In brief, all 30 datasets (10× ligand only, 10× ligand + G_i, 10× ligand + β-arr) were analysed simultaneously assuming a variable number of two to seven Gaussians whose mean positions and widths (global fitting parameters) were constrained in the range of 20–100 Å, and 2–20 Å, respectively. For each individual condition the sum of populations (local fitting parameters) was normalized to 1. Each of the thirty datasets was allowed a unique modulation depth (range 0.3–0.7) and each transducer condition allowed for a unique receptor concentration in the range of 25–150 μM. Model-based distance distributions and background corrected dipolar kernels were calculated using DEERlab functions and fit simultaneously to all 30 datasets using the fitparamodel.m routine (Multistart = 10). Post hoc model selection was performed using the Akaike information criterion corrected (AICc) and the more restrictive Bayesian information criterion (BIC) which were both evaluated globally for all DEER datasets and both yielded 6 Gaussians as most parsimonious model. Error analysis using 1,000 bootstrap iterations was performed for all fitting parameters, the dipolar fits and the parametric distance distributions, and evaluated at the 95% confidence level. Significant population changes between different transducer conditions were determined by disjunct 95% confidence intervals and are marked with * (star).

Comparison of model-based and model-free analysis

As a control, we also analysed all DEER data using Tikhonov regularization (TR) and model-free based analysis in DEERlab and LongDistances (v.946; http://www.biochemistry.ucla.edu/Faculty/Hubbell/software.html). Regularization or smoothness parameters were determined via AICc and L-curve criterion, respectively. The results from both analyses were superimposable. For comparison, the distance distributions derived from the model-based (6 Gaussian) best fit and model-free DEERlab fits are shown in Extended Data Fig. 5. Both methods yield almost identical distance distributions and reveal all ligand or transducer-dependent distance changes supporting the validity of the model-based fit. Most apparent differences appear in the 35–45-Å distance range, where model-based analysis was able to differentiate two peaks, namely at 39 Å and 43 Å, of different width, namely 3.8 Å and 2 Å. This finding exemplifies one of the inherent advantages of the global, GMM-based fitting approach over Tikhonov regularization or model-free analysis. While Tikhonov regularization or model-free based analyses apply a single regularization or smoothness parameter to the full distance range, the chosen GMM allows different widths for individual distance peaks, as they may exist for different conformational states. Other advantages of the model-based approach include straightforward quantification of each population (Gaussian area) and a rigorous error analysis for each fitting parameter using covariance matrix or bootstrapping based approaches.

We conducted biological repeats for naloxone and lofentanil with and without G protein. These conditions represent the most distinct ligand/transducer conditions investigated and we observe good reproducibility. In particular, for both ligands, the smaller G_i-induced shifts are accurately reproduced (Extended Data Fig. 8d).

Radioligand binding

Membranes of Sf9 cells expressing µOR were used for saturation binding and competition binding. Saturation binding was performed by incubating Sf9 membrane with increasing concentrations of the antagonist [³H]diprenorphine (³H-DPN, Perkin Elmer) for 2 h at room temperature in 0.5 ml of binding buffer containing 50 mM Tris-HCl pH 7.5, 100 mM NaCl, 0.1% BSA. Nonspecific binding of ³H-DPN was measured by adding 10 µM naloxone in the binding reaction. To separate unbound ³H-DPN, binding reactions were rapidly filtered over GF/B Brandel filters. The filters were then washed three times with 5 ml ice-cold binding buffer. Radioactivity was assayed by liquid scintillation counting.

For competition binding with ³H-DPN, Sf9 cell membrane was incubated with 2.9 nM ³H-DPN and increasing concentrations of DAMGO in 0.5 ml of binding buffer. Binding reactions were incubated for 2 h at room temperature. The free ligand was separated by rapid filtration onto a GF/B Brandel filter with the aid of a 48-well harvester (Brandel). Radioactivity was assayed by liquid scintillation counting.

The resulting data were analysed using Prism 9.0 (GraphPad Software). The dissociation constant (K_d) of ³H-DPN was calculated by fitting the saturation data in a one-site (total and nonspecific binding) model. The K_i of DAMGO was calculated by fitting the competition binding data in a one-site (fit K_i) model.

For competition binding with [³H]naloxone, mouse µOR-containing insect cell membranes prepared above were diluted to normalize expression levels between wild-type (1:1,000) and minimal-cysteine mouse µOR (1:100) in 20 mM HEPES pH 7.4, 100 mM NaCl, and 0.05% BSA. Membranes were then incubated with 3 nM [³H]naloxone and serially-diluted orthosteric ligands at their respective final concentrations. Tested ligands were diluted into the buffer above to a final concentration of 100 µM with a fourfold serial dilution series for 10 total concentrations. The only exception is BU72, which was diluted to 1.3 µM final concentration before the same serial dilution. All ligands include independent ‘no ligand’ controls (100% binding) and excess cold naloxone (200 µM) controls (0% binding) to which points were normalized. The mixtures were shaken for 1 h at room temperature before collection onto Filtermat B (Perkin Elmer) and washed with cold binding buffer (20 mM HEPES pH 7.4, 100 mM NaCl). The filters were then dried at 60 °C before adding a sheet of MultiLex B/HS melt-on scintillator sheets (Perkin Elmer) and counts read on a MicroBeta Counter (Perkin Elmer). Quadruplicate data values were plotted and normalized as described above.

BRET-based assays with TRUPATH and arrestin signalling

The BRET-based assays were based on TRUPATH⁵⁹ and arrestin signalling⁴⁸. To measure µOR’s coupling with G_i1, HEK 293 T cells (ATCC CRL-3216, authenticated by the supplier, routinely tested for mycoplasma) were plated in 10 cm dishes at 3–4 million cells per dish in Dulbecco′s Modified Eagle′s Medium (DMEM) supplemented with 10% FBS. The next day, cell medium was replaced with fresh DMEM + 10% FBS medium. Cells were transfected 2 h later, using a 1:1:1:1 DNA ratio of receptor:Gα-RLuc8:Gβ1:Gγ2-GFP2 (500 ng per construct). Transit 2020 (Mirus Biosciences) was used to complex the DNA at a ratio of 3 µl Transit per µg DNA, in OptiMEM (Gibco-ThermoFisher) at a concentration of 10 ng DNA per µl OptiMEM. The next day, cells were collected from the plate using Versene (0.1 M PBS  +  0.5 mM EDTA, pH 7.4) and plated in poly-d-lysine-coated white, clear-bottom 96-well assay plates (Greiner Bio-One) at a density of 50,000 cells in 200 µl culture medium (DMEM + 1% dialysed FBS) per well. The next day, white backings (Perkin Elmer) were applied to the plate bottoms, and growth medium was carefully aspirated and replaced immediately with 60 µl of assay buffer (1× Hank’s balanced salt solution (1× HBSS, Gibco), 20 mM HEPES, pH 7.4), supplemented with 5 µM (final concentration) coelenterazine 400a (Nanolight Technologies). After a 5 min equilibration period, cells were treated with 30 µl of drug (3×) prepared in assay buffer for an additional 5 min. Plates were then read in an LB940 Mithras plate reader (Berthold Technologies) with 395 nm (RLuc8-coelenterazine 400a) and 510 nm (GFP2) emission filters, at integration times of 1 s per well. Plates were read serially four times, and measurements from the fourth read were used in all analyses. BRET ratios were computed as the ratio of the GFP2 emission to RLuc8 emission.

To measure coupling of µOR coupling with β-arrestin-1, the procedures are mostly similar to those in BRET-G-protein assays except: HEK 293 T cells were co-transfected in a 1:5 ratio with µOR-Rluc8 and Venus–β-arrestin-1. Before the addition of tested drugs, white backings (Perkin Elmer) were applied to the plate bottoms, and growth medium was carefully aspirated and replaced immediately with 60 µl of assay buffer (1× HBSS, 20 mM HEPES, pH 7.4), supplemented with 5 µM (final concentration in assay buffer) coelenterazine h (Nanolight Technologies). After a 5 min equilibration period, cells were treated with 30 µl of drug (3×) prepared in assay buffer for an additional 5 min. Plates were then read in an LB940 Mithras plate reader (Berthold Technologies) with 485 nm (RLuc8-coelenterazine h) and 530 nm (Venus) emission filters, at integration times of 1 s per well. Plates were read serially four times, and measurements from the fourth read were used in all analyses. BRET ratios were computed as the ratio of the Venus emission to RLuc8 emission. The BRET ratio from G-protein or arrestin assays was plotted using nonlinear regression and Dose-response stimulation equation in Prism 9 (Graphpad).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.