High-mass MALDI-MS unravels ligand-mediated G protein–coupling selectivity to GPCRs

Significance G protein–coupled receptors (GPCRs) are important pharmaceutical targets for the treatment of a broad spectrum of diseases. Upon ligand binding, GPCRs initiate intracellular signaling pathways by interacting with partner proteins. Assays that quantify the interplay between ligand binding and initiation of downstream signaling cascades are critical in the early stages of drug development. We have developed a high-throughput mass spectrometry method to unravel GPCR–protein complex interplay and demonstrated its use with three GPCRs to provide quantitative information about ligand-modulated coupling selectivity. This method provides insights into the molecular details of GPCR interactions and could serve as an approach for discovery of drugs that initiate specific cell-signaling pathways.

Expression and purification of β1AR. Expression and purification of turkey apo-β1AR were performed as described previously (2), but with some modifications. Turkey β1AR was cloned into pcDNA4/TO vector and expressed in stable HEK293 GntI-cells (1). Cells were grown to density of 3-4 x 10 6 cells/mL, induced with tetracycline and sodium butyrate, and harvested 48-72 h postinduction. Detergent screening was performed to optimization the solubilization step for apoprotein. Membranes were prepared and solubilized for 1 hour in LMNG detergent, followed by removal of insoluble fraction by centrifugation at 40,000 rpm for 1.5 hours. Solubilised fraction was bound to 1D4 resin for 3 hours. The resin was washed with 20 mM Hepes pH 7.5, 300 mM NaCl, 0.01% LMNG and 0.001% CHS, and β1AR was eluted through cleavage buffer with HRV3C protease containing 20 mM Hepes 7.5, 100 mM NaCl, 0.01% LMNG and 0.001% CHS.
Eluted β1AR protein was concentrated and subject to SEC on Superdex 200 Increase column (GE) equilibrated in 20 mM Hepes pH 7.5, 100 mM NaCl, 0.01% LMNG and 0.001% CHS. Pure β1AR protein in the peak fractions was collected, concentrated, flash frozen in liquid nitrogen and stored at -80 o C.
Protein was eluted with 3 CV of buffer C (20 mM Hepes pH 7.5, 100 mM NaCl, 500 mM imidazole, 10% glycerol, 1 mM MgCl2, 50 µM GDP). His-tag was cleaved off with TEV protease in presence of 1 mM DTT and the sample was simultaneously dialysed overnight against 2 L of buffer D (20 mM Hepes pH 7.5, 100 mM NaCl, 10% glycerol, 1 mM MgCl2, 10 µM GDP). Sample was mixed for 2h after addition of 4 ml Ni-NTA resin and 20 mM imidazole and collected on a column containing additional 1 ml Ni-NTA resin. Flow-through was collected and the resin was washed with 2 CV of buffer D. Both fractions were pooled, concentrated to 1.5 ml and loaded onto a Superdex-200 26/600 gel filtration column equilibrated with buffer E (10 mM Hepes pH 7.5, 100 mM NaCl, 10% glycerol, 1 mM MgCl2, 1 µM GDP, 0.1 mM TCEP). Peak fractions were pooled, concentrated, flash-frozen in liquid nitrogen and stored at -80°C. The method for expression and purification of the full-length mGα proteins and the truncated versions, mGo_Δ5 and mGi_Δ5, followed the same protocol.
Expression and purification of Nb80. Nanobody 80 was expressed and purified using modified previously described method (5). The Nb80 with a C-terminal His-tag was expressed in the periplasm of E. coli strain WK6. The cells were cultured in TB media supplemented with 0.1% (w/v) glucose and 2 mM MgCl2. The expression was induced with 1 mM IPTG after OD600 of 0.70 was reached and the temperature was reduced to 28°C. The cells were harvested the next day and lyzed with ice-cold buffer (50 mM Tris-HCl pH 8.0, 12.5 mM EDTA, 0.125 M sucrose). After centrifugation, the protein was purified by nickel affinity chromatography followed by size-exclusion chromatography on a Superdex75 60/300 column equilibrated with buffer 20 mM Hepes pH 7.45, 100 mM NaCl. The protein was concentrated to 100 mg/mL, flash frozen in liquid nitrogen and stored at -80°C.
Preparation of heterotrimeric G protein subunits. The full-length human Gαi subunit (Gαi1) with an N-terminal deca-histidine tag was prepared by heterologous expression in E. coli strain BL21(DE3) and purified as described (6). The transducin heterotrimer (Gαtβ1γ1) was isolated from the rod outer segment of bovine retina (W. L. Lawson Company) and Gβ1γ1 was separated from Gαt by Blue Sepharose 6 Fast Flow (GE Healthcare) (6).
Preparation of β-arrestin-1. The wild-type full-length human β-arrestin-1 with an N-terminal hexa-histidine tag was expressed and purified as described previously (7) High-mass MALDI mass spectrometry of protein complexes. A sandwich method (matrix/sample/matrix) was used to deposit the protein samples on a 384-spot MALDI plate (AB Sciex). First, the MALDI plate was cleaned by flushing/wiping alternately with methanol and Milli-Q water and finally with methanol. Then, a 1:500:500 (v/v/v) TFA/water/acetonitrile (TWA) solution and a 2:1:3 (v/v/v) formic acid/water/isopropanol (FWI) solution were prepared, respectively. A saturated sinapinic acid solution was prepared by dissolving SA in TWA, following by centrifuging at 14,000 rpm for 5 min to separate undissolved SA from the supernatant. Then a half-saturated SA solution was prepared by evenly mixing with FWI in a 1:1 ratio (v/v), which was deposited on the MALDI plate (0.5 μL/spot) and dried in air. This formed a uniform matrix crystal film on the surface of the plate, and such a matrix coated MALDI plate could be stored in clean air for several weeks if desired. Then, the protein samples were deposited on the plate (three spots/sample), followed by drying in air. Finally, a saturated TWA solution of SA was deposited on top of the sample (0.5 μL/spot) and dried in air for high-mass MALDI-MS analysis. All mass spectrometric measurements were performed in linear positive ion mode on a MALDI-TOF/TOF mass spectrometer (model 4800 plus, AB Sciex, Darmstadt, Germany) equipped with a high-mass detector (HM2, CovalX AG, Zurich, Switzerland). The HV1 and HV2 voltages of the HM2 detector were set to −3.5 kV and −20.0 kV, respectively. MALDI was initiated by a Nd:YAG laser pulse (355 nm), and 500 shots per spectrum were accumulated. The Origin software was used to process the data, including plotting and normalisation of the mass spectra, calculation of the peak area, plotting of all the curves, fitting of the data and calculation of dissociation constants (Kds).

Determination of the number of mono-, intra-and inter-molecular links.
During the crosslinking, the three types of links (mono-, intra-and inter-molecular) are formed simultaneously according to the relative equilibrium of the species in the reaction. Once formed, these links act as locks on the protein. The unwanted links on the single protein components (receptor alone and partner protein alone) help to maintain the equilibrium between the formed complexes and single protein components by preventing further complex formation. As a result, there should be no further conformational changes due to complex formation after crosslinking.
The types of crosslinks on the protein after the crosslinking reaction was calculated using the following formulae, where a is monolink and b is intramolecular crosslink in the single protein components (GPCR or partner protein monomer), and a' is monolink, b' is intramolecular crosslink, and x is intermolecular crosslink in a GPCR•partner complex (based on the assumption that the number of reacted lysine residues is same in the protein monomers and the protein complex): [1] The difference in the molecular weight (∆MW=∆m/z) between m/z of the crosslinked complex and the total m/z of crosslinked GPCR and crosslinked partner protein was calculated: i) if the NHS groups in the crosslinks are not hydrolyzed ∆m/z=a*MW(crosslinker lost one -NHS)+b*MW(crosslinker lost two -NHS)-a'*MW(crosslinker lost one -NHS)-(b'+x)* MW(crosslinker lost two -NHS) [2] ii) and if all the NHS groups are hydrolyzed ∆m/z=a*MW(crosslinker lost two -NHS)+b*MW(crosslinker lost two -NHS)-a'*MW(crosslinker lost two -NHS)-(b'+x)*

Establishment of concentration standard curve.
Protein samples (Rho, β1AR, AT1R, Gαi, or mGo) were pre-treated with BS(PEG)9 crosslinker for 1 hour at room temperature with the goal to block all lysine residues, and then diluted to gradient concentrations (SI Appendix, Fig. S5) using Hepes buffer. After being precooled on ice, the gradient concentrations of protein solutions were evenly and quickly mixed with precooled β-Galactosidase dissolved in Hepes buffer at a volume ratio of 1: 1, v/ v, respectively, and immediately deposited on the MALDI plate (three spots/sample) for further mass spectrometry analysis. Note that for obtaining the standard curves, different concentrations of β-Gal were chosen for different target proteins, i.e., 2.0, 0.6, 1.0, and 0.6 μM for rho, β1AR, AT1R, and mGo, respectively. Finally, the mass spectra were normalized by the intensity of β-Gal, and a standard curve was obtained by plotting the relative intensity (= analyte peak area/β-Gal peak area) of the analyte signals against the concentration.

Supplementary Fig. 4 | Selectivity in the complex formation of apo-and ligand-bound GPCRs with partner proteins assayed by high-mass MALDI-MS, and binding constants of the formed complexes. a, Raw data
showing complex formation of three GPCRs, rhodopsin (Rho), beta-1 adrenergic receptor (β1AR), and angiotensin II type 1 receptor (AT1R) in the presence or absence of agonists, antagonists or inverse agonists, with their partner proteins mGs, mGo, mGi, mGq and Nb80. The ligands used were the following: atr = all trans-retinal, iso = isoprenaline, pro = propranolol, nad = nadolol, s32212, car = carvedilol, angII = angiotensin II, and azi =azilsartan (see Supplementary Table 4). b, Comparison of the binding constants (Kas) of the GPCR•partner protein complexes formed (shown in Fig. S4a). Ka values were calculated using Kd (Fig. 3c, Fig. 4b, and table S5) and the relationship Ka=1/Kd. Each measurement performed using at least 3 different incubations; error bars represent standard deviations. Supplementary Fig. 7 | a, Mass spectra showing titration of Nb80 to the mixture of mGo and β1AR, without the reference protein (left), and with the reference protein (right). b, Mass spectra normalised to the peak intensity of mGo, without the reference protein (left), and with the reference protein (right). c, Comparison of the changes in the peak intensity ratio of Nb80 to mGo (left) and the change in the peak intensity ratio of β1AR•mGo to mGo (right) in the mass spectra without reference protein (pink) and with reference protein (light blue). Supplementary Fig. 15 | a, Mass spectra of the concentration of the conversion between β1AR•mGo and β1AR•Nb80 modulated by isoprenaline. b, Intensity ratio of β1AR•Nb80 to β1AR•mGo as a function of added isoprenaline, based on the data shown in a. c, Intensity ratio of mGo to β1AR•mGo, as a function of added isoprenaline, based on the data shown in a. Each measurement was replicated three times; error bars represent standard deviations.
Supplementary Fig. 16 | Binding of partner proteins to 1AR as a function of ligand competition. a, Schematic of GPCR conformational ensembles induced by the competition between antagonist and agonist. b, Mass spectra of the selective assembly of ligand-bound GPCR (first column: β1AR/S32212/iso at 2.5/50/50 μM, respectively; second column: β1AR/pro/iso at 2.5/50/50 μM, respectively; third column: β1AR/car/iso at 2.5/50/50 μM, respectively; fourth column: β1AR/nad/iso at 2.5/50/50 μM respectively) with various partner proteins (from top to bottom: mGs, mGo, mGq, and Nb80). Structure of a heterotrimeric G protein (PDB code: 1GP2) and the measured distances in Å between lysine residues in the G protein subunits. Lysine residues are highlighted as green, blue, and red spheres. c, Mass spectra of complexes formed between G protein subunits following incubation and crosslinking with DSS, BS(PEG)5 or BS(PEG)9.