Dominant Negative G Proteins Enhance Formation and Purification of Agonist-GPCR-G Protein Complexes for Structure Determination

Abstract

Advances in structural biology have yielded exponential growth in G protein-coupled receptor (GPCR) structure solution. Nonetheless, the instability of fully active GPCR complexes with cognate heterotrimeric G proteins has made them elusive. Existing structures have been limited to nanobody-stabilized GPCR:Gs complexes. Here we present methods for enhanced GPCR:G protein complex stabilization via engineering G proteins with reduced nucleotide affinity, limiting Gα:Gβγ dissociation. We illustrate the application of dominant negative G proteins of Gαs and Gαi2 to the purification of stable complexes where this was not possible with wild-type G protein. Active state complexes of adenosine:A1 receptor:Gαi2βγ and calcitonin gene-related peptide (CGRP):CLR:RAMP1:Gαsβγ:Nb35 were purified to homogeneity and were stable in negative stain electron microscopy. These were suitable for structure determination by cryo-electron microscopy at 3.6 and 3.3 Å resolution, respectively. The dominant negative Gα-proteins are thus high value tools for structure determination of agonist:GPCR:G protein complexes that are critical for informed translational drug discovery.

Figure 1

Alignment of human Gα isoforms. Clustalw omega alignment of reference sequences for human Gα isoforms manually adjusted to take into account secondary elements from deposited PDB structures: 1SVK, 3FFB, 1ZCA, 1ZCB, 2BCJ, 1AZT, and 3SN6. α-Helices (zig-zags) from the α-helical domain are indicated in light blue, in dark blue are those outside either core domain and in green are those from the Ras-like domain. β-Strands are indicated in the same color scheme with wavy arrows. Secondary structure elements from the α-helical domain are indicated with letters and the Ras-like domain with numbers. The position and substitution for common DN-substitutions are highlighted in purple with the CGN numbering shown below. Highlighted in yellow are Gαi residues that are substituted into Gαs, which improve the dominant negative effect.

Figure 2

Comparison of purification of WT-Gs and DN-Gs containing heterotrimers with CTR and GLP-1R. (a) Coomassie stained gel showing relative abundance of various components of CTR (with salmon calcitonin) and GLP-1R (with exendin-P5)–G protein complexes following FLAG purification in the presence of either WT or DN-Gs but not Nb35. (b) 2D class averages from negative stained EM micrographs of the CTR:DN Gs and GLP-1R:DN Gs complexes in the absence of Nb35. (c) Western blot against His-tag (Gβ, Nb35, and CTR:GLP-1R) and Gαs from purified active complex with DN-Gs in the presence or absence of Nb35. (d,e), Western blot and coomassie stained gel of WT- (d) and DN- (e) containing GLP-1R:Gs complexes in the presence of Nb35 and corresponding SEC traces (f), illustrating that both complexes can be purified but that the proportion of complex in the void/aggregate is slightly higher with the WT-Gαs (c).

Figure 3

Formation and purification of WT-Gs and DN-Gs containing heterotrimers by oxyntomodulin with GLP-1R. (a,b) Size exclusion chromatography of complexes initiated with 1 μM oxyntomodulin revealed higher yield of complexes with the DN-Gs (b) relative to that seen with WT-Gs (a), although both had relatively low yields for complex formation (the elution position of the complexes is illustrated by the red arrow). (b) Increasing the concentration of oxyntomodulin to 50 μM in the Oxyn:GLP-1R:DN-Gs:Nb35 preparation during the initiation step markedly improved the yield of complexes, which could be purified to homogeneity by an additional size exclusion chromatography step (d). (e) Coomassie stained gel from the peak in panel d, illustrating high purity and recovery of each of the component proteins. (f) 2D class averages from negative stain EM demonstrate the presence of complexes suitable for single-particle cryo-EM structure determination.

Figure 4

DN-Gαs allows purification and structure determination of the fully active heteromeric CGRP receptor. (a) Western blots of purification fractions of the CGRP active complex using WT-Gαs showing purification is possible but yield is poor (see text): (left) anti-His antibody against His tags on Gβ₁ and Nb-35; (middle left) anti-Gαs antibody for Gαs detection; (middle right) anti-CLR antibody for CLR detection; (right) anti-FLAG antibody against the FLAG tag on RAMP1. (b) (Left) Western blot of the final purified fraction of the CGPR complex formed using DN-Gαs showing that all components are present; (right) coomassie stained gel showing stoichiometric recovery of proteins. (c) Monodisperse peak of the purified complex following size exclusion chromatography. (d) Negative stain EM micrograph. (e) 2D class averages from the negative stain EM data. (f) 3D cryo-EM map from single particle cryo-EM determination of structure, colored according to each protein subunit (the 3D map was adapted from Liang et al.).

Figure 5

DN-Gαi₂ allows purification and structure determination of the fully active A1-AR. (a) Size exclusion chromatography trace showing a broad peak for ADO:A1-AR:G protein complex in the presence of WT Gi₂, and a monodisperse peak in the presence of DN Gi₂. (b) Western blot of final purified fraction showing (left) limited Gαi₂ recovery in the ADO:A1-AR:WT G protein complex and (right) stoichiometric recovery of G protein in the ADO:A1-AR:DN G protein complex (modified from Ext. Data Figure 1c). (c) Negative stain EM micrograph of the ADO:A1-AR:DN Gi₂ protein complex. (d) 2D class averages from the negative stain EM micrographs; (e) 3D cryo-EM map, A1-AR is in red, heterotrimeric DN-Gi₂ is in blue, gold, and green for α, β, and γ, respectively. The DN-Gi₂ data were adapted from Draper-Joyce et al. 2018. (f) Size exclusion chromatography illustrating purification of the ADO:A1-AR:DN-Gi₃ complex; the peak fractions (gray box) were pooled and assessed by negative stain EM (g), with 2D class averages shown in panel h.

Figure 6

Functional analysis of DN-G proteins. (a) Time-course for ligand-induced changes in BRET of WT-Gαs or DN-Gαs (nanoLuc tagged) and Gγ (Venus) at increasing concentrations of salmon calcitonin (sCT) (arrow, A) followed by the addition of 30 μM GTP (arrow, B) on membranes from HEK293 cells that lack endogenous Gαs and were stably transfected with the human CTR; the rate and magnitude of agonist-induced structural rearrangement are similar, but the DN-Gαs is resistant to GTP-induced conformational rearrangement. (b) Comparison of time-courses for agonist-induced changes in BRET of WT-Gαs/Gγ and DN-Gαs/Gγ are shown for GLP-1(7–36)NH₂ activated GLP-1R; both WT- and DN-Gαs report the same difference in rate and extent of ligand-induced G protein conformational change (A); however, after the addition of 30 μM GTP (B), the DN-Gαs is less susceptible to conformational change. (c) time-course for GLP-1(7–36)NH₂-induced changes in BRET of WT-Gαi₂ or DN-Gαi₂ (nanoLuc tagged) and Gγ (Venus) (A), followed by the addition of 30 μM GTP (B), on membranes from HEK293 cells that lack endogenous Gαs/Gαq/11/Gα12/13 and were stably transfected with the human GLP-1 receptor. The rate and magnitude of structural rearrangement are similar but the DN-Gαi₂ is resistant to GTP-induced conformational rearrangement. All data are N = 3 + SEM.