The structural basis of the dominant negative phenotype of the Gαi1β1γ2 G203A/A326S heterotrimer

Abstract

Aim: Dominant negative mutant G proteins have provided critical insight into the mechanisms of G protein-coupled receptor (GPCR) signaling, but the mechanisms underlying the dominant negative characteristics are not completely understood. The aim of this study was to determine the structure of the dominant negative Gαi1β1γ2 G203A/A326S complex (Gi-DN) and to reveal the structural basis of the mutation-induced phenotype of Gαi1β1γ2.

Methods: The three subunits of the Gi-DN complex were co-expressed with a baculovirus expression system. The Gi-DN heterotrimer was purified, and the structure of its complex with GDP was determined through X-ray crystallography.

Results: The Gi-DN heterotrimer structure revealed a dual mechanism underlying the dominant negative characteristics. The mutations weakened the hydrogen bonding network between GDP/GTP and the binding pocket residues, and increased the interactions in the Gα-Gβγ interface. Concomitantly, the Gi-DN heterotrimer adopted a conformation, in which the C-terminus of Gαi and the N-termini of both the Gβ and Gγ subunits were more similar to the GPCR-bound state compared with the wild type complex. From these structural observations, two additional mutations (T48F and D272F) were designed that completely abolish the GDP binding of the Gi-DN heterotrimer.

Conclusion: Overall, the results suggest that the mutations impede guanine nucleotide binding and Gα-Gβγ protein dissociation and favor the formation of the G protein/GPCR complex, thus blocking signal propagation. In addition, the structure provides a rationale for the design of other mutations that cause dominant negative effects in the G protein, as exemplified by the T48F and D272F mutations.

Figure 1

Sequence alignment of Gα_i-DN and homologous Gα_i. 1BH2, 1GP2, and 3UMS are PDB codes of already published Gα_i structures. 1BH2, A326S mutant of Gα_i; 1GP2, wild type Gα_i; 3UMS, G202A mutant of Gα_i. Secondary structures are shown schematically above the sequences. The sequences of G1–G5 box are marked red, and the Switch I–III (SW I–III) are marked blue under the sequences. G203A and A326S mutations are highlighted by magenta arrows.

Figure 2

Expression, purification and characterization of Gi-DN heterotrimer. Schematic depiction of the expression constructs for Gγ-Gα_i and Gβ. (B) A representative size-exclusion chromatography (SEC) elution profile for the Gi-DN heterotrimer. Pooled SEC elution fractions were analyzed by SDS-PAGE (B), Multi-Angle Static Light Scattering (C), and Thermal Shift Assay (D).

Figure 3

Overall structure of the Gi-DN heterotrimer and its comparison with GDP bound wild-type Gi heterotrimer (PDB code 1GP2). (A) The asymmetric unit contains two Gi-DN heterotrimers, one of which is composed of chain A (Gα_i, green), chain B (Gβ, cyan), chain G (Gγ, magenta), and the other one of chain H (Gα_i, yellow), chain J (Gβ, salmon), chain K (Gγ, gray). GDP is presented as spheres. (B) Comparison of Gi-DN and wild-type (PDB code 1GP2) heterotrimer structures in different views. The Gα_i, Gβ, and Gγ subunits of the Gi-DN heterotrimer are shown in green, cyan and magenta, and those of wild-type are shown in yellow, pink and orange, respectively. GDP is presented as spheres.

Figure 4

Nucleotide-binding site of GDP. (A) Alignment of Gα_i in the Gi-DN and wild-type (PDB code 1GP2) Gi heterotrimers. Gα_i-DN is colored green and the wild-type protein in yellow. GDP is shown in stick presentation. (B, C) Ligplot+ interaction map of GDP and the binding pocket residues from Gi-DN (B) and wild-type Gi (PDB code 1GP2) (C). E43, S44, G45, K46, S47, and T48 are located in the G1 box, N269, K270, and D272 in the G4 box, C325, and A326 in the G5 box. (D) and (E) The A326S mutation destabilizes the G5 box. (D) Both DN simulations exhibit decreased stability of the G5 box with average displacements of 3.77±1.42 Å compared to 2.62 ± 0.58 Å for WT simulations. (E) Structural comparision between the crystal structure (green) and deformed G5 box (cyan; position in DN1 trajectory marked with asterisk in Figure 4D).

Figure 5

Effects of Gα_i G203A and A326S mutations on guanine nucleotide binding. (A) Structural alignment of the GTPγS-bound A326S Gα_i1 (blue) (PDB code 1BH2) and the model of G203A mutation (green). The Van-der Waals spheres for G203A and GTP atoms are shown as dotted spheres, whose overlapping indicates space clash. (B) A203 forms a hydrogen bond with T181 located in the G2 box, resulting in a conformational change of the G2 box. (C) The geometry for the hydrogen bonds between the carboxylate group of D272 and the two amine groups (N1 and N2) of the guanine base. Left, Gi-DN heterotrimer; Right, wild type Gi heterotrimer (PDB code 1GP2). (D) The hydroxyl group of S326 introduces steric constraints for GDP, which causes S326 to move away from GDP. In addition, hydrogen bonds between S326 and N269 lead to conformational changes of their neighboring residues and is accompanied by loss of hydrogen bonds between GDP and N269, K270, and D272. In (B) and (D), the Gi-DN heterotrimer is shown in green and wild-type one is in yellow. The additional formed hydrogen bonds are shown as black dashed lines, and lost hydrogen bonds in Gi-DN heterotrimer are shown as yellow dashed lines.

Figure 6

The interface and polar contacts between Gα_i and Gβ subunits. (A) The two interfaces between Gα_i and Gβ subunits of the Gi-DN and wild-type (PDB code 1GP2) heterotrimers. The α subunits of DN-Gi and wild type Gi are colored in green and yellow, and the β subunits in blue and pink, respectively. (B–D) Polar contacts in interface 1 of DN-Gα_i and Gβ are rotated and displaced from blade 7 towards blade 1 (D). The number of polar contacts in Gi-DN (B) is five and in the wild-type complex (C) four. (E–G) Polar contacts in interface 2 of the Gα_i and Gβ subunits (G). The number of polar contacts in Gi-DN (E) is fourteen and in wild-type (F) is twelve.

Figure 7

The Gi-DN heterotrimer shares conformational features with G protein in the receptor-bound state. (A) Structure alignment of the Gi-DN and wild type heterotrimers (PDB code 1GP2). The Gi-DN Gα and Gβ subunits are shown in green and the Gγ subunit in light magenta. For the wild-type Gi heterotrimer, Gα and Gβ are shown in yellow and Gγ in orange. (B) Alignment of Gi-DN heterotrimer with the Gα_sβγ trimer in the β₂AR-bound state (PDB code 3SN6), with Gα and Gβ shown in blue and Gγ in salmon. The subunits of the Gi-DN heterotrimer are shown with same colors as in (A). (C) Hydrogen bond between T221 in Gβ and E22 in Gγ. The Gβ subunit is shown in green and Gγ in light magenta. (D) Hydrogen bond between Q259 and R22 in the Gβ subunit. (E) Superposition of the α5 helices from the Gi-DN heterotrimer (green), wild type heterotrimer (yellow) and the heterotrimer from the β₂AR-Gα_sβγ complex (blue).

Figure 8

Effects of additional mutations introduced into the GDP binding pocket on thermal stability of the Gi-DN heterotrimer. (A) GDP in the Gα_i-DN binding pocket. Pocket residues selected for mutagenesis are shown in stick presentation. (B–F) Thermal shift and corresponding SDS-PAGE profiles of mutant Gi-DN heterotrimers. Negative control (Gi-DN without additional mutations; B), Gi-DN D150L (C), Gi-DN C325I (D), Gi-DN T48F (E), and Gi-DN D272F (F). A GDP-induced thermos-shift was observed for Gi-DN, D150L, and C325I but not for T48F and D272F mutants. (G) Structure modeling of the T48F mutation. (H) Structure modeling of the D272F mutation.