NMR experiments and analyses

All experiments were performed at 20 °C on Bruker Avance 500 or 600 spectrometers, equipped with a cryogenic probe. All spectra were processed by the Bruker TopSpin 2.1 or 3.1 software, and the data were analyzed using Sparky (T. D. Goddard and D. G. Kneller, Sparky 3, University of California, San Francisco, CA). Protein samples were dissolved in NMR buffer (20 mM HEPES-NaOH, pH 7.0, 150 mM KCl, 0.5 mM GDP, 0.1 mM DSS, and 99.5% D₂O), containing 5 mM DTT (for assignment) or 0.1 mM tris(2-carboxyethyl)phosphine (TCEP) (for the other experiment).

Resonance assignments of the Alaβ, Ileδ1, Leuδ, and Valγ methyl groups in Gαi3βγ were obtained by combining mutagenesis and nuclear Overhauser effect (NOE) analyses, based on the crystal structure. To observe the methyl-backbone amide and methyl-methyl NOEs, we acquired a set of three-dimensional NOESY spectra. The [¹H–¹H] NOESY-[¹H–¹⁵N] TROSY, [¹H–¹H] NOESY-[¹H–¹³C] HMQC, and [¹H–¹³C] HMQC-[¹H–¹H] NOESY-[¹H–¹⁵N] TROSY spectra were recorded on a {uniform(ul)-[²H, ¹⁵N]; Alaβ, Ileδ1-[¹³CH₃]} Gαi3-[non-labeled]βγ sample, with mixing times of 150–200 ms. The [¹H–¹H] NOESY-[¹H–¹³C] HMQC and [¹H–¹³C] HMQC-[¹H–¹H] NOESY-[¹H–¹³C] HMQC spectra were recorded on {ul-[²H, ¹⁵N]; Alaβ, Ileδ1, Leu/Val ^proS-[¹³CH₃]} Gαi3-[non-labeled]βγ and {ul-[²H, ¹⁵N]; Leu/Val-[¹³CH₃, ¹³CH₃]} Gαi3-[non-labeled]βγ samples, with a mixing time of 100 ms. The identified NOEs were assigned, based on the crystal structure of Gαi1β₁γ₂ (PDB ID: 1GP2)³³. For mutagenesis, we constructed 32 mutants of Gαi3 (L5I, A7V, A11V, A12V, V13A, I19V, L23I, A30V, A31V, V50I, V71A, V73I, A98S, A99S, A101V, A111S, A114V, V118F, I162V, V174I, V179I, V185I, V201I, L232I, V233I, L234I, A235V, L273I, L310I, V342I, L348I, and L353I), recorded the ¹H–¹³C HMQC spectrum of each mutant in the presence of an excess amount of Gβγ, and compared each spectrum with that of the wild type. We established 96% of the Alaβ (25/25), Ileδ1 (25/26), Leuδ (52/54), and Valγ (39/42) assignments for Gαi3 in complex with Gβγ.

About two-thirds of the resonance assignments of the Alaβ, Ileδ1, Leuδ, and Valγ methyl groups in Gαiqiβγ were readily transferred from those of Gαi3βγ, since the signals from the Gαi3 moiety overlapped. The other signals were assigned by NOE analyses, based on the crystal structure. We recorded the [¹H–¹³C] HMQC-[¹H–¹H] NOESY-[¹H–¹³C] HMQC spectra with a mixing time of 100 ms for the {ul-[²H, ¹⁵N]; Alaβ, Ileδ1, Leuδ, Valγ-[¹³CH₃]} Gαiqi-[non-labeled]βγ and {ul-[²H, ¹⁵N]; Leu/Val-[¹³CH₃, ¹³CH₃]} Gαiqi-[non-labeled]βγ samples. The ¹H–¹³C HMQC spectrum was recorded on a {[ul-²H,¹⁵N], Ileδ1, Leu/Val^proS-[¹³CH₃]} Gαiqi-[non-labeled]βγ sample, to obtain the stereospecific assignments for the Leu/Val^proS and Leu/Val^proR signals. We used the crystal structures of Gαi1β₁γ₂ (PDB ID: 1GP2) and Gαqβ₁γ₂ (PDB ID: 3AH8)⁵¹ as references. We established 95% of the Alaβ (20/23), Ileδ1 (22/22), Leuδ (52/54), and Valγ (40/42) assignments for Gαiqi in complex with Gβγ.

As for Gαi3-q(αA)βγ, the resonance assignments of the signals overlapping with those of Gαi3βγ, were transferred from those of Gαi3βγ.

To examine the chemical shift changes of Gαi3 upon forming the Gαi3βγ complex, we prepared NMR samples containing 100 μM {ul-[²H, ¹⁵N]; Alaβ, Ileδ1, Leuδ, Valγ-[¹³CH₃]} Gαi3 in the GDP-bound form, or 120 μM {ul-[²H, ¹⁵N]; Alaβ, Ileδ1, Leuδ, Valγ-[¹³CH₃]} Gαi3-[non-labeled]βγ (hereafter referred to as Gαi3[ILVA]βγ), and obtained the ¹H–¹³C HMQC spectra for each sample. The chemical shift differences (Δδ) were calculated using the equation Δδ = [(Δδ_H)² + (Δδ_C/5.9)²]^0.5.

To examine the spectral changes of Gαi3βγ, Gαiqiβγ, and Gαi3-q(αA)βγ induced by the addition of the GIRK chimera-nanodiscs, we prepared NMR samples containing Gα[ILVA]βγ (11 μM), with or without the GIRK chimera-nanodiscs (22 μM), and obtained the ¹H–¹³C HMQC spectra for each sample. We also performed experiments using the empty nanodiscs at the concentration that gives a lipid amount similar to that of the GIRK chimera-nanodiscs.

For site-specific spin-labeling, we first prepared the GIRK chimera with the C53S/C310T mutations, as it lacks reactive cysteine residues. Using this mutant as a template, cysteine substitutions were separately introduced to Q344, V351, and L366. MSP1E3, another protein component of the GIRK chimera-nanodisc, has no cysteine residue. After the GIRK chimera-nanodiscs were purified, spin-labeling was performed in buffer, composed of 50 mM Tris-HCl, pH 7.0, 100 mM NaCl, 50 mM KCl, and 0.1 mM TCEP. The GIRK chimera-nanodiscs were incubated with 0.9 mM 4-maleimido-2,2,6,6-tetramethylpiperidine-1-oxyl (4-maleimido-TEMPO) (Sigma-Aldrich) at room temperature for 4 h. Single cysteine labeling was confirmed by MALDI-TOF mass spectrometry on an Axima TOF² mass spectrometer (Shimadzu Biotech). The excess 4-maleimido-TEMPO was removed by passage through NAP-5 or PD-10 desalting columns (GE healthcare).

In the PRE experiments examining the paramagnetic state, ¹H-¹³C HMQC spectra were recorded for samples containing 20 μM Gαi3[ILVA]βγ and 25 μM 4-maleimido-TEMPO-labeled GIRK chimera-nanodiscs. Subsequently, the samples were reduced to the diamagnetic state by an incubation at 30 °C for 1 h in the presence of 0.3 mM ascorbic acid, and the ¹H-¹³C HMQC spectra were recorded. Using the signal intensities in the paramagnetic state (I^para) and the diamagnetic state (I^dia), the PRE contribution to the transverse relaxation rate, Γ₂, was calculated by the following equation⁵²:

where R₂^diaH and R₂^diaHC are the transverse relaxation rates of the ¹H single quantum coherence and the ¹H-¹³C multiple quantum coherence of the side chain methyl groups in the diamagnetic state, respectively. The R₂^diaH and R₂^diaHC rates were measured using NMR samples containing 200 μM Gαi3[ILVA]βγ or 250 μM Gαiqi[ILVA]βγ, with the pulse sequences, which create ¹H single quantum or ¹H–¹³C multiple quantum coherences during the relaxation periods^53,54. The magnetization transfer time in HMQC, t_HMQC, was set to 6.9 ms.