G protein-coupled receptors (GPCRs) perceive many extracellular signals and transduce them to heterotrimeric G proteins, which in turn initiate a multitude of signaling cascades that alter cellular function. The β₂ adrenergic receptor (β₂AR) is a prototypical class A GPCRs [1], which preferentially couples to the stimulatory G protein (G_αs) compared to inhibitory G protein (G_αi) [2-4]. The first crystal structure of the quaternary complex composed of agonist-activated monomeric β₂AR and heterotrimeric G protein [5] opened a new avenue in the study of the interaction between GPCRs and G proteins. These high-resolution complex structures would provide structure basis for the downstream signaling-selective GPCR-targeting drug development. However, the static crystal structure and cryo-EM structure could not provide sufficient information about the dynamic process of G protein coupling to GPCRs. Furthermore, the G_αi coupling is much less efficient than G_αs coupling in most cytological and biochemical analysis making the complex of β₂AR-G_i is less stable than β₂AR-G_s [6]. Therefore, the current research on the mechanism of the interactions between β₂AR and G_αi is limited [7].

G proteins transduce signals from GPCRs to a variety of effector proteins, which might be localized to membrane microdomains with different lipid composition [8]. Membrane association and binding have been proved to play a crucial role in G protein function [9-11]. Direct membrane contact of the photoreceptor G protein heterotrimer has been observed by low resolution (~30 Å) electron crystallography [12]. Although there have been numerous studies of G protein-membrane interactions using mutagenesis, biochemical and functional approaches [13, 14], the molecular basis underlying the interaction of the distinct G proteins subunits with the plasma membrane are still not fully understood [15]. This may in part be due to the fact that the structural heterogeneity of lipid-GPCRs ternary complex hindered the high-resolution structure determination by Cryo-EM or X-ray crystallography [16-18].

Nuclear magnetic resonance (NMR) spectroscopy in solution is one of the key approaches to explore dynamic features, which can be accessed at physiological temperature and with minimal modification of proteins [19-21]. ¹³C-methylated lysine, ¹³CH₃-labelled methionine and cysteine-based modification approach such as 2, 2, 2-trifluoroethanethiol (TET) or 3-bromo-1, 1, 1-trifluoroacetone (BTFA) labelling have been extensively performed for conformational dynamics study of GPCRs [22-24]. Previous spectroscopic strategies have shown that β₂AR adopts multiple conformational states activated by agonist in the presence or absence of G protein [25]. However, the dynamic conformational change of G protein coupling to GPCRs is remained elusive.

Herein, we utilize ¹⁹F NMR spectroscopy to monitor the conformational change of the G_α proteins coupling to the β₂AR. The ¹⁹F atoms are introduced by site-specific incorporation method with high specificity, efficiency and fidelity [26]. Compared with other methods for the fluorine introduction [27-30], site-specific incorporation of a ¹⁹F labelled amino acids into proteins is a powerful tool for monitoring protein conformational changes, protein processing and protein stability studies in vivo and in vitro [19, 31]. Our data show that the apo-form of G_α protein adopts two conformations. Further structural changes were observed when G_α protein interacted with detergent mimicking cell membrane, which confirmed that interaction with the cell membrane is pivotal for G protein activation. The conformation of both G_αs and G_αi proteins were further stabilized by coupling to activated β₂AR. Our studies provided insights into the dynamic process of GPCR-G protein signaling pathway.

To obtain ¹⁹F labelled target proteins, we used two plasmids in the system: One plasmid expressing the Methanococcus Jannaschii tyrosyl amber suppressor tRNA (MjtRNA-Tyr_CUA)/tyrosyl-tRNA synthetase (MjTyrRS) pair, and the other plasmid containing a TAG mutant of the G protein gene (Fig. S1 in Supporting information) [32]. E. coli cells carrying both plasmids are grown in the presence of l-4-trifluoromethyl-phenylalanine (tfmF) (Fig. 1a) to produce the ¹⁹F-labelled G_α protein. Due to the rapid rotation of the CF₃ group, the sensitivity of NMR signals can be improved by tfmF labelling.

We individually introduced unnatural amino acid tfmF to the receptor coupling interface of the G_αs and G_αi that can be used as ¹⁹F NMR reporters to monitor structural changes in response to β₂AR. One-dimensional NMR spectra were collected with the proton coil tuned to the ¹⁹F frequency for 1-3 h to further confirm that the tfmF was successfully incorporated into target proteins. Finally, the ¹⁹F-labelled proteins with high labelling efficiency and friendly thermal stability were selected for the following NMR experiment, such as G_αs376 tfmF, G_αs383 tfmF and G_αi351 tfmF (Figs. 1b-d). All these residues located in the C- terminal of α5 helix, which is a key determinant of GPCR-G protein coupling selectivity [33]. Previous fluorescence lifetime based FRET studies have revealed that both G_αs and G_αi presented two distinct conformations in the apo-state [34]. Consistent with previous studies, our results show that there are two conformations (hereinafter called states S1 and S2) for each tfmF incorporation site (Fig. 1e). In the apo-form, we observed two obvious peaks for G_αs376 tfmF, suggesting the existence of conformational heterogeneity in this location with slow exchange between two conformational states on slow time scale (k < < 1.5 ms⁻¹) at room temperature. Similar to the structural heterogeneity in ¹⁹F NMR spectra of G_αs376 tfmF, G_αs383 tfmF show a broad line shape, derived from the overlap of two signals. The resulting ¹⁹F NMR spectra identify two states of G_αs protein that likely correspond to the free-binding states S1 and S2. Compared to the 1D ¹⁹F NMR spectra of G_αs site-specifically incorporation proteins, there is only one ¹⁹F peak in the 1D NMR spectra of G_αi351 tfmF. The difference in the ¹⁹F NMR spectra of G_αs and G_αi proteins is probably due to the G_αi351 located at the C-tail of α5 helix in G_αi containing a total of 354 amino acids, which is highly flexible and the two free-binding states S1/S2 are in fast exchange timescale resulting only one sharp peak on the NMR spectrum.

Recently, nanodisc has been widely used for structural and functional studies of membrane proteins [35]. Membrane scaffold protein MSP1D1 and the lipid mixture of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-(1′-rac-glycerol) (POPG) were used to assemble the nanodisc with the suitable diameter for titration analysis (Fig. S2 in Supporting information). May due to the invisibility of giant size G_α-nanodisc complex, we could mainly observed the gradually peak intensity decreases of G_α protein titrated with nanodiscs from the ¹⁹F NMR spectroscopy (Fig. S3 in Supporting information). Fortunately, we observed a new broad peak on the G_αs376 tfmF spectrum titrated with nanodisc, indicating the new G_α-nanodisc complex state (Fig. S3a in Supporting information). In order to determine the dynamic process of G_α protein interacting with plasma membrane, we use micelle composed of small molecule detergents mimicking membrane to monitoring the conformational changes of different G_α proteins compared of their apo state. The mixed micelles of n-dodecyl-β-d-maltoside (DDM) and cholesteryl hemisuccinate (CHS), with DDM/CHS (2:1), was gradually titrated into G_α protein. The structural changes that occurred upon the addition of detergent to the G_α protein were performed using ¹⁹F NMR. Interestingly, the ¹⁹F NMR spectrum of G_α protein changes accompany with the detergent concentration increases. Starting with the two weakly peaks on the spectrum of G_αs376 tfmF, addition of the detergent resulted in the appearance of a broad line shape signal peak, indicated that the rate of chemical exchange between S1 and S2 has increased induced by detergent monomers (Fig. 2a). Subsequent addition of an excess of the detergent made the peak shape of the resonance sharp and the signal stronger. Therefore, it can be reasonably deduced based on the peak shape and intensity, there covered another new state of G_αs376 tfmF apart from S1-S2, which we referred as state S3. Since the new broad peak of G_α-micelle complex state shows the same chemical shift with G_α-nanodisc complex, we consider that the mixed micelles could mimic the native plasma membrane in our case (Fig. S3b in Supporting information). Similarly, as detergent is gradually added to the G_αs383 tfmF, the broad line shape getting sharper at low detergent concentration, indicating faster chemical exchange of state S1 and S2. The binding of detergent micelles promotes the generation of equilibrium between S1 and S2, driving the G_αs from S1-S2 towards S3 (Fig. 2b). In addition, observation of the ¹⁹F NMR signals originating from G_αi provided similar results to those obtained from observation of G_αs. In contrast to G_αs376 tfmF and G_αs383 tfmF, addition of the detergent micelles resulted in a pronounced shift towards S3 for G_αi351 tfmF, making position of the NMR peaks is clearly separated from the S1-S2 (Fig. 2c). Collectively, detergent will accelerate the rate of equilibrium between S1-S2 and induce a new state S3, which is associated with the membrane interaction. This state probably represents the transition state before the G protein and receptor interaction.

To characterize the dynamic properties of different G proteins coupling to GPCR, we acquired ¹⁹F NMR spectra of G_α proteins bound to β₂AR in the presence of the BI167107, an ultra-high-affinity full agonist that increase basal G_αs coupling activity [36]. Due to the activation of the β₂AR results in TM6 movement that lead to a receptor-binding site for intracellular effector or regulatory proteins, a given concentration of agonist is added throughout the experiment [37]. We observed spectral changes in G_α protein upon detergent binding, and further changes were observed following β₂AR coupling (Fig. 3, Fig. 4). No significant chemical shift changes of G_αs376 tfmF was observed upon β₂AR binding. However, the interaction with β₂AR slow down the global tumbling of G_αs376 tfmF leading to a slightly broad line shape of the NMR peaks (Fig. 4a). This shows that G_αs376 tfmF has produced a receptor-binding conformation, named state S4. Moreover, the G_α protein and β₂AR were indeed coupled into a complex which was verified through size exclusion chromatography (SEC) (Fig. S4 in Supporting information).

In the presence of the β₂AR, the signal peaks representing S1-S2 of G_αs383 tfmF disappear almost completely. The newly up-field peak almost completely replaces the downfield peak, illustrated the population of states shifted towards the binding state S4 at the expense of S1-S2 and S3. Similar to G_αs383 tfmF, upon coupling to β₂AR we see the G_αi351 shifted the S1-S2 equilibrium towards the S4 state by an increment in intensity comparing with the G_αi-detergent binding state (Fig. 4c). However, same amount of β₂AR and detergent can only induce a small fraction of β₂AR-G_αi complex (~36% based on the integrated peak area). This may be due to the dynamic nature of the G_αi351 tfmF that a small fraction of G_αi351 existing in a state is capable of coupling with the β₂AR, which confirmed the weak coupling specificity of β₂AR-G_αi complex. In summary, except for the widths and intensity of the NMR peaks, chemical shift of G_α protein remain practically unchanged after coupling to the β₂AR compared to detergent binding state and confirmed that there may be a shallow energy barrier between S3 and S4, allowing rapid chemical exchange of the two state populations. In this study, the relatively weak interaction between G_αi351 and β₂AR was detected using ¹⁹F NMR spectroscopy (Fig. 4c). Compared with G_αs376 and G_αs383, less population of G_αi351 could form complex S4 states with β₂AR. These differences in one-dimensional NMR spectra consistent with the result of a previous study that β₂AR can interact with both G_αi and G_αs and it may have a preference over coupling G_αs.

As noted above, the tfmF was site-specific incorporated into G_α proteins with high fidelity and used to monitor G protein-membrane interactions and G protein-receptor binding conformational changes. In 1D NMR, each peak defines a given conformation or state. Our data suggests that G_α proteins adopt allosteric conformational changes upon interacting with membrane and coupling with GPCRs (Fig. S5 in Supporting information). One-dimensional ¹⁹F NMR studies provide evidence for conformational heterogeneity as we observe more than one peak for F376, I383 in apo G_αs and C351 in apo G_αi. Previous study showed that the G_α protein was anchored on the cell membrane through palmitoylated post-translational modification. Based on our data, the C-terminal α5 helix also involve in the G_α-membrane interaction. The ¹⁹F NMR spectra of G_αs and G_αi in the presence of detergent may represented the membrane binding state of G_αs and G_αi, which is ready to coupling to GPCR. The membrane interactions with G_α simulated by nanodisc or micelle both induce a transition from the equilibrium state of the two free conformations to the membrane-binding state, which may be similar to the transition state before the G protein coupling to the receptor. Combine the results of G_α-β₂AR coupling, we suspect that the interaction with membrane may be indispensable for the activation of different G_α proteins. These data provide evidence for the importance of G protein and membrane interaction for GPCR coupling. Additionally, we captured the protein conformational changes and differences in the interactions between β₂AR -G_αi and the β₂AR-G_αs complex by monitoring of the ¹⁹F NMR chemical shifts. The ¹⁹F NMR may thus prove to be a convenient and effective tool for the studies of dynamic process of GPCR-G protein signaling pathway.

Declaration of competing interest

The authors declare no conflict of interest.

Acknowledgments

This work was supported by the National Key Research and Development Project of China (Nos. 2019YFA0904100 and 2017YFA0505400), the National Natural Science Foundation of China (Nos. 22077117 and 31971152) and the USTC Research Funds of the Double First-Class Initiative.

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2021.07.042.