Strike a Pose: Gα_q Complexes at the Membrane

Abstract

The heterotrimeric G protein Gα_q is a central player in signal transduction, relaying signals from activated G protein-coupled receptors (GPCRs) to effectors and other proteins to elicit changes in intracellular Ca²⁺, the actin cytoskeleton, and gene transcription. Gα_q functions at the surface of the plasma membrane, as do its best characterized targets phospholipase C-β, p63RhoGEF, and GPCR kinase 2 (GRK2). Recent insights into the structure and function of these signaling complexes reveals several recurring themes, including complex multivalent interactions between Gα_q, its protein target, and the membrane, that are likely essential for allosteric control and maximum efficiency in signal transduction. Thus, the plasma membrane is not only a source of substrates, but also a key player in the scaffolding of Gα_q-dependent signaling pathways.

Signaling by heterotrimeric G protein Gα subunits

The proteins involved in GPCR signaling pathways are essential for vision, blood pressure regulation, cardiac contractility, and numerous other processes ¹. Activation of a GPCR by an extracellular signal, such as a photon, hormone, or drug, induces conformational changes that allow the GPCR to function as a guanine nucleotide exchange factor (GEF) for heterotrimeric G protein Gα subunits ², promoting the binding of GTP by Gα. Once activated, Gα·GTP dissociates from the heterodimeric Gβγ subunits, and both components can bind to and activate downstream effectors at the plasma membrane to elicit changes in the concentration of second messengers such as inositol-1,4,5-triphosphate (IP₃) and diacylglycerol (DAG) ³ (Figure 1).

Figure 1. Interplay of Gα_q-mediated signaling processes.

GPCR activation catalyzes the binding of GTP to the Gα_q subunit, leading to its activation and dissociation from the receptor and the Gβγ heterodimer. Gα_q·GTP then interacts with PLCβ or p63RhoGEF, leading to production of the second messengers IP₃ and DAG or activation of RhoA, respectively. GRK2 phosphorylation of G_q-coupled receptors initiates homologous desensitization, and GRK2 binding to Gα_q·GTP sequesters it from its downstream targets, resulting in phosphorylation-independent desensitization. Meanwhile, RGS proteins catalyze the hydrolysis of GTP on Gα_q, leading to its deactivation, and can act on Gα_q while it is in complex with p63RhoGEF or GRK2 ⁵⁰, or block its interactions with PLCβ. In addition to a helix-turn-helix motif (orange elbow), Gα_q must interact with multiple domains of PLCβ and p63RhoGEF for full functionality. In the case of PLCβ, the N-terminal helix (blue rectangle) and the Ras-like domain of Gα_q contact the distal CTD (orange), and the C2 and EF hands domains of the catalytic core, respectively. In the case of p63RhoGEF, the Ras-like domain and C-terminal helix of Gα_q bridge the DH and PH domains of p63RhoGEF. Gα_q must also simultaneously interact with the membrane via palmitoylation sites at its N-terminus. Disordered regions within the proteins, primarily those bearing the lipid modifications of Gα_q, Gβγ, p63RhoGEF, and RhoA to the membrane are shown as black lines. Palmitoyl and geranylgeranyl groups are modeled as orange sticks embedded within the membrane.

The Gα subunit consists of a Ras-like domain with a large α helical domain insertion ³. The guanine nucleotide binding pocket is formed at the interface of the Ras-like and α helical domains, where three “switch” regions (SwI-III) adopt unique conformations dependent on the identity of the bound nucleotide. In the GTP-bound state, the switches help coordinate the catalytic Mg²⁺ ion, participate in catalysis, and contribute to the effector binding site (Figure 2a, b). In the GDP-bound state, SwII contributes instead to the Gβγ binding site ³. The N-terminus of Gα is lipid-modified (e.g myristoylated and/or palmitoylated) and can form an extended helix that interacts with GPCRs ², Gβγ subunits ^{4, 5}, and some effectors ⁶.

Figure 2. Structural and functional features of activated Gα_q.

(a) Structure of Gα_q in its activated state. Gα_q (shown as a gray cartoon with green β strands and red switch regions) is trapped in a GTP hydrolysis transition-like state upon addition of GDP and AlF₄⁻, (blue ball and sticks). Mg²⁺ is shown as a black sphere. The effector binding site is formed in part by the cleft between SwII and the α3 helix. In most structures of activated Gα_q, ~30 N-terminal and ~5 C-terminal residues are disordered (not shown). (b) Accessible surface of activated Gα_q, in the same orientation as in (a). Residues that primarily contribute to the effector binding site, as defined by the contacts made by the helix-turn-helix motif found in PLCβ and p63RhoGEF, are shown in orange. Note that in each case interactions are also formed with regions of Gα_q outside this cleft, which are likely critical for high affinity binding. Residues that primarily contribute to regulator of G protein signaling (RGS) binding, as defined by the structure of the RGS2–Gα_q complex ⁵², are shown in purple. Some residues, particularly in SwII, contribute to both interaction sites. The asterisk demarks the position of a hydrophobic pocket between the SwII and α3 helices that is exploited by the leucine in the ALXXPI motifs of PLCβ and p63RhoGEF. (c, d, e) RGS proteins (green) and effectors (cyan) can often bind their Gα targets (surface representation as in panel b) simultaneously, such that they allosterically regulate the interaction and function of each other. Ternary complexes formed between p63RhoGEF (c) or GRK2 (d) with the RGS2–Gα_q complex have been confirmed biochemically ⁶⁴. The RGS9–Gα_i/t–PDEγ complex (e) is the only ternary complex that has been structurally characterized ⁶⁵. To create the models shown in panels c and d, the RGS2–Gα_q complex ⁶⁷ was superimposed on the p63RhoGEF ¹¹ and GRK2 ¹⁴ complexes, respectively, using the Gα subunit for alignment. Only the helical bundle subdomain of the GRK2 RH domain is shown for clarity. Yellow regions correspond to helical elements that engage the effector binding site as defined in panel b.

The focus of this review is on the interactions formed by activated Gα_q, which represents one of the four canonical subfamilies of Gα subunits (Gα_s, Gα_i/o, Gα_q/11 and Gα_12/13). The other members of this subfamily include Gα₁₁, Gα₁₄, and Gα_15/16. Gα_q subfamily members are activated by GPCRs that respond to hormones and neurotransmitters such as norepinephrine, endothelin, and glutamate ⁷, and play important roles throughout the cardiovascular and nervous systems. Classically, Gα_q activates phospholipase C-β (PLCβ), which hydrolyzes phosphatidylinositol-4,5-bisphosphate (PIP₂) to generate the second messengers inositol-1,4,5-triphosphate (IP₃) and diacylglycerol (DAG), leading to increases in intracellular Ca²⁺ and activation of protein kinase C ^{8, 9}. More recently, it was discovered that Gα_q also activates the guanine nucleotide exchange factor p63RhoGEF (ARHGEF25) and its homologs Trio and Duet ^10–13, thereby linking GPCRs to the activation of RhoA and control of the actin cytoskeleton. In addition, Gα_q binds to GPCR kinase 2 (GRK2) ¹⁴, an enzyme that initiates homologous desensitization by phosphorylating activated GPCRs ¹⁵. However, GRK2 is not activated via binding Gα_q. Instead, it seems to sequester Gα_q from its other targets in a form of phosphorylation-independent desensitization ^16–19 (Figure 1). Not all members of the Gα_q subfamily interact with these targets in the same manner. Gα₁₆ does not bind appreciably to GRK2 ²⁰, and binds to but does not activate p63RhoGEF ²¹

Structural and functional studies of activated Gα_q in complex with PLCβ3 ^{6, 22}, p63RhoGEF ¹¹, and GRK2 ¹⁴ reveal some general themes. PLCβ, p63RhoGEF, and GRK2 all must interact with the membrane and, in some cases, with other peripheral membrane proteins for optimal catalytic efficiency. A specific orientation of these enzymes with respect to the membrane plane is essential for their function. All three enzymes engage the effector binding site of Gα_q using exposed hydrophobic residues on helical elements. Unexpectedly, comparison of the likely membrane-bound orientation of Gα_q in complex with PLCβ, p63RhoGEF, and GRK2 reveals that Gα_q adopts a similar orientation with respect to the membrane plane in each assembly (Figure 3). It is also clear that the membrane itself plays one or more active roles in the function of these complexes: it can enhance the affinity of signaling components by co-localization, serve as a scaffold for the assembly of high order assemblies, and/or optimize the orientation of these proteins for maximum signaling efficiency.

Figure 3. Gα_q complexes modeled at the cell membrane.

(a) The Gα_q–PLCβ3 complex. PLCβ3 (cyan) is oriented such that its active site is in close proximity to the lipid bilayer. The catalytic Ca²⁺ (black sphere) defines the location of the active site. Residues that constitute the hydrophobic ridge, which may anchor the active site in the membrane, are shown in green spheres, and the conserved basic region of the distal CTD, thought to form the primary membrane binding site, is shown in blue spheres. The helix-turn-helix (Hα1/Hα2) in the proximal CTD of PLCβ (yellow) docks to the effector binding site of Gα_q (shown as in Figure 2b). The end of the proximal CTD and start of the distal CTD are denoted by blue asterisks. The palmitoyl groups at the N-terminus of Gα_q are shown as red sticks inserted within the membrane (grey rectangle). All disordered regions, including the X–Y linker (black) proposed to be displaced upon interaction with the membrane, are shown as dashed lines. To generate this model, the distal CTD in complex with the Gα_q N-terminal helix and the Gα_q–catalytic core complex from PDB entry 4GNK were each independently docked at a shared surface. The Gα_q N-terminus-distal CTD complex was then rotated in the plane of this surface to allow the N-terminal helix to join the Ras-like domain of Gα_q. (b) The Gα_q–PLCβ3 complex as viewed from the perspective of the putative membrane surface. (c) The Gα_q–p63RhoGEF RhoA–complex. The DH (light green) and PH (cyan) domains form the primary binding sites for RhoA (yellow accessible surface with pink switch regions) and activated Gα_q (shown as in Figure 2b). As in PLCβ, the p63RhoGEF catalytic core has an extension that forms a helix-turn-helix (yellow) which docks in the Gα_q effector binding site. The orientation of Gα_q–p63RhoGEF–RhoA complex at the membrane is less well determined, as the membrane binding determinants are all found in disordered elements: the C-terminus of RhoA (13 disordered residues containing a polybasic region ending in a geranylgeranyl group), the N-terminus of p63RhoGEF (three palmitoylation sites positioned ~125 residues N-terminal to the DH domain), and the N-terminus of Gα_q (two palmitoylation sites ~25 residues N-terminal to the beginning of the Ras-like domain). (d) The Gα_q–p63RhoGEF–RhoA complex viewed from the perspective of the membrane. Red asterisks mark the last ordered residues at the C-terminus of RhoA and N-terminus of Gα_q, which are attached to flexible sequences that are lipid modified. (e) The Gα_q–GRK2–Gβγ complex. The RH domain of GRK2 (cyan) binds to the effector binding site of Gα_q using a helix (yellow) analogous to the first helix of helix-turn-helix motif found in PLCβ and p63RhoGEF. GRK2 binds to the membrane primarily via its PH domain, which binds to both PIP₂ and heterotrimeric Gβγ subunits (blue and green), which are prenylated. GRK2 is modeled in an active configuration with an ordered N- terminal helix (demarked by “N”) based on a recent structure of GRK6 ⁴⁶. This helix has amphipathic character and is believed to interact with either activated GPCRs in the membrane or with the phospholipid bilayer itself ⁴⁵. (f) The Gα_q–GRK2–Gβγ complex viewed from the perspective of the membrane surface.

Scaffolding interactions in the activation of PLCβ by Gα_q

PLCβ enzymes comprise two major regions: a catalytic core responsible for PIP₂ hydrolysis and a ~400 residue C-terminal extension required for Gα_q regulation and membrane association (Figure 3a, b). The catalytic core is conserved across all PLC isozymes ²³ and contains an autoinhibitory element called the X–Y linker ²⁴, which is immediately adjacent to the active site. This linker is poorly conserved but contains highly acidic stretches of amino acids in human PLCβ isoforms. In all PLCβ crystal structures reported to date, the C-terminal residues of the linker are ordered and occlude the active site ⁹. It has been proposed this linker is dislodged by electrostatic repulsion when the PLCβ catalytic core is brought in close proximity to the negatively charged inner leaflet of the plasma membrane ²⁴. Also adjacent to the active site is the so-called hydrophobic ridge, a series of loops with exposed hydrophobic residues believed to insert into the membrane during catalysis ²⁵ (Figure 3a, b).

The C-terminal extension can be subdivided into two domains. The first 45 amino acids form the proximal C-terminal domain (CTD), comprising the primary Gα_q binding site and an autoinhibitory helix (Hα2′). In the absence of Gα_q, the first 25 amino acids of the proximal CTD are disordered ²⁶. Upon Gα_q binding, these residues form a helix-turn-helix (Hα1/Hα2) that docks into the effector binding site of Gα_q via a conserved motif, ALXXPI (where X represents any amino acid) ^{6, 22}. An additional contact between Gα_q and PLCβ is made by an asparagine residue within the catalytic core that interacts with the active site of Gα_q to increase the rate of GTP hydrolysis. Thus, PLCβ functions as its own GTPase activating protein (GAP), regulating the duration of its own signal ²². Under basal conditions, the Hα2′ helix of the proximal CTD interacts with a conserved cleft within the PLCβ catalytic core. Disruption of the Hα2′-core interaction in human PLCβ3 increases basal activity up to ~50–fold over wild-type and decreases the efficacy of Gα_q activation, suggesting that Gα_q activation of PLCβ includes an allosteric component ²⁶.

The last ~300 residues of the C-terminal extension form the distal CTD, a long coiled-coil domain ^{6, 27} required for maximum basal and Gα_q-stimulated activity and for membrane association ^{22, 26, 28–31} (Figure 3a, b). The distal CTD features two conserved surfaces. The first is an extended positively charged region formed by clusters of basic residues along one face of the domain. Mutation of these residues decreases Gα_q-stimulated activity and particulate association, suggesting they represent the primary membrane binding surface ^{28, 31, 32}. The second surface contains a cluster of conserved hydrophobic residues that binds the N-terminal helix of Gα_q in the crystal structure of the Gα_q–PLCβ3 complex ⁶. Abrogation of this interaction diminished maximal Gα_q-stimulated activity, as did loss of Gα_q palmitoylation.

Direct interactions between PLCβ and Gα_q increases the rate of PIP₂ hydrolysis by up to ~80–fold over basal. However, although Gα_q is palmitoylated at its N-terminus and localized at the membrane ³³, there are no quantitative studies demonstrating that interactions with Gα_q increase PLCβ membrane association ^{34, 35}, and thus membrane recruitment on its own does not provide a holistic model for activation. Instead, it has been proposed that in the absence of Gα_q, PLCβ enzymes are in equilibrium between cytosolic and membrane-associated populations ^{6, 35}, which may be regulated in part through direct interactions between the distal CTD and the catalytic core, as observed in crystals and in solution by cryo-electron microscopy ⁶. Gα_q binds to the Hα1/Hα2 helix-turn-helix and displaces the auto-inhibitory Hα2′ helix from the catalytic core, allosterically activating PLCβ. Meanwhile, the distal CTD simultaneously interacts with the membrane and with the N-terminal helix of Gα_q in a manner that helps the Gα_q–catalytic core complex adopt a specific orientation with respect to the membrane surface ⁶. The resulting electrostatic repulsion between the membrane and acidic regions in the X–Y linker helps to open the active site, allowing insertion of the hydrophobic ridge into the membrane and binding of the substrate PIP₂ (Figure 3a). Thus, the distal CTD and the membrane can be thought of as scaffolds that not only help optimize the orientation of the Gα_q–catalytic core complex, but also allosterically activate it. Higher order scaffolding interactions are known to occur. Complexes that include GPCRs can be mediated by proteins such as WD40 repeat protein 36 (WDR36), which has been reported to interact with Gα_q, PLCβ and the thromboxane A₂ receptor ³⁶, and well-characterized scaffolding proteins such as InaD and Na⁺-H⁺ exchange regulatory factor (NHERF) ³⁷.

Despite recent progress in understanding the molecular mechanisms of Gα_q-dependent activation, many questions remain. It is unclear whether membrane binding induces conformational changes within the catalytic core of PLCβ and whether displacement of the segment of the X–Y linker that occludes the active site also contributes to structural rearrangements. The mechanism by which Hα2′ mediates autoinhibition is currently unknown. Finally, given its similarity to Bin–Amphiphysin–Rvs domains, the distal CTD may play other roles in regulating Gα_q-stimulated activity, such as by altering the local structure of the membrane itself ⁶.

Membrane scaffolding of the Gα_q–p63RhoGEF–RhoA signaling complex

Although Gα_q has been reported to interact with other RhoGEFs ^38–40, p63RhoGEF is the best characterized direct link between G_q-coupled receptors and RhoA ¹⁰. p63RhoGEF plays a role in neurite outgrowth ^{41, 42}, and in smooth muscle cells it contributes to muscle contraction by promoting GTP binding by RhoA in response to activation of angiotensin II, endothelin-1 and phenylephrine-responsive GPCRs ^43–45. Recent studies also suggest that p63RhoGEF plays an important role in breast carcinoma cell chemotaxis ^{46, 47} and osteoblast diffentiation ⁴⁸. Two related RhoGEFs, Trio and Duet, have also been found to activate RhoA in response to binding Gα_q ^{11, 13}. In C. elegans, the Trio homolog is required for proper acetylcholine-stimulated locomotion and egg-laying ¹². Interestingly, Gα₁₆ binds to, but does not activate p63RhoGEF, which can lead to inhibition of Gα₁₆-mediated activation of PLCβ ²¹. Conversely, Gα₁₆ could also inhibit the activation of p63RhoGEF mediated by other Gα_q subfamily members, although this remains to be demonstrated experimentally.

As a member of the Dbl family of RhoGEFs, the catalytic core of p63RhoGEF contains the characteristic Dbl homology (DH) domain, which harbors the principal binding site for RhoA, immediately followed by a pleckstrin homology (PH) domain that often plays a regulatory role in these enzymes. In the case of p63RhoGEF, the PH domain inhibits the catalytic activity of the DH domain, but the molecular mechanism for this is not fully understood. Binding of activated Gα_q leads to allosteric activation of p63RhoGEF, the most obvious manifestation of which is an increase in the affinity of p63RhoGEF for apo-RhoA ⁴⁹.

The crystal structure of the Gα_q–p63RhoGEF–RhoA complex reveals that Gα_q binds in a manner that bridges the DH and PH domains of p63RhoGEF ¹¹ (Figure 3c, d). Like PLCβ, the DH/PH catalytic core of p63RhoGEF features a C-terminal extension, also conserved in Trio and Duet ¹³, which forms a helix-turn-helix element that docks into the effector-binding site of Gα_q. Strikingly, this extension contains the same ALXXPI motif used in PLCβ isozymes to bind the effector binding site on Gα_q. Mutation of the conserved leucine in this motif (L859 in human PLCβ3, L475 in human p63RhoGEF) is sufficient to eliminate Gα_q binding ^{11, 13, 22, 31}. Consequently, PLCβ and p63RhoGEF cannot simultaneously interact with a single Gα_q molecule ¹⁰, and would therefore appear to represent competitive pathways in cells where both enzymes are expressed. However, in C. elegans, the activity of both enzymes is required for proper function of the neuromuscular junction, suggesting there may be distinct pools of Gα_q scaffolded in close proximity to each effector enzyme at the cell membrane ¹².

The interactions between the C-terminal extension of p63RhoGEF and Gα_q are further stabilized by bridging interactions of the DH domain with the Ras-like domain of Gα_q. Mutation of residues involved in this latter interface decreases Gα_q-mediated activation but had little affect on affinity ¹¹, including those converting positions at the C-terminus of Gα_q to their equivalents in Gα₁₆ ⁴⁹. These results suggest that the bridging interaction of Gα_q serves to constrain the relative orientation of the DH and PH domains ⁴⁹. Indeed, comparison with structures of closely related DH/PH domains indicates that the relative orientation of the DH and PH domains in the Gα_q–p63RhoGEF complex is atypical. Thus, in the basal state, either the p63RhoGEF PH domain occludes the RhoA binding site, or the linker helix connecting the DH and PH domains, which forms an important portion of the interface with apo-RhoA (Figure 3d), is not in an optimal configuration. When the Gα_q–p63RhoGEF–RhoA complex is positioned such that all the lipid modifications (geranylgeranyl in RhoA, and palmitoyl in Gα_q and p63RhoGEF) are positioned to engage the membrane (Figure 3c), the orientation of Gα_q is similar to that observed in the Gα_q–PLCβ complex ^{6, 22}. However, unlike PLCβ, the entire complex is likely offset from the membrane due to the length of disordered elements bearing the lipid modifications (Figure 3c, d). Such an offset is consistent with the lack of an obvious basic electrostatic surface on the Gα_q–p63RhoGEF–RhoA complex that would interact with the negatively charged inner leaflet of the membrane.

In cells, p63RhoGEF is constitutively localized to the cell membrane by virtue of three palmitoylated cysteines at its N-terminus. Mutation or deletion of these cysteines abrogates membrane association and activity of p63RhoGEF in cells, but not in vitro ⁵⁰. This result indicates the interaction between p63RhoGEF and RhoA is not of sufficient affinity in vivo to drive activation in the absence of co-localization mediated by the membrane. A recent study showed that GEFT, a splice variant of p63RhoGEF that lacks N-terminal palmitoylation sites, is not membrane associated and does not efficiently couple to Gα_q in cells unless it is artificially targeted to the plasma membrane. This study further concluded that the membrane plays an important role in increasing local concentration, but not encounter times ⁵¹. Lipid modifications in the p63RhoGEF complex might also drive these proteins into raft-like domains, where their effective affinity would be even higher and result in highly localized signal amplification ⁵².

GRK2: a non-canonical Gα_q interaction partner and G protein scaffold

The catalytic core of GRK2 consists of an N-terminal RGS homology (RH) domain and a kinase domain ⁵³. The GRK2 RH domain binds activated Gα_q with high affinity ^{14, 20, 54, 55}, but this ability seems unique to the GRK2 subfamily of GRKs. GRK2 also has a C-terminal PH domain that binds to Gβγ subunits and PIP₂ (Figure 3e, f), which are essential for GRK2 membrane recruitment ⁵⁶ and activity in vivo ⁵⁷, but not in vitro ⁵⁸, echoing the role of N-terminal palmitoylation in p63RhoGEF. Crystal structures of GRK2 alone and in complex with Gβγ, sequence conservation, and biochemical studies identified an extended basic surface along one face of GRK2 that likely interacts with the membrane ^{53, 57, 59}, which includes the prenyl modification of Gβγ, the lipid binding surface of the GRK2 PH domain, and the N-terminal helix of GRK2, which has been proposed to either interact directly with activated GPCRs ⁶⁰ or phospholipid bilayers ⁶¹ (Figure 3e, f). Recently, sum frequency generation (SFG) spectroscopy has provided experimental support for the proposed orientation of the GRK2–Gβγ complex on model phospholipid bilayers ⁶².

The Gα_q interaction with GRK2 shares several structural similarities with those formed with PLCβ and p63RhoGEF. The RH domain of GRK2 binds the effector binding site of Gα_q using a helix analogous to the Hα1 helix of PLCβ3 and the first helix of the p63RhoGEF C-terminal extension ¹⁴. Although it lacks an ALXXPI motif, a conserved leucine (human GRK2 L118) plays a structurally equivalent role to the leucine found in this motif and is likewise essential for Gα_q binding ⁵⁴. There are significant differences in the GRK2 RH domain interface that confer its specificity among Gα_q subfamily members²⁰. Gα_q-Tyr261 forms multiple hydrogen bonds with GRK2 that cannot be formed by the leucine at this position in Gα₁₆, and mutation of Gα_q-Tyr261 to leucine abrogates activation-dependent binding to GRK2 ¹⁴. Thus, the GRK2 RH domain can be thought of as a scaffolding domain that presents the associated protein kinase and PH domains to the membrane in a manner that allows simultaneous binding of activated receptors, Gβγ subunits, and Gα_q (Figure 3e, f). For its part, the membrane co-localizes these proteins, thereby enhancing their effective affinity, and allosterically activates GRK2 via the binding of PIP₂ to the PH domain. When the Gα_q–GRK2 complex is positioned in its expected membrane binding orientation, Gα_q once again adopts a similar orientation to that observed in the PLCβ and p63RhoGEF complexes.

A conserved pose in Gα_q signaling

In all published structures of its protein complexes, the predicted orientation of activated Gα_q at the membrane is similar (Figure 3). Its position in the p63RhoGEF complex is the most distinct, but in this case there is a greater degree of uncertainty about the membrane orientation of the complex. Is this common pose a coincidence, or is there a functional basis? All the substrates of known Gα_q targets are integral membrane components (e.g. PIP₂, RhoA, or GPCRs). The N-terminus of Gα_q is disordered in the absence of a protein interaction partner and palmitoylated, which allows the Ras-like domain of the Gα_q core to remain tethered to the membrane following activation and provides flexibility for Gα to engage its targets in a manner that does not interfere with their ability to interact with the membrane, their substrates, or other peripheral or integral membrane protein partners. Indeed, PLCβ, p63RhoGEF, GRK2 and their complexes must all engage the membrane via multiple sites for efficient signal transduction, and having Gα_q offset from the membrane and on the periphery of the complex likely facilitates these interactions. It is also possible that the pose of Gα_q also serves to accommodate higher order scaffolding interactions.

PLCβ is unique among the Gα_q binding partners in that it has significant intrinsic GAP activity towards Gα_q. However, the interactions of p63RhoGEF and GRK2 with Gα_q do not preclude the binding of RGS proteins like RGS2, which are also membrane localized ⁶³ and can readily fit in the space between the membrane surface and these complexes when docked on Gα_q ⁶⁴ (Figure 2c, d). Thus, the orientation of Gα_q in these latter complexes may also permit the cell to regulate the duration of these interactions, analogous to how RGS9 modulates the duration of the Gα_t–phosphodiesterase γ (PDEγ) complex during phototransduction ⁶⁵ (Figure 2e). In the case of the Gα_q–GRK2 complex, the binding of RGS proteins could also contribute to phosphorylation-independent inhibition of Gα_q signaling, wherein GRK2 not only sequesters activated Gα_q, but also presents it to RGS proteins for rapid deactivation. In the case of PLCβ, RGS proteins such as RGS2 likely regulate signaling by serving as an effector antagonist ⁶⁶ because the PLCβ and RGS2 binding sites on Gα_q extensively overlap ⁶⁷.

Concluding remarks

As the Gα_q complexes described above reveal, the interactions between G protein, effector, and membrane can be quite complex, as befitting pathways of central importance to cell physiology. The importance of the membrane in these interactions is evident, whether it simply supplies the enzyme substrate or regulatory lipids for the target enzyme, enhances the local concentration of the interacting groups, or helps to orchestrate an allosteric change in the Gα_q target. Yet the membrane is clearly not the only player. In the future it will be important to describe the molecular basis for how scaffolding proteins such as WDR36 organize the Gα_q signaling pathway in a manner that also promotes efficient interactions with GPCRs and/or downstream second messenger targets such as protein kinase C ^{36, 37}. It is possible that such higher order scaffolding complexes also help to segregate the activation of distinct pools of Gα_q effectors, as suggested by the requirement for both p63RhoGEF and PLCβin proper locomotion and egg laying in C. elegans ¹² (Figure 1). The complexity of these interactions, however, allows for many layers of regulation and many opportunities for selective therapeutic intervention. For example, the binding site for the Hα2′ helix on the PLC catalytic core may represent a allosteric regulatory pocket unique to the PLCβ isoforms ⁹, and compounds that inhibit the palmitoylation of p63RhoGEF will likey inhibit chemotaxis in cancer cells ⁵¹.

The membrane orientation of other Gα subfamily complexes, and consequently the role of the membrane in regulating their activity, remains poorly understood. This gap in our knowledge occurs because the membrane binding determinants in crystal structures of their complexes are typically either truncated and/or not resolved. For example, the role of the transmembrane domains of mammalian adenylyl cyclase (AC) remains controversial ⁶⁸ despite examples of accessory domains in ACs playing a critical role in dictating the conformation of its catalytic core ⁶⁹. The orientation of the Gα_s–adenylyl cyclase catalytic core structure ⁷⁰ relative to its integral membrane domains and the functional consequence of this orientation is not known. Such knowledge will be essential for a comprehensive understanding of how scaffolding interactions, mediated by the membrane and/or other proteins, help dictate the strength, duration, and fidelity of GPCR signaling.

• Gα_q and its target proteins signal in a highly membrane-dependent manner

• The membrane plays a key role in organizing and regulating high order Gα_q assemblies

• Gα_q adopts a similar pose with respect to the membrane when bound to structurally diverse proteins

• The orientation of Gα_q is enforced by multivalent interactions with its targets