Non-canonical Golgi-compartmentalized Gβγ signaling: mechanisms, functions and therapeutic targets

Abstract

G protein Gβγ subunits are key mediators of G protein-coupled receptor signaling under physiologic and pathological conditions; their inhibitors have been tested for treatment of human diseases. Conventional wisdom is that the Gβγ complex is activated and subsequently exerts its functions at plasma membrane. Recent studies have revealed non-canonical activation of Gβγ at intracellular organelles, with Golgi apparatus being a major locale, via translocation or local activation. Golgi-localized Gβγ activates specific signaling cascades and regulates fundamental cell processes, such as membrane trafficking, proliferation and migration. More recent studies have shown that inhibiting Golgi-compartmentalized Gβγ signaling ameliorates cardiomyocyte hypertrophy and prostate tumorigenesis, implicating new therapeutic targets. Herein, we will review novel activation mechanisms and non-canonical functions of Gβγ at the Golgi, and discuss potential therapeutic interventions by targeting Golgi-biased Gβγ-directed signaling.

Gβγ-mediated signaling as a potential therapeutic target

Gβγ subunits of heterotrimeric G proteins are important mediators for the signaling and functions of G protein-coupled receptors (GPCRs) under physiological and pathological conditions. The diversity, post-translational modifications such as prenylation (see Glossary), structures and downstream effectors of Gβγ subunits have been extensively reviewed elsewhere [1–6]. Gβγ-mediated signaling is a therapeutic target for human diseases and several Gβγ inhibitors, such as the C-terminal portion of G protein-coupled receptor kinase 2 (GRK2ct) and the small molecule inhibitors gallein and M119, have been tested for the treatment of heart failure, hypertension, addiction, pain and cancers [7,8].

After activation by GPCRs and dissociation from Gα subunits, the Gβγ complex is conventionally considered to remain plasma membrane (PM)-bound and exert its functions at the cytoplasmic surface of the PM [9–11]. A body of emerging evidence indicates that free Gβγ may translocate from the PM to intracellular compartments, particularly the Golgi [12–24]. More recent studies have also shown that Gβγ can be directly activated by Golgi-localized GPCRs [25,26]. Gβγ at the Golgi activates specific signaling cascades, including oncogenic pathways [21], in which the Golgi apparatus provides a spatial platform to coordinate the activation events. Activated Gβγ at the Golgi regulates a number of fundamental cell processes, such as membrane-mediated trafficking and cell proliferation, invasion and migration, and importantly, targeted inhibition of Gβγ signaling at the Golgi ameliorates cardiac myocyte hypertrophic growth [27] and prostate cancer progression [21]. These non-canonical functions of Gβγ at the Golgi compartment have revealed previously unappreciated therapeutic targets. In this article, we will first review novel activation mechanisms of Gβγ at the Golgi and update non-canonical Gβγ functions which are mediated through activating protein kinase D (PKD) or mitogen-activated protein kinases (MAPK). We will then summarize recent research findings indicating that Golgi-compartmentalized Gβγ-mediated signaling is a potential target for therapeutic interventions.

Activation mechanisms of Gβγ at the Golgi apparatus

In the classical GPCR signaling systems, GPCR activation by ligands results in the dissociation of GTP-bound Gα and Gβγ subunits. At the cytoplasmic surface of the PM, active Gα and free Gβγ complex can separately interact with and activate downstream effectors, such as adenylyl cyclases, phospholipases (PLCs), G protein-coupled receptor kinases (GRKs), phosphoinositide 3-kinases (PI3Ks), P-Rex1, and synaptosome-associated protein 25 [3,4,9–11] (Figure 1A). This PM-restricted signaling of G proteins has been challenged by numerous studies demonstrating that both Gα and Gβγ may perform their functions in various intracellular organelles that provide non-canonical spatial domains for the actions of G proteins, with the Golgi being extensively studied [28,29]. In this section, we will review the current understanding of activation mechanisms of Gβγ at the Golgi, involving 1) free Gβγ translocation from the PM to the Golgi and 2) local activation of Gβγ by Golgi-localized GPCRs as revealed recently.

Figure 1. Non-canonical activation mechanisms of Gβγ at the Golgi apparatus.

(A) Gβγ translocation from the plasma membrane (PM) to the Golgi. After G protein activation by GPCRs, PM-associated Gβγ may activate cytosolic effectors (E1) through recruitment or PM effectors (E2) by allosteric activation [3,4]. The Gβγ complex may translocate from the PM to the Golgi to activate non-canonical signaling. (B) Gβγ activation at the Golgi by Golgi-localized GPCRs. Nascent GPCRs may be delivered from the endoplasmic reticulum (ER) to the Golgi via anterograte transport and the PM-localized receptors may be transported to the Golgi via internalization and retrograde transport. Golgi-localized GPCRs are functional to activate heterotrimeric G proteins to release free Gβγ dimers.

Gβγ translocation from the PM to the Golgi

It has been known that after GPCR activation by agonists, Gβγ dimers can translocate from the PM to intracellular organelles, with the Golgi being a dominant locale, and that receptor antagonists can cause Gβγ translocation back to the PM [12–24] (Figure 1A). In a recent elegant study using optical biosensors, Masuho et al. have shown that all theoretically possible 60 Gβγ dimers are able to translocate to various intracellular compartments with different transport kinetics and efficiency which take place rapidly on a timescale from milliseconds to minutes and are mediated through simple diffusion and active transport [24].

Gβγ translocation kinetics is negatively correlated with their PM binding affinity which is determined by Gγ subunits, rather than Gβ subunits. For example, mutation of KK and/or FF in the CT of Gγ3 significantly reduces its ability to associate with the PM, but enhances its translocation to the Golgi, whereas addition of basic and/or hydrophobic residues in the CT of Gγ9 enhances its PM association but inhibits its Golgi translocation [16,23] (Table 1). In addition to interacting with Gα subunits and downstream effectors, Gβγ may also associate with GPCRs [30–32] and the receptor-Gβγ interaction affects Gβγ translocation. Based upon their translocation kinetics, 12 Gγ subunits can be divided into fast, intermediate and slow Golgi-translocating groups (Table 1). The slow translocating Gγ subunits have relatively higher affinity for the PM as their CTs are geranylgeranylated and contain multiple basic and hydrophobic residues. In contrast, the top three fast translocating Gγ subunits are farnesylated with less basic and hydrophobic residues in the CTs, thus have lower affinity for the PM. In all cases studied, Gγ9 is the fastest Golgi-translocating subunit with a t_1/2 of less than 10 seconds in response to activation by different GPCRs in different cell types. In prostate cancer DU145 and PC3 cells, Gγ9 in complex with Gβ1 has the highest magnitude with approximately 80% of total Gγ9 being translocated from the PM to the Golgi, whereas less than 20% of Gγ3 translocated [21]. It should be pointed out that Gβγ translocation studies have heavily relied on the overexpression of fluorescence-labeled Gβγ subunits, which may not reflect the real translocation properties of Gβγ at the endogenous levels.

Table 1.

Gβγ translocation to the Golgi after GPCR activation^a

GPCRs and cell types	Gγ	CT and CAAXb	Prenylation	Translocation rate
α2-AR: CHO, HeLa [14,16,18] β1-AR: HeLa [18,19] β2-AR: CHO [13] C5aR: RAW [18] CXCR4: HeLa, HEK293, PC3, DU145 [16,18,19,21,22] D1R: HeLa [16] D2R: HeLa, HEK293 [18,24] KOR and MOR: HeLa [18] mAChR2: CHO, J774 [12–14] mAChR3: CHO, HeLa, A549 [12–15] mAChR4: HeLa [18] Opsin: HeLa, RAW [18–20,23] PAR2: HEK293 [17]	Gγ9	KEKGG-CLIS	Farnesylation	Fast
	Gγ1	KELKGGCVIS
	Gγ11	KEKGS-CVIS
	Gγ7	KDKKP-CIIL	Geranyl-geranylation	Intermediate
	Gγ5	RPQKV-CSFL
	Gγ12	KDKKT-CIIL
	Gγ13	VEKGK-CTIL
	Gγ10	REPRS-CALL
	Gγ4	REKKFFCTIL	Geranyl-geranylation	Slow
	Gγ8	RDKRLFCVLL
	Gγ2	REKKFFCAIL
	Gγ3	REKKFFCALL

^aAbbreviations: GPCR, G protein-coupled receptor; AR, adrenergic receptor; C5aR, complement component 5a receptor; D1R, dopamine D1 receptor; KOR, κ opioid receptor; MOR, μ opioid receptor; mAChR, muscarinic acetylcholine receptor; PAR2, protease-activated-receptor 2; CT, C-terminus.

^bBasic and hydrophobic residues in the CT are blue- and red-colored, respectively. In the CAAX motif, AA are any two aliphatic amino acid residues and X is any amino acid residue.

Local Gβγ activation at the Golgi

In addition to translocation from the PM, recent studies have revealed another mechanism by which Gβγ subunits are directly activated at the Golgi through Golgi-localized GPCRs. It has been known that some GPCRs, including β₁-AR [33–35], dopamine D1 receptor (D1R) [36], δ opioid receptor (DOR) [37,38] and the orphan receptor GPRC5A (GPCR class C group 5 member A) [25], partially localize at the Golgi. These receptors can locally activate heterotrimeric G proteins and their downstream effectors (Figure 1B). For β₁-AR, its hydrophilic ligands can be transported to intracellular compartments through organic cation transporter 3 (OCT3) and ligand activation of Golgi-localized pool of β₁-AR enhances Gs-mediated cAMP production and PKA activation [34]. Similarly, the activation of Golgi-localized D1R by dopamine which is transported via OCT2 activates Gs and enhances local cAMP production and PKA activation [36]. DOR localizes primarily to the Golgi in neuronal cells and neurons, with a small fraction at the PM, and the activation of DOR at the Golgi and the PM may have discrete signaling outputs and distinct functional consequences [38]. Collectively, these studies demonstrate that Golgi-localized receptors are fully functional. However, it remains not fully understood whether these Golgi-localized receptors are nascent receptors that are transported from the ER where the receptor are synthesized or are delivered from the PM to the Golgi via internalization and retrograde transport. Similarly, although different Gα and Gβγ subunits are found at the Golgi [39], where Golgi-localized G proteins come from is not well known. Nevertheless, all three subunits of G proteins are synthesized in the cytosol on free polysomes and they form heterotrimers on the Golgi before transport to the PM [40].

Direct evidence indicating local Gβγ activation at the Golgi came from two recent publications studying GPRC5A [25] and sphingosine 1-phosphate (S1P) receptors (S1PR) [26]. In the case of GPRC5A, the authors demonstrate that cargo molecules that are ready for export from the trans-Golgi network (TGN) interact with and activate the receptor at the TGN where it in turn activates Gαi3βγ to release the Gβγ complex [25]. In the study of S1PR, vesicular stomatitis virus glycoprotein (VSVG) accumulation in the Golgi can induce S1PR from the PM to the TGN in a GRK2- and β-arrestin 2dependent but ligand-independent fashion. Continuous activation of S1PR at the TGN by S1P generated via sphingosine kinase 1 activates Gβγ [26]. In both studies, local Gβγ activation was shown to play an important role in post-Golgi protein trafficking.

In addition to GPCRs, a number of accessory proteins have been identified to regulate heterotrimeric G protein activation independent of GPCRs [41,42]. Among these proteins, activator of G protein signaling 3 (AGS3) and Gα-interacting vesicle-associated protein (GIV, also known as Girdin) have been well studied and both are localized at the Golgi [43–45]. GIV is a prototypic member of non-GPCR guanine nucleotide exchange factors (GEFs) and was shown to activate heterotrimeric G proteins to release Gβγ and regulate ARF1 activation and function at the Golgi [45]. Gβγ activation at the Golgi achieved by either translocation or local activation has been shown to spatiotemporally regulate receptor signaling and cell functions. There are two novel, non-canonical signaling pathways that are activated by Golgi-localized Gβγ, leading to the activation of PKD and MAPK, which are discussed separately below.

Non-canonical functions of Gβγ at the Golgi through activating PKD

Novel mechanisms of PKD activation by Gβγ at the Golgi

PKD consists of a family of three serine-threonine kinases that are activated by a variety of cellular stimuli, including GPCR agonists [46,47]. In the classic PKD activation pathway, GPCR activation by agonists at the PM activates PLCβ via Gα and Gβγ subunits. PLCβ activation by Gβγ may be mediated through direct interaction of Gβγ with both the pleckstrin homology (PH) and catalytic domains of PLCβ [3] and PLCβ activates PKD via PKC [48] (Figure 2A). Activated PKD disassociates from the PM and translocates to distinct cellular locations (including nucleus) where it activates multiple signal transduction pathways and nuclear transcriptional factors, which work together to modulate many fundamental biological processes [47].

Figure 2. Non-canonical Gβγ signaling and functions at the Golgi through PKD.

(A) In the canonical pathway, phospholipase β (PLCβ) hydrolyzes phosphatidylinositol 4,5-bisphosphate (PIP2) to generate diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (IP3). DAG activates novel PKCs and IP3 mobilizes internal calcium. DAG together with calcium activates conventional PKCs. DAG interacts with and recruits cytosolic PKD to the PM where PKCs phosphorylate PKD, leading to PKD activation. In the non-canonical pathway, Gβγ at the Golgi activates PKD via PLCβ-mediated PIP2 hydrolysis or PLCε-mediated phosphatidylinositol-4-phosphate (PI4P) hydrolysis and PKD regulates a number of cellular function. Insert 1 shows three PKD substrates at the Golgi. Phosphatidylinositol 4-kinase III β (PI4KIIIβ) phosphorylates phosphatidylinositol (PI) in the 4-position to yield PI4P which is enriched in the Golgi. Ceramide transport protein (CERT) mediates ceramide transport from the ER to the TGN that is required for DAG production and PKD activation at the TGN. Oxysterol-binding protein (OSBP) mediates the transfer of sphingomyelin and cholesterol from the ER to the TGN. Both CERT and OSBP are recruited to the TGN via PI4P and their phosphorylation by PKD results in their dissociation from the TGN. Insert 2 represents a mechanism of PKD-mediated gene expression in the nucleus of cardiac myocytes. Histone deacetylases (HDACs) interact with and suppress myocyte enhancer factor 2 (MEF2), thus functioning as negative regulators of gene expression in cardiac hypertrophy. HDAC phosphorylation by PKD promotes HDAC interaction with 14–3-3 proteins and then nuclear exportation, leading to MEF2 activation and hypertrophic gene expression. (B) Specific GPRC5A/Gβγ signaling at the Golgi involved in basolateral cargo transport. (C) S1PR/Gβγ-mediated signaling at the Golgi in cargo TGN-PM traffic. See text for detail.

In addition to the PM, PKD can be non-canonically activated at intracellular organelles. GPCR activation may induce Gβγ translocation to the Golgi where it activates PLCβ or PLCε to generate diacylglycerol (DAG) which further recruits PKD to the Golgi, leading to PKD activation at the Golgi (Figure 2A).The Smrcka group has demonstrated that Gβγ directly interacts with and activates PLCε [49]. They have also shown that PLCε activates PKD through hydrolyzing phosphatidylinositol-4-phosphate (PI4P) to generate DAG (Figure 2A) and that PLC-mediated hydrolysis of PI4P, rather than PIP2, in response to GPCR activation is the major source of DAG production at both the PM and the Golgi [27,50,51]. Some Golgi-localized substrates of PKD have been identified, including phosphatidylinositol 4-kinase III β (PI4KIIIβ), ceramide transport protein (CERT) and oxysterol-binding protein (OSBP) (Figure 2A, insert 1) [52].

Role of Gβγ signaling at the Golgi in the maintenance of Golgi structure and insulin secretion

The Golgi complex has a unique structural feature with a series of flattened membrane-bound cisternae, including cis, medial and trans cisternae; it disassembles and reassembles rapidly during the mitosis of the cell cycle. A series of elegant work from the Malhotra group has demonstrated that free Gβγ complex, but not Gα subunits, can trigger the fragmentation of the Golgi stacks, which involves PKD recruitment to the Golgi and subsequent activation by PKCŋ [53]. The Wedegaertner group has shown that chemically induced Gβγ recruitment to the Golgi activates PKD and causes Golgi vesiculation [54] and that Gβγ regulates mitotic Golgi fragmentation in the late G2 phase which is mediated through PKD and the phosphorylation of GRASP55, a Golgi structural protein [55]. Altogether, these studies indicate an important role for the Gβγ complex in the Golgi fragmentation under physiological conditions.

Gβγ-induced Golgi fragmentation reflects a membrane fission increase at the Golgi and PKD activation by Gβγ dimers at the Golgi is an important step in the fission of PM-destined vesicles from the TGN [56,57]. Saini et al. have demonstrated that Gβγ translocation to the Golgi induced by M3 muscarinic acetylcholine (Ach) receptor (mAChR3) causes Golgi fragmentation through PLCβ and PKD in the human lung epithelial cell line A549 and that enhanced expression of the non-translocating Gγ3 subunit (which likely blocks endogenous Gβγ translocation) markedly inhibited insulin secretion in response to mAChR3 activation in the pancreatic β cell line NIT-1 [15]. These data suggest a possible function for Gβγ translocation to the Golgi in regulated insulin secretion in β cells. Recent studies have also shown that VU0119498, a positive allosteric modulator of mAChR3, enhances Ach-induced insulin secretion from β cells and animal and human islets, and improves glucose tolerance and homeostasis in mouse modes of diabetes, indicating that β cell mAChR3 is potential therapeutic target for diabetes [58,59].

Control of post-Golgi trafficking and sorting of nascent proteins, including GPCRs, by Gβγ signaling at the Golgi

The Golgi complex plays a critical role in the trafficking, processing and sorting of the newly synthesized membrane and secretory proteins and lipids. After being synthesized in the ER, cargo molecules export from the ER, pass through the Golgi cisternae, and reach the TGN. The TGN is considered as a final sorting station where different cargoes may be segregated into different transport carriers (such as clathrin-coated vesicles, pleomorphic tubular-vesicular carriers) that are targeted to distinct final destinations. Gβγ-mediated PKD activation at the Golgi is well described to regulate forward cargo transport. Inhibition of Gβγ localization at the Golgi blocks the TGN-PM transport of VSVG and the secretion of signal sequence horseradish peroxidase [54,60].

As mentioned above, Golgi-localized GPRC5A and S1PR can locally activate Gβγ. Golgi-accumulated cargo molecules activate GPRC5a and subsequent Gαi3βγ and Gβγ. Free Gβγ further activates the PLCβ3-ŋPKC-PKD cascade, leading to the formation of transport carriers for basolateral cargo delivery (Figure 2B) [25]. This TGN-localized GPRC5A/Gβγ-mediated signaling mechanism for basolateral cargo transport was termed autoregulation of TGN export (ARTG) [25]. Similarly, autoregulation of ER export (AREX) was revealed to regulate cargo export from the ER via Gα12 [28], in which binding of unfolded cargo molecules to Sec24, a component of COPII vesicles, activates Gα12 which then activates multiple kinases at the ER exit sites, resulting in cargo export from the ER via COPII vesicles and attenuation of protein synthesis. Similarly, Golgi accumulation of nascent VSVG induces S1PR relocation to the TGN where it activates Gβγ and likely PKD, and facilitates cargo TGN-PM transport (Figure 2C) [26].

GPCR trafficking along the biosynthetic pathway is also subject to regulation by Gβγ. As compared with well-understood internalization and recycling, the molecular mechanisms underlying anterograde transport of nascent GPCRs from the ER to the PM en route through the Golgi are comparatively less well defined [61]. A number of specific structural determinants and regulatory proteins have been identified to regulate the ER-Golgi-PM traffic of GPCRs [62–65]. The Golgi-PM transport of protease-activated-receptor 2 (PAR2) is under control by its activation to induce Gβγ translocation to the Golgi and PKD activation. Inhibition of Gβγ or PKD attenuates the Golgi-PM transport of PAR2, mobilization of intracellular pool of PAR2 to rapidly replenish the PM with functional receptors, and sustained nociceptive signaling [17,66]. These studies indicate that GPCRs may control their own biosynthetic trafficking through activating the Golgi-localized Gβγ-PKD pathway.

It is apparent that Gβγ activation at the Golgi regulates post-Golgi trafficking of different proteins including important signaling molecules like GPCRs. As dysregulated trafficking and mis-targeting of GPCRs are well known to be associated with pathogenesis of a number of human diseases [67,68] and functional rescue using different strategies such as pharmacochaperones has been tested for the treatment of such diseases [69,70], to further study the functional roles of Gβγ-mediated signaling in protein trafficking may provide additional avenues for the treatment of trafficking-associated diseases.

Possible contribution of Gβγ signaling at the Golgi to cardiac hypertrophy

GPCRs and Gβγ are well known to play crucial roles in the development of cardiac hypertrophy and congestive heart failure. GPCRs, particularly β-adrenergic receptors, are actual therapeutic targets of the diseases, and inhibiting Gβγ function, via GRK2ct expression and small molecule inhibitors, has been widely tested for the treatment of cardiovascular diseases in animal models [71–74]. Recent studies have demonstrated that GPCR-mediated Gβγ activation at the Golgi significantly affects cardiomyocyte hypertrophic growth. The Smrcka group has demonstrated that after endothelin-1 (ET-1) stimulation, PLCε, via direct interaction with the scaffold mAKAP (muscle-specific A kinase anchoring protein), localizes to the nuclear envelope where it hydrolyzes PI4P to generate IP2 and DAG in the perinuclear Golgi region in cardiac myocytes [50]. As the Golgi is in close proximity to the nucleus, DAG generated at the Golgi can activate nuclear PKD [50]. Their subsequent studies have shown that Gβγ translocation to the Golgi is necessary for ET-1-dependent PI4P hydrolysis, nuclear PKD activation and hypertrophic growth in neonatal rat ventricular myocytes [27]. Blocking Gβγ function specifically at the Golgi, but not at the PM, inhibits ET-1-stimulated PI4P hydrolysis and nuclear PKD activation in cardiac myocytes, whereas forced Gβγ expression at the Golgi produces opposing effects. It is also interesting to note that the activation of Golgi-localized β₁-AR regulates cardiac hypertrophy through stimulating PLCε-mediated PI4P hydrolysis at the Golgi [33]. These data demonstrate that GPCR activation induces Gβγ translocation to the Golgi where it activates PLCε, leading to nuclear PKD activation and cardiomyocyte hypertrophic growth (Figure 2A). These data indicate that Gβγ-mediated signaling at the Golgi should be considered for drug design for the treatment of cardiac diseases.

Non-canonical functions of Gβγ at the Golgi through activating oncogenic MAPK

MAPK activation by Golgi-localized Gβγ

The MAPK Raf-MEK-ERK1/2 pathway regulates many fundamental cellular processes and its activation by GPCRs involves a variety of signaling molecules. In particular, β-arrestins are well defined to function as scaffolds to coordinate MAPK activation by some GPCRs at endosomes [75–77]. The functional roles of Gβγ subunits in GPCR-mediated MAPK activation are generally considered at the PM [74]. Our recent studies have shown that Gγ9 knockout abolishes ERK1/2 activation by GPCRs, whereas Gγ3 knockout is ineffective [21,22,78]. Importantly, chemically-induced Golgi translocation of different Gβγ dimers directly activates ERK1/2 and this activation can be inhibited by the Gβγ inhibitor gallein, Golgi-targeted GRK2ct (Golgi-GRK2ct) and Golgi disruptors [21]. These data demonstrate a novel function for the Gβγ complex at the Golgi to activate ERK1/2 (Figure 3).

Figure 3. Non-canonical Gβγ signaling to oncogenic ERK1/2 at the Golgi and potential novel therapeutic targets for prostate cancer.

Gβγ at the Golgi activates ERK1/2 via PI3Kγ heterodimer p110γ-p101 and ARF1. Mutations shown are representatives in the molecules. Inhibitors of general Gβγ (GRK2ct and gallein), specific Golgi-localized Gβγ (Golgi-GRK2ct), PI3Kγ (AS-604850), ARF1 (GCA and Exo2), Raf (GW5074) and MEK (U0126 and PD98059), as well as knockout of Gγ9, p110γ and ARF1, have been shown to inhibit MAPK activation in prostate cancer cells and suppress prostate cancer cell phenotypes and prostate tumor progression. * and # indicate genetic mutation and over-expression, respectively, which likely contribute to the enhanced MAPK activation and/or tumorigenesis. EGFR, epidermal growth factor receptor; PDGFRA, platelet derived growth factor receptor α. See text for detail.

Unique Gβγ pathway activating MAPK at the Golgi

PI3Ks are a family of lipid kinases that specifically phosphorylate the inositol moiety of phospholipids at the 3’ position. PI3Ks are involved in regulation of fundamental cellular functions such as cell proliferation, differentiation and migration and are important therapeutic targets of a variety of cancers [79]. PI3Ks are divided into three classes and class I includes PI3Kα, β, γ and δ isoforms which are heterodimers consisting of a catalytic subunit and a regulatory subunit. For PI3Kγ, the catalytic subunit p110γ can form a complex with the regulatory subunit p101 or p87. It is known that Gβγ activates both phosphoinositide 3-kinase (PI3K) γ and β via direct interaction [3,4]. Pharmacological inhibition of PI3Kγ and knockout of p110γ or p101 abrogate ERK1/2 activation induced by GPCR activation and Gβγ expression at the Golgi, whereas knockout of p87 and inhibition of PI3Kβ have no effect [21], indicating that activation of PI3Kγ p110γ-p101 heterodimers is required for ERK1/2 activation by Golgi-localized Gβγ (Figure 3).

ADP-ribosylation factor 1 (ARF1) is a crucial small GTPase involved in the regulation of Golgi structure and membrane trafficking [80]. It has been known for years that some GPCRs can activate ARF1 but the functional consequences remain elusive. In parallel with ERK1/2 activation, Gβγ translocation to the Golgi induced by GPCR activation or inducible targeting activates ARF1 that can be reversed by inhibition and depletion of Gβγ and PI3Kγ. Importantly, ARF1 knockout and inhibition of Golgi-localized ARF1 pool abolish ERK1/2 activation by GPCRs and Gβγ [22]. Collectively, these data reveal a novel signaling pathway in which GPCR activation at the PM induces Gβγ translocation to the Golgi where the Gβγ complex activates PI3Kγ which in turn activates ARF1, leading to MAPK activation (Figure 3). So far, CXCR4, α₂-AR and OR51E2 have been shown to utilize this pathway to activate ERK1/2 [21,22,78]. This Golgi-localized signaling pathway is supported by the facts that the Raf-MEK-ERK pathway can be activated at the Golgi by signaling molecules, such as Ras, and that the Golgi-resident protein Sef functions as a scaffold to bind active MEK/ERK complexes, leading to inhibition of ERK signaling [81]. In addition, both p110γ and p101 subunits of PI3Kγ are partially localized at the Golgi [21]. It is interesting to note that inhibition of AKT, a well-known effector acting downstream of PI3Kγ at the PM does not affect ERK1/2 activation by Gβγ translocation to the Golgi [21]. These data suggest that ERK1/2 activation by PI3Kγ at the Golgi is independent of AKT and that PI3Kγ can activate distinct effectors at different subcellular locations.

GPCR/Gβγ-mediated activation of MAPK in prostate cancer

The hyper-activation of the Raf-MEK-ERK1/2 pathway is directly associated with the pathogenesis of many different types of cancers, and a number of mutations have been identified in different signaling molecules upstream of ERK1/2 (Figure 3), which are partially responsible for super-activation of the MAPK pathway and cancer progression. Importantly, these signaling molecules serve as targets in cancer therapy. In marked contrast, prostate cancer patients do not frequently carry these oncogenic mutations, suggesting that specific mechanisms exist to activate MAPK in prostate cancer [82,83].

A number of GPCRs are highly expressed in prostate cancer and regulate prostate cancer cell phenotypes and prostate tumor progression [84]. These GPCRs include CXCR4 and OR51E2 which activate ERK1/2 via the Gβγ-PI3Kγ-ARF1 pathway at the Golgi in prostate cancer cells [21,22,78]. OR51E2, also known as a prostate-specific GPCR (PSGR), is suggested as a prostate cancer biomarker. Among several Gγ subunits detected in prostate cancer cells, only Gγ9 is highly expressed. ARF1 is also overexpressed in prostate cancer cells and patients [85]. Inhibition of Gβγ, PI3Kγ and ARF1 via genetic and pharmacological approaches significantly suppresses prostate cancer cell phenotypes and tumor progression (Figure 3) [21,22,86,87]. Collectively, these data imply that the exaggerated activation of PM-localized GPCRs and Golgi-localized Gβγ-PI3Kγ-ARF1 cascade represent crucial mechanisms responsible for enhanced MAPK activation in prostate cancer patients (Figure 3) and that targeting this Gβγ-mediated oncogenic MAPK pathway may offer opportunities for prostate cancer therapy as discussed below.

Targeting compartmentalized Gβγ signaling for therapeutics

A number of specific Gβγ inhibitors have been tested for their potentials as drugs for the treatment of human disorders [7,8]. Among these inhibitors, GRK2ct, gallein and M119 specifically block Gβγ signaling without disrupting Gα-mediated signaling and have been widely used to investigate the functional roles of Gβγ in various cellular and animal settings [17,21,22,26,66,74,86,88–93]. A recent study has identified a nanobody (Nb5) that binds Gβγ subunits and inhibits Gβγ-mediated signaling, but has no effect on Gαq- and Gαs-mediated signaling [94].

The Golgi complex undergoes disassembly and fragmentation under stress and pathological circumstances, such as cancer, and is becoming increasingly important as an anticancer target. In prostate cancer, some Golgi-associated proteins are used as biomarkers for diagnosis or prognosis and Golgi dysfunction is related to the pathogenesis. As discussed above, Golgi may function as a spatial platform to regulate GPCR/Gβγ-mediated signaling and functions. In particular, inhibition of Gβγ signaling at the Golgi using Golgi-GRK2ct blocks cardiomyocyte hypertrophic growth [27]. Pharmacological inhibition of signaling molecules in the Gβγ-PI3Kγ-ARF1-MAPK pathway have been shown to inhibit prostate cancer cell proliferation, invasion and migration in vitro and tumor metastasis in vivo [21,22,85,86,95] (Figure 3). It is worth pointing out that expression of Golgi-GRK2ct inhibits ERK1/2 activation by GPCRs and Gβγ at the Golgi, KO of Golgi-translocating Gγ9 (but not non-translocating Gγ3) suppresses prostate tumor progression, and specific inhibition of Golgi-localized ARF1 by GCA and Exo2 (but not PM-localized ARF1 by SecinH3) inhibits prostate cancer cell invasion and migration in vitro (Figure 3) [21,22]. These data provide evidence implicating a possibility of targeting the components of Golgi-localized Gβγ-PI3Kγ-ARF1-MAPK pathway for prostate cancer therapy.

In addition to the Golgi apparatus, Gβγ subunits are also found to signal in the ER, endosomes, mitochondria, and nucleus. GRK2ct, fused to specific intracellular compartment targeting sequences, is the only tool now available to inhibit location-specific Gβγ-mediated signaling, particularly at the PM and the Golgi [21,22,27,54,60]. Development of more experimental reagents and tools, including small molecule modulators and nanobodies, for dissecting Gβγ functions at different intracellular organelles is crucial for better understanding of spatiotemporal regulation of GPCR/G protein-mediated signaling and for drug discovery by targeting location-biased signaling molecules.

Concluding remarks and future perspectives

It is becoming increasingly appreciated that the Gβγ complex mediates GPCR signaling both at the PM and endomembranes and regulate discrete cellular functions. It is the Gγ subunits that dictate Gβγ association with the PM and translocation and thus, control Gβγ functions at different subcellular locales, suggesting an important role for Gγ subunits in regulating the functionality of the GPCR members at different cellular locations. It should be pointed out that the functions of Gβγ subunits at the Golgi apparatus have just begun to be revealed. Of many well-characterized signaling pathways that are activated by the Gβγ complex at the PM, the PLC-PKC-PKD cascade is only one identified thus far which is activated by Gβγ at the Golgi apparatus and has important functional consequences. Further research should continue to search for more Gβγ-mediated pathways at the Golgi and to identify the entities involved in regulation of compartmentalized Gβγ signaling (see Outstanding questions).

Outstanding questions:

1. In addition to PLC and PI3K, does the Gβγ complex at the Golgi activate other well-characterized signaling molecules that are activated by Gβγ at the PM?

2. What are the detailed mechanisms underlying Gβγ-mediated MAPK activation at the Golgi and its functional importance in tumor progression? In particular, how does PI3Kγ activate ARF1, how does ARF1 activate Raf-MEK-ERK1/2, do activated ERK1/2 translocate to the nucleus to activate transcriptional factors or stay in the cytoplasm to phosphorylate cytosolic substrates?

3. Besides the Golgi apparatus, Gβγ subunits are also activated and signal in other intracellular organelles and membrane domains. How do Gβγ-mediated intracellular signaling events spatiotemporally regulate GPCR functions and control cell responses to drugs and hormones?

4. Will compartmentalized Gβγ signaling at the Golgi or other subcellular locales be targeted for therapeutic interventions?

Gβγ-mediated activation of the PI3Kγ-ARF1-MAPK pathway likely occurs at the Golgi, not at the PM, suggesting selective Golgi-compartmentalized Gβγ signaling. Further studies are needed to prove such location-specific MAPK signaling by using other tools, such as genetically encoded fluorescent biosensors [77,96–98]. Indeed, a recent study has utilized ERK activity biosensors targeted to different subcellular locations to demonstrate that ERK1/2 activation by β₂-AR occurs at endosomes, but not at the plasma membrane [99]. Nanobody-based fluorescent biosensors have also been used to study compartmentalized GPCR signaling at the Golgi and endosomes [33,100]. Nevertheless, this pathway may be commonly used by multiple GPCRs to activate MAPK under physiologic and/or pathological conditions, implicating a novel strategy for cancer therapy by targeting location-biased oncogenic signaling.

Highlights:

1. GPCR activation induces Gβγ translocation from the plasma membrane to the Golgi and Gγ subunits dictate the translocation kinetics and efficiency. Gβγ can also be activated at the Golgi by Golgi-localized GPCRs.

2. Free Gβγ at the Golgi activates the PLC-PKC-PKD pathway and regulates a number of cellular processes.

3. Golgi-localized Gβγ activates oncogenic MAPK via PI3Kγ and ARF1 that may represent an important mechanism responsible for MAPK hyper-activation in prostate cancer.

4. Golgi-compartmentalized Gβγ-mediated signaling is a potential therapeutic target.

Glossary:

Prenylation

a post-translational modification with prenyl group by which 15 carbon farnesyl or 20 carbon geranylgeranyl group is transferred from farnesyl diphosphate or geranylgeranyl phosphate to the Cys residue of the C-terminal CAAX motif. After prenylation, the last three amino acid residues (-AAX) will be removed by endopeptidase and methyl group added to the C-terminal Cys residue by carboxymethyl transferase

GRK2ct

the C-terminal portion of G protein-coupled receptor kinase 2 (also shown as βARKct). It contains the PH domain of GRK2 which interacts with Gβγ, thus its expression has been used as strategy to inhibit Gβγ-mediated signaling

Golgi-GRK2ct

GRK2ct fused with the Golgi targeting sequence which delivers GRK2ct specifically at the Golgi compartment

COPII vesicles

a type of transport vesicles that are derived from the ER and exclusively mediate the export of newly synthesized cargoes from the ER

ER exit sites

specialized ER regions (also called transitional ER) where COPII vesicles are formed