G protein-coupled receptor signaling: transducers and effectors

Abstract

G protein-coupled receptors (GPCRs) are of considerable interest due to their importance in a wide range of physiological functions and in a large number of Food and Drug Administration (FDA)-approved drugs as therapeutic entities. With continued study of their function and mechanism of action, there is a greater understanding of how effector molecules interact with a receptor to initiate downstream effector signaling. This review aims to explore the signaling pathways, dynamic structures, and physiological relevance in the cardiovascular system of the three most important GPCR signaling effectors: heterotrimeric G proteins, GPCR kinases (GRKs), and β-arrestins. We will first summarize their prominent roles in GPCR pharmacology before transitioning into less well-explored areas. As new technologies are developed and applied to studying GPCR structure and their downstream effectors, there is increasing appreciation for the elegance of the regulatory mechanisms that mediate intracellular signaling and function.

INTRODUCTION

Receptor theory was first postulated at the beginning of the 20th century by Paul Ehrlich’s study of agent selectivity (1) and John Newport Langely’s study of receptive substance (2). However, the concept of receptors as physical entities was controversial and it became clear that new technologies were needed to better understand the interaction between ligands and receptors. Successful radioligand binding on both β-adrenergic receptors (βAR) (3) and α-adrenergic receptors (αAR) (4) paved the road for the functional characterization of GPCRs and the eventual discovery of the superfamily of seven-transmembrane (TM) receptors (5).

Although it had been established that hormones could stimulate the production of cyclic AMP (cAMP) (6), precise mechanisms were unknown. Initial studies on coupling between receptors and adenyl cyclase (AC) demonstrated the necessity of guanosine triphosphate (GTP) (7) and GTPase (8). These studies were followed by a series of reconstitution experiments that led to the identification of a novel protein: the Gα_s subunit, a heterotrimeric G protein complex component that coupled the receptor to its effectors (9). It is now well-known that heterotrimeric complexes are composed of α, β, and γ subunits. Gα subunit consists of four major families, including Gα_s, Gα_i, Gα_q/11, and Gα_12/13, ranging in size from 39 to 52 kDa (10) (Fig. 1).

Figure 1.

G protein-coupled receptor activation. Seven-transmembrane GPCRs are a group of diverse receptors in eukaryotes. When an external agonist binds to the GPCR, the receptor will undergo a conformational change, triggering a subsequent interaction with the heterotrimeric G protein complex. The G protein is considered inactive when the α subunit is bound to GDP; however, upon an agonist stimulation, the GDP will physically be displaced by the GTP, rendering the complex active. Both the Gα and the Gβγ subunits will then go on to activate distinct downstream effectors. Figure created with BioRender and published with permission. GPCR, G protein-coupled receptor; GTP, guanosine triphosphate.

Understanding how receptors become desensitized following stimulation has also been a vigorous area of study. First observed was the light-induced phosphorylation of visual rhodopsin (11), which led to the subsequent discovery and purification of rhodopsin kinase (GRK1) as the enzyme catalyzing the phosphorylation of rhodopsin (12). In addition to the demonstration of receptor phosphorylation via the cAMP-dependent protein kinase A (PKA), another mechanism whereby agonist-occupied receptors undergo phosphorylation was identified and termed, homologous desensitization, and led to the purification of the β-adrenergic receptor kinase GRK2 (13).

Although the mechanism of receptor desensitization by GPCR kinase (GRK) phosphorylation was initially observed using crude enzyme preparations, it was realized that another factor was necessary for desensitization when highly purified GRK enzyme was used(14, 15). This factor, a functionally analogous and structurally homologous protein to visual arrestin, was identified as a regulator of the β-adrenergic receptor and named β-arrestin (16, 17). Today, they are known as the ubiquitously expressed β-arrestin-1 (arrestin-2) and β-arrestin-2 (arrestin-3).

G PROTEINS

G Protein Structure and Function

G proteins bind to both GTP and GDP, playing a central role in GPCR signaling and mediating a wide range of physiological responses (Fig. 1). G proteins differ in size, ranging from the small signaling protein Ras to the large heterotrimeric G protein complexes composed of α, β, and γ subunits.

The Gα subunit has three distinct structural components: a Ras-like GTPase domain, an α-helical domain, and an N-terminal helical domain (18). The GTPase and helical domains are conserved across all Gα subunits (19). The GTPase domain is structurally homologous to other monomeric G proteins and is the site of GTP hydrolysis. The GTPase also exhibits conformational plasticity, demonstrated by its highly dynamic state while bound to GDP as well as its rigid, locked structure while bound to GTP (20). Furthermore, the helical domain is credited for enhancing the Gα subunit affinity for guanine nucleotides (21) and for increasing the rate of GTP hydrolysis (22). In addition, changes in the carboxyl end, including single point mutations, can alter the type of the G protein (23, 24). The N-terminus serves as a critical site for Gα and the Gβγ subunit interaction (25), with myristoylation increasing the affinity between the Gα and Gβγ subunits (26), further stabilizing the N-terminal helix (27). Earlier structural studies showed that both a “switch interface” and an “N-terminal interface” of Gα are involved in heterotrimeric G protein interactions, resulting in conformational changes (18, 28). The crystal structure of an agonist-occupied monomeric β2 adrenergic receptor (β2AR) and a nucleotide-free Gs subunit revealed displacement of the helical domain relative to the Ras-like GTPase domain (29). Gα subunits may undergo modification by myristoylation, phosphorylation, and palmitoylation (30). Fatty acid modification is prevalent within the N-terminus, exemplified by the thioester bond between cysteines and the saturated 16-carbon palmitate (31).

Gβ and Gγ subunits, anchored to the plasma membrane by lipids, are obligate heterodimers and work as a single Gβγ unit. The Gβ subunit comprises a seven-bladed β-propeller that has seven WD40 repeats, and its N-terminal helix forms extensive, primarily hydrophobic, parallel interactions with the N-terminal helix of the Gγ subunit (32). The Gβ subunit may also interact with the receptor as shown by studies using a synthetic peptide derived from the third intracellular loop of the α2-adrenergic receptor (33). The Gγ subunit, composed of two α helices, is much smaller, between 7 and 8 kDa (34), and undergoes either isoprenylation or geranylgeranylation (35, 36). In the absence of Gγ, Gβ subunits are unstable and tend to form aggregates (37).

Fatty acid modification is vital for membrane localization since mutations that result in the improper fatty acid modification may lead to impaired cellular localization (38, 39). In particular, Gγ interacts with the plasma membrane and the Gβ subunit through its highly variable C-terminal posttranslational modification (40). Furthermore, studies have shown that the Gγ subunits are predominately found on the endoplasmic reticulum (ER), but upon Gα activation, are subsequently directed to the plasma membrane (41). There are multiple possible Gβγ subunit combinations, with some specificity among the different subtypes of Gβ and Gγ subunits (42). Gβγ subunits are around 36 kDa in size (34) and are frequently subjected to prenylation and phosphorylation (19).

G Protein Activation

The heterotrimeric G protein complex couples the activation of the receptor to downstream effector signaling. Early kinetic models suggest a collision coupling theory between the GPCR and the G protein, but more recent fluorescence resonance energy transfer (FRET) studies demonstrate a conformational rearrangement of a preexisting GPCR-G protein complex upon agonist binding (43, 44). Early studies using the retinal receptor rhodopsin showed that the light-activated conformational change of rhodopsin creates a critical interaction between the cytoplasmic hydrophobic cleft of the receptor and the C-terminus of the transducing Gα subunit (45), consistent with previous studies using synthetic peptides (46). Based on site-directed spin-labeling experiments, an outward movement of the receptor TM helix VI creates a binding pocket for the Gα subunit on the receptor cytoplasmic site (47).

After ligand binding, the GPCR, acting as a guanine nucleotide exchange factor, initiates exchange of GTP for GDP on the G protein. Release of GDP from the G protein occurs by the allosteric disruption of the receptor on the nucleotide-binding site, resulting in low binding affinity for GDP (48). The unoccupied binding site is then filled with GTP due to its high intracellular concentration (49). The GDP-GTP switch occurs in the Ras-like G domain of the Gα subunit (18), and linkers between the Ras-like G domain and the helical domain exhibit dramatic changes in conformation (50). Structurally, an N/TKXD motif and a highly conserved P-loop are critical for the binding specificity (51). Binding between the Gα subunit and GTP produces an unstable heterotrimeric complex resulting in the dissociation of Gβγ from the Gα subunit. Once dissociated, the Gα and the Gβγ subunits diffuse freely, interacting with downstream effector proteins to elicit cellular and physiological responses (52).

G Protein Signaling

Both the Gα and the Gβγ subunits exert downstream effects on signaling and different families of Gα activate distinct downstream signaling profiles (Fig. 2). Gα_s and Gα_i subunits regulate adenylyl cyclase (AC) activity generating the second messenger cAMP. Gα_s stimulates AC resulting in increased intracellular concentrations of cAMP, whereas coupling of the inhibitory G protein, Gα_i, reduces intracellular cAMP concentration (10). Gα_q/11 activates β-type phospholipase C (PLC-β) to hydrolyze membrane-bound phosphatidylinositol 4,5-bisphosphate (PIP₂) into diacylglycerol (DAG) and inositol triphosphate (IP₃) (53). After cleavage, DAG remains membrane-bound, whereas IP₃ can diffuse throughout the cytoplasm. As second messengers, IP₃ and DAG continue to exert their effects downstream, including calcium release through the calcium channel IP₃ receptor from the ER and protein kinase C (PKC) activation, respectively (54). Gα_12/13 can target regulator of G protein signaling homology (RH)-RhoGEFs, activating GTPase Rho (55), including proteins such as leukemia-associated RhoGEF (LARG) and p115RhosGEF (56, 57).

Figure 2.

The heterotrimeric G protein signaling. There are four major families in the α subunit: Gαs, Gαi, Gαq/11, and Gα12/13. Although Gαs activates the AC, Gαi inhibits it, both of which will regulate the intracellular concentration of cAMP. Gαq/11 activates PLC-β, hydrolyzing PIP2 into DAG and IP3. Gα12/13 also has a variety of downstream targets, including RhoGEF. AC, adenyl cyclase; ATP, adenosine triphosphate; cAMP, cyclic adenosine monophosphate; DAG, diacylglycerol; GEF, guanine nucleotide exchange factor; GTP, guanosine triphosphate; IP3, inositol trisphosphate; PIP2, phosphatidylinositol 4,5-bisphosphate; PKA, protein kinase A; PKC, protein kinase C; PLC-β: phospholipase C-β. Figure created with BioRender and published with permission.

Early studies assumed a passive role for the Gβγ subunit; however, the molecular basis for an interaction of Gβγ with effectors was demonstrated using a mutational analysis of Gβ residues that makes contact with Gα-GDP to identify Gβγ subunits as critical regulators of signal transduction (58). With the loss of the subunit Gβγ, GDP-GTP exchange does not proceed and effector signaling is impaired (59). Gβγ interacting effectors include various isoforms of phospholipases (60), ion channels (61), Src-dependent phosphorylation of the epidermal growth factor receptor (EGFR) (62), and GRK2 (63), to name a few (64–69). Therefore, once the GPCR is activated, both the Gα and the Gβγ subunits can act as regulators of a wide range of effectors.

Many other noncanonical signaling pathways also involve G proteins. In the absence of a GPCR, a group of activators of G protein signaling (AGS) may promote GTP binding or stabilize the heterotrimeric G protein complex (70). Gα_q/11 can interact with p115RhoGEF, PDZ-RhoGEF, and LARG (71), all of which are key regulators in cellular growth and differentiation that are frequently impaired in cancer and metabolic diseases (72). Heterotrimeric G proteins can also be regulated by non-GPCR entities, such as Ric-8A to promote GTP-GDP exchange (73) and AGS3 to inhibit guanine nucleotide dissociation (70).

G Protein Signal Termination

Following Gβγ dissociation, intrinsic GTPase activity of GTP-bound Gα will hydrolyze GTP back to GDP and reform the heterotrimer, which can be accelerated by GTPase-activating proteins (GAP) that are regulators of G protein signaling (RGS) (74). For example, RGS3 can bind to Gβ₁γ₂, limiting its ability to produce IP₃ and activate Akt and PKA (75). Gβγ subunits can also terminate signaling by reassociating with Gα. Receptor-G protein interaction may also be inhibited through GRK-mediated receptor phosphorylation and subsequent arrestin binding, which will be discussed in GRK and β-arrestin sections.

G Proteins: Cardiovascular Physiological Functions and Possible Therapeutic Intervention

In the heart, overexpression of the Gα_s subunit can depress heart rate variability and the baroreflex (76), blunt βAR desensitization (77), and cause dilated cardiomyopathy (78). A common synonymous polymorphism (ATT → ATC, Ile131) of exon 5 in GNAS1 has been found to be associated with hypertension (79).

Upregulation of Gα_i has been observed in patients with heart failure (80), and may be cardioprotective under conditions of chronic adrenergic signaling (81). Overexpression of Gα_i in the swine heart by viral gene transfer can modulate electrical currents in a model persistent atrial fibrillation (82). In contrast to Gα_i, moderate Gα_q overexpression induces cardiac hypertrophy, whereas a high level of overexpression results in cardiomyocyte apoptosis (83), contractile failure (84, 85), and cardiomyopathy (86). Gα_q inhibition reduces the hypertrophic response induced by mechanical pressure overload along with preserved cardiac function despite the increase in wall stress (87–89). Gα_q deficient mice lack response from platelet activators and have increased bleeding time (90). A noncoding polymorphism in the GNAQ promoter which codes for the Gα_q subunit is also associated with cardiac hypertrophy and adverse outcomes in heart failure (91, 92).

In cultured neonatal rat ventricular cardiomyocytes, overexpression of a constitutive active Gα₁₃ increases cell size (93). Constitutive active Gα_12/13 will induce a RhoA-mediated vascular smooth muscle cell contraction in vitro (94). Although Gα_12/13 deficiency is embryonic lethal (95), mice with cardiomyocyte-specific deletion of Gα₁₃ are protected against pressure overload-induced hypertrophy and fibrosis (96).

Agonist-induced βAR desensitization requires GRK2 that is recruited to the membrane through its interaction with Gβγ subunits (97). When Gβγ subunits are sequestered through overexpression of the C-terminal of βARK1 (βARKct), there is enhanced catecholamine responsiveness (98), preserved cardiac function in a model of genetic cardiomyopathy (99), and improved survival (100). Among the βARKct transgenic animals, a linear positive correlation is further observed between the level of βARKct protein expression and fractional shortening after chronic pressure overload (101). The salutary findings from the βARKct animal model study are confirmed by M119 (102) and its analog galleon (103), a Gβγ subunit inhibitor.

Polymorphisms in the Gβγ subunit have been identified. In particular, a C825T polymorphism in Gβ₃ has been shown to increase G protein signaling, reduce the risk for myocardial infarction, and increase efficacy of statin therapy to lower cholesterol (104, 105). In patients with hypertension, the role of C825T is less clear with studies showing an association with elevated blood pressure (106, 107) and others showing no increased risk for hypertension (108, 109).

G PROTEIN-COUPLED RECEPTOR KINASES

GRK Characteristics and Structure

The termination of agonist-induced GPCR signaling is primarily mediated by the phosphorylation of GRKs, a group of serine and threonine kinases (110). There are seven members in the GRK family, which are grouped into three subfamilies: GRK1 and GRK7; GRK2 and GRK3; GRK4, GRK5, and GRK6 (111), with each class of GRK having distinct N- and C-terminal characteristics (112). GRK1 and GRK7 are known as visual GRKs because they are largely restricted to retinal cells (113). GRK2 and GRK3 are also referred to as the β-AR kinases because of their close association with the βARs (114). Due to their high degree of structural similarities, GRK4, GRK5, and GRK6 are grouped together (115). GRK4 is commonly found in the testes (116), whereas GRK2, GRK3, GRK5, and GRK6 can be found throughout the human body with relatively even distribution (112, 117).

Unlike other kinases, GRKs do not require autophosphorylation for full activity (118). Although there are seven different isoforms of GRKs proteins, they share a conserved structure among its over 500 residue assembly: a protein kinase catalytic domain inserted into a loop of the regulator of G protein signaling (RGS) homology (RH) domains (119, 120). Besides interacting with the kinase domain, the RH domain has two additional domains: one that interacts with the C-terminal pleckstrin homology (PH) domain and one with the Gαq subunit (121). The PH domain interacts with membrane lipids and free Gβγ subunits (122).

GRKs also contain an N-terminal α-helical domain that has an antiparallel β-sheet with five strands and a variable C-terminus featuring multiple α-helices, which acts as the primary binding site (123). Mutation of any residue within the N-terminus can lead to a loss of function in receptor phosphorylation (124, 125). The N-terminus may play a stabilization role—its hydrophobic residues may interact with the membrane, whereas polar residues interact with the kinase domain (126, 127). When the kinase catalytic domain is stabilized in a closed conformation, the N-terminus may also interact with the C-terminus to form a docking site for the activated GPCRs (128). Similarly, mutations in the C-terminal region can inhibit receptor phosphorylation (129, 130), with the C-terminus playing a critical role in mediating cytosol membrane localization (131).

GRK1/7.

GRK1 was first crystalized in 2008 (132). A more recent cryo-electron microscopy single-particle reconstruction of rhodopsin and GRK1 shows the N-terminus forming a helix that interacts with rhodopsin’s cytoplasmic cleft and stabilizes the activated configured kinase domain (133). Thus, receptor recognition and kinase activation are closely coupled together for GRKs. Even though both GRK1 and GRK7 share a signature CaaX motif on the C-terminus that is used in membrane localization, GRK1 is farnesylated (134), whereas GRK7 is geranylgeranylated (135).

GRK2/3.

In the βARK family, the crystal structure of bovine GRK2 was first published in 2003 in complex with the Gβ₁γ₂ subunit and shown to simultaneously inhibit Gα and Gβγ signaling (136). The Gα_q-GRK2-Gβγ complex has further demonstrated that GRK2 can serve as a scaffold protein, reorienting the subunits and allowing diverse effectors to recognize specific regions on the subunits (137). To date, GRK3 has not been crystallized. Both GRK2 and GRK3 have a Pleckstrin homology domain, critical for membrane localization (138).

GRK4/5/6.

There are four splice variants in human GRK4, namely, GRK4α, GRK4β, GRK4γ, and GRK4δ (116). One such polymorphism, GRK4α A486V, associated with hypertension, has been crystallized (139). GRK5 is a monomer and its active-site forms part of the nucleotide-binding pocket, a feature that is not observed among other GRKs (140). The crystal structure of GRK5 in complex with an inhibitor identified a localization function for the C-terminus (141). More recently, the binding interface of the β₂AR-GRK5 complex has been mapped by mass spectrometry, in which a three-dimensional model was generated that confirms the importance of the cytoplasmic loops (142). The leucine and threonine on N-terminus of GRK5 are also implicated in receptor phosphorylation (126).

The crystal structure of GRK6 reveals a phospholipid-binding site near its N-terminus and other structural elements in the kinase domain that may affect GPCR access and specificity (143). Furthermore, hydrophobic residues in the N-terminus are thought to interact directly with the receptor based on the closed conformation of GRK in crystal structure (128).

GRK Cellular Physiology

Short-term receptor desensitization is divided into two mechanisms: heterologous desensitization, and refers to the involvement of cAMP and PKA in the absence of an agonist (144), and homologous desensitization, which is agonist-dependent and typically mediated by GRKs (145). As serine/threonine kinases, GRKs phosphorylate the third intracellular loop or the C-terminal tail of the activated GPCR (146) (Fig. 3). When phosphorylation sites are removed from the receptor, a delay in desensitization rate is observed (149). Recruitment of arrestins to agonist-occupied phosphorylated receptors will uncouple the heterotrimeric G protein from the receptor (150). In addition, specific anionic phospholipids work in concert with the activated GPCR to regulate GRK activity (151).

Figure 3.

GPCR desensitization and internalization. Two critical factors are involved in receptor desensitization and internalization after stimulation. The first is GRK phosphorylation of the C-terminal tail of the GPCR. The second is β-arrestin recruitment to the phosphorylated receptor followed by clathrin-dependent internalization. Depending on the GPCR-β-arrestin complex conformation, β-arrestin may either bind to the core (core conformation) or the tail (tail conformation). Because the tail conformation presumably exposes the core for further G protein interactions, it is possible to form a GPCR-G protein-β-arrestin “megaplex” that can exhibit sustained G protein signaling as demonstrated for the β2AR and cytoplasmic cAMP generation (147, 148). Figure created with BioRender and published with permission. β2AR, β2-adrenergic receptor; cAMP, cyclic AMP; GPCR, G protein-coupled receptor; GRK, GPCR kinase.

GRK1/7.

The GRK activation mechanism was first studied in the rhodopsin system. Phosphorylation of rhodopsin by GRK1 was first identified in 1972 (152). When light strikes the receptor, subsequent activation of GRK1 can phosphorylate many sites in close proximity including exogenous peptides (153). Using truncated photolyzed rhodopsin, it was noted that GRK1 binds to cytoplasmic loops of the receptor to stimulate catalytic function (153). Receptor mutagenesis has further confirmed the importance of cytoplasmic loops for GRK docking (154). Similar to GRK1, GRK7 can phosphorylate the cone photopsins and terminate light-activated signaling pathways (155).

GRK2/3.

GRK2 was initially named βAR kinase or βARK due to its ability to phosphorylate β2ARs (13, 120). Synthetic peptides that are composed of β₂AR domains demonstrate the potential of multiple interaction sites between the receptor and the kinase (156). For example, the first intracellular loop plays an inhibitory role (157), whereas the second and third intracellular loop are critical for binding affinity between receptor and kinase (158, 159).

GRK2’s C-terminal proline-rich motif is involved in the recruitment of agonist-occupied receptors through its binding of Gβγ subunits on the pleckstrin homology domain (160, 161). GRK2 can also promote phosphorylation-independent desensitization through sequestration of Gα_q via its RGS homology domain (162). A noncanonical role for GRK2 has been shown by its direct protein-protein interaction between the RGS domain of GRK2 and Gα_q/11 (163). GRK2 can also phosphorylate non-GPCRs, exemplified by its negative regulation on the platelet-derived growth factor receptor β (PDGFRβ) (164).

Although GRK2 and GRK3 share a high degree of structural similarity (165), some key differences do exist between the two isoforms. For example, although GRK2 is the predominating isoform, the olfactory epithelium only contains GRK3 (166, 167). GRK3 can also enhance angiogenesis through its downregulation on thrombospondin-1 and plasminogen activator inhibitor-2 (168). For the chemokine receptor C-X-C chemokine receptor type 4 (CXCR4), GRK3 overexpression can restore the desensitization of the receptor and normalize chemotaxis (169).

GRK4/5/6.

Unlike other GRKs, GRK4 demonstrates a constitutive phosphorylation activity in the absence of agonist stimulation (170), achieved through palmitoylation (116). GRK4 can also directly interact with caveolin-1, preventing further phosphorylation of the dopamine 1-like (D1) receptors (171).

In addition to its canonical role in phosphorylation, GRK5 binds calmodulin and is necessary for nuclear localization and maladaptive cardiomyocyte growth (172). Due to its large array of targets, it is thought that nuclear targeting following agonist stimulation may decrease cell survival, whereas GRK5 localized to the myocyte membrane may be protective (173). Beyond nuclear accumulation, GRK5 can also phosphorylate numerous other substrates, such as class II histone deacetylase 5 (174), the platelet-derived growth factor receptor-β (175), the insulin-like growth factor-1 receptor (176), and moesin (177). GRK5 polymorphisms also exist, such as the Leu41Glu variant that may promote a cardioprotective effect (178, 179).

Similar to GRK4, palmitoylated GRK6 can only be found in the membrane (180) and its activity not only depends on membrane association (181), but also on the rate of palmitoylation (182). GRK6 can also target non-GPCR low-density lipoprotein receptor-related protein 6 (183) and the signaling protein Na⁺/H⁺ exchange regulatory factor (184). These nonreceptor roles of GRKs indicate that GRKs play a broad role in signaling and disease (185).

GRK Physiological Functions and Possible Therapeutic Intervention

Maladaptive GRK activity is frequently observed in cardiovascular pathophysiology and therefore has potential as a therapeutic target (110).

GRK2/5.

As GRK2 desensitizes β adrenergic receptors, they are pharmaceutically relevant targets for potential heart failure therapies (186). In hypertrophied hearts, high GRK2 activity leads to marked desensitization of βARs (187). In patients with heart failure, there is a threefold increase in GRK2 protein expression (188) and levels of GRK2 may be a prognostic marker for heart failure mortality (189). Treatment with β-blockers in patients with heart failure has been shown to decrease GRK2 activity (190). Although homozygous knockout of GRKs is embryonic-lethal (191), heterozygous knockout show enhanced contractility (192). Mice that overexpress GRK2 show blunted catecholamine responsiveness (98) and increased susceptibility to ischemia-reperfusion injury (193), likely involving Akt signaling (194). In contrast, targeted ablation of GRK2 has been shown to improve cardiac function in murine heart failure models (195).

Inhibition of GRK2 has been tested as a potential therapy for heart failure. Rats with chronic heart failure injected with adeno-associated virus type 6 (AAV6) vectors that express βARKct showed improved cardiac function and reversal of reverse cardiac remodeling (196). Similar results were also achieved in large animal heart failure models (197). Thus, there is a considerable interest in developing highly selective GRK inhibitors. The crystal structure of a GRK2 complex with an RNA aptamer demonstrates how a hairpin loop can mimic ATP interaction in the active site (198). Other compounds, such as balanol (199), balanol-related compounds (200), and paroxetine (201), offer structural hints about the development of more potent and selective inhibitors.

GRK5, another highly expressed GRK in the heart, is also implicated in heart failure (202, 203). Transgenic mice with elevated GRK5 expression develop more hypertrophy than its wild-type counterparts after pressure overload (174). When GRK5 is knocked-out, mice show attenuated hypertrophy and preserved cardiac function in response to pressure overload by transverse aortic constriction (204) as well as enhanced muscarinic acetylcholine receptor signaling, manifesting in an exaggeration of cholinergic responses (205). The RGS homology domain within the N-terminus (GRK5-NT) can also inhibit NF-κB transcription activity and ameliorate cardiac hypertrophy (206).

GRK1/3/4/6/7.

Because of their roles in phosphorylation within the visual system, the in vivo ablation of GRK1 (207) or GRK7 (208) will lead to an abnormal response to light. Cardiac-specific GRK3 inhibition leads to elevated systolic pressure (209), and in the renal proximal convoluted tubules, reduced GRK4 expression can alleviate the severity of hypertension (210). GRK6, a key regulator in dopaminergic signaling, has been studied as a pharmaceutical target for addiction and Parkinson’s disease (211).

β-ARRESTINS

β-Arrestin Characteristics and Structure

Arrestins are multifunctional adaptor proteins that function to regulate a wide array of receptors and cellular signaling cascades (212). First discovered in 1978 (213), four isoforms of arrestins are currently known in vertebrates: the visual arrestins (Arrestin-1 and Arrestin-4) and the nonvisual arrestins: (Arrestin-2 and Arrestin-3) also known as β-arrestin-1 and β-arrestin-2.

Arrestin-1 and 4 are restricted to ocular rod and cone cells, binding almost exclusively to rhodopsin and color opsins, respectively (214). The nonvisual arrestins, β-arrestin-1 and β-arrestin-2, are ubiquitously expressed and interact with Class A and B GPCRs (16). Class A GPCRs, such as the β2AR, are characterized by β2AR·β-arrestin complexes that dissociate at or near the plasma membrane, whereas class B GPCRs, typified by the angiotensin type I receptor (AT1R), form a stable AT1R·β-arrestin complex and remain associated in endocytic vesicles (215). Despite sharing highly similar amino acid sequences, the two β-arrestin isoforms have distinct structural and functional identities. For example, β-arrestin-2 contains a leucine-rich nuclear export signal that restricts its location to the cytoplasm, whereas β-arrestin-1 can be found in both the nucleus and cytoplasm (216).

The structure of arrestin remains largely conserved across all four isoforms with an overall architecture that contains a central polar core flanked by N- and C-domains and a C tail connecting the two domains (217, 218). Structural variation is largely restricted to the distal C-domain. The polar core is made up of charged side chains (219) and is involved in its activation (220), as well as interaction between the N- and the C-terminal domain (221). Multiple loops are also critical for the stabilization of the molecule, including the “finger loop,” “middle loop,” and “lariat loop,” which are respectively defined by residues 63–75, 129–140, and 274–300 (222). The clathrin-binding domain on β-arrestin is located between residues 376–380 (223), whereas the binding domain of c-Src is located within the proline-rich residue regions of 88–96 and 120–124 (224).

β-Arrestin Activation

Phosphorylated GPCRs are the most common activators of β-arrestin and the binding to the receptor is proposed to be a two-step process that uses a phosphate and a conformational sensor (225). Structural studies of β-arrestin-1 (222) and β-arrestin-2 (226) have shown that a lysine and arginine on the N-domain form salt bridges with phosphates on the C-terminal tail. Phosphate binding will lead to structural changes, such as physical disruption of the polar core through a 20° rotation around the central axis of the N- and C- domains and displacement of the lariat loop (222). In addition to the phosphorylated c-tail of a GPCR, β-arrestins can be activated by IP6 (226) and an arginine to glutamic acid mutation at residue 175 (227).

The “finger loop” is thought to be the active sensor on the arrestin that will insert itself into a newly emerged binding pocket (228, 229). The C-edge loop of β-arrestin contributes to formation of a stable receptor complex through its engagement with the lipid bilayer (230, 231). Thus, when in complex with its receptor, β-arrestin has also been found to undergo a series of conformational changes distinct from that in its basal inactive state.

β-arrestins and G proteins can additionally bind to the same interhelical cavity on the cytoplasmic side of GPCRs, with arrestin showing a greater binding affinity for the site (232). With the ability of β-arrestin to bind to the phosphorylated C-terminal tail of the GPCR and promote receptor internalization (233), a GPCR-β-arrestin-G protein super complex was shown to occur in endocytic vesicles following agonist stimulation that can promote sustained intracellular G protein signaling (147).

β-Arrestin in Cellular Physiology

Although bound to GPCRs, β-arrestins can fulfill a variety of cellular functions that range from receptor desensitization, receptor trafficking and internalization, and activation of downstream signaling (234). A key element in the process of β-arrestin recruitment to activated GPCRs is phosphorylation of amino acid residues along the C-terminal tail of the receptor. Distinct phosphorylation patterns on the C-tail of the receptor have led to the concept of a “barcode” of phosphorylated serine and threonine residues as a driver of β-arrestin-GPCR interaction stability (235, 236). The importance of a receptor tail phosphorylation barcode is that it stabilizes distinct active conformations of β-arrestin (237, 238), thereby directing unique signaling cascades (239), and physiological responses (240). Specific receptor phosphorylation sites by different GRKs may further define the cellular function carried out by β-arrestin. For example, β-arrestin-mediated extracellular signal-regulated kinases (ERK) activation is associated with GRK 6 sites, whereas β2AR internalization is mediated by GRK2 sites, as well as both GRK sites resulting in receptor desensitization (236). Other studies have postulated a “flute model” of receptor tail phosphorylation—β-arrestin conformation that directs selective signaling (241, 242).

β-Arrestin in Receptor Desensitization

Among the wide array of functions ascribed to β-arrestin, its role in receptor desensitization, defined as the rapid loss of GPCR responsiveness with agonist stimulation, is of critical importance. Although kinase phosphorylation results in partial attenuation of receptor signaling, β-arrestin recruitment to the phosphorylated receptor leads to complete termination of G protein signaling through G protein uncoupling (243). Although the precise mechanism by which a GPCR is desensitized depends on the receptor subtype and its cellular context, the physical steric inhibition that uncouples G protein from the receptor is commonly observed (244). Furthermore, based on its capability of scaffolding phosphodiesterase 4, β-arrestin can promote the negative regulation on the receptor through both receptor desensitization and an accelerated degradation rate of cAMP (245). β-arrestin-2 is potentially more effective in mediating GPCR desensitization compared with β-arrestin-1, emphasizing an underlying correlation between structural conformation and cellular function (246).

β-Arrestin in Receptor Trafficking

Following receptor desensitization, GPCRs are internalized and trafficked from the plasma membrane to endocytic vesicles by β-arrestin (233). GPCR internalization is contingent upon β-arrestin binding to the phosphorylated receptor followed by binding of clathrin and its adaptor protein (AP2) to by β-arrestin (247). Clathrin-dependent endocytosis is achieved through the interaction between a clathrin-binding box on β-arrestin and blades in the clathrin β-propeller domain (248, 249). Based on a dynamin-dependent mechanism (250), β-arrestin may recruit other proteins, such as ADP-ribosylation factor 6 (251) and N-ethylmaleimide-sensitive fusion protein (252).

Through GPCR internalization and the formation of a megaplex, receptors can further maintain their signaling properties in internalized compartments (147, 148) (Fig. 3). This function is believed to be native to both the tail and core binding conformations of β-arrestin and therefore is common to both Class A and Class B GPCRs (253). After GPCR internalization to endosomes, mediated by Rab proteins (254), receptors are directed either toward the lysosomal pathway for degradation, the trans-Golgi-Network for recycling, or the plasma membrane for repopulation and resensitization (255). Therefore, β-arrestin maintains an integral role in the mechanics of the G protein-coupled receptor, either by its effects on receptor signaling or trafficking.

β-Arrestin as a Signal Transducer

In 1999, it was discovered that the role of β-arrestin extends beyond “arresting” (224). In fact, β-arrestin maintains the ability to scaffold molecules and initiate receptor-dependent and -independent signaling cascades (212). Its vast interactome and involvement in a wide array of signaling pathways allow it to play a central role in a variety of cellular functions (Fig. 4).

Figure 4.

β-Arrestin-mediated signaling. β-arrestins have been shown to regulate a diverse set of signaling pathways involved in a variety of cellular processes. The transducer plays a crucial role in the desensitization, trafficking, internalization, and translocation of its receptor, directing crucial stages of the GPCR life span. β-arrestin has also been shown to activate many signaling cascades, particularly due to its ability to scaffold additional proteins. The wide array of signaling cascades consequently regulate diverse aspects of cellular function, including but not limited to EGFR transactivation, cell growth, apoptosis, cell differentiation, inflammatory responses, and DNA damage repair. ASK1, apoptosis signal-regulating kinase 1; EGFR, epidermal growth factor receptor; ERK, extracellular signal-regulated kinase; GPCR, G protein-coupled receptor; HB-EFG, heparin-binding EGF-like epidermal growth factor; IkBα, NF-κB inhibitor; JNK3, Jun N-terminal kinase 3; MDM2, murine double minute 2 gene (negatively regulates p53); MEK, mitogen-activated protein kinase; MMK4/7, mitogen-activated protein kinase 4/7; MMP, matrix metalloproteinase; NF-κB, nuclear factor-kappa B; RaF, rapidly accelerated fibrosarcoma kinase; Src, protooncogene tyrosine-kinase Src. Figure created with BioRender and published with permission.

c-Src Recruitment and Signaling.

GRK2 phosphorylation residues along the C-terminal tail of a GPCR lead to the recruitment and binding of β-arrestin to the phosphorylated receptor, followed by the β-arrestin-mediated protooncogene tyrosine-protein kinase Src (c-Src) recruitment (224). The two main interactions between β-arrestin-1 and c-Src occur between the Src homology 3 domain and PXXP motifs at resides 121–124 as well as at the terminal catalytic SH1 Src domain and epitopes within residue 185 of the N-terminal (256). β2AR, β-arrestin-1, and c-Src then form a complex upon agonist activation of the receptor, leading to receptor-dependent activation of ERK. GPCR activation and engagement of β-arrestin with the phosphorylated receptor tail allosterically activate Src (257, 258). Src-dependent GPCR-mediated signaling effects include tyrosine phosphorylation of dynamin (256), ERK MAP kinase cascade activation (259), and EGFR transactivation (260).

GPCR transactivation.

Receptor transactivation is defined as agonist-stimulation of a receptor leading to the activation and cytosolic generation of downstream products of a second distinct receptor, particularly in the context of cell surface receptors (261). More specifically, ligand activation of βARs leads to recruitment of β-arrestin-1-Src to the receptor complex initiating matrix metalloproteinase (MMP) activation, heparin-binding epidermal growth factor (HB-EGF) shedding and EGFR binding (260). EGFR transactivation by β1-adrenergic receptor (β1AR) has also been shown to confer cardioprotective effects under conditions of sustained catecholamine stimulation (260). Interestingly, the β1AR antagonists, alprenolol and carvedilol, have been found to stimulate βAR-EGFR transactivation and EGFR-dependent ERK activation (262).

ERK activation.

As a scaffolding protein, β-arrestin may also bring two or more proteins closer together for downstream signaling cascades. For example, GPCRs can partially transduce signals via the ERK/β-arrestin signaling complex, as exemplified by β-arrestin-dependent ERK activation (263). Upon formation, the receptor complex undergoes endocytosis mediated by clathrin-coated pits, whereby activated ERK is translocated to endocytic vehicles (264). For AT1aR, there are two distinct pathways for ERK activation: a G protein-dependent pathway that is rapid, transient, and will lead to nuclear translocation, and a β-arrestin-dependent pathway that is slow and sustained (265). Furthermore, β-arrestin-activated ERK functions plays a bigger role in cellular chemotaxis, cytoskeletal rearrangements (266, 267), cell growth and proliferation (268), mobility, and apoptosis (269).

JNK3 activation.

c-Jun N-terminal kinases (JNK) signaling is involved in cellular proliferation, apoptosis, differentiation, and migration (270). β-arrestin-2 acts as a scaffold for JNK3 along with the upstream JNK activators apoptosis signal-regulating kinase 1 (ASK1) and mitogen-activated protein kinase MAPK kinase 4 (MKK4) (271). Stimulation of the AT1R results in JNK3 activation and colocalization of β-arrestin-2 and JNK3 in endocytic vesicles (271). JNK3 has been shown to interact with the β1 strand of β-arrestin-2, which requires a C-terminal truncation (272). In addition, an active JNK3 molecule can also replace an inactive JNK3 molecule in the β-arrestin-2/MKK complex leading to signal amplification (273).

Biased Agonism

Biased agonism refers to the ability of a ligand to activate a distinct subset of a receptor’s signaling portfolio by allowing the receptor to engage distinct transducers (274–276) (Fig. 5). In the context of a GPCR, transducers can be Gα and Gβγ, GRKs, and β-arrestin resulting in “bias” toward either β-arrestin-mediated signaling or G protein-mediated signaling.

Figure 5.

Biased agonism and signaling selectivity. Biased agonism is a recently emerging concept in which a ligand preferentially activates one signaling pathway over another. In balanced agonism, binding of a ligand potentiates both G protein and β-arrestin signaling. However, when a β-arrestin-biased-ligand binds to a GPCR, conformational changes in the receptor favor β-arrestin-mediated signaling while silencing of G protein-mediated pathways. Similarly, when a G protein-biased ligand binds to the GPCR, conformational changes in the receptor favor G protein-mediated signaling while silencing β-arrestin signaling. Figure created with BioRender.com. GPCR, G protein-coupled receptor; GTP, guanosine triphosphate.

Ligand bias.

Different ligands can promote the stabilization of different conformations of the receptor-transducer complex (277). β-arrestin-biased ligands stabilize receptor conformations favoring the engagement β-arrestin as the signaling transducer resulting in distinct downstream cellular signaling (278). For the AT1R, β-arrestin-biased ligands, such as [Sar1,Ile4,Ile8]-AngII (SII) and TRV120023, can lead to β-arrestin-dependent signaling, including the ERK activation signaling cascade (275, 279, 280). βARs also exhibit ligand bias. For example, carvedilol, a βAR antagonist, can stabilize distinct receptor conformations to stimulate β-arrestin-mediated signaling (281, 282).

Allosteric modulation.

Allosteric modulators are small molecules that can preferentially bind to allosteric sites which differ from the orthosteric site at which endogenous agonists bind (283). Allosteric modulators bound to receptors may then either increase or decrease the binding affinity of orthosteric ligands, and therefore are described as positive or negative allosteric modulators, respectively (284). Because of their enhanced selectivity and intrinsic properties in modulating receptor function they hold great clinical promise (285).

Recently, it was shown that the βAR allosteric modulator Cmpd-6 positively cooperates with carvedilol to augment the properties of carvedilol by enhancing β-arrestin signaling (286, 287) and cardioprotection to ischemia reperfusion (287). Thus, GPCR allosteric modulators are of interest as novel therapeutics due to associated increases in receptor subtype specificity and potential for a decrease in side effects (276, 287).

β-Arrestin Cardiovascular Physiological Function and Possible Therapeutics

In the heart, β-arrestin-1 is the predominated form (288) and has been shown to be associated with adverse remodeling post myocardial infarction (289). In contrast, β-arrestin-2 can be cardioprotective following myocardial infarction due to its anti-apoptotic and anti-inflammatory effects (290, 291). Knockout of β-arrestin-2 in aged mice shows signs of cardiac dysfunction (292). β-arrestin activation of the AT1R through the biased ligand TRV120023 is cardioprotective in ischemia-reperfusion injury by promoting cardiomyocyte survival (293). Interestingly, β-arrestin-dependent signaling activated by the β-blocker carvedilol confers cardioprotection in experimental ischemia-reperfusion injury through a β-arrestin-dependent mechanism (287). Finally, the βAR allosteric modulator, Cmpd-6, can potentiate the prosurvival/antiapoptotic effect of carvedilol in vivo by enhancing the binding affinity of carvedilol to the βAR (287). Therefore, β-arrestin’s multifunctionality allows it to promote signaling cascades that can be protective or detrimental within the context of the cellular function or disease process. For instance, β-arrestin-1 has been implicated in the cellular process of T-cell activation (294).

CONCLUSION

With advancement in understanding of the mechanisms that drive GPCR transducer interactions, we can appreciate the magnitude and versatility of their cellular effects. With continued investigation into mechanisms of ligand–receptor interaction, effector-receptor engagement, conformational analysis, and signaling profiles should yield new approaches to the development of novel therapeutics.

GRANTS

This work was supported by the National Institutes of Health Grants HL056687 and HL075443 (to H.A.R.).

DISCLOSURES

H. A. Rockman is a scientific cofounder of Trevena Inc., a company that is developing new GPCR ligands. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

This article is part of the special collection “Advances in GPCRs: Structure, Mechanisms, Disease, and Pharmacology.” Wei Kong, MD, PhD, and Jinpeng Sun, PhD, served as Guest Editors of this collection.

AUTHOR CONTRIBUTIONS

H.J., D.G., and J.W. prepared figures; H.J., D.G., and J.W. drafted manuscript; H.J., D.G., J.W., and H.A.R. edited and revised manuscript; H.J., D.G., J.W., and H.A.R. approved final version of manuscript.