Targeting G protein-coupled receptor kinases (GRKs) to G protein-coupled receptors

Abstract

Introduction

G protein-coupled receptors (GPCRs) transmit extracellular stimuli into intracellular responses via their ability to activate heterotrimeric G proteins. GPCR signaling is then modulated by agonist-dependent phosphorylation of the receptor by G protein-coupled receptor kinases (GRKs) [1]. This promotes arrestin binding which inhibits G protein coupling, a process termed desensitization. Non-visual arrestins (β-arrestins) also initiate receptor endocytosis where the active GPCR can continue to signal via G protein- and/or arrestin-dependent pathways. The role of GRKs in regulating GPCR signaling has been extensively studied while aberrant GRK activity has been attributed to a range of disease pathologies and serves as a viable target for intervention [2]. This review provides a concise overview of GRKs with an update on notable advances in GRK function and regulation of GPCRs over the past few years (Figure 1).

Open in a separate window

Figure 1

Overview of GRK targeting to GPCRs and nodes of GRK regulation. (1) GRKs are recruited to agonist-activated GPCRs at the plasma membrane. (2) Upon binding GPCRs, GRKs undergo a conformational change that promotes kinase activation. (3) The active GRK phosphorylates the GPCR C-terminus. (4) GPCR phosphorylation by GRKs promotes β-arrestin (βarr) recruitment to the receptor, mediating receptor desensitization, endocytosis and arrestin-dependent signaling pathways. GRKs and GRK regulation can play a key role in determining receptor signaling. Recruitment factors such as G protein subunits or acidic lipids can mediate GRK recruitment to the receptor. Regulatory factors can directly bind to or post-translationally modify GRKs to regulate kinase recruitment and activity. Crosstalk receptors or the activated GPCR can signal these recruitment or regulatory factors to modulate GRK function. GRK inhibitors block kinase activity, inhibiting GPCR phosphorylation and phosphorylation-dependent pathways.

GRK Family

GRKs were initially discovered in the early 1970s when rhodopsin kinase (now GRK1) was first identified as the kinase responsible for phosphorylating photoactivated rhodopsin [3]. Next, the β-adrenergic receptor kinase (βARK, now GRK2) was identified as a kinase that phosphorylates the active β-adrenergic receptor (βAR) and suggested to be a part of a larger family of GPCR-specific kinases [4]. Following the cloning of GRK1 and GRK2 and the discovery of βARK2 (now GRK3), sequence analysis revealed that βARKs and GRK1 belong to distinct subfamilies [5–7]. GRK4 (originally called IT11), GRK5 and GRK6 were next discovered, establishing a third GRK subfamily [8–10]. Finally, GRK7 was identified as a second member of the GRK1 subfamily although primarily restricted to cone photoreceptor cells [11].

GRKs belong to the AGC kinase superfamily and are characterized by a distinct regulator of G protein signaling homology (RH) domain that flanks the kinase domain and regulates GPCR phosphorylation (Figure 2). The RH domain is composed of a terminal subdomain, whose peptide sequence feeds into the kinase N-lobe, and a bundle subdomain that forms intramolecular contacts with the kinase C-lobe. In addition, GRKs contain an N-terminal α-helix (αN-helix) that is critical for kinase function. The GRK C-terminus diverges between subfamilies and mediates lipid binding for membrane localization and GPCR targeting [12]. The GRK1 subfamily is prenylated and the GRK2 subfamily contains a pleckstrin homology (PH) domain that facilitates recruitment to the membrane via interaction with Gβɣ subunits and acidic lipids. Within the GRK4 subfamily, GRK4 and GRK6 are palmitoylated. In addition, a C-terminal amphiphilic α-helix directly mediates GRK5 binding to the plasma membrane and is conserved throughout the GRK4 subfamily [2]. This motif, in conjunction with an N-terminal PIP₂ binding site, is predicted to form a basic interface for membrane association [13].

Open in a separate window

Figure 2

GRK domain architecture. GRKs contain a conserved catalytic kinase domain, regulatory N-terminus and regulator of G protein signaling homology (RH) domain. Each GRK subfamily is characterized by a distinct, divergent C-terminal motif that mediates membrane binding. The GRK1 subfamily is prenylated and the GRK2 subfamily contains a PH domain that binds acidic lipids and Gβɣ. At the C-terminus, the GRK4 subfamily is palmitoylated (except for GRK5) and contains a lipid-binding motif. In addition, the GRK4 subfamily has a PIP₂ binding site near the N-terminus. Thick cylinders represent defined structural domains. Domains were defined using previously determined structures for GRK5 (PDB ID: 4TNB), GRK6 (PDB ID: 3NYO), GRK2 (PDB ID: 6U7C) and GRK1 (PDB ID: 3C4Z) and GRK1–7 multiple sequence alignment.

GPCR Targeting by GRKs

Previous studies have relied on functional assays to probe GRK regions that mediate GPCR binding, though multiple factors contribute to GRK function at the receptor. Thus, it is difficult to discern whether identified regions are true GPCR binding determinants using these functional readouts. In addition, the relatively low affinity of the GRK/GPCR complex has made traditional techniques such as X-ray crystallography and in vitro binding assays challenging [12]. Two recent studies utilized several innovative methods in parallel to identify key regions responsible for GRK binding to GPCRs (Figure 3A) [13,14]. In the first study, the N- and C-terminal lipid binding sites were identified as regions involved in initial GRK5 recruitment to the β₂AR while the GRK5 RH bundle subdomain was also identified as a site of receptor interaction [13]. Interestingly, the N-terminal site is analogous to a region in GRK1 that mediates recruitment to rhodopsin cluster of residues within the GRK RH terminal subdomain was also suggested in GRK1 binding to rhodopsin [14]. This evolutionarily conserved region was previously implicated in regulating GPCR phosphorylation by GRKs and is adjacent to the kinase N-lobe [15]. Therefore, GPCR binding and GRK activity may be intimately related via an allosteric activation network.

Open in a separate window

Figure 3

Structural model of the GRK/GPCR complex and GPCR-mediated GRK activation. Models were adapted from docked β₂AR/GRK5 structures based on cross-linking generated by Komolov et al. [13]. (a) Inactive GRK conformation upon recruitment to the active GPCR. The GRK ionic lock establishes contact between the RH and kinase domain to maintain the closed, inactive state. The N-terminal lipid-binding domain (royal blue) [13] and the RH terminal domain (lilac) [14], regions thought to mediate GPCR targeting, are contiguous with the receptor. (b) Domain movements for kinase activation. Arrows indicate GRK domain changes when transitioning from an inactive conformation (as shown) to the active state. Model from (a) is rotated 55° counterclockwise to better visualize domain movements upon activation. The RH domain opens outward and backward from the kinase domain and the kinase N-lobe undergoes closure towards the kinase C-lobe. The entire kinase rotates clockwise relative to the GPCR. (c) GRK active conformation for GPCR phosphorylation.

GPCR/GRK complex formation is coupled to interaction of the GRK kinase domain with the receptor C-terminus for phosphorylation [13]. A recent study introduced a reciprocal relationship where the μ-opioid receptor (MOR) C-terminal phosphorylation sites mediate GRK2 recruitment to the receptor. In addition, a cluster of phosphorylation sites (TSST, residues 354–357) dictate robust downstream multisite phosphorylation. While this upstream motif does not mediate GRK recruitment, it appears to enhance GRK2 stability at MOR, suggesting that multisite phosphorylation is dependent on sustained GRK binding [16]. This indicates a potential role for GPCR phosphorylation sites in reinforcing GRK binding following initial recruitment mechanisms.

Recent work identified a high-energy electrostatic interaction, an “ionic lock”, between the GRK RH bundle and kinase C-lobe that maintains the kinase in an inactive conformation (Figure 3A). RH domain dissociation from the kinase domain (RH domain opening) is coupled to kinase domain closure and catalytic activation (Figure 3B and ​andC)C) [13,17]. GPCR-mediated disruption of the GRK5 ionic lock is associated with RH domain opening. Elongation of the kinase affords the RH bundle subdomain to further interact with the receptor, potentially stabilizing this active conformation [13]. Interestingly, disruption of the GRK1 ionic lock did not enhance rhodopsin recruitment [14]. This suggests that RH domain opening does not directly mediate GPCR binding but instead is likely a response to receptor interaction. Ultimately, these findings further illustrate how GPCRs promote GRK activation following initial recruitment.

The GRK αN-helix has been shown to be critical for GPCR phosphorylation [18] although its direct role in complex formation is unclear since it was not identified as a binding determinant in recent structural studies [13,14]. This suggests that it might be involved in GPCR-mediated kinase activation. Indeed, the ordered GRK6 N-terminus coordinates the kinase C-tail at the active site, a motif critical for catalysis [19]. In addition, a cytoplasmic hydrophobic pocket in rhodopsin is required for phosphorylation by GRK1 and modeling of the GRK1 αN-helix within this pocket suggests that the hydrophobic environment stabilizes the α-helix structure [20]. Thus, transient interfacing of the GRK N-terminus with the receptor cytoplasmic pocket during GRK recruitment may be a key step in the mechanism of GPCR-mediated kinase activation.

Regulation of GRKs

Various post-translational modifications including phosphorylation, ubiquitination and nitrosylation have been shown to regulate GRK activity, localization and degradation [21–47] (Table 1). In addition, phospholipids mediate several aspects of GRK-mediated receptor phosphorylation and acidic lipids have been characterized as a key regulator, as reviewed previously [2]. PIP₂ binding has been shown to orient GRKs at a model membrane plane, suggesting that PIP₂ positions the kinase to effectively phosphorylate the receptor [48,49]. In fact, acidic lipids are necessary to form a stable GRK5/β₂AR complex in vitro [13]. Additionally, Gβɣ mediates GRK2 subfamily phosphorylation of GPCRs, primarily by promoting membrane association via the PH domain [50]. The GRK2 subfamily RH domain also binds Gα_q, sequestering G protein activity [51]. Recent evidence suggests that this Gα_q interaction may also mediate GRK2 membrane binding to promote stable kinase association with the M3-muscarinic acetylcholine receptor (M₃AChR) and subsequent arrestin recruitment [52].

A recent study also established copper metabolism MURR1 domain–containing (COMMD) 3 and COMMD8 as adaptor proteins that mediate GRK6-specific recruitment to active GPCRs [53]. In this study, Nakai et al. identified GRK2/3-mediated phosphorylation of CXCR4 as a key determinant for recruitment of the COMMD3/8 complex. COMMD3/8 binding to the receptor subsequently recruits GRK6 and mediates GRK6-specific phosphorylation of CXCR4, initiating β-arrestin binding. Importantly, the authors demonstrate that this COMMD3/8-dependent CXCR4 signaling pathway mediates B-cell trafficking and the humoral immune response [53]. COMMD3/8 are ubiquitously expressed, therefore this complex may serve a general role for GRK6 recruitment to a range of GPCRs.

While GRKs are regulated by various post-translational modifications that can alter kinase stability and function (Table 1), several proteins have also been shown to directly inhibit GRK receptor phosphorylation. For example, raf kinase inhibitor protein (RKIP) is an established physiological inhibitor of GRK2-mediated GPCR phosphorylation and internalization, attenuating receptor desensitization. Phosphorylation of RKIP by protein kinase C (PKC) promotes RKIP binding to the GRK2 N-terminal region (residues 1–185). Thus, GPCR-mediated PKC activation tempers GRK2 receptor phosphorylation via RKIP and serves as a feedback mechanism to inhibit GPCR desensitization [54]. Recently, RKIP regulation of GRK2 was identified as the mechanism in which inflammatory signals sensitize the δ-opioid receptor (DOR). In peripheral neurons, GRK2 is constitutively bound to DOR, blocking G protein coupling. B2 bradykinin receptor stimulation via the inflammatory signal bradykinin activates PKC, switching RKIP binding to GRK2 and restoring DOR responsiveness [55]. RKIP regulation of GRK2 may also enhance GPCR signaling independent of receptor phosphorylation. For example, RKIP overexpression in mice induces heart failure progression that is partially dependent on angiotensin II receptor type 1 (AT1)-mediated G_q signaling. RKIP sequestration of the GRK2 RH domain is proposed to block GRK2 inhibition of G_q, restoring AT1-G_q coupling for cardiotoxic signaling [56]. Together, these recent studies establish that in addition to GPCR feedback signaling, RKIP regulation of GRK2 can mediate receptor crosstalk pathways as well as G protein coupling.

Several calcium binding proteins such as recoverin and calmodulin (CaM) were previously established as calcium-dependent inhibitors of GRK-mediated GPCR phosphorylation [57]. Recoverin binds GRK1 via the αN-helix and sequesters the kinase from phosphorylating rhodopsin [58]. CaM is selective for the GRK4 subfamily with the highest efficacy towards GRK5 although it can also inhibit GRK2. In addition, GRKs are thought to contain two distant and distinct CaM binding sites [57]. Despite CaM inhibition of GRK-mediated GPCR phosphorylation, CaM has been shown to stimulate GRK5 phosphorylation of soluble substrates [59]. Several structural and biophysical methods were recently employed to help define the structural basis of GRK5 regulation by CaM. While CaM was determined to form an equimolar complex with GRK5 [60], the precise structural mechanism in which CaM occupies two binding sites remains unclear. Interestingly, CaM also mediates non-canonical GRK5 activity implicated in cardiac hypertrophy [61], although its role in regulating GPCR signaling has not been determined. A recent study suggests that CaM binding inhibits GRK5 phosphorylation of MOR upon calcium channel influx [62]. Similar to RKIP, calcium-dependent CaM regulation of GRKs may serve to potentiate GPCR signaling when coupled to calcium activation, though additional exploration is necessary.

Role of GRKs in Arrestin Recruitment and GPCR Signaling

While arrestins play a key role in GPCR desensitization, they also mediate GPCR signaling via a range of arrestin-dependent pathways. Differential GPCR phosphorylation patterns by distinct GRK isoforms, or “barcoding”, drives the diversity of GPCR signaling by dictating arrestin conformation following recruitment. Barcoding by GRKs has been shown to modulate signaling across a range of GPCR substrates, as reviewed previously [63]. Conversely, it has been difficult to reconcile how a highly variable GPCR C-terminal phosphorylation sequence mediates a conserved mechanism of arrestin recruitment. The structural basis of how specific patterns of GPCR phosphorylation mediate arrestin recruitment and signaling has not been well characterized. Thus, two recent studies set out to define general GPCR phosphorylation codes and how they relate to arrestin affinity, activation, specificity and conformation using several structural and biochemical techniques [64,65]. Their efforts to generalize these phosphorylation patterns and their corresponding functions are critical towards understanding the complex nature of arrestin-dependent GPCR signaling. Ultimately, defining these barcodes will further inform how different GRKs regulate arrestin recruitment and diversify signaling.

Canonical GPCR signaling consists of balanced G protein activation, GRK-mediated phosphorylation and arrestin-dependent processes. In certain instances, either G protein or arrestin activation is dominant resulting in signaling that is biased towards a particular pathway. Prior studies have focused on circumstances that modulate G protein or arrestin recruitment as determinants for GPCR bias. Recently, GRKs have emerged as modulators of biased signaling due to their crucial role in arrestin recruitment through GPCR phosphorylation [66]. GPCR mutants that exhibit bias are useful models to tease apart G protein- or arrestin-specific pathways and are also valuable for understanding the mechanisms of biased signaling. For example, the G protein-biased β₂AR mutant Y129A is unable to recruit arrestin. Further characterization of the receptor identified that the G protein-bias was due to a deficiency in phosphorylation by GRKs, ultimately inhibiting arrestin-dependent function [67]. Similarly, a G protein-biased dopamine D₂ receptor (D2R) mutant deficient in arrestin recruitment also exhibited reduced GRK2 recruitment [68]. In addition, mutation of M₃AChR phosphorylation sites elicits G protein bias [69]. This study identified receptor phosphorylation as a requirement for bronchial smooth muscle contraction, a response implicated in asthma. While GRKs were not specifically probed, the study demonstrates how a model receptor can inform the therapeutic outcomes of a potential ligand biased against receptor phosphorylation. Alternatively, an arrestin-biased D2R mutant unable to engage G proteins was dependent on GRK2 activity for receptor phosphorylation and arrestin recruitment [68]. Furthermore, GRK2 recruitment and function at an arrestin-biased D2R is G protein-independent, a component intrinsic to wildtype D2R balanced signaling. Corroborating results with a D2R arrestin-biased agonist (UNC9994) indicate that D2R may directly recruit GRK2 to mediate arrestin-biased signaling [68]. These studies suggest a larger, understudied contribution of GRKs to mechanisms of GPCR bias and underscore the necessity of GRK assessment when characterizing biased ligands.

GRK function can also dictate physiological GPCR bias due to its role in receptor signaling and trafficking pathways. It has been shown that protease activated receptor-1 (PAR1) activation by Activated Protein C results in arrestin-biased signaling that mediates a cytoprotective response in endothelial cells [70]. GRK5 is required for this cytoprotective pathway, suggesting that it mediates arrestin-biased signaling through PAR1 [71]. Regulation of GRK recruitment to the receptor can also modulate arrestin-mediated pathways, introducing receptor bias. For example, calcium influx via calcium channel activation inhibits the phosphorylation and internalization of MOR and is correlated with reduced GRK5 membrane localization. The authors suggest that Ca²⁺/CaM blocks receptor internalization by sequestering GRK5 to prevent MOR phosphorylation and subsequent arrestin recruitment [62]. Gastric inhibitory polypeptide receptor (GIPR) is a constitutively internalized and recycled receptor that mediates incretin signaling in adipocytes. As a result, the receptor undergoes a noncanonical desensitization pathway. Upon stimulation, GIPR is regulated via trafficking of the internalized endosomal receptor to the Golgi, ultimately slowing receptor recycling back to the plasma membrane. Analogous to the canonical desensitization mechanism, arrestin mediates GIPR targeting to the Golgi, effectively sequestering rapid receptor recycling. This pathway is dependent on both GRK2 and GRK5, likely via phosphorylation of multiple GIPR C-terminal sites. Interestingly, a GIPR variant that is associated with obesity shows enhanced sequestration upon stimulation. This variant exhibits increased arrestin binding that is independent of GRK receptor phosphorylation. Phosphorylation-independent binding is speculated to prolong arrestin at GIPR, ultimately biasing the receptor towards Golgi trafficking rather than immediate recycling [72]. Clearly, GRK activity serves a key role in initiating and moderating receptor regulatory pathways, biasing GPCR signaling.

Phosphorylation-independent regulation of GPCRs by GRKs is an emerging concept in receptor desensitization, primarily characterized by GRK2 sequestration of downstream effectors, such as Gα_q or MEK [73]. Transient, phosphorylation-independent desensitization by GRK2 was observed for MOR, though via a distinct and novel mechanism [16]. MOR desensitization in the absence of GRK2/3 kinase activity was initially observed in neurons [74]. Additional mechanistic studies revealed that GRK2 and arrestin are both recruited to active MOR in the absence of receptor phosphorylation. This mode of arrestin recruitment is rapid and transient, initiating desensitization, but not endocytosis. Miess et al. suggest that GRK2 still regulates this rapid component of arrestin recruitment and desensitization, independent of kinase activity [16]. This function of GRK may immediately modulate MOR activity, serving as a short-term response until the receptor is phosphorylated by GRK2 for sustained regulation. In sensory neurons, GRK2 is basally associated with DOR, constitutively desensitizing the receptor by inhibiting G protein coupling in a kinase-independent manner. It is suggested that neither DOR phosphorylation nor arrestin regulate receptor signaling in this setting. Ultimately, inhibition of GRK2 binding to DOR sensitizes the receptor to stimulation [55]. Direct, phosphorylation-independent regulation of GPCRs by GRKs could constitute an alternate and more rapid form of desensitization that is also readily reversible. It is still unclear exactly how GRKs mediate this mechanism of GPCR regulation and whether it occurs outside the opioid receptor family, warranting further study.

GRK Inhibitors

Previous studies (reviewed in [75]) identified a number of GRK-specific inhibitors that have proven useful in dissecting GRK function and potentially biasing GPCR signaling. For example, GRK inhibition via Paroxetine or Cmpd101 attenuates βAR and MOR phosphorylation, blunting arrestin recruitment and receptor endocytosis in heterologous systems [74,76]. In addition, Cmpd101 effectively blocks MOR desensitization in rat neurons [74]. Paroxetine and Cmpd101 inhibition of GRK2 also prevent P2Y₂ and histamine H1 receptor desensitization in rat and human smooth muscle cells, respectively. Ultimately, blocking GPCR desensitization via GRK2 inhibition maintains arterial contractions, suggesting that GRKs serve as a viable pharmacological target to mediate arterial tone and blood pressure [77].

Recent efforts in inhibitor development have focused on understanding the determinants for GRK potency and selectivity to improve on the therapeutic potential of existing inhibitors. Structural analysis revealed an active site hydrophobic pocket that can be targeted to enhance GRK2 inhibitor potency and selectivity [78]. It is thought that inhibitors with minimal hydrogen bonding at the active site hinge exhibit additional selectivity by affording structural flexibility. These compounds might avoid locking the kinase in the highly conserved active state and instead can access and stabilize a unique conformation. This strategy could circumvent the lack of structural diversity at the active site by instead targeting conformational diversity to achieve GRK and GRK subfamily selectivity [75,79]. Despite the success in identifying GRK2-selective compounds, it has been particularly difficult to identify a potent, GRK4 subfamily-specific inhibitor [80]. In one study, high-throughput screening and substituent modification yielded a GRK5 inhibitor with nanomolar potency, but with minimal selectivity over GRK2 [81]. Recently, Rowlands et al. utilized a unique GRK5 C-tail cysteine residue (Cys474) positioned at the active site to selectively target GRK5. The addition of an electrophilic moiety to an existing GRK inhibitor scaffold resulted in the covalent modification of Cys474. This electrophilic substituent is associated with enhanced GRK5 potency and specificity over GRK2. Therefore, irreversible GRK5 modification via Cys474 is a potentially effective method of gaining subfamily specificity in inhibitor development [82].

Conclusions

GRK function and regulation is fundamental to GPCR signaling. Over the past several years, structural detail has contributed to our understanding of the mechanisms of GPCR targeting and kinase activation. In addition, various forms of physiological and pharmacological GRK inhibition have been detailed, ultimately blocking GPCR desensitization. Several novel modes of GRK regulation of GPCR signaling were recently identified and GRK receptor phosphorylation was established as a determinant of receptor bias. Thus, continued insight into the mechanisms in which GRKs modulate GPCRs is critical to our understanding of receptor signaling.

Acknowledgements

Footnotes

References