Phosphorylation Barcoding as a Mechanism of Directing GPCR Signaling

See allHide authors and affiliations

Science Signaling  09 Aug 2011:
Vol. 4, Issue 185, pp. pe36
DOI: 10.1126/scisignal.2002331


A unifying mechanism by which G protein–coupled receptors (GPCRs) signal in cell type–dependent and G protein–independent ways has developed over the past decade. GPCR kinases (GRKs) are mediators of homologous desensitization: GRK phosphorylation of the receptors leads to the subsequent binding of β-arrestins, which partially quenches receptor coupling to G proteins. For some receptors, this GRK-mediated phosphorylation stimulates additional signaling through the scaffolding action of β-arrestin. These downstream signals are configured by β-arrestin conformation, which is dictated by the GRK phosphoacceptors on the receptors in a barcode-like fashion. Furthermore, each of the GRKs can potentially phosphorylate different serine and threonine residues on a given receptor, and the phosphorylation pattern can be biased by the receptor conformation established by bound ligand. Finally, the arrangement of potential GRK phosphorylation sites—and thus the conformation of β-arrestin and its effect on downstream signaling—can differ substantially between even closely related GPCRs stimulated by the same agonist. The diversity of the barcoding to flexible β-arrestin explains the multidimensional nature of signaling in the superfamily and represents new opportunities for drug discovery.

Seven transmembrane–spanning, G protein–coupled receptors (GPCRs) represent the largest signaling family in the human genome. Many cellular processes and most organ function involve GPCR signaling, placing the superfamily as a master regulator of human physiology. Activating, suppressing, or mimicking GPCR signaling represents one of the most common themes of currently prescribed therapeutics. Early concepts of GPCR signaling were similar to a simple on/off switch: Agonist binds the receptor, which induces an active state (often denoted R*); heterotrimeric G protein binds to the agonist-occupied receptor; and the dissociated α subunit or βγ subunits signal to effectors, such as adenylyl cyclase. Thus, agonist stimulation resulted in a change in an intracellular “second messenger,” such as cyclic adenosine monophosphate (cAMP), that affected cell function. Ligands were categorized as full or partial agonists, which activated the receptor, or as antagonists, which bound to the receptor and neutrally blocked access of the agonist.

It soon became clear that the switch was not simple, or even binary, in nature. With continuous agonist activation, receptor signaling waned, a process termed desensitization (1). Nevertheless, the basic concepts could still be described by adding a few attenuators and timers to the switch model (2). The most common mechanism of agonist-promoted desensitization of GPCRs is due to receptor phosphorylation by GPCR kinases (GRKs) (1). Such phosphorylation often occurs at Ser or Thr residues within the third intracellular loop or cytoplasmic tail of the receptor, which are regions of G protein interaction. However, this phosphorylation alone is not sufficient to promote desensitization. Rather, the phosphorylated receptor becomes a substrate for the binding of β-arrestins, which interdict between the receptor and G protein, resulting in partial uncoupling and loss of function (3). With the cloning and heterologous expression of GPCRs, particularly the adrenergic receptors (ARs) such as the prototypical β2AR, it became evident that receptors were signaling in ways that were not consistent with previous notions.

An early observation that led to the unraveling of traditional receptor theory was that of receptor signaling in the absence of agonist by unoccupied receptors (Fig. 1). When cells were transfected to express different amounts of β2AR, basal adenylyl cyclase activities, in the absence of agonist, were directly proportional to the amount of the receptor, in the case of wild-type receptors as well as a muted receptor that had impaired agonist-stimulated coupling to Gαs (4). Subsequent studies revealed that in the absence of agonist, most GPCRs oscillate between the inactive state and the active state, which evokes signaling (5). The time spent in the active state in the absence of agonist is short (the equilibrium favors the inactive state), but the activity of the receptor is measurable. Agonists do not “force” the receptor to undergo a conformational change, but rather stabilize and maintain the conformation of the spontaneously achieved conformation (6, 7). Furthermore, receptors appear able to adopt many different conformations (Fig. 1, denoted as R*1, R*2, R*3, and so on). Upon agonist stabilization, it is conceivable that only one major conformation and its signal arises (such as A1R*1 of Fig. 1) or that more than one receptor conformation and signal is realized (such as A1R*1 and A1R*2). With the same receptor, another agonist (A2 in Fig. 1) could stabilize one of those conformations but not the other, be a partial agonist for another conformation, and even be an inverse agonist that decreases the spontaneous activity of a third receptor conformation. In a situation where one pathway is deleterious to the cell, one can readily appreciate the value of ligands that are biased toward activation of the therapeutic pathway and inhibit or are neutral toward others.

Fig. 1

Multifunctional signaling by GPCRs. Multiple active conformations of a single GPCR (R) in the absence (R*) or presence (AR*) of an agonist (A1 or A2) are depicted. In the absence of agonist, receptor oscillates to multiple active forms with an accompanying signal, although the equilibrium (depicted by length of the arrows) favors the inactive state. Bound agonist stabilizes one or more of the active conformations (green); others may be unaffected (blue). Differences in agonist structure can lead to stabilization of different conformations and different signals, can be less effective in stabilizing a particular conformation (yellow), or, in the case of an inverse agonist, move the equilibrium toward the inactive state (purple).


Despite all this whirling around receptor conformations and signaling, mechanisms that account for this diversity of signaling with GPCRs have been in short supply. The idea that an altered agonist-promoted conformation of a receptor triggered the G protein to signal to a different effector was not supported. Nor was it clear how two closely related receptor subtypes, which couple to the same G protein and are activated by the same agonist, could have different effects on the cell. In some cases, such as when a receptor couples to two classes of G proteins (8, 9) or when a receptor directly interacts with another protein (10), one can envision such multifunctional signaling and how it might be manipulated by different agonists. However, for the majority of these “other” signals, a unifying mechanism has been lacking. In this issue, Nobles et al. (11) reveal a detailed mechanism whereby a “barcode” of phosphorylated Ser and Thr residues of the β2AR evoked by the GRK and read by β-arrestin directs distinct signaling. In earlier studies by Lefkowitz and colleagues, β-arrestins were found to serve as scaffolds and thereby evoke signaling that was independent of the actions of G protein (12). As one example, β2AR-specific signaling, independent of cAMP, to extracellular signal-regulated kinases (ERKs) requires β-arrestin binding (13). These studies set the stage for the concept that β-arrestins can promote signaling, as opposed to the traditional view that they only act to decrease signaling.

Although GRK phosphorylation sites have been mapped on various receptors, including the four Ser residues identified in the third intracellular loop of the α2AAR (14), a consensus sequence for GRK phosphorylation has not been established. Most cells have multiple GRKs, and there are many possible arrangements of Ser and Thr residues on receptor intracellular domains that could be phosphoryl­ated by GRKs. Layered on top of this diversity is the ability of specific GRK-induced phosphorylation patterns to trigger agonist-specific conformations of β-arrestin at the receptor, which is a key to GPCR multifunctional signaling (Fig. 2).

Fig. 2

Phosphorylation barcoding controls β-arrestin conformation, leading to multiple signaling options. (A) One agonist acting at two closely related receptor subtypes (R1, R2) results in two distinct signals (S1, S2) because the arrangement of phosphoacceptors molds distinct conformations of β-arrestin. (B) One agonist acting on the same receptor present in two different cell types in which either GRK2 or GRK6 predominates results in two different signals, because each kinase phosphorylates a unique set of residues and directs two different β-arrestin conformations. (C) Two structurally different agonists acting on the same receptors in the same cell evoke two different signals, because the receptor conformation favors GRK2- or GRK6- selective phosphorylation, thus directing agonist-specific β-arrestin conformations.


Nobles et al. (11) showed that β2AR phosphorylation by GRK2 or GRK6 is important for agonist-promoted desensitization of the receptor’s cAMP (Gs-coupled) signaling. However, when the ERK1 and ERK2 (ERK1/2) signal was determined, only GRK6 phosphoryl­ation was required. In similar experiments, cells were exposed to carvedilol, a biased ligand for the βAR that does not stimulate cAMP but does activate ERK1/2 (15). Again, ERK1/2 activation was markedly diminished by GRK6 knockdown, but not by GRK2 knockdown, when this biased ligand was used as the activator. These results pointed toward the possibility that GRK2 and GRK6 phosphorylate distinct sets of Ser and Thr residues of the β2AR, thereby directing the conformation of bound β-arrestin and its subsequent signaling, which was confirmed by intramolecular bioluminescence resonance energy transfer (BRET). A total of 11 Ser and Thr residues are present in the cytoplasmic tail of the β2AR, a region previously identified to be phosphorylated by GRKs (16). Combining stable isotope labeling with amino acids in cell culture, receptor purification, and identification of phosphopeptide residues by mass spectroscopy, Nobles et al. identified eight of these residues that exhibited an increase in phosphorylation when cells were exposed to isoproterenol. Four Ser residues within PKA consensus sequences (located in the proximal cytoplasmic tail and third intracellular loop) were also phosphorylated, which was expected because isoproterenol stimulates β2AR signaling through Gs to promote cAMP accumulation, which activates PKA. With selective knockdowns of GRK2 or GRK6, a complex phosphorylation pattern became evident. GRK2 phosphorylates six residues (five Ser and one Thr), and GRK6 phosphorylates a different set of two serines. Furthermore, GRK6-mediated phosphoryl­ation of these latter residues is partially inhibited by GRK2. Carvedilol exposure evoked phosphorylation of only the two GRK6-associated serines, which confirmed the functional studies showing that carvedilol-mediated stimulation of ERK1/2 was dependent on GRK6 as well as the BRET studies that implied a unique β-arrestin conformation was directed by GRK6. This confluence of results also shows how different agonists can lead to selective signaling: The ligand-stabilized conformation of the receptor can be specific for phosphoryl­ation of the receptor by one GRK, which phosphorylates unique residues, leading to a distinct conformation of bound β-arrestin, and a specialized signal. Phosphorylation encoding is not unique to the β2AR; biased ligands, differential phosphorylation, or cell type–dependent phosphorylation have been described for several GPCRs (1722).

On the basis of the barcoding by GRKs, it is intriguing to reconsider GRK and β-arrestin actions in terms of drug development, with an emphasis on the encoded β-arrestin signaling rather than short-term desensitization. Mechanisms that are involved with GPCR agonist–dependent desensitization are time-dependent. Chronic agonist exposure causes desensitization, which can be manifested clinically as tachyphylaxis, often involving a decrease in the net number of receptors; this may involve β-arrestins or may be mediated through other distinct processes. In contrast, the second-to-second regulation of GPCR function through the rapid desensitization and resensitization mediated by GRK and β-arrestin is essential for cell, organ, and whole organism homeostasis and may be considered an integral component of, and inseparable from, the initial physiologic response to exogenously administered agonists with effective actions of hours given to a naïve individual (23). Thus, receptor phosphoryl­ation by GRKs and subsequent β-arrestin binding should be considered in a potentially positive light, rather than as a requisite harbinger of an undesirable therapeutic profile such as tachyphylaxis. In addition, one then has to consider the complement of the GRKs and β-arrestins present in the on-target and off-target cell types of interest, because each GRK appears to direct a different β-arrestin conformation (and signal) according to which sites are phosphorylated on the receptor. To take advantage of the barcode to further understand GPCR signaling and to develop novel therapeutics will require taking a much broader look at the signaling consequences of ligand-receptor interactions. Impartial metabolomic and proteomic investigations of whole cell “signalomes” in response to structurally distinct ligands will undoubtedly reveal many previously unrecognized events mediated by G proteins, β-arrestins, and possibly as-yet-undiscovered partners.


View Abstract

Navigate This Article