β-Arrestins 1 and 2 are ubiquitously expressed intracellular adaptor and scaffolding proteins that play important roles in GPCR (G-protein-coupled receptor) desensitization, internalization, intracellular trafficking and G-protein-independent signalling. Recent developments in BRET (bioluminescence resonance energy transfer) technology enable novel insights to be gained from real-time monitoring of GPCR–β-arrestin complexes in live cells for prolonged periods. In concert with confocal microscopy, assays for studying internalization and recycling kinetics such as ELISAs, and techniques for measuring downstream signalling pathways such as those involving MAPKs (mitogen-activated protein kinases), investigators can now use a range of experimental tools to elucidate the ever-expanding roles of β-arrestins in mediating GPCR function.
GPCRs (G-protein-coupled receptors) interact with a range of proteins, either directly or indirectly, in order to function as agonist-modulated signalling complexes. This mini-review will focus on the interaction between GPCRs and the β-arrestin intracellular adaptor and scaffolding proteins, why this interaction is of interest and why it is desirable to carry out prolonged real-time monitoring of this interaction in live cells using BRET (bioluminescence resonance energy transfer) in addition to utilizing more established techniques such as confocal microscopy and ELISAs.
Agonist binding leads to stabilization of an active receptor conformation that couples to a G-protein(s), thereby initiating downstream signalling cascades (G-protein-dependent signalling). One active receptor can activate multiple individual G-protein heterotrimers and so the regulation of this process is critically important. One of the major mechanisms of ‘desensitizing’ GPCRs to further G-protein coupling involves a two-step process of agonist-dependent receptor serine/threonine residue phosphorylation by GRKs (GPCR kinases), followed by binding of β-arrestin to the active and phosphorylated receptor . β-Arrestin sterically hinders further coupling to G-protein, thereby resulting in a substantial reduction in G-protein-dependent signalling.
GPCR internalization and trafficking
Once β-arrestin is bound to the activated GPCR, it can target the receptor to clathrin-coated pits by interacting with clathrin and adaptor protein 2. Receptors utilizing this pathway are internalized into early endosomes in a dynamin-dependent manner. Sorting follows and at a simplistic level, receptors appear to enter either recycling endosomes to be trafficked back to the plasma membrane, or late endosomes to be trafficked to lysosomes for degradation . Such an apparent separation of GPCR types has led to a classification scheme being proposed. Class A GPCRs, exemplified by the β2-adrenergic receptor, appear to form relatively unstable complexes with β-arrestins (preferably β-arrestin 2) that dissociate at or near the plasma membrane before rapid receptor recycling. In contrast, Class B GPCRs exemplified by the AT1AR (angiotensin II receptor type IA), appear to form stable complexes (with similar affinity for β-arrestin 1 and 2) that internalize together into endocytic vesicles . Class B GPCRs appear to be characterized further by the presence of serine/threonine clusters in the C-terminal tail (serine or threonine residues in a minimum of three out of four consecutive positions), which seem to be lacking in Class A GPCRs . These clusters are believed to form particularly high-affinity β-arrestin-binding sites upon phosphorylation by GRKs. Furthermore, the degree of β-arrestin ubiquitination also appears to define the different classes, with transient or persistent ubiquitination correlating with the Class A or B classification respectively .
There are clearly important differences between types of GPCRs, and subdivision according to function is a useful approach to assist our understanding. However, as our knowledge of β-arrestin roles increases and the number of receptors, accessory proteins and cellular environments investigated expands, it is likely that the two-class framework for β-arrestin usage will require expansion, both in terms of what defines Class A and B and in terms of certain receptors not fitting the classification under some or all circumstances. For example, certain somatostatin (sst) receptors seem to exhibit characteristics of both classes, with the sst2A receptor forming stable receptor–β-arrestin complexes that internalize in a Class B manner, but recycle rapidly to the plasma membrane. In contrast, sst3 receptors seem to form relatively unstable complexes that dissociate at or near the plasma membrane in a Class A manner, but a large proportion appears to undergo ubiquitin-dependent lysosomal degradation rather than being rapidly recycled . Our findings indicate that both OxRs (orexin receptors) are Class B for β-arrestin usage, but they have distinct kinetic profiles for internalization, recycling and β-arrestin interactions as demonstrated using ELISAs, confocal microscopy, BRET dose–response assays and eBRET (extended BRET). In particular, we believe that OxR1 is another example of a Class B GPCR that recycles rapidly to the plasma membrane, with similarities to the sst2A receptor.
There is increasing evidence for β-arrestin scaffolding components of signalling cascades such as those of the MAPK (mitogen-activated protein kinase) pathways, leading to their cytoplasmic localization and activation (see  for a review). Not only is there the potential for a whole plethora of possible signalling pathways being activated independently of and/or in addition to G-protein signalling, there is also the possibility of GPCR agonist-activated signalling occurring on a time scale beyond that normally associated with G-protein-mediated signalling. For example, ERK (extracellular-signal-regulated kinases) 1 and 2 have been shown to be activated by AT1AR via kinetically distinct pathways dependent upon either G-protein or β-arrestin 2. G-protein-dependent activation was rapid (peak <2 min) and transient (t½ approx. 2 min), whereas β-arrestin-2-dependent activation was slower (peak 5–10 min) and prolonged (likely to persist for well beyond 90 min) . In this example, the distinct pathways were distinguished with the aid of β-arrestin 2 siRNA (small interfering RNA) to inhibit G-protein-independent activation and the PKC (protein kinase C) inhibitor Ro-31-8425 to inhibit G-protein-dependent signalling.
Monitoring GPCR–β-arrestin interactions in live cells using BRET
The BRET technique for monitoring protein–protein interactions has been reviewed recently [9,10]. BRET requires the heterologous expression of fusion proteins, with a donor Rluc (Renilla luciferase) or acceptor variant of GFP (green fluorescent protein) genetically fused to the protein of interest, in this case either the GPCR or β-arrestin. It is critical that these fusion proteins are functionally validated prior to BRET analysis, as illustrated with OxR2 and β-arrestin 1 (Figure 1). The proximity of the donor and acceptor molecules resulting from agonist-induced interaction of the GPCR and β-arrestin enables resonance energy transfer to occur in the presence of a suitable Rluc substrate. This energy transfer only occurs if donor and acceptor are within 10 nm, is inversely proportional to distance by the sixth power, and results in increased light emission at wavelengths characteristic of the acceptor molecule .
Functional validation of BRET fusion proteins
A number of studies have now utilized BRET to investigate interactions between GPCRs and β-arrestins (see  for a review). β-Arrestin has been shown to be particularly amenable to BRET analysis, including multiplexing BRET to investigate GPCR–β-arrestin and β-arrestin–ubiquitin interactions in parallel  and intramolecular BRET to investigate GPCR agonist-dependent conformational changes in the β-arrestin by monitoring energy transfer between donor and acceptor attached to either end of the molecule . Indeed, the potential utility of the GPCR–β-arrestin interaction for high-throughput screening of GPCR-specific drugs has been highlighted . The advantage of BRET compared with a number of techniques is that it can monitor interactions in live cells in real time, at 37 °C. It is also possible to add reagents, including agonists and antagonists, and continue monitoring in real time to assess their effects (Figure 2). Indeed, we have generated data demonstrating the effect of OxR1-selective antagonist SB334867 on the orexin A-induced interaction between OxR1 and β-arrestins in real time.
Real-time eBRET kinetics compared with original BRET time points
GPCR–β-arrestin interactions occur for prolonged periods compared with GPCR–G-protein interactions, particularly those involving Class B GPCRs. Furthermore, a prolonged agonist's presence, as is usually the case upon administration of therapeutic drugs to patients, results in GPCRs and β-arrestins associating, dissociating and re-associating over time depending on recycling kinetics and the degree of lysosomal degradation. Therefore a steady-state level of GPCR–β-arrestin interactions results over time. As our understanding of the importance of β-arrestins increases, particularly with respect to β-arrestin-mediated signalling, it will be important to monitor GPCR–β-arrestin interactions in real time for time scales in which β-arrestins are functioning following agonist stimulation. In the case of β-arrestin-mediated signalling to ERK as a result of activating AT1AR, that would appear to be beyond 90 min . The eBRET technique that has recently been validated in comparison with the original BRET technique (Figure 2)  enables real-time monitoring over this time scale and consequently has substantial utility for investigating GPCR–β-arrestin interactions in combination with complementary technologies.
Family Resemblances? Ligand Binding and Activation of Family A and B G-Protein-Coupled Receptors: A Biochemical Society Focused Meeting held at GlaxoSmithKline, Stevenage, U.K., 24–25 April 2007. Organized and Edited by D. Poyner (Aston, U.K.) and M. Wheatley (Birmingham, U.K.).
angiotensin II receptor type 1A
bioluminescence resonance energy transfer
green fluorescent protein
mitogen-activated protein kinase
small interfering RNA, sst, somatostatin
This work was supported by the National Health and Medical Research Council (NHMRC) of Australia (project grant no. 404087). K.D.G.P. was supported by an NHMRC Peter Doherty Fellowship (no. 353709), M.B.D. by a Keogh Institute for Medical Research scholarship, J.R.D. by a University of Western Australia Postgraduate Scholarship and K.A.E. by an NHMRC Principal Research Fellowship (no. 212064).