Allostery at G protein-coupled receptor homo- and heteromers: uncharted pharmacological landscapes

Abstract

For many years seven transmembrane domain G protein-coupled receptors (GPCRs) were thought to exist and function exclusively as monomeric units. However, evidence both from native cells and heterologous expression systems has demonstrated that GPCRs can both traffic and signal within higher-order complexes. As for other protein-protein interactions, conformational changes in one polypeptide, including those resulting from binding of pharmacological ligands, have the capacity to alter the conformation and therefore the response of the interacting protein(s), a process known as allosterism. For GPCRs, allosterism across homo- or heteromers, whether dimers or higher-order oligomers, represents an additional topographical landscape that must now be considered pharmacologically. Such effects may offer the opportunity for novel therapeutic approaches. Allosterism at GPCR heteromers is particularly exciting in that it offers additional scope to provide receptor subtype selectivity and tissue specificity as well as fine-tuning of receptor signal strength. Herein, we introduce the concept of allosterism at both GPCR homomers and heteromers and discuss the various questions that must be addressed before significant advances can be made in drug discovery at these GPCR complexes.

Fig. 1.

On-target allosteric effects on binding and function at a monomeric GPCR. A, binding of an allosteric modulator (red) to a monomeric GPCR can result in reciprocal modulation of an orthosteric ligand (yellow) binding to a nonoverlapping site on the receptor. The simulated example provided illustrates the effect of changing concentrations of a modulator (x-axis) with different α values (the cooperativity factor that reflects the influence of a modulator on affinity) on the binding of a fixed concentration of radioligand at the orthosteric site. Where α > 1, the modulator is a PAM for orthosteric binding, whereas α < 1 represents NAMs at the receptor. Thus, it is apparent how a NAM with strong negative allosteric properties can prevent binding of an orthosteric ligand at either a monomer or indeed dimer. Curves were simulated using the simple allosteric Ternary Complex Model described by Christopoulos and Kenakin (2002) using Prism 5.02 (GraphPad Software, San Diego, CA) with the following parameters for the orthosteric radioligand: K_D, 1 nM; radioligand concentration, 1 nM. The modulator was assigned a K_B of 10⁻⁶ M, and α was varied as indicated. B, in cases where the orthosteric ligand possesses efficacy (i.e., can generate a measurable response), represented here by fractional response (F/F₀), an allosteric modulator can alter the measured potency of the orthosteric ligand. In this simulation, the arrows indicate the shift in EC₅₀ with increasing concentrations of allosteric modulator with an α value fixed to either 20 (this would make the allosteric modulator a PAM, therefore less orthosteric agonist is required for an equivalent effect) or 0.05 (a NAM, where more orthosteric agonist is required to achieve an equivalent effect). This simulation illustrates one of the principal tenets of allosterism: despite increasing concentrations of allosteric modulator, the shift in orthosteric EC₅₀ is saturable. If the same experiment was performed with a simple competitive antagonist, the right shift seen for the NAM would theoretically continue infinitely as the concentration of antagonist is increased. The curves were simulated as for (A) with the following parameters: EC₅₀ = 10⁻⁶ M, K_B = 10⁻⁹ M, basal = 0, E_max = 1, Hill slope = 1, and the concentration of allosteric modulator, B, was varied from 10 μM to 0.1 pM. C, in addition to modulating the binding or signaling of an orthosteric ligand, an allosteric modulator can itself possess efficacy and is thus both an allosteric modulator and an allosteric agonist, subsequently referred to as an ago-allosteric modulator. In the accompanying example, concentration-response curves were simulated for an orthosteric agonist in the presence of increasing concentrations of ago-allosteric modulator according to the operational model of allosteric modulation and allosteric agonism (Leach et al., 2007). Because the allosteric modulator possesses efficacy, increasing concentrations of the coadministered ligand increases the basal signaling in the system, yet unlike a partial agonist, this ligand is also a PAM for the signaling of the orthosteric ligand. Simulation parameters were as follows: τ_A = 20, τ_B = 1 [where τ represents the capacity of either orthosteric (A) or allosteric (B) ligands to act as agonists], K_A = 10⁻⁶ M, K_B = 10⁻⁷ M, α = 20, β = 10 (where β represents the allosteric effect on efficacy), slope factor = 1, basal = 0, and E_max = 1. The concentration range for ago-allosteric modulator was 3 mM to 0.1 nM. In theory, there is no reason why the same effects on orthosteric ligand binding and efficacy would not exist at a homo- or heterodimer.

Fig. 2.

Allosteric possibilities at GPCR homomers and heteromers. Homo- or heteromerization provides the opportunity for both on-target (as seen in Fig. 1) and off-target allosterism. For simplicity, allosteric modulation of ligand binding across a dimer is illustrated (without accounting for effects on signaling), although allosterism will also occur across higher order oligomers. A, at a GPCR homodimer, an orthosteric ligand (yellow) can bind to one or both protomers, and binding of the first ligand can lead to a conformational change in the homodimer such that the affinity of the second identical protomer for the second identical copy of the ligand is altered either positively or negatively. This is referred to as cooperativity and is indicated in A by orange arrows. Cooperativity can also occur between two identical allosteric ligands (red) binding to the same site on different protomers. In addition to cooperative effects on affinity, the allosteric ligand can influence binding of the orthosteric ligand on the same protomer (on-target allosterism) and on the opposing protomer (off-target allosterism) and this modulation is reciprocal. B, for a heterodimer that, by definition, must comprise two different GPCR protomers, each of the allosteric and orthosteric binding sites is unique, and any energy transfer between them must be allosteric in nature. If we assume a single allosteric binding site for each of protomer A (blue) and protomer B (green), then up to four different ligands are capable of concurrently occupying the heterodimer and allosterically influencing each other. Note: for the purpose of illustration we have depicted the allosteric modulators binding within the transmembrane region of the protomers, although allosteric ligands are able to interact at numerous sites on a GPCR, including the extracellular loops and even the intracellular surface of the receptor.

Fig. 3.

Experimental approaches to the identification of allosteric or heteromer-specific ligands at heteromers. A, reconstitution of a functional receptor through heterodimerization. i, protomer A is a GPCR-G protein fusion product that is able to bind ligand but contains a mutation within a conserved region in intracellular loop 2 that prevents G protein activation. ii, protomer B is also able to bind ligand and can transmit signal to the G protein, but the fused G protein is mutated such that guanine nucleotide binding is prevented, thus no signal is generated. iii, if protomers A and B are able to form a functional signaling unit (i.e., a heterodimer), it is possible for ligand binding at the functional receptor (protomer B) to result in signal rescue via the G protein of protomer A. Because signal is generated only upon functional reconstitution, the signal-to-noise ratio of the assay is high and particularly amenable to high-throughput screening. B, the use of BRET assays to examine signaling at GPCR heteromers. i and ii, BRET signal is generated by energy transfer from Renilla reniformis luciferase (Rluc, herein fused to the C terminus of Protomer B) that has oxidized the exogenously applied substrate, coelenterazine, to a fluorescent protein [such as yellow fluorescent protein (YFP), here fused to the C terminus of β-arrestin]. However, given the limited distance that the Rluc signal can travel, the YFP moiety must be within 100 Å to receive and subsequently transmit energy in the form of fluorescence. Thus, if YFP signal is generated upon stimulation of Rluc, the two proteins to which they are fused must be in close proximity, such as would be expected of a protein-protein interaction such as a homo- or heterodimer. By combining BRET with β-arrestin recruitment to an activated receptor, it is possible to monitor changes in BRET ratio upon ligand stimulation. i, for most GPCRs, ligand binding to a receptor facilitates translocation of β-arrestins from the cytoplasm to the activated receptor at the plasma membrane. In this case, protomer A is activated and recruits β-arrestin-YFP but no change in BRET signal is observed when Rluc is absent. ii, stimulation of Rluc-fused protomer B with agonist also leads to recruitment of β-arrestin-YFP to the activated receptor, yet in contrast to i, Rluc and YFP are now in close proximity and Rluc is able to excite YFP. iii, by combining the above scenarios, it is possible to determine whether two protomers are in close proximity and therefore likely to exist as a heteromer. By coexpressing β-arrestin-YFP and protomers A and B-Rluc, stimulation of protomer A will lead to β-arrestin recruitment. However, only when protomers A and B are heteromers will a change in BRET ratio be apparent.

Fig. 4.

Effect of ligand occupancy and intrinsic efficacy at dopamine receptor homomers and implications for bivalent ligand design. A, Han et al. (2009) examined the signaling of dopamine D₂ receptor homomers using a variety of mutations at either protomer A or protomer B. The authors demonstrated that protomer B occupancy and the intrinsic efficacy of the ligand at protomer B led to allosteric modulation of the signal generated by protomer A, represented here graphically as protomer A “activation.” When both dopamine D₂ protomers are unbound, they generate a basal signal illustrated by i. Agonist binding to protomer A in the absence of protomer B occupancy leads to receptor activation (iii) that is greater than when protomer B possesses constitutive activity (ii), indicating that the active state of protomer B negatively allosterically modulates the signaling of protomer A. Consistent with this observation, Han et al. (2009) were able to demonstrate that maximal signaling from protomer A is achieved when protomer B is in a completely inactive conformation, such as when an inverse agonist is bound (iv). [Adapted from Han Y, Moreira IS, Urizar E, Weinstein H, and Javitch JA (2009) Allosteric communication between protomers of dopamine class A GPCR dimers modulates activation. Nat Chem Biol 5:688–695. Copyright © 2009 the Nature Publishing Group. Used with permission.] B, extrapolation of the findings of Han et al. (2009) to the rational design of bivalent ligands highlights the importance of the intrinsic efficacy of the individual moieties attached to the linker. For example, attachment of two agonist moieties (yellow and green) would lead to moderate receptor activation, as per the example of A, ii. The signal generated by protomer A would be enhanced if the second moiety was instead a neutral antagonist (pink), reflecting the example of single occupancy in A, iii where there is no allosteric modulation of protomer A by protomer B. If maximal signal via protomer A was required, the moiety interacting with protomer B should be an inverse agonist (red), as per A, iv.