Research ArticleGPCR SIGNALING

Mutations in the NPxxY motif stabilize pharmacologically distinct conformational states of the α1B- and β2-adrenoceptors

See allHide authors and affiliations

Science Signaling  12 Mar 2019:
Vol. 12, Issue 572, eaas9485
DOI: 10.1126/scisignal.aas9485

Pharmacological heterogeneity in GPCRs

G protein–coupled receptors (GPCRs) share a common overall structure that switches between inactive and active conformations, with the latter being stabilized by ligand binding. Ragnarsson et al. investigated the role of a switch formed by hydrogen bonds between the NPxxY motif in transmembrane helix 7 (TMH7) and a tyrosine residue in TMH5 in activation of the α1B- and β2-adrenoceptors (ARs). Mutating the switch stabilized the inactive receptor conformations, resulting in reduced signaling. Although the mutations reduced the agonist affinity of the β2-AR without affecting signaling efficacy, they enhanced the agonist affinity and reduced the signaling efficacy of the α1B-AR. These findings show that inactive conformations of individual GPCRs have different pharmacological properties, which may help with designing new drugs or improving the efficacy and reducing the side effects of existing drugs.

Abstract

G protein–coupled receptors (GPCRs) convert extracellular stimuli to intracellular responses that regulate numerous physiological processes. Crystallographic and biophysical advances in GPCR structural analysis have aided investigations of structure-function relationships that clarify our understanding of these dynamic receptors, but the molecular mechanisms associated with activation and signaling for individual GPCRs may be more complex than was previously appreciated. Here, we investigated the proposed water-mediated, hydrogen-bonded activation switch between the conserved NPxxY motif on transmembrane helix 7 (TMH7) and a conserved tyrosine in TMH5, which contributes to α1B-adrenoceptor (α1B-AR) and β2-AR activation. Disrupting this bond by mutagenesis stabilized the α1B-AR and the β2-AR in inactive-state conformations, which displayed decreased agonist potency for stimulating downstream IP1 and cAMP signaling, respectively. Compared to that for wild-type receptors, agonist-mediated β-arrestin recruitment was substantially reduced or abolished for all α1B-AR and β2-AR inactive-state mutants. However, the inactive-state β2-ARs exhibited decreased agonist affinity, whereas the inactive-state α1B-ARs had enhanced agonist affinity. Conversely, antagonist affinity was unchanged for inactive-state conformations of both α1B-AR and β2-AR. Removing the influence of agonist affinity on agonist potency gave a measure of signaling efficacy, which was markedly decreased for the α1B-AR mutants but little altered for the β2-AR mutants. These findings highlight the pharmacological heterogeneity of inactive-state GPCR conformations, which may facilitate the rational design of drugs that target distinct conformational states of GPCRs.

INTRODUCTION

G protein–coupled receptors (GPCRs) control major physiological functions by coupling extracellular signaling events to intracellular effectors in response to hormones, neurotransmitters, and sensory stimuli such as vision, olfaction, and taste (1). Despite GPCRs already being targeted in cancer, cardiac dysfunction, diabetes, central nervous system disorders, obesity, inflammation, and pain, there remains a need for drugs that are more selective for specific receptor subtypes or conformational states to reduce dose-limiting side effects and enhance efficacy. Advances in our understanding of GPCR structure-function relationships suggest that targeting active, intermediate, or inactive conformational states (25) of these highly dynamic structures may be advantageous. GPCRs exist in multiple conformations including the nonsignaling (resting) state that predominates in the absence of agonist. In contrast, agonist binding stabilizes the active state of the receptor, which promotes G protein binding and intracellular signaling. However, the molecular basis of activation and subsequent receptor inactivation remain to be fully elucidated.

GPCR crystal structures of the inactive and active states of the β2-adrenoceptor (β2-AR) (610), M2 muscarinic receptor (11, 12), rhodopsin (13, 14), δ-opioid receptor (15, 16), and μ-opioid receptor (17, 18), as well as the inactive state of the A2A adenosine receptor (19), reveal an extended water-mediated polar network that connects the orthosteric binding pocket to cytoplasmic G protein–coupling and activation. Activation of GPCRs is directed by ligand-induced conformational changes in the transmembrane helices (TMHs) and associated extracellular and intracellular loops, culminating in the formation of a Tyr5.58-Tyr7.53 [superscripts denote Ballesteros-Weinstein numbering (20)] hydrogen-bonded “activation switch” through water molecules, linking the conserved NPxxY7.53 motif on TMH7 with the conserved Tyr5.58 in TMH5. This coupling favors an outward movement in the cytoplasmic end of TMH6, which facilitates binding of the G protein and intracellular signaling, as seen in crystal structures of active-state β2-AR, rhodopsin, M2 muscarinic, and μ-opioid receptors (7, 8, 12, 17, 2123). However, molecular modeling and pharmacological evaluation are required to better understand the molecular mechanism of activation for GPCRs, especially those that are yet to be crystallized, including the α1B-AR.

The α1B-AR is a member of family A GPCRs that mediates smooth and cardiac muscle contraction in various tissues, including the heart, prostate, gut, and gall bladder (2427). Although the α1B-AR shares common conserved motifs with other family A GPCRs, including the ionic lock switch (E/DRY) motif on TMH3, the rotamer toggle switch (CWxP) motif on TMH6, and the tyrosine toggle switch motif (NPxxY) on TMH7 (28), the contribution of these motifs for α1B-AR function has not been fully investigated experimentally. A study by Venkatakrishnan et al. (29) suggested that Tyr7.53 is pivotal for receptor activation in all class A GPCRs, supporting mutational studies at the NPxxY motif that decreased or abolished signaling by the β2-AR (30), adenosine A2B receptor (31), type 1 angiotensin II receptor (32), M3 muscarinic acetylcholine receptor (33), N-formyl peptide receptor (34), cholecystokinin B receptor (35), B2 bradykinin receptor (36), rhodopsin (37), and the V2 vasopressin receptor (29). This decrease in signaling arises directly from reduced or abolished activation and not impaired Gα protein binding to the receptor (33, 35, 36). In addition to reduced signaling for inactive-state GPCRs, several studies report reduced agonist affinity in inactive-state receptors, including the angiotensin II receptor (38) and the β2-AR (30, 39). In contrast, active-state receptors exhibit high agonist affinity, as shown for the β2-AR (8, 39), κ-opioid receptor (40), and μ-opioid receptor (17), as do receptors stabilized in the active state by a nanobody (9) or Gαs protein (8), which has facilitated their study using crystallographic techniques. However, the N-formyl cholecystokinin B receptor (35), the B2 bradykinin receptor (36), and rhodopsin (37) showed no change in agonist affinity when the NPxxY motif was mutated, whereas the M3 muscarinic acetylcholine receptor showed higher affinity for the agonist acetylcholine when Asn7.49 was mutated to alanine (33).

To investigate whether the α1B-AR shares the same activation mechanism and a similar inactive conformational state (involving the NPxxY motif) as the closely related and well-studied β2-AR, we stabilized inactive-state conformations of α1B- and β2-ARs by mutating the conserved NPxxY motif and the interacting residue Tyr5.58 in TMH5 to destabilize their active-state conformations (68). Thus, to stabilize the inactive α1B-AR and β2-AR conformational states, the conserved Tyr5.58 and Tyr7.53 residues were mutated to phenylalanine, either singly or as a double mutant to break the hydrogen-bonded activation switch and lock the receptors in inactive conformations. We pharmacologically characterized these inactive-state mutants using α1B-AR orthosteric [prazosin (41)] and allosteric [ρ-TIA (41, 42)] inhibitors and the β2-AR inhibitor propranolol and the partial agonist CGP-12177 (43) for radioligand binding assays. In addition, we determined alterations in norepinephrine (NE) signaling through α1B-AR to inositol 1-phosphate (IP1), cyclic adenosine monophosphate (cAMP), or β-arrestin, as well as alterations in isoproterenol signaling through β2-AR to cAMP and β-arrestin. In contrast to most family A GPCRs that show enhanced agonist affinity for active-state receptors compared to inactive- and resting-state receptors, the α1B-AR showed enhanced agonist affinity for both inactive- and active-state (44) receptors compared to the resting-state receptor. Our data suggest that closely related GPCRs can adopt different agonist affinity profiles despite sharing conserved signaling motifs and a common activation mechanism.

RESULTS

Constructing molecular models of the active- and inactive-state α1B-AR

We constructed the molecular model of active-state α1B-AR using the crystal structure of the β2-AR-Gs protein complex [Protein Data Bank (PDB) code, 3SN6 (8)] as a template and the automated mode of the SWISSMODEL online server. We visualized and aligned the α1B-AR and β2-AR models in PyMOL to generate a structural superposition of the two models with a root mean square deviation (RMSD) of 0.211. Similarly, we modeled the inactive state of the α1B-AR using the inactive β2-AR-T4 lysozyme fusion protein complex [PDB code, 2RH1 (6)], visualized and aligned it in PyMOL with the β2-AR template, and generated a structural superposition of the two models with an RMSD of 0.247. These models suggest that the α1B-AR can adopt active and inactive conformations that are structurally similar to the active and inactive β2-AR conformations, in particular in regard to the NPxxY motif (Fig. 1). Furthermore, the water-mediated polar network seen in the NPxxY region in the active μ-opioid receptor crystal structure [μ-OR; PDB code, 5C1M] (17), and suggested in the active β2-AR crystal structure (10), is likely to exist in the active-state α1B-AR model because the residues involved in this polar network in the μ-opioid receptor and the β2-AR are the same in the α1B-AR and also positioned within equivalent distances in respective receptor (Fig. 1).

Fig. 1 Cytoplasmic view of the active- and inactive-state structures around the conserved NPxxY motif.

Comparisons of active- and inactive-state homology models of the α1B-AR with the β2-AR and μ-OR structures. When the inactive-state receptors are activated by agonist, TMH6 moves outward and TMH7 moves inward to allow the asparagine (Asn7.49) and the tyrosine (Tyr7.53) in the N7.49PxxY7.53 motif to form a hydrogen-bonded activation switch involving water molecules (red dots), which also includes Tyr5.58 in TMH5 and residues in TMH3 and TMH6. In contrast, in the inactive-state receptors, TMH6 obstructs the formation of the activation switch.

Prazosin and CGP-12177 affinity and Bmax for resting- and inactive-state α1B-ARs and β2-ARs

The following point mutations were created in the α1B-AR to disrupt the suggested activation switch between TMH7 and TMH5 and thus stabilize the receptor in the inactive state: Y223F5.58, Y348F7.53, and the double-mutant Y223F5.58/Y348F7.53. Analogous point mutations in the β2-AR were created to stabilize it in the inactive state: Y219F5.58, Y326F7.53, and the double-mutant Y219F5.58/Y326F7.53 to stabilize inactive-state β2-AR. Wild-type receptor is not stabilized in any conformation, which means that it is flexible and dynamic and thus represented resting-state receptor in this study. We used the orthosteric antagonist prazosin to evaluate the structural integrity and expression of wild-type and mutant α1B-ARs and the β2-AR agonist CGP-12177 to evaluate the β2-ARs in membranes from COS-1 cells transiently transfected with the α1B-ARs and β2-ARs, respectively. None of the α1B-AR mutations statistically significantly affected prazosin affinity, whereas the CGP-12177 affinity was decreased threefold in the Y219F5.58 mutant compared with wild-type β2-AR (Tables 1 and 2). Bmax was decreased nearly twofold for the Y223F5.58/Y348F7.53 mutant compared with wild-type α1B-AR, whereas Bmax increased over twofold for the Y219F5.58 and Y326F7.53 mutants compared with wild-type β2-AR (Tables 1 and 2).

Table 1 Pharmacological characterization of WT and mutant α1B-ARs showing Bmax % of WT determined from saturation binding assays; NE EC50 and NE Emax % of WT determined measuring IP1 accumulation in response to increasing concentrations of NE in an IP1 HTRF assay; prazosin Ki, NE Ki, and ρ-TIA IC50 determined from radioligand binding assays, with NE efficacy calculated from NE pEC50 − NE pKi; NE EC50 determined measuring cAMP; and β-arrestin recruitment and logτ determined from the operational model.

Values are the mean ± SEM for the number of (n) separate experiments. nd, not detected.

View this table:
Table 2 Pharmacological characterization of WT and mutant β2-ARs showing CGP-12177 Ki and Bmax % of WT determined from saturation binding assays; isoproterenol (Iso) EC50 and Iso Emax % of WT determined measuring cAMP accumulation in response to increasing concentrations of Iso in a LANCE Ultra cAMP assay; Iso Ki, CGP-12177 Ki, and propranolol Ki determined from radioligand binding assays, with Iso efficacy calculated from Iso pEC50 − Iso pKi; Iso EC50 determined measuring β-arrestin recruitment; and logτ determined from the operational model.

Values are the mean ± SEM for the number of (n) separate experiments.

View this table:

Gq signaling by inactive-state α1B-ARs in response to NE

To characterize the ability of the α1B-AR mutants to signal through Gq, we measured NE-stimulated accumulation of myo-IP1, a downstream metabolite of the second messenger inositol 1,4,5-trisphosphate (IP3), which is generated by Gq activation, using a homogeneous time-resolved fluorescence assay for IP1 (IP1 HTRF) in transiently transfected COS-1 cells. All mutants statistically significantly reduced NE-activated signaling potency [half-maximal excitatory concentration (EC50)] even after adjusting the EC50 for the lower Bmax (~85% of wild type for Y223F5.58, ~80% of wild type for Y348F7.53, and ~52% of wild type for Y223F5.58/Y348F7.53). The Y223F5.58, Y348F7.53, and Y223F5.58/Y348F7.53 mutants significantly decreased NE potency 83-, 17-, and 39-fold, respectively, compared with the wild-type receptor (14.44 ± 4.02 nM) (Fig. 2, A and B, and Table 1). The maximal response (Emax) for all the mutants was also statistically significantly reduced compared to the wild-type receptor (Table 1).

Fig. 2 Effect of mutating the activation switch involving the NPxxY motif on EC50 in the α1B-AR and β2-AR.

(A) Representative NE concentration-response curve measuring IP1 accumulation for wild-type (WT) (control) and mutant α1B-ARs using an IP1 signaling assay in transiently transfected COS-1 cells. Data are means ± SEM of one experiment performed in triplicate. (B) Comparison of NE EC50 values for WT and α1B-AR mutants measuring IP1 accumulation in response to increasing concentrations of NE in transiently transfected COS-1 cells. Values are means ± SEM of six independent experiments for WT and four to five independent experiments for each mutant (each performed in triplicate). Data were analyzed by Dunnett’s multiple comparisons test after one-way analysis of variance (ANOVA). **P < 0.01, ***P < 0.001, and ****P < 0.0001 compared to WT. (C) Representative isoproterenol concentration-response curve measuring cAMP accumulation for WT (control) and mutant β2-ARs using a cAMP signaling assay in transiently transfected CHO-1 cells. Data are means ± SEM of one experiment performed in triplicate. (D) Comparison of isoproterenol EC50 values for WT and β2-AR mutants measuring cAMP in response to increasing concentrations of isoproterenol in transiently transfected CHO-1 cells. Values are means ± SEM of three independent experiments (each performed in triplicate) and were analyzed by Dunnett’s multiple comparisons test after one-way ANOVA. **P < 0.01 compared to WT.

Gs signaling by inactive-state β2-ARs in response to isoproterenol

To characterize the ability of the β2-AR mutants to signal through Gs, we performed assays to measure isoproterenol-stimulated accumulation of cAMP in transiently transfected Chinese hamster ovary (CHO) cells using a LANCE Ultra cAMP kit, which is a homogeneous time-resolved fluorescence resonance energy transfer immunoassay designed to measure cAMP accumulation. After adjusting for the increase in Bmax (~248% of wild type for Y219F5.58, ~261% of wild type for Y326F7.53, and ~244% of wild type for Y219F5.58/Y326F7.53), all the β2-AR mutants reduced isoproterenol-activated signaling potency (EC50). The Y219F5.58, Y326F7.53, and Y219F5.58/Y326F7.53 mutants decreased isoproterenol potency 9-, 5-, and 47-fold, respectively, compared with the wild-type receptor (0.086 ± 0.024 nM) (Fig. 2, C and D, and Table 2), although this change only reached significance for the Y219F5.58/Y326F7.53 mutant. The maximal response (Emax) was statistically significantly reduced for the Y219F5.58 and Y2195.58F/Y326F7.53 mutants compared to the wild-type receptor (Table 2), whereas the reduction in Emax for the Y226F mutant was not statistically significant.

NE affinity of inactive-state α1B-ARs

To determine the affinity [inhibition constant (Ki)] of the α1B-AR mutants for NE, we measured displacement of [3H]prazosin by NE in membranes from COS-1 cells transiently transfected with wild-type and mutant α1B-ARs. The α1B-AR mutants Y348F7.53 (7-fold) and Y223F5.58/Y348F7.53 (16-fold) significantly enhanced the affinity for NE, whereas the 3-fold increase in affinity in the Y223F5.58 mutant was not significant compared with wild-type α1B-AR (Ki = 18.71 ± 1.76 μM) (Fig. 3, A and B, and Table 1).

Fig. 3 Effect of mutating the activation switch involving the NPxxY motif on Ki in the α1B-AR and β2-AR.

(A) Representative NE concentration-response curve determined by displacement of the radiolabeled α1-AR antagonist [3H]prazosin using membranes from α1B-AR–transfected COS-1 cells and increasing concentrations of NE for WT and mutant α1B-ARs. Data are means ± SEM of one experiment performed in triplicate. (B) Comparison of NE Ki values for WT and α1B-AR mutants determined by displacement of [3H]prazosin by increasing concentrations of NE. Values are means ± SEM of 12 independent experiments for WT and 3 to 5 independent experiments for each mutant (each performed in triplicate). Data were analyzed by Dunnett’s multiple comparisons test after one-way ANOVA. ****P < 0.0001 compared to WT. (C) Representative isoproterenol concentration-response curve determined by displacement of the radiolabeled β2-AR agonist [3H]CGP-12177 using membranes from β2-AR–transfected COS-1 cells and increasing concentrations of isoproterenol for WT and mutant β2-ARs. Data are means ± SEM of one experiment performed in triplicate. (D) Comparison of isoproterenol Ki values for WT and β2-AR mutants determined by displacement of [3H]CGP-12177 by increasing concentrations of isoproterenol. Values are means ± SEM of four independent experiments for WT and three independent experiments for each mutant (each performed in triplicate). Data were analyzed by Dunnett’s multiple comparisons test after one-way ANOVA. *P < 0.05 compared to WT.

Isoproterenol affinity of inactive-state β2-ARs

We determined the affinity (Ki) of the β2-AR mutants for isoproterenol by measuring the displacement of [3H]CGP-12177 by isoproterenol in membranes from COS-1 cells transiently transfected with wild-type and mutant β2-ARs. As expected, the Y219F5.58, Y326F7.53, and Y219F5.58/Y326F7.53 mutants decreased the affinity for isoproterenol nine-, five-, and sevenfold, respectively, compared with the wild-type receptor (0.83 ± 0.088 μM), although this change was not significant for the Y326F7.53 mutant (Fig. 3, C and D, and Table 2).

NE signaling by inactive-state α1B-AR

To establish how effectively NE activates the wild-type and inactive-state forms of α1B-ARs, we removed the contribution of NE affinity (pKi) from our measure of NE potency (pEC50) to obtain a measure of NE signaling efficacy (NE efficacy = pEC50 − pKi). Each of the α1B-AR mutants significantly reduced (234-fold for Y223F5.58, 120-fold for Y348F7.53, and 616-fold for Y223F5.58/Y348F7.53) the efficacy of NE signaling compared with the wild-type receptor (3.11 ± 0.37) (Fig. 4A and Table 1). In addition, we used the operational model of agonism to calculate an operational measure of efficacy (logτ) for wild-type and inactive-state α1B-ARs, which was corrected for relative receptor abundance (45). In agreement with our log(NE efficacy) values, the α1B-AR mutants Y223F5.58 (165-fold), Y348F7.53 (69-fold), and Y223F5.58/Y348F7.53 (181-fold) showed statistically significantly reduced logτ compared with the wild-type receptor (2.96 ± 0.25) (Table 1).

Fig. 4 Effect of mutating the activation switch involving the NPxxY motif on signaling efficiency (efficacy) of the α1B-AR and β2-AR.

(A) To characterize how effectively NE activated WT and mutant α1B-ARs, the NE efficacy was calculated as the NE pEC50 value minus the NE pKi value (pEC50 − pKi) for NE on WT and α1B-AR mutants. Values are means ± SEM of six independent experiments for WT and three to four independent experiments for each mutant (each performed in triplicate). Data were analyzed by Dunnett’s multiple comparisons test after one-way ANOVA. **P < 0.01 and ***P < 0.001 compared to WT. (B) To characterize how effectively isoproterenol activated WT and mutant β2-ARs, the isoproterenol efficacy was calculated as the isoproterenol pEC50 value minus the isoproterenol pKi value (pEC50 − pKi) for isoproterenol on WT and β2-AR mutants. Values are means ± SEM of three independent experiments (each performed in triplicate) and were analyzed by Dunnett’s multiple comparisons test after one-way ANOVA.

Isoproterenol signaling by inactive-state β2-AR

To establish how effectively isoproterenol activates the wild-type and inactive-state β2-ARs, we removed the contribution of isoproterenol affinity (pKi) from our measure of isoproterenol potency (pEC50) to obtain a measure of isoproterenol signaling efficacy (NE efficacy = pEC50 − pKi). The β2-AR mutants Y219F5.58, Y326F7.53, and Y219F5.58/Y326F7.53 did not show significantly reduced isoproterenol signaling efficacy compared with the wild-type receptor (3.98 ± 0.39) (Fig. 4B and Table 2). In addition, we used the operational model of agonism to calculate an operational measure of efficacy (logτ) for wild-type and inactive-state β2-ARs, which was corrected for relative receptor abundance (45). In agreement with our log(isoproterenol efficacy) values, the β2-AR mutants Y219F5.58 (5-fold), Y326F7.53 (4-fold), and Y219F5.58/Y326F7.53 (14-fold) did not statistically significantly reduce logτ compared with the wild-type receptor (3.97 ± 0.30) (Table 2).

Allosteric antagonist ρ-TIA affinity of inactive-state α1B-ARs

To determine whether the affinity changes observed for NE extended to allosteric antagonists acting on the inactive-state α1B-AR, we determined the affinity of the allosteric antagonist ρ-TIA for wild-type and inactive-state α1B-ARs by displacement of [3H]prazosin in membranes from COS-1 cells transiently transfected with α1B-ARs. The Y223F5.58, Y348F7.53, and Y223F5.58/Y348F7.53 mutants showed no change in ρ-TIA affinity compared with the wild-type receptor (0.96 ± 0.035 nM) (Fig. 5A and Table 1).

Fig. 5 Characterization of antagonist binding to inactive-state α1B-ARs and β2-ARs.

(A) Comparison of WT and α1B-AR mutant IC50 for ρ-TIA determined using the radiolabeled α1B-AR antagonist [3H]prazosin and increasing concentrations of ρ-TIA. Values are means ± SEM of three independent experiments (each performed in triplicate) and were analyzed by Dunnett’s multiple comparisons test after one-way ANOVA. (B) Comparison of WT and β2-AR mutant IC50 for propranolol determined using the radiolabeled β2-AR agonist [3H]CGP-12177 and increasing concentrations of propranolol. Values are means ± SEM of three independent experiments (each performed in triplicate) and were analyzed by Dunnett’s multiple comparisons test after one-way ANOVA. *P < 0.05 compared to WT.

Antagonist propranolol binding to inactive-state β2-ARs

Propranolol affinity (Ki) for wild-type and inactive-state β2-ARs was determined from displacement of [3H]CGP-12177 in membranes from COS-1 cells transiently transfected with wild-type and mutant β2-ARs. There was no change in propranolol affinity for the single mutants, Y219F5.58 and Y326F7.53, whereas the Y219F5.58/Y326F7.53 mutant showed a fourfold decrease in affinity compared with the wild-type receptor (1.46 ± 0.23 nM) (Fig. 5B and Table 2).

Biased signaling to cAMP and β-arrestin-2 recruitment by the inactive-state α1B-ARs

Although the α1B-AR predominantly couples to Gq, we investigated whether the activation switch mutations had any effect on Gs signaling and cAMP production or a change in β-arrestin recruitment. Thus, we determined the EC50 for NE-activated increases in intracellular cAMP for the activation switch mutants and the wild-type receptor using a LANCE Ultra cAMP assay in COS-1 cells transiently transfected with α1B-ARs. The Y223F5.58, Y348F7.53, and Y223F5.58/Y348F7.53 mutants decreased NE potency 118-, 70-, and 30-fold, respectively, compared with the wild-type receptor (0.34 ± 0.076 μM) (Fig. 6, A and B). The EC50 for NE-induced β-arrestin-2 recruitment using a bioluminescence resonance energy transfer (BRET) assay in human embryonic kidney (HEK) 293 cells transiently transfected with α1B-ARs could only be determined for the Y348F7.53 mutant, which was not different from the wild-type receptor (36.7 ± 5.71 nM), whereas no β-arrestin-2 recruitment could be detected for the other mutants (Fig. 6, C and D). However, although the potency for NE-induced β-arrestin-2 recruitment for the Y348F7.53 mutant was not different from the wild-type receptor, the absolute accumulation of β-arrestin-2 decreased (Fig. 6C).

Fig. 6 Characterization of biased signaling and β-arrestin recruitment by the α1B-AR and β2-AR mutants.

(A) Representative NE concentration-response curve measuring cAMP accumulation for WT and mutant α1B-ARs using a LANCE Ultra cAMP signaling assay. Data are means ± SEM of one experiment performed in triplicate. (B) Comparison of NE EC50 values for WT and α1B-AR mutants measuring cAMP accumulation in response to increasing concentrations of NE in transiently transfected COS-1 cells. Values are means ± SEM of five independent experiments for WT and three to four independent experiments for each mutant (each performed in triplicate). Data were analyzed by Dunnett’s multiple comparisons test after one-way ANOVA. **P < 0.01 and ***P < 0.001 compared to WT. (C) Representative NE concentration-response curve measuring β-arrestin-2 recruitment for WT and mutant α1B-ARs using a BRET assay. Data are means ± SEM of one experiment performed in triplicate. (D) Comparison of EC50 values for NE-induced β-arrestin-2 recruitment by WT and mutant α1B-ARs in transiently transfected HEK293 cells. Values are means ± SEM of four independent experiments for WT and three independent experiments for each mutant (each performed in triplicate). Data were analyzed by Dunnett’s multiple comparisons test after one-way ANOVA. (E) Representative isoproterenol concentration-response curve measuring β-arrestin-2 recruitment for WT and mutant β2-ARs using a BRET assay. Data are means ± SEM of one experiment performed in triplicate. (F) Comparison of EC50 values for isoproterenol-induced β-arrestin-2 recruitment by WT and mutant β2-ARs in transiently transfected HEK293 cells. Values are means ± SEM of three independent experiments (each performed in triplicate) and were analyzed by Dunnett’s multiple comparisons test after one-way ANOVA.

Biased signaling to IP1 and β-arrestin-2 recruitment by the inactive-state β2-ARs

Although the β2-AR predominantly couples to Gs, we investigated whether the activation switch mutants had any effect on Gq signaling and IP1 production using an IP1 HTRF assay. Confirming this signaling bias, isoproterenol failed to stimulate the accumulation of IP1 for wild-type β2-AR or its activation switch mutants in CHO cells transiently transfected with β2-ARs (fig. S1). To investigate whether inactive-state β2-ARs had an effect on β-arrestin recruitment, we determined the EC50 for isoproterenol-induced increases in β-arrestin-2 recruitment for the activation switch mutants and wild-type β2-ARs using a BRET assay in HEK293 cells transiently transfected with β2-ARs. The EC50 for isoproterenol-induced β-arrestin-2 recruitment could only be determined for the Y219F5.58 mutant, which was not different from the wild-type receptor (23.3 ± 0.39 nM), whereas no β-arrestin-2 recruitment could be detected for the other mutants (Fig. 6, E and F). However, although the potency for NE-induced β-arrestin-2 recruitment for the Y219F5.58 mutant was not different from the wild-type receptor, the absolute accumulation was nearly eliminated (Fig. 6E).

DISCUSSION

Despite advances in structural analysis of GPCRs in active and inactive conformations, pharmacological characterization continues to augment our understanding of GPCR structure-function relationships and signaling. In this study, we hypothesized that the hydrogen-bonded activation switch network between the NPxxY motif and TMH5, seen in several active-state GPCRs (7, 8, 12, 17, 2123), plays a similar role in the α1B-AR. To test our hypothesis, we investigated the role played by the Tyr5.58-Tyr7.53 hydrogen bond in α1B-AR activation by mutating these residues in the α1B-AR and the related β2-AR. These studies revealed that the NPxxY motif played an important role in the activation of the Gq, Gs, and β-arrestin signaling pathways for both α1B- and β2-ARs. Consistent with previous studies (30, 39), the inactive state of the β2-AR had reduced agonist affinity compared to the resting state of the receptor. However, the inactive state of the α1B-AR had enhanced agonist affinity compared to the resting state, suggesting that inactive-state conformations can affect agonist affinities in a receptor-specific manner.

In the absence of an α1B-AR crystal structure, we built molecular homology models of the α1B-AR using the active (8) and inactive (6) β2-AR structures. Our active α1B-AR model aligned with the β2-AR-Gs protein complex with an RMSD of 0.211, and our inactive α1B-AR model aligned with the β2-AR-T4 lysozyme protein complex with an RMSD of 0.247. These models revealed a similar movement of Tyr7.53 going from the inactive to the active state as shown for the β2-AR (7, 8) and the μ-opioid receptor (17), suggesting that they have a similar activation mechanism (Fig. 1). To establish the role of the NPxxY motif and the importance of the Tyr5.58-Tyr7.53 hydrogen bond in the α1B-AR, we mutated the corresponding tyrosines, Tyr5.58 and Tyr7.53, to phenylalanine, generating both the single mutants Y223F5.58 and Y348F7.53 as well as the double-mutant Y223F5.58/Y348F7.53, to determine whether removing these hydrogen bond interactions inhibited receptor activation by destabilizing the active-state conformation of the receptor, as shown previously for the β2-AR (30). To support our pharmacological characterization of the α1B-AR, we mutated the equivalent residues in the β2-AR and used the wild-type receptor to reflect the pharmacology of the resting-state conformation. Mutating the Tyr5.58-Tyr7.53 activation switch residues decreased NE and isoproterenol potency at the α1B-AR and β2-AR, respectively, and decreased the maximal responses (Emax) of both the single and double mutants of these receptors, which was adjusted for any changes in expression for either α1B-ARs or β2-ARs compared with the corresponding wild-type receptors (Fig. 2, A to D, and Tables 1 and 2). These results are consistent with previous studies showing that disrupting the Tyr5.58-Tyr7.53 hydrogen bond in the β2-AR (30), adenosine A2B (31), type 1 angiotensin II (32), rhodopsin (37), and V2 vasopressin (29) receptors produced an inactive receptor conformation, based on reduced G protein activation and signaling.

To investigate whether the inactive-state mutants shifted agonist signaling bias, we determined how the α1B-AR inactive-state mutants affected NE-induced cAMP accumulation and β-arrestin recruitment and how the β2-AR inactive-state mutants affected isoproterenol-activated IP1 accumulation and β-arrestin recruitment. The wild-type α1B-AR is able to couple to the Gs protein and thus showed NE-activated cAMP signaling. However, the α1B-AR inactive-state mutants showed much decreased NE-activated cAMP potency and decreased signaling efficacy compared with the wild-type receptor (Fig. 6, A and B). In contrast, we did not detect any Gq protein coupling and IP1 signaling for wild-type or mutant β2-ARs (fig. S1). In addition, β-arrestin recruitment was strongly reduced or abolished in the α1B-AR inactive-state mutants, although the NE-activated β-arrestin recruitment potency for the Y348F7.53 mutant was not unchanged from the wild-type α1B-AR (Fig. 6, C and D). Similarly, isoproterenol stimulation also showed strongly reduced or abolished β-arrestin signaling in the β2-AR mutants, but the measurable agonist potency for the Y219F5.58 mutant was not different from the wild-type receptor (Fig. 6, E and F). Thus, the decreased signaling observed in these activation switch mutants, for both the α1B-ARs and β2-ARs, is likely due to the receptors being stabilized in an inactive-state conformation. In agreement with our study, which shows that mutating the Tyr5.58-Tyr7.53 activation switch residues in two different GPCRs decreases or abolishes both G protein and β-arrestin signaling, a study by Venkatakrishnan et al. (29) reported that Tyr7.53 is critical for receptor activation in all class A GPCRs.

To further pharmacologically characterize these inactive-state receptors, we measured agonist and antagonist affinities for inactive-state forms of α1B-AR and β2-AR. In agreement with previous studies of inactive-state β2-AR, antagonist affinity was little altered in the inactive state (30, 39). Specifically, propranolol affinity of inactive-state β2-ARs was unaltered, except for a small fourfold decrease in affinity for the Y219F5.58/Y326F7.53 β2-AR mutant (Fig. 5B and Table 2). Similarly, inactive-state α1B-ARs showed no change in affinity for the orthosteric antagonist prazosin or the allosteric antagonist ρ-TIA (41, 42) (Fig. 5A and Table 1). In agreement with previous studies of inactive-state β2-AR (30, 39), the agonist isoproterenol had weaker affinity for inactive-state β2-ARs mutants Y219F5.58 (ninefold), Y326F7.53 (fivefold), and Y219F5.58/Y326F7.53 (sevenfold) compared with the resting-state (wild-type) receptor (Fig. 3, C and D, and Table 2). In contrast, the Nb60 nanobody–stabilized inactive state of the β2-AR had a 70-fold lower isoproterenol affinity compared to the wild-type receptor (39). This discrepancy in affinity shift for mutated and Nb60-stabilized inactive-state receptors suggests that pharmacologically distinct inactive and active receptor states exist with specific ligand binding profiles. In contrast to the inactive-state β2-ARs, inactive-state α1B-ARs had high affinity for agonist compared with the resting-state receptor. The α1B-AR Y223F5.58/Y348F7.53 double mutant enhanced NE affinity 16-fold, whereas the single mutants Y223F5.58 and Y348F7.53 enhanced NE affinity three- and sevenfold, respectively, compared with wild-type receptor (Fig. 3, A and B, Table 1). In addition, active-state α1B-ARs also have high agonist affinity when the ionic lock is broken by mutating Glu289 (Glu6.30) (44).

In addition to measuring ligand potency and affinity, we determined the signaling efficacy of NE and isoproterenol for the inactive-state α1B-ARs and β2-ARs, respectively, to reveal how effectively agonists can activate the inactive-state receptors when the influence of any changes in agonist affinity are removed (46). Inactive-state α1B-ARs had statistically significantly reduced NE signaling efficacy in the single mutants Y223F5.58 (234-fold) and Y348F7.53 (120-fold) and in the double-mutant Y223F5.58/Y348F7.53 (616-fold). In contrast, the signaling efficacy of isoproterenol for inactive-state β2-ARs was unchanged for the two single mutants (Y219F5.58 and Y326F7.53) and decreased only sevenfold for the double mutant (Y219F5.58/Y326F7.53). In support, using the operational model of agonism to calculate efficacy, after adjusting for decreased or increased receptor levels compared with wild type, gave equivalent results. Both methods independently suggest that it is harder for the agonist to activate the inactive-state α1B-ARs than for them to activate the inactive-state β2-ARs.

Our data suggest that the α1B-AR likely uses an extended network of water molecules between the NPxxY motif in TMH7 and TMH5 to form an activation switch that stabilizes the active receptor conformation, as shown for other GPCRs (7, 8, 12, 17, 2123). Thus, abolishing this hydrogen-bonded activation switch impairs signaling for both the α1B- and β2-AR by stabilizing inactive-state conformations. The α1B-AR had higher affinity for its orthosteric agonist NE when stabilized in the inactive state in contrast to the β2-AR, which displayed reduced agonist affinity in the inactive state. In addition, these inactive-state α1B-ARs show a much greater decrease in agonist signaling efficacy compared to the inactive-state β2-ARs. Therefore, we hypothesize that the affinity and energy landscapes for the α1B-ARs and β2-ARs differ when the receptors transition from the inactive to the active state. We propose that the α1B-AR has increased affinity for agonist in both the inactive and active receptor conformations compared to the resting state, whereas the β2-AR has low affinity in the inactive state but high affinity in the active state (Fig. 7).

Fig. 7 Relative affinity landscape models for the α1B-AR and β2-AR conformational states.

Proposed affinity changes for the α1B-AR and β2-AR during the transition from the inactive to the active state in the presence of the agonists NE and isoproterenol, respectively.

The Kobilka group has shown how the energy barriers change for the β2-AR when the receptor transitions from the inactive to the active state, emphasizing the G protein (Gs)– or nanobody (Nb80)–bound active receptor as the lowest energy state (4). They also suggest that the β2-AR easily transitions between different states because the relatively low energy barriers allow for rapid exchange between conformational states (4, 21). Our results show significantly decreased agonist signaling efficacy for inactive-state α1B-ARs, but unchanged efficacy for inactive-state β2-ARs (Fig. 4, A and B). Thus, we hypothesize that more energy is needed to transition the α1B-AR to an active conformation compared with the β2-AR, despite the hydrogen network involving the NPxxY motif following a common activation mechanism for the α1B-AR and β2-AR. To explain this difference, we suggest that stabilizing the inactive state by mutating the hydrogen-bonded activation switch between Tyr5.58 and Tyr7.53 results in different inactive-state conformations for these two receptors. This is in agreement with Staus et al. (39), who suggest that there are multiple distinct active and inactive conformational states with different ligand binding properties that modulate downstream cellular responses. The importance of key conserved residues in activation and function is reinforced in diseases wherein these residues are mutated. For example, mutations at Tyr7.53 in the melanocortin-4 receptor cause early-onset obesity (47), and mutations in the melatonin receptor 1B increase the risk of type 2 diabetes (48), whereas a mutation at Tyr7.53 in the gonadotropin-releasing hormone receptor causes hypogonadotropic hypogonadism (49).

In conclusion, the high-affinity, inactive-state α1B-AR conformation emphasizes the importance of pharmacological characterizations of GPCRs in different conformational states to understand their distinct energy landscapes and pharmacological profiles. Some GPCRs, such as the α1B-AR investigated here, have a small range of agonist affinity between active- and inactive-state receptors, whereas the β2-AR shows a 15,000-fold difference in agonist affinity (isoproterenol) between active- and inactive-state receptors stabilized with nanobodies or G protein binding (39). Our findings highlight the pharmacological heterogeneity across inactive-state GPCR conformations using mutational studies that may help guide the rational development of conformational state–dependent therapeutics.

MATERIALS AND METHODS

Molecular modeling

A molecular homology model of the active-state α1B-AR was generated using the crystal structure of the β2-AR-Gs protein complex [PDB code, 3SN6 (8)] as a template, using the automated mode of the SWISSMODEL online server. In the same way, a molecular homology model of the inactive-state α1B-AR was generated using the crystal structure of the β2-AR-T4 lysozyme fusion protein complex [PDB code, 2RH1 (6)]. Briefly, the FASTA sequence for the α1B-AR was loaded together with the crystal structure of the β2-AR-Gs protein complex [PDB code, 3SN6 (8)] or the β2-AR-T4 lysozyme fusion protein complex [PDB code, 2RH1 (6)], their sequences were automatically aligned, and α1B-AR models were generated. The resulting models were energy-minimized using the GROMOS force field in the program DEEPVIEW, and the resulting models were visualized and aligned in PyMOL with their respective β2-AR structure to generate a structural superposition of the two models to generate an RMSD measurement.

Site-directed mutagenesis

The hamster α1B-AR complementary DNA (cDNA) in the pMT2′ vector was provided by R. M. Graham (Victor Chang Cardiac Research Institute, Sydney, Australia) and the human β2-AR cDNA in the pcDNA3.1 vector was provided by B. Kobilka (Stanford University, CA, USA). The α1B-AR and β2-AR subunit cDNA was subjected to in vitro site-directed mutagenesis using the QuikChange Mutagenesis Kit (Stratagene) following the manufacturer’s instructions. The following point mutations were created in the α1B-AR: Y223F5.58, Y348F7.53, and the double-mutant Y223F5.58/Y348F7.53. The following point mutations in the β2-AR were created: Y219F5.58, Y326F7.53, and the double-mutant Y219F5.58/Y326F7.53. Primers used to generate the mutants were from Sigma-Aldrich. TOP10 Escherichia coli (Invitrogen) were transformed with wild-type or mutant cDNA and subsequently used for plasmid preparation using a PureLink Quick Plasmid Miniprep kit (Invitrogen) or a High Speed Maxi kit (Qiagen). Purified cDNA was used to confirm all mutations by sequencing by the Australian Genome Research Facility.

Cell culture methods, transient expression of α1B- and β2-ARs, and membrane preparation

COS-1 and HEK293 cells [American Type Culture Collection (ATCC)] were cultured in Dulbecco’s modified Eagle’s medium containing glutamine and 5 and 10% fetal bovine serum, respectively. CHO-1 cells (ATCC) were cultured in F-12 medium containing 10% fetal bovine serum. Cells were transiently transfected with purified plasmid DNA encoding wild-type or mutant α1B- and β2-ARs, using FuGENE HD in a 1:3 ratio of DNA and FuGENE, following the manufacturer’s protocol. Cell membranes were prepared 48 hours after transfection as described previously for α1B-ARs (41) and prepared the same way for the β2-ARs but in β2 assay buffer [50 mM tris-HCl and 3 mM MgCl2 (pH 7.4)].

IP1 HTRF assays

The IP1 HTRF assay was performed as previously described for wild-type or mutant α1B-ARs (41). Assays measuring IP1 accumulation for wild-type and mutant β2-ARs were tested in CHO cells transiently transfected with β2-ARs and performed in the same way except for stimulating with increasing concentrations of isoproterenol (100 pM to 100 μM) instead of NE.

LANCE Ultra cAMP assays

CHO cells (ATCC) were transiently transfected with wild-type or mutant β2-AR, and COS-1 cells (ATCC) were transiently transfected with wild-type or mutant α1B-AR following the manufacturer’s protocol (FuGENE HD, Roche). Assays measuring cAMP accumulation were performed 48 hours after transfection following the manufacturer’s instructions (LANCE Ultra cAMP kit, PerkinElmer). In brief, increasing concentrations of isoproterenol (100 pM to 100 μM) (β2-AR) or increasing concentrations of NE (100 pM to 100 μM) in the presence of a saturating concentration (10 μM) of propranolol to remove β-AR responses (α1B-AR) were added to 1000 transfected cells in stimulation buffer in a white 384-well plate (OptiPlate, PerkinElmer Life Sciences). The plates were incubated for 30 min at room temperature. Cells were then lysed by the addition of the europium (Eu) chelate–labeled cAMP tracer and the cAMP-specific monoclonal antibodies labeled with the ULight dye, diluted in cAMP detection buffer (LANCE Ultra cAMP kit, PerkinElmer), followed by incubation for 1 hour at room temperature. The emission signals were measured at 615 and 665 nm after excitation at 340 nm using a Tecan microplate reader (Tecan).

Radioligand binding assays

The affinity of NE and the allosteric antagonist ρ-TIA at the α1B-ARs was determined from [3H]prazosin (0.5 nM) displacement. Reactions containing membranes from α1B-AR–transfected COS-1 cells (5 μg of protein), radioligand, and increasing concentrations of NE (10 nM to 10 mM) or ρ-TIA (10 pM to 10 μM), respectively, in HEM buffer [20 mM Hepes, 1.5 mM EGTA, 12.5 mM MgCl2 (pH 7.4)] were established in clear round-bottom 96-well plates, as previously described (41). Each assay was performed in triplicate in a total reaction volume of 150 μl. After incubation for 60 min at room temperature, the membranes were harvested onto Whatman GF/B filtermats (PerkinElmer) pretreated with 0.6% polyethylenimine using a Tomtec harvester. Betaplate scintillant (PerkinElmer) was then applied, and the filter-bound radioactivity was counted on a Wallac MicroBeta counter (PerkinElmer). Scintillation proximity assay (SPA) beads (PerkinElmer) were used to determine the affinity of prazosin at α1B-ARs from displacement of the radiolabeled α1B-AR antagonist [3H]prazosin (0.5 nM), and saturation binding experiments were performed to determine receptor density (Bmax) for the different mutants at α1B-ARs in the presence of 100 μM phentolamine to determine nonspecific binding. Reactions containing membranes from α1B-AR–transfected COS-1 cells (5 μg of protein), radioligand, increasing concentrations of prazosin (10 pM to 10 μM), and SPA beads (100 μg per well) in HEM buffer were established in 96-well white polystyrene plates with clear flat bottoms. Specific binding was calculated as the difference of total and nonspecific binding. The assays were performed in triplicate in a total reaction volume of 80 μl. The plates were sealed with TopSeal-A film and incubated with shaking for 1 hour at room temperature. Radioligand binding was detected using a Wallac MicroBeta counter (PerkinElmer).

The β2-AR radioligand binding assays were performed using SPA beads (PerkinElmer). The affinity of isoproterenol and propranolol at β2-ARs was determined from displacement of the radiolabeled β-AR partial agonist [3H]CGP-12177 (1 nM), whereas saturation binding experiments were performed to determine receptor density (Bmax) for the different mutants at β2-ARs in the presence of 10 μM propranolol to determine nonspecific binding. Reactions containing membranes from β2-AR–transfected COS-1 cells (5 μg of protein), radioligand, increasing concentrations of isoproterenol (0.1 nM to 1 mM) or propranolol (1 pM to 1 μM), and SPA beads (100 μg per well) in assay buffer [50 mM tris-HCl and 3 mM MgCl2 (pH 7.4)] were established in 96-well white polystyrene plates with clear flat bottoms. Specific binding was calculated as the difference of total and nonspecific binding. The assays were performed in triplicate in a total reaction volume of 80 μl. The plates were sealed with TopSeal-A film and incubated with shaking for 1 hour at room temperature. Radioligand binding was detected using a Wallac MicroBeta counter (PerkinElmer).

β-Arrestin signaling and trafficking

We investigated NE’s and isoproterenol’s effects on β-arrestin trafficking at wild-type and mutant α1B-ARs and β2-ARs, respectively, using a BRET assay. The yellow fluorescent protein (YFP)–β-arrestin-2 construct was provided by M. Caron (Duke University, NC, USA), and the Rluc8/pcDNA3.1 construct for cloning of the α1B- and β2-AR cDNA was provided by A. Christopoulos (Monash Institute of Pharmaceutical Sciences, Melbourne, Australia). HEK293 cells were transfected with Rluc8-tagged wild-type or mutant receptors and YFP–β-arrestin-2. Agonist-induced BRET responses were performed 48 hours after transfection. On the day of the assay, cells were allowed to equilibrate in Hank’s balanced salt solution containing 0.05% (w/v) bovine serum albumin at 37°C before incubating with coelenterazine-h (5 μM) for 10 min, before reading BRET signals at 475 ± 30 nM and 535 ± 30 nM in response to increasing concentrations of NE (α1B-AR) or isoproterenol (β2-AR) (100 pM to 100 μM). The BRET signal was determined by calculating the ratio of the light emitted at 515 to 555 nm over the light emitted at 465 to 505 nm. Net BRET signals were determined by subtracting the BRET signal obtained with cells only expressing Rluc8-tagged α1B-AR or β2-AR, respectively, from BRET signals obtained with cells co-expressing Rluc8-tagged α1B-AR or β2-AR, respectively, and YFP-tagged β-arrestin-2.

Statistics and data analysis

Sigmoidal curves for the calculation of the EC50 and half-maximal inhibitory concentration (IC50) values were fitted to individual data points by nonlinear regression using the software package Prism (GraphPad Software). The IC50 values were transformed into Ki values according to the equation of Cheng and Prusoff (50). Emax was calculated as the difference between the maximal and minimal response to agonist and presented as percent of maximal wild-type response on the day of the assay for normalization. The agonist signaling efficacy (NE efficacy and isoproterenol efficacy, respectively) was calculated as the agonist pEC50 value minus the agonist pKi value and then adjusted for changes in expression levels compared with wild type. An operational model was fitted to the NE concentration response curves for the IP1 accumulation assays and the isoproterenol concentration response curves for the cAMP accumulation assays in Prism (GraphPad Software) using the following equation (45):Y= Basal+EmBasal1+[(KA+[A])/(τ×[A])]nwhere Em is the maximal system response, A is the agonist concentration, τ is the index of coupling efficiency (efficacy) of the agonist and is defined as RT/KE [where RT is the total concentration of receptors (i.e., Bmax) and KE is the concentration of agonist-receptor complex that yields half the Em], and n is the slope of the transducer function that links occupancy to response, with n constrained to be shared across all data sets. KA values were constrained to their respective Ki values estimated from competition binding studies. To account for effects of the expression level of each of the different mutant receptors on the observed efficacy of each agonist, the Bmax values determined from saturation bindings were used to normalize the τ values derived from the operational model. Data are presented as means ± SEM of results obtained from three to six separate experiments as indicated. A two-tailed single-sample t test was performed on the Bmax and Emax values. For multiple comparisons, one-way ANOVA was used with post hoc t tests performed by Dunnett’s method using Prism (GraphPad Software). Values of P < 0.05 were considered statistically significant. The ANOVA on EC50, Ki, and efficacy data was performed on the log values.

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/12/572/eaas9485/DC1

Fig. S1. Characterization of Gq coupling and IP1 accumulation for the β2-ARs.

REFERENCES AND NOTES

Acknowledgments: We thank N. Abraham (University of Queensland, Australia) for invaluable assistance with the modeling, R. Graham (Victor Chang Cardiac Research Institute, Australia) for the α1B-AR construct, B. Kobilka (Stanford University, USA) for the β2-AR construct, M. Caron (Duke University, USA) for the YFP–β-arrestin-2 construct, A. Christopoulos (Monash University, Australia) for the Rluc8/pcDNA3.1 construct, and L. May (Monash University, Australia) for invaluable assistance with the operational model. Funding: This work was supported by an NHMRC project grant (011246), an NHMRC program grant (351446), and NHMRC Principal Research Fellowship (to R.J.L.). Author contributions: L.R. conceived the research. L.R. and Å.A. developed methods and performed experiments and data analysis in this study. L.R., Å.A., W.G.T., and R.J.L. contributed to the interpretation of data and co-wrote the manuscript. R.J.L. provided funding and research facilities. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.
View Abstract

Stay Connected to Science Signaling

Navigate This Article