Abstract
Fluorescence studies with purified human β2-adrenoceptor (β2AR) revealed that the endogenous catecholamines, (-)-epinephrine (EPI), (-)-norepinephrine (NE), and dopamine (DOP), stabilize distinct active receptor conformations. However, the functional relevance of these ligand-specific conformations is as yet poorly understood. We addressed this question by studying fusion proteins of the β1-adrenoceptor (β1AR) and β2AR with the short and long splice variants of Gsα (GsαS and GsαL), respectively. Fusion proteins ensure efficient receptor/G-protein coupling and defined stoichiometry of the coupling partners. EPI, NE, DOP, and the prototypical synthetic βAR agonist, (-)-isoproterenol (ISO), showed marked differences in their efficacies at stabilizing the high-affinity ternary complex at β1AR-Gsα and β2AR-Gsα fusion proteins. Ternary complex formation was more sensitive to disruption by GTP with the β2AR than with the β1AR. Generally, in steady-state GTPase assays, ISO, EPI, and NE were full agonists, and DOP was a partial agonist. Exceptionally, at β1AR-GsαL, NE was only a partial agonist. Generally, in adenylyl cyclase assays, ISO, EPI, and NE were full agonists, and DOP was a partial agonist. At β2AR-GsαL, NE was only a partial agonist. There was no correlation between efficacy at stabilizing the ternary complex and activating GTPase, and there were also dissociations between Ki values for high-affinity agonist binding and EC50 values for GTPase activation. In contrast to synthetic partial agonists, DOP did not exhibit increased efficacy at βAR-GsαL versus βAR-GsαS fusion proteins. In conclusion, our data with βAR-Gsα fusion proteins show that endogenous catecholamines and ISO stabilize distinct conformations in the β1AR and β2AR.
The β1AR and β2AR are GPCRs, play crucial roles in the regulation of cardiovascular functions, and are activated by the catecholamines EPI, NE, DOP, and ISO (Rohrer and Kobilka, 1998; Rockman et al., 2002). ISO is a prototypical synthetic agonist at βARs, serving as a reference compound for the analysis of agonist potencies and efficacies (Rousseau et al., 1996; Hoffmann et al., 2004; Kobilka, 2007). βARs couple to the G-protein Gs (Gilman, 1987; Birnbaumer et al., 1990). Gs consists of a Gsα-subunit and a Gβγ-complex. In the resting state, the Gsα-subunit is bound to GDP. Binding of an agonist to a βAR stabilizes a GPCR conformation that allows it to promote the dissociation of GDP from Gsα, the rate-limiting step of the G-protein cycle (Gilman, 1987; Kobilka, 2007). Gsα exists as a short splice variant (GsαS) and a long splice variant (GsαL), GsαS possessing a higher GDP affinity than GsαL (Seifert et al., 1998b). Subsequently to GDP dissociation from Gsα, a ternary complex consisting of the agonist, βAR, and nucleotide-free GsαGβγ forms (De Lean et al., 1980; Kent et al., 1980; Seifert et al., 1998b). This complex possesses high agonist affinity. Upon binding of GTP to Gsα, the ternary complex is disrupted, resulting in a decrease of the βAR affinity for agonist and dissociation of Gs into Gsα and the Gβγ-complex. The ternary complex model of GPCR activation assumes that there is a correlation between the efficacy of an agonist at promoting high-affinity agonist binding and its efficacy at promoting GDP/GTP exchange and downstream effector activation (De Lean et al., 1980; Kent et al., 1980). GTP-bound Gsα activates the effector AC that catalyzes the conversion of ATP into the second messenger cAMP (Sunahara et al., 1996). cAMP, through interaction with specific protein kinases, ion channels, and nucleotide exchange factors, changes cell functions (Rehmann et al., 2007). Termination of G-protein activation is achieved by the high-affinity GTPase activity of Gsα, cleaving GTP to GDP and Pi (Gilman, 1987). Subsequently, GDP-bound Gsα and Gβγ reassociate, closing the G-protein cycle.
Although the ternary complex model is capable of describing basic aspects of GPCR/G-protein/effector interactions, the model is not sufficient at fully explaining GPCR-mediated signal transduction (De Lean et al., 1980; Kent et al., 1980; Kobilka, 2007). For example, by studying a panel of synthetic agonists with different efficacies at β2AR in the presence of inosine 5′-triphosphate, we observed dissociations in the efficacies of agonists at inosine 5′-triphosphate hydrolysis and AC activation (Seifert et al., 1999a). Furthermore, by comparing β2AR with a constitutively active β2AR mutant (β2ARCAM), we noticed dissociations in the efficacies of agonists at stabilizing the ternary complex and promoting GTP hydrolysis (Seifert et al., 2001). Moreover, the pharmacological profile of the β2AR depends on the specific G-protein coupling partner (Wenzel-Seifert and Seifert, 2000). These data suggest that ligands stabilize unique β2AR conformations that differ from each other in their efficacy at promoting ternary complex formation on one hand and the overall G-protein cycle on the other hand. Studies with other GPCRs, including the β1AR, further corroborate the concept of ligand-specific GPCR conformations (Granneman, 2001; Kenakin, 2007; Galandrin et al., 2008).
So far, most studies concerning ligand-specific βAR conformations were conducted with synthetic ligands, whereas relatively little attention has been paid to the endogenous catecholamines EPI, NE, and DOP. Intriguingly, fluorescence studies with purified β2AR revealed that endogenous catecholamines stabilize two kinetically distinct active conformational states (Swaminath et al., 2004). Specifically, DOP induces only the rapid conformational change, whereas ISO, EPI, and NE induce both the rapid and slow conformational change. ISO and EPI are more efficient than NE at promoting fluorescence changes, whereas ISO, EPI, and NE are similarly efficient at activating cAMP accumulation and β2AR internalization. Computational analysis of the β2AR confirms the existence of NE- and DOP-specific conformations (Bhattacharya et al., 2008). However, the functional consequences of the distinct β2AR conformations stabilized by endogenous catecholamines and ISO are still poorly understood. Therefore, the aim of our present study was to comprehensively characterize the interactions of the endogenous catecholamines EPI, NE, and DOP in comparison with ISO at β1AR and β2AR fused to either GsαL or GsαS. Specific advantages and disadvantages of the fusion protein approach are outlined under Materials and Methods.
Materials and Methods
Materials. The construction of baculoviruses encoding β1AR-GsαL, β1AR-GsαS, β2AR-GsαL, and β2AR-GsαS was described earlier (Seifert et al., 1998a,b; Wenzel-Seifert et al., 2002). In our present study, we analyzed the Gly389 isoform of the β1AR. A previous study revealed no differences between β1AR-Gly389-Gsα and the corresponding β1AR-Arg389-Gsα fusion proteins (Wenzel-Seifert and Seifert, 2003). [32P]Pi (8500–9100 Ci/mmol), [α-32P]ATP (3000 Ci/mmol), and [3H]DHA (85–90 Ci/mmol) were from PerkinElmer Life and Analytical Sciences, Waltham, MA). [γ-32P]GTP was synthesized enzymatically from GDP and [32P]Pi as described previously (Walseth and Johnson, 1979). Unlabeled nucleotides were from Roche Diagnostics (Mannheim, Germany). ISO, EPI, NE, and DOP were from Sigma-Aldrich (St. Louis, MO). Stock solutions of catecholamines (10 mM each) were prepared fresh daily in 1 mM HCl. Catecholamine dilutions were also prepared in 1 mM HCl. All catecholamine solutions were kept at 4°C and under light protection until experiments were performed. Binding experiments, lasting 90 min, were also conducted under light protection. Glass fiber filters (GF/C) were from Whatman Schleicher and Schuell (Keene, NH).
Cell Culture and Membrane Preparation. βAR-Gsα fusion proteins were expressed in Sf9 inset cells. Sf9 cells were cultured in 250-ml disposable Erlenmeyer flasks at 28°C under rotation at 125 rpm in SF 900 II medium (Invitrogen, Carlsbad, CA) supplemented with 5% (v/v) fetal calf serum (Cambrex, East Rutherford, NJ) and 0.1 mg/ml gentamicin (Cambrex) (Seifert et al., 1998a,b). Cells were maintained at a density of 1.0 to 6.0 × 106 cells/ml. Sf9 cells were seeded at 3.0 × 106 cells/ml and infected with 1:100 to 1:1000 dilutions of high-titer baculovirus stocks encoding βAR-Gsα fusion proteins. Cells were cultured for 48 h before membrane preparation. Sf9 membranes were prepared as described previously (Seifert et al., 1998a,b), using 1 mM EDTA, 0.2 mM phenylmethylsulfonyl fluoride, 10 μg/ml benzamidine, and 10 μg/ml leupeptin as protease inhibitors. Membranes were suspended in binding buffer (12.5 mM MgCl2, 1 mM EDTA, and 75 mM Tris/HCl, pH 7.4). Membranes were stored at -80°C for periods of up to 6 months (longer periods of time were not analyzed in this study) without loss of functional activity in the various assays employed.
[3H]DHA Binding Assay. Membranes were thawed and sedimented by a 10-min centrifugation at 4°C and 15,000g to remove residual endogenous guanine nucleotides as far as possible and resuspended in binding buffer. Expression levels of fusion proteins were determined by incubating Sf9 membranes (20–25 μg of protein/tube) in the presence of 10 nM [3H]DHA. The total volume of the binding reaction was 500 μl. Incubations were performed for 90 min at 25°C and shaking at 250 rpm. Nonspecific [3H]DHA binding was determined in the presence of 10 μM (±)-alprenolol. Nonspecific [3H]DHA binding amounted to less than 10–15% of total [3H]DHA binding. For agonist competition binding experiments, membranes expressing βAR-Gsα at levels between 3.8 and 5.2 pmol/mg membrane protein were used. Tubes contained Sf9 membranes (20–25 μg of protein/tube), 1 nM [3H]DHA, and agonists at increasing concentrations. Reaction mixtures additionally contained solvent (control) or GTP (1 mM). Binding experiments were conducted under light protection. Bound [3H]DHA was separated from free [3H]DHA by filtration through GF/C filters using a 48-well harvester (Brandel, Gaithersburg, MD), followed by three washes with 2 ml of binding buffer (4°C). Filter-bound radioactivity was determined by liquid scintillation counting using Rotiszint eco plus cocktail (Roth, Karlsruhe, Germany).
Steady-State GTPase Activity Assay. The GTPase assay was performed as described previously, with minor modifications (Seifert et al., 1998a,b). In brief, Sf9 membranes were thawed, sedimented by centrifugation at 15,000g for 10 min at 4°C, and resuspended in 10 mM Tris/HCl, pH 7.4. For GTPase assays, membranes expressing βAR-Gsα at levels between 1.7 and 5.2 pmol/mg membrane protein were used. Assay tubes contained Sf9 membranes (10 μg of protein/tube), 1.0 mM MgCl2, 0.1 mM EDTA, 0.1 mM ATP, 100 nM GTP, 0.1 mM adenylyl imidodiphosphate, 5 mM creatine phosphate, 40 μg of creatine kinase, and 0.2% (w/v) bovine serum albumin in 50 mM Tris/HCl, pH 7.4, and catecholamines at various concentrations. Reaction mixtures (80 μl) were incubated for 2 min at 25°C before the addition of 20 μl of [γ-32P]GTP (0.1 μCi/tube). All stock and work dilutions of [γ-32P]GTP were prepared in 20 mM Tris/HCl, pH 7.4. Reactions were conducted for 20 min at 25°C. Reactions were terminated by the addition of 900 μl of slurry consisting of 5% (w/v) activated charcoal and 50 mM NaH2PO4, pH 2.0. Charcoal absorbs nucleotides but not Pi. Charcoal-quenched reaction mixtures were centrifuged for 7 min at room temperature at 15,000g. Six hundred microliters of supernatant fluid of reaction mixtures was removed, and 32Pi was determined by Čerenkov radiation in 3 ml water. Enzyme activities were corrected for spontaneous degradation of [γ-32P]GTP. Spontaneous [γ-32P]GTP degradation was determined in tubes containing all of the above described components plus a high concentration of unlabeled GTP (1 mM) that, by competition with [γ-32P]GTP, prevents [γ-32P]GTP hydrolysis by enzymatic activities present in Sf9 membranes. Spontaneous [γ-32P]GTP degradation amounted to <1% of the total amount of radioactivity. The experimental conditions chosen ensured that not more than 10% of the total amount of [γ-32P]GTP added was converted to 32Pi.
AC Activity Assay. The AC assay was performed as described previously, with minor modifications (Seifert et al., 1998a,b). Membranes were thawed and sedimented by a 10-min centrifugation at 4°C and 15,000g to remove residual endogenous guanine nucleotides as far as possible and resuspended in binding buffer. For AC assays, membranes expressing βAR-Gsα at levels between 1.1 and 2.0 pmol/mg membrane protein were used. Tubes contained Sf9 membranes expressing fusion proteins (30 μg of protein/tube), 5 mM MgCl2, 0.4 mM EDTA, 30 mM Tris/HCl, pH 7.4, GTP (10 μM), and catecholamines at various concentrations. Assay tubes containing membranes and additions in a total volume of 30 μl were incubated for 3 min at 37°C before starting reactions with 20 μl of reaction mixture containing (final) [α-32P]ATP (0.5–1.0 μCi/tube) plus 40 μM unlabeled ATP, 0.1 mM cAMP, and a regenerating system consisting of 2.7 mM mono(cyclohexyl)ammonium phosphoenolpyruvate, 0.125 IU of pyruvate kinase, and 1 IU of myokinase. Reactions were conducted for 20 min at 37°C and terminated by the addition of 20 μl of 2.2 N HCl. Denatured protein was sedimented by a 3-min centrifugation at 25°C and 15,000g. Sixty-five microliters of the supernatant fluid was applied onto disposable columns filled with 1.3 g of neutral alumina (MP alumina N Super I; MP Biomedicals, Irvine, CA). [32P]cAMP was separated from [α-32P]ATP by elution of [32P]cAMP with 4 ml of 0.1 M ammonium acetate, pH 7.0. Recovery of [32P]cAMP was ∼80%. Blank values were routinely ∼0.01% of the total amount of [α-32P]ATP added. [32P]cAMP was determined by Čerenkov radiation in 10 ml of water.
Advantages and Disadvantages of Using βAR-Gsα Fusion Proteins as Model Systems in the [3H]DHA Binding, GTPase, and AC Assay. GPCR-Gα fusion proteins ensure close proximity between, and defined 1:1 stoichiometry of, the coupling partners (Seifert et al., 1999b; Milligan et al., 2007). The fusion facilitates efficient GPCR/G-protein coupling under defined experimental conditions. This effect of the fusion is most prominent for Gs-coupled GPCRs. Specifically, in conventional βAR-Gsα coexpression systems, ternary complex formation is less efficient than in fusion proteins, rendering a detailed comparison of various ligands very difficult (Seifert et al., 1998a). Moreover, in fusion proteins, the efficiency of ternary complex formation is independent of the specific expression level of the construct, thereby greatly facilitating comparison of various receptors coupled to different G-proteins (Seifert et al., 1998b; Wenzel-Seifert et al., 1999, 2001, 2002). It has been observed repeatedly that ternary complex formation in βAR-Gα fusion proteins can be (partially) insensitive to disruption by guanine nucleotides (Seifert et al., 1998a, 2001; Wenzel-Seifert and Seifert, 2000). This is not a peculiar property of fusion proteins but also a property of conventional coexpression systems (Seifert et al., 1998a).
With respect to steady-state GTP hydrolysis, it is exceedingly difficult to perform a detailed pharmacological analysis in a conventional βAR-Gsα coexpression system because the signal/noise ratio is very low (Seifert et al., 1998b). However, with fusion proteins, a detailed pharmacological analysis is feasible because of the high signal/noise ratio (Seifert et al., 1998a,b, 2001). Moreover, ligand potencies and efficacies are independent of the expression level of constructs (Seifert et al., 1998a,b, 2001).
Concerning the analysis of AC activity, such studies can be performed with conventional βAR-Gsα coexpression systems because signal amplification at the effector level is sufficiently large (Seifert et al., 1998a). However, one has to keep in mind that the number of available AC molecules is the limiting factor in the system so that high βAR expression levels should be avoided (Seifert et al., 1998a). βAR-Gsα fusion proteins are also capable of mediating efficient AC activation (Seifert et al., 1998a,b, 1999a). To avoid depletion of AC molecules, we only used membranes with relatively low βAR-Gsα expression levels, i.e., in the range between 1.1 and 2.0 pmol/mg. These conditions ensure good signals while avoiding AC depletion (Seifert et al., 1998a,b, 1999a; Wenzel-Seifert et al., 2002).
Evidently, βAR-Gsα fusion proteins are artificial and do not occur physiologically. However, in previous studies, we had carefully compared the properties of fusion proteins with those of nonfused βAR and Gsα proteins and did not reveal large differences in terms of ligand affinities/potencies and efficacies, validating the approach (Seifert et al., 1999b). Previous studies also revealed that the functional integrity of β1AR-Gsα and β2AR-Gsα fusion proteins as assessed by [35S]guanosine 5′-[γ-thio]triphosphate saturation binding is similar (Seifert et al., 1998a,b; Wenzel-Seifert and Seifert, 2000, 2003; Wenzel-Seifert et al., 2002). Thus, it is unlikely that there are large differences in functionally relevant membrane insertion between the various fusion proteins. Based on these considerations, the conclusions obtained with GPCR-Gα fusion proteins can be cautiously transferred to nonfused systems.
Miscellaneous. Protein concentrations were determined using the Bio-Rad DC protein assay kit (Bio-Rad, Hercules, CA). Data shown in Figs. 1, 2, 3, 4, 5, 6 were analyzed by nonlinear regression (Prism 5.0 software; GraphPad Software Inc., San Diego, CA). Statistical comparisons of the efficacies of ISO versus endogenous catecholamines in the GTPase assay and AC assay were performed using ANOVA followed by Dunnett's multiple comparison post test.
Results
Competition by ISO, EPI, NE, and DOP of [3H]DHA Binding in Sf9 Membranes Expressing βAR-Gsα Fusion Proteins: Effect of GTP. At β1AR-GsαL, all catecholamines studied (ISO, EPI, NE, and DOP) inhibited [3H]DHA binding according to biphasic competition isotherms. DOP was the least effective catecholamine at stabilizing the ternary complex as reflected by the low Rh (percentage) value (Fig. 1; Table 1). GTP shifted the agonist-competition curves to the right, with the shift being most pronounced for NE. However, with all catecholamines studied, GTP did not achieve a complete conversion into a monophasic competition isotherm at β1AR-GsαL. At β1AR-GsαS, like at β1AR-GsαL, DOP was the least effective catecholamine at stabilizing the ternary complex (Fig. 2; Table 1). At β1AR-GsαS, ISO was a less effective stabilizer of the ternary complex than at β1AR-GsαL. Like at β1AR-GsαL, GTP did not convert the biphasic competition isotherms at β1AR-GsαS into monophasic isotherms. At both β1AR-Gsα fusion proteins, the order of affinity (Kih and Kil) of catecholamines was ISO ∼ NE > EPI >> DOP. At the β1AR expressed in CHO cells, the order of affinity was ISO > NE ∼ EPI (DOP was not studied) (Hoffmann et al., 2004), but in this system, only Kil values in the presence of GTP were determined, annihilating the impact of Gsα on agonist affinity.
At β2AR-GsαL, EPI was the most effective ternary complex stabilizer, and DOP was the least effective agonist in this respect (Fig. 3; Table 1). With the exception of DOP, GTP converted the biphasic agonist competition isotherms into monophasic isotherms. Similar to β2AR-GsαL, EPI was the most effective ternary complex stabilizer at β2AR-GsαS, and DOP was the least effective agonist in this regard (Fig. 4). GTP converted the biphasic competition isotherms for EPI, NE, and DOP at β2AR-GsαS into monophasic isotherms, whereas the competition isotherm for ISO remained biphasic. At both β2AR-Gsα fusion proteins, the order of affinity (Kih and Kil) of catecholamines was ISO ∼ EPI > NE > DOP. At the β2AR expressed in CHO cells, the order of affinity was ISO ∼ EPI >> NE (DOP was not studied) (Hoffmann et al., 2004), but in this system, only Kil values in the presence of GTP were determined, annihilating the impact of Gsα on agonist affinity.
Agonist Potencies and Efficacies at βAR-Gsα Fusion Proteins in the GTPase Assay. At β1AR-GsαL and β1AR-GsαS, the order of potency of catecholamines in the GTPase assay was ISO ∼ NE > EPI >> DOP (Figs. 5 and 6; Table 2). At β1AR-GsαL, ISO and EPI were full agonists, whereas NE and DOP were only partial agonists. At β1AR-GsαS, ISO, EPI, and NE were all full agonists, whereas DOP was only a partial agonist.
At β2AR-GsαL and β2AR-GsαS, the order of potency of catecholamines in the GTPase assay was ISO ∼ EPI > NE >> DOP. At both fusion proteins, ISO, EPI, and NE were full agonists, whereas DOP was a partial agonist (Figs. 5 and 6; Table 2). There was a trend toward strong partial agonism of NE versus ISO at both β2AR-Gsα fusion proteins, but the difference did not reach significance.
Agonist Potencies and Efficacies at βAR-Gsα Fusion Proteins in the AC Assay. At β1AR-GsαL and β1AR-GsαS, catecholamines activated AC in the presence of GTP in the order of potency ISO ∼ NE ∼ EPI >> DOP (Figs. 5 and 6; Table 3). At both fusion proteins, ISO, EPI, and NE were full agonists, whereas DOP was a partial agonist. There was a trend toward strong partial agonism of EPI and NE versus ISO at both β1AR-Gsα fusion proteins, but the difference did not reach significance.
At β2AR-GsαL and β2AR-GsαS, catecholamines activated AC in the presence of GTP in the order of potency ISO ∼ EPI > NE >> DOP (Figs. 5 and 6; Table 3). At β2AR-GsαL, ISO and EPI were full agonists, and NE and DOP were partial agonists. At β2AR-GsαS, ISO and EPI exhibited full agonism, NE exhibited a nonsignificant trend toward partial agonism, and DOP was a partial agonist.
Correlations between Agonist Affinities in the Agonist Competition Binding Assay and Potencies for GTPase Activation and Correlation between Potencies in the GTPase and AC Assays. The ternary complex model predicts that here should be a correlation between the Kih values obtained in the agonist competition binding assay and the EC50 values for GTPase and AC activation (Kent et al., 1980; Seifert et al., 2001). Our experimental data do not fulfill the prediction. Specifically, just for one ligand, i.e., DOP at β1AR-GsαL, Kih and EC50 for GTPase activation differed by no more than 2-fold (Table 4). For the other ligands, EC50 values were 4 to 460-fold higher than Kih values, DOP at β2AR-GsαS showing the most extreme difference between the two parameters. These differences indicate that efficient stimulation of GDP/GTP exchange by the high-affinity agonist state of a GPCR is rather an exception than the rule. In accordance with the data on the Gs-coupled β1AR and β2AR, there is evidence for the Gi-coupled formyl peptide receptor that the high-affinity agonist state does not mediate GDP/GTP exchange (Gierschik et al., 1989; Wenzel-Seifert et al., 1999). Rather, in the case of the formyl peptide receptor, the low-affinity agonist state mediates GDP/GTP exchange. However, when we correlated Kil values with EC50 values for GTPase activation at βAR-Gsα fusion proteins, we found a match between the two values only for DOP at β2AR-GsαS (Table 4). For the other ligands, Kil values were approximately 3 to 20-fold higher than EC50 values for GTPase activity. These data indicate that there is considerable variation in the agonist affinity of the β1AR and β2AR state that mediates GDP/GTP exchange, depending on the specific ligand studied. In most cases, an intermediate agonist affinity state that is not easily distinguished in radioligand competition binding studies appears to mediate GDP/GTP exchange. Furthermore, it would also have been predicted that the EC50 values of agonists for activation of GTPase and AC had been identical. For EPI at β2AR-GsαS, the two EC50 values matched, but at β2AR-GsαL, EPI was more than 3-fold more potent in the GTPase assay than in the AC assay (Table 4). Less pronounced differences between these parameters were also observed for other ligands.
Discussion
The ternary complex model is of fundamental importance for understanding the mechanisms of GPCR/G-protein/effector interactions (De Lean et al., 1980; Kent et al., 1980; Kobilka, 2007). A central paradigm of this model is a correlation between the efficacy of agonists at stabilizing the ternary complex and the efficacy at activating GDP/GTP exchange and effector activation. At the turkey βAR and at the β2AR-GsαL fusion protein, using a panel of synthetic ligands, such a correlation was, indeed, observed (Kent et al., 1980; Seifert et al., 2001). In contrast, with β2ARCAM, no such correlation was apparent (Seifert et al., 2001), indicating that the ternary complex model is not applicable to all GPCRs. However, so far, little attention has been paid to the analysis of the endogenous catecholamines EPI, NE, and DOP. It is intriguing that fluorescence studies with purified β2AR and computational studies indicate that these catecholamines stabilize ligand-specific GPCR conformations (Swaminath et al., 2004; Bhattacharya et al., 2008). Moreover, there is evidence that the conformational state of a GPCR depends on the structure and efficacy of a ligand for a specific G-protein and/or signaling pathway (Wenzel-Seifert and Seifert, 2000; Bhattacharya et al., 2008; Galandrin et al., 2008). Therefore, we comprehensively examined the interactions of ISO, EPI, NE, and DOP with β1AR-Gsα and β2AR-Gsα fusion proteins. We included the two splice variants of Gsα, GsαS and GsαL in our analysis to account for the G-protein specificity aspect of GPCR conformations.
We made several observations that are not compatible with the ternary complex model. First, at β1AR-GsαS, β2AR-GsαL, and β2AR-GsαS, ISO was a full agonist in the GTPase and AC assays, but in terms of ternary complex formation, ISO was surpassed by EPI and/or NE. These data can be explained by the existence of “frozen” ternary complexes (Seifert et al., 2001; Kenakin, 2007), i.e., complexes that are inefficient at stimulating GTP binding and subsequent effector activation. Second, as a general rule, ternary complex formation with β1AR-Gsα fusion proteins was less sensitive to disruption by GTP than ternary complex formation with the corresponding β2AR-Gsα fusion proteins. These data indicate that the β1AR couples to Gsα more tightly than the β2AR, i.e., even the activated Gsα protein is still in physical contact with the GPCR, conferring to it high agonist affinity. Such observations are also not without precedence in the literature (Seifert et al., 1998a, 2001). Third, at β2AR-GsαL, NE and DOP were more effective at stimulating steady-state GTPase activity than AC activity. The ternary complex model predicts a correlation between those parameters (De Lean et al., 1980; Kent et al., 1980; Seifert et al., 1999a). Thus, certain ligand-specific β2AR conformations couple less efficiently to AC than to GDP/GTP exchange. These observations imply that a quaternary complex consisting of agonist, GPCR, G-protein, and effector exists. Differential regulation of AC by formyl peptide-Giα and β2AR-Giα fusion proteins supports this interpretation (Seifert et al., 2002). Fourth, the ternary complex model predicts a correlation between Kih values and EC50 values for G-protein activation (Kent et al., 1980; Seifert et al., 2002), but we did not observe such a correlation. Moreover, we did not observe the predicted correlation between EC50 values for GTPase and AC activation. All these data support the concept of ligand-specific βAR conformations.
The ternary complex model was extended to account for the observation of agonist-independent, i.e., constitutive activity (Samama et al., 1993; Chidiac et al., 1994; Kenakin, 1996). Inverse agonists reduce the constitutive activity of GPCRs. Using the wild-type β2AR and β2ARCAM as model, increased potency and efficacy of synthetic partial agonists and increased efficacy of inverse agonists emerged as the hallmark of high constitutive activity (Samama et al., 1993; Seifert et al., 2001). The comparison of βAR-GsαS and βAR-GsαL fusion proteins provides another model system for probing the extended ternary complex model. Specifically, when fused to GsαL, but not when fused to GsαS, both the β1AR and β2AR exhibit the properties of high constitutive activity as assessed by increased potency and efficacy of a series of synthetic partial agonists including salbutamol, dobutamine, (-)-ephedrine, dichloroisoproterenol, and alprenolol, as well as increased efficacy of inverse agonists (Seifert et al., 1998b; Wenzel-Seifert and Seifert, 2000, 2003; Wenzel-Seifert et al., 2002). These data are explained by the fact that GsαL possesses a lower GDP affinity than GsαS (Seifert et al., 1998b), facilitating GDP dissociation by the agonist-free GPCR or GPCR bound to a partial agonist.
The fact that at all four fusion proteins studied, DOP was a partial agonist both in the GTPase and AC assays, offered a unique opportunity to probe the validity of the extended ternary complex model using a natural partial agonist. In agreement with the model, both in the GTPase and in the AC assay, DOP exhibited increased potency at a βAR-GsαL fusion protein compared with the corresponding βAR-GsαS fusion protein. However, with respect to efficacy, no significant differences emerged for DOP at βAR-GsαS versus βAR-GsαL fusion proteins, although for synthetic ligands with comparable efficacy, i.e., salbutamol and dobutamine, such differences were apparent (Seifert et al., 1998b; Wenzel-Seifert and Seifert, 2000; Wenzel-Seifert et al., 2002). Intriguingly, deviations from the predictions of the extended ternary complex model were also observed for certain partial agonists at the human histamine H2-receptor fused to GsαS and GsαL (Wenzel-Seifert et al., 2001) and the highly constitutively active canine H2-receptor fused to GsαS (Preuss et al., 2007).
It is a convention to use the synthetic agonist ISO as reference compound for assessing ligand efficacy at βARs (Rousseau et al., 1996; Hoffmann et al., 2004; Kobilka, 2007). Our data on four β1AR-Gsα and β2AR-Gsα fusion proteins in two different assays (GTPase and AC) corroborate the validity of this convention, i.e., in none of the systems studied did EPI or NE surpass ISO in terms of efficacy. Rather, there was a trend, more pronounced for NE than for EPI, toward reduced efficacy of endogenous catecholamines. DOP was a partial agonist under all conditions studied. Our data on ISO, EPI, NE, and DOP data fit to fluorescence studies with purified β2AR and β2AR internalization studies in HEK293 cells (Swaminath et al., 2004). In contrast, studying AC activation by β1AR and β2AR expressed in CHO cells, there was a trend of reduced efficacy of ISO compared with EPI and NE at most GPCR expression levels studied (Hoffmann et al., 2004). The high efficacy of NE in the latter study contrasts with fluorescence and GPCR internalization data (Swaminath et al., 2004) and our data. The molecular basis for these differences in efficacy is unknown, but the aggregate data show that the efficacy of ISO, EPI, and NE is sensitive to the specific experimental conditions. Previous studies revealed that the efficacy of ligands at βARs depends on the specific G-protein and the specific signal transduction pathway studied (Wenzel-Seifert and Seifert, 2000; Kenakin, 2007; Bhattacharya et al., 2008; Galandrin et al., 2008). Thus, any given ligand actually possesses multiple efficacies.
In the present study, we only considered coupling of βARs to Gs-proteins. However, given the fact that various synthetic βAR ligands show distinct pharmacological properties when studying different G-protein families such as Gi and Gq and signaling pathways such as the mitogen-activated protein kinase pathway (Wenzel-Seifert and Seifert, 2000; Galandrin et al., 2008), it is possible that those ligand-specific signaling differences extend to endogenous catecholamines. Previous studies reported intriguing differences between endogenous catecholamines in organ systems and the intact organism that could not be explained satisfactorily (McNay and Goldberg, 1966; Mueller, 1978). Our data suggest that differential activation of signaling pathways by various catecholamines contribute to their complex in vivo effects. An in depth-analysis of the unexpectedly complex effects of endogenous catecholamines on βARs will also help us understand why the clinical effects of catecholamines in septic shock and renal failure are so controversial and identify subpopulations of patients that may benefit from therapy with defined catecholamines (Beale et al., 2004; Myburgh, 2007). By analogy to the developments in the nuclear receptor field (Regitz-Zagrosek et al., 2007), catecholamines may exhibit organ- and/or cell-type-specific effects that would facilitate a more specific treatment of disease states while reducing toxicity. The development of organ- and/or cell-type-specific catecholamine therapy is feasible in view of the fact that different agonist binding sites can be exploited in βARs, i.e., the aryloxypropanolamine binding site in the β1AR (Granneman, 2001) and the catechol and noncatechol binding site in the β2AR (Swaminath et al., 2005). Thus, when studying βARs at the molecular, cellular, tissue, organ, or intact organism level, one should keep in mind that there is no “standard” catecholamine but that the analysis of multiple agonists from different chemical classes is required.
Collectively, by applying classic pharmacological methods, i.e., ternary complex formation and GTPase and AC activity assays, we corroborate the concept developed with sophisticated methods, i.e., fluorescence spectroscopy and molecular modeling (Swaminath et al., 2004; Bhattacharya et al., 2008), that endogenous catecholamines stabilize distinct active β2AR conformations. Moreover, previous pharmacological studies revealed that various synthetic ligands stabilize distinct β1AR conformations (Granneman, 2001; Galandrin et al., 2008). Our present study extends the concept of ligand-specific β1AR conformations to endogenous catecholamines.
Acknowledgments
We thank G. Wilberg for expert help with the cell culture, Drs. S. Dove (Department of Medicinal Chemistry II, University of Regensburg, Regensburg, Germany) and G. Swaminath (Amgen, South San Francisco, CA) for stimulating discussions, and the reviewers for helpful critique.
Footnotes
-
This study was supported by the Research Training Program (Graduiertenkolleg) GRK760 “Medicinal Chemistry: Molecular Recognition: Ligand-Receptor Interactions” of the Deutsche Forschungsgemeinschaft.
-
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
-
doi:10.1124/jpet.108.143412.
-
ABBREVIATIONS: β1AR, β1-adrenoceptor; β2AR, human β2-adrenoceptor; βAR, nonspecified human β-adrenoceptor; Gα, nonspecified G-protein α-subunit; GPCR, G-protein-coupled receptor; EPI, (-)-epinephrine; NE, (-)-norepinephrine; DOP, dopamine; ISO, (-)-isoproterenol; GsαS, short splice variant of Gsα;GsαL, long splice variant of Gsα; AC, adenylyl cyclase; β2ARCAM, constitutively active mutant of the β2AR; DHA, dihydroalprenolol; ANOVA, analysis of variance; CHO, Chinese hamster ovary.
-
↵1 Current affiliation: Institute of Pharmacology, Medical School of Hannover, Hannover, Germany.
- Received July 10, 2008.
- Accepted September 3, 2008.
- The American Society for Pharmacology and Experimental Therapeutics