Abstract
Chimeric receptors of the β2-adrenergic receptor in which the extracellular loops were replaced with the corresponding amino acids of the α1a-adrenergic receptor were generated to measure changes in α1-antagonist affinity. Although no changes in α1-antagonist affinity were measured in the β2/α1a chimeras, a decreased IC50 (10-fold) for agonists as compared with wild type β2 control was found because of the replacement of the third extracellular loop (EX3). These agonist high affinity changes were because of a greater proportion of high affinity sites (2-fold) that were convertible to low affinity sites with guanosine 5′-3-O-(thio)triphosphate. Adenylate cyclase activity evoked by the EX3 chimera showed commensurate increases in the basal signal transduction as well as the isoproterenol-stimulated potency, suggesting constitutive activity. However, unlike other constitutively active adrenergic receptor mutants in which the mutation causes G protein-independent changes, the mechanism of the EX3 chimera seems to be attributable to a greater ease with which the active ternary complex is formed because of a higher affinity/coupling of the G protein. Although the changes because of EX3 are indirect and most likely affect helical packing, they support an emerging hypothesis that G protein-coupled receptors have evolved their structure-function relationships to constrain the receptor in an inactive state.
Many hormone receptors mediate their intracellular responses by interacting with one or more of a family of G proteins. The process of receptor activation and G protein coupling is described by the widely accepted revised TCM (Samma et al., 1993). In this allosteric model, the active conformation of the native receptor, which leads to signaling, is the cornerstone of the ternary agonist-receptor-G protein complex. Without agonist present, the model predicts spontaneous receptor isomerization between inactive (R) and the active conformation (R*), with equilibrium under native conditions shifted toward R. At any given time, even though most receptors reside in R, a small population will reside in R*, permitting formation of the HR*G protein complex that causes effector activation. Receptor mutations that induce an agonist-independent shift in isomerization toward the R* conformation are termed CAMs and couple to and evoke second messenger responses greater than the WT receptor in the absence of agonist.
Receptor-mediated activation of G proteins involves the exchange of complexed GDP for GTP and the GTP-bound form of the G protein interacts with the specific effector system. The receptor-G protein complex has a higher affinity for agonists than does the free receptor, resulting in the appearance in membrane preparations of a high affinity agonist site in addition to low affinity sites characteristic of the free or uncoupled receptor. Addition of GTP or GTP analogs, which uncouple the receptor-G protein complex, can convert the high affinity site to the low affinity site. Although receptor activation models and GTP-evoked shifts in affinity (which are dependent on receptor/G protein ratios) have not been correlated (i.e., 30% high affinity sites does not mean a 30% “active” or R* population), the revised TCM model assumes that both hormone and G protein bind with greater affinity to the R* form. Given this condition, there is cooperativity in binding of the G protein and hormone to the receptor which explains the GTP shift and, therefore, can be used as a relative but not absolute measure of the active ternary complex, HR*G.
Of the G protein-coupled receptors, the β2-AR is perhaps the best characterized and most widely used to formulate activation theories. They display the characteristic G protein-coupled receptor topography of seven membrane-spanning domains. They form a family of related receptors (α1, α2, β1, β2, and β3 subtypes) that mediate the sympathetic actions of epinephrine and norepinephrine. Whereas the agonist-binding pocket for these receptors is formed within a hydrophilic pocket formed by the circular array of the TM domains, we have obtained evidence that members of the α1-adrenergic family may bind antagonists partly on the extracellular surface of the receptor (Zhao et al., 1996). To explore this hypothesis further, we made a series of chimeras in which we replaced the extracellular loops of the β2-AR with the corresponding loops of the α1a subtype, thinking that we could impart to the β2-AR a higher affinity for α1-antagonists. Although we did not observe this phenotype, we serendipitously found that replacement of the third extracellular loop imparted to the β2-AR a higher binding affinity for agonists, an increased potency, and an agonist-independent higher basal activity of adenylate cyclase. Unlike other previously characterized CAMs, where the high affinity binding of agonists was intrinsic to the receptor and independent of G protein modulation, the β2/α1achimeric phenotype of constitutive activity seems to be attributable to a higher affinity/coupling of the G protein for the receptor promoting the formation of the active ternary complex. This suggests a possible role of the third extracellular loop of the β2-AR in controlling receptor/G protein affinity; however, this is most likely caused indirectly by influencing the arrangement of TM 6 and 7 in the membrane bilayer.
Experimental Procedures
Construction of β2/α1a chimeric receptors.
The constructs used were the synthetic gene of the hamster lung β2-AR (Noda et al., 1994) or the rat α1a-AR cDNA (Perez et al., 1994). The synthetic β2-AR has unique restriction sites spaced approximately 50 base pairs apart to facilitate mutagenesis via oligonucleotide-cassette replacement. The synthetic gene displays WT pharmacology and signal transduction properties of the β2-AR cDNA (Noda et al., 1994). Generation of the β2/α1a chimeric receptors was accomplished via double restriction digestion of the synthetic WT β2-AR, removal of the cassette by agarose gel electrophoresis, ligation of the annealed oligonucleotide replacement cassette containing the appropriate sticky ends, and encoding of the rat α1a sequence in the corresponding region to the β2-AR as described previously (Noda et al., 1994). cDNAs were sequenced by the dideoxy method (Sequenase; Amersham, Arlington Heights, IL) to confirm the chimeras. Fig. 1 shows the amino acids that were replaced in the β2-AR with the rat α1a sequence (in black). Location of the start and stop points was determined by the location of the restriction sites in the synthetic gene. The designations, EX1, EX2, and EX3 (see Table 1) correspond to β2/α1a chimeric receptors in which the EX1, EX2, or EX3 regions were replaced individually with α1a-residues. The EX1+2+3 designation refers to the β2-AR with replacement of all three extracellular loops with α1a-residues.
Cell culture and transfection.
COS-1 cells (American Type Culture Collection) were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum. cDNAs encoding the WT β2-AR and various chimeras were subcloned into the mammalian expression vector pMT2′, as previously described (Perezet al., 1991). Plasmid DNA, purified by Wizardprep (Fisher Scientific, Pittsburgh, PA) and Bio-Gel A-50m column chromatography (BioRad, Hercules, CA), was used to transfect cells. Transient expression in COS-1 cells was accomplished by the DEAE-dextran method (Sambrook et al., 1989) using the same mastermix of reagents on the same day with all pertinent constructs on equalized cell numbers. Efficiency of transfection is 20% as determined by β-galactosidase staining. Cells were harvested 60 hr post-transfection.
Membrane preparation and radioligand binding.
COS-1 membranes were prepared as previously described (Perez et al., 1991). The protein concentration was measured using the method of Bradford (1976). The ligand-binding characteristics of the expressed receptors were determined in a series of radioligand binding studies using the β2-antagonist radioligand [125I]CYP as previously described at 37° (Noda et al., 1994). Competition reactions (total volume, 0.25 ml) contained 20 mm HEPES, pH 7.5, 1.4 mmEGTA, 12.5 mm MgCl2, 150 pm [125I]CYP, COS-1 membranes, and increasing amounts of unlabeled ligands known to interact with β2-ARs. Nonspecific binding was determined in the presence of 10-5 m propranolol. Reactions were stopped by the addition of cold HEPES buffer and were filtered onto Whatman GF/C glass fiber filters with a Brandel cell harvester (Brandel, Gaithersburg, MD). Filters were washed five times with HEPES buffer, and bound radioactivity was determined using a Packard Auto-γ 500 counter (Hewlett Packard, Avondale, PA). Binding data were analyzed by the curve-fitting program GraphPad Prism (GraphPAD Software, San Diego, CA). Hill coefficients were determined from the slope of the log-logit curve. For saturation binding studies, [125I]CYP concentrations ranging from 5 to 200 pm were used. Saturation curves were obtained by incubating cell membranes with increasing amounts of [125I]CYP in the same buffer system used for the competition studies. To reduce interassay variation, binding assays were performed simultaneously with all constructs.
Coupling analysis.
Coupling analysis is based upon the conversion of a two-site into a one-site model upon the addition of GTPγS (0.1 mm) in the competition binding of agonists. The high affinity site is taken to be the fraction of receptors precoupled to the G protein, whereas the low affinity site is the uncoupled state. One-site versus two-site models were tested for best fit by the F test, and significance was determined for ap value of 0.05 or lower using the program GraphPad Prism. Two site model conforms to the equation
cAMP assay.
Accumulation of cAMP in WT and chimeric transfectants was measured using a commercially available cAMP assay system (Amersham) according to the directions supplied by the manufacturer. Cell extracts were derived from cultures in 60-mm dishes that were preincubated for 30 min with 5 mm theophylline and then for 30 min with both theophylline and increasing concentrations of agonist at 37°. Cells were lysed with 0.1m HCl and diluted to obtain values within a standard curve supplied with the kit. Receptor density was determined by saturation binding on plates that were cotransfected in parallel studies to the signal transduction. All constructs were measured in the same assay to reduce interassay variations. Statistical significance (p < 0.05) in both binding and function were determined by a Student’s t test or a one-way ANOVA followed by a Student-Newman-Keuls multiple comparison test.
Materials.
Drugs were obtained from the following manufacturers: isoproterenol, epinephrine, WB4101, phentolamine, propranolol, albuterol, and alprenolol, Sigma (St. Louis, MO); ICI 118551, Research Biochemicals Inc. (Natick, MA); [125I]CYP and cAMP kit were from Amersham Corp. Bio-Gel A-50m resin was from BioRad.
Results and Discussion
The adrenergic subtypes, although structurally similar in their transmembrane domains, have significant differences in their ligand-binding affinities for a number of synthetic agonists and antagonists. Determination of the critical amino acids responsible for these differences may assist in the design of better subtype-selective drugs. Previous work exploring these determinants in α1-subtypes found that three adjacent residues in the putative second extracellular loop region near the fifth TM were fully responsible for the higher binding affinity (10–100-fold) of the α1-antagonists, WB4101 and phentolamine, seen in the α1a-subtype as compared with the α1b subtype (Zhao et al., 1996). The data indicated that α1 antagonists may bind either nearer the surface of the receptor than the binding of agonists or that the antagonists may bind either partly or mostly in the extracellular loop regions. To address these possibilities, we constructed a series of β2/α1a chimeras in which we replaced the extracellular loop regions of the β2-AR with the corresponding regions of the α1a-AR. The regions replaced are shown in black in Fig. 1. Residues shown in gray represent amino acids in the β2-AR that are identical to similar positions in the α1a-AR. Transmembrane boundaries are tenative based upon no structural data. Given the extracellular changes and the high degree of homology in the transmembrane regions between the β2-AR and the α1a-AR, we rationalized that, if most of the antagonist binding determinants for α1-ARs were localized in the extracellular loop regions, we could impart to the β2-AR a higher binding affinity for α1-antagonists via generation of these chimeras, representing a gain of function paradigm.
All of the chimeric constructs expressed functional receptors in transiently transfected COS-1 cells. [125I]CYP labeled an apparently homogenous population of binding sites with similar affinities in membranes prepared from the chimeras or WT β2-AR (Table 1). Binding of [125I]CYP was statistically best fit to a one-site model. As shown in Fig. 2, replacement of individual or of all three extracellular loop regions did not change the affinity of either α1a-antagonist, WB4101 or phentolamine, for the β2-AR. This finding contradicts our hypothesis that our chimeric receptors would exhibit higher α1-antagonist binding affinity than would the WT β2-AR. However, the binding pocket of the β2-AR is quite distinct from the α1-AR, and there is precedent for even similar receptors (α2-AR versus β2-AR) showing distinct helical orientations and packing arrangements such that only functional receptors can be generated from certain chimeric combinations (Mizobe et al., 1996). This would suggest that the adrenergic receptors are quite distinct from each other in the arrangement of the helices and ultimately, the binding pocket, even though they bind the endogenous ligands with similar affinity and have many amino acids in common. As evidence for this, mutagenesis studies have shown that the serines in TM 5 that are utilized in the binding of the catechol hydroxyls in the α1a-AR are quite distinct in both usage and orientation from the β2-AR paradigm (Hwa and Perez, 1996). Therefore, our chimeras may not have worked because the binding of phentolamine and WB4101 in the WT β2-AR may be in a quite distinct orientation from the α1-AR.
Even though our α1-antagonist binding hypothesis seems incorrect, a routine assessment of these chimeras in their ligand binding properties revealed an unexpected phenotype. Table1 indicates the pharmacological characteristics of the WT β2-ARs and the chimeric ARs. Competition binding studies with various β2-agonists and antagonists revealed higher IC50 values (10-fold) for agonists in some of the chimeras than the WT receptor with no changes in the antagonist values. The higher affinity for agonists was specific for the chimeras with changes in EX3 because EX3, EX2+3 and EX1+2+3 all displayed higher affinity, whereas EX2 alone showed no change in affinity. These results suggest that perhaps the EX3 replacement rendered this β2-AR chimera constitutively active, because a higher binding affinity for agonists and not antagonists is a common characteristic for CAM receptors (Sammaet al., 1993).
However, higher agonist binding affinity is also seen in membrane preparations of β-ARs that partition into high and low affinity sites based upon G protein precoupling. Although high affinity sites are typically difficult to obtain in transiently transfected COS-1 cell membranes because of receptor overexpression in a limited number of cells and, thus, a disproportionate amount of receptors are in an uncoupled state, reduction of receptor density to the low femtomolar range can sometimes reveal mixed populations. To explore this possibility, a detailed analysis of isoproterenol binding with and without GTPγS was performed at low receptor number. As shown in Table2 and Fig.3, binding of isoproterenol was statistically best fit to a two-site model in both WT receptor and the EX3 chimera. To assure that receptor number and thus, the receptor/G protein ratio was similar in both preparations, the binding studies were performed under a similar receptor density. Because the EX3 chimera had a fixed expression under 100 fmol/mg of membrane protein that could not be altered significantly with larger amounts of DNA, we lowered the amount of the WT construct used in the transfection until a similar expression was achieved. Nevertheless, the fraction of high affinity sites in EX3 (63%) was twice that of the WT control (30%), and all of the high affinity sites were converted to low affinity sites with GTPγS (Fig. 3). This suggests that the agonist high binding affinity seen in EX3 was not an intrinsic property of the receptor but rather because of precoupling of the receptor to G protein(s). Although the G protein precoupling phenomenon in binding studies has never been directly and proportionately correlated to the formation of an active ternary complex, it is assumed to represent some relationship of the affinity of the receptor for the G protein in the presence of agonist. Given this, the results are consistent with the interpretation that the extracellular loop replacement altered G protein coupling in which EX3 formed a ternary complex much easier than WT control because of its higher affinity/coupling to the G protein.
If the rate of formation of an active ternary complex in EX3 was indeed greater than the WT control and not because of some artifact or alteration in receptor/G protein ratio in membrane preparations, we should observe concomitant increases in both the basal and agonist-dependent changes in signal transduction in whole cell studies. Therefore, basal cAMP production was measured in COS-1 cells transfected with the chimeric or WT receptor. As shown in Fig.4, even though the WT and EX2 expresses more receptor, EX3 and EX1+2+3, displayed a significantly higher basal cAMP activity (70% over WT, p < 0.001), which is commensurate with the 100% increased fraction of high affinity sites in EX3. This constitutively active receptor is not fully active but represents a basal activity that is 20% of a fully activated WT receptor (Fig. 5). The constitutive activity of EX3 was reversed upon the addition of the inverse agonist, ICI 118551. Agonist-dependent changes in cAMP production is shown in Fig.5. Even though EX3 expressed half the receptors of WT, the isoproterenol-stimulated EC50 of the cAMP response for EX3 was 4-fold lower (a significantly higher potency) than WT control (7.4 × 10-7 m versus 2.8 × 10-6 m, p < 0.03). Even though the degree of constitutive activity was not proportional to the amount of high affinity sites, both of these whole cell studies are consistent with the EX3 chimera being able to generate more active ternary complex than the WT control.
All in all, the data indicate that replacement of EX3 by corresponding α1a-AR residues generates a constitutively active receptor, because signal transduction can occur without the presence of agonist. However, the mechanism behind this constitutive activity is distinct from the mechanism described by all other CAM models previously characterized in adrenergic receptors (Samma et al., 1993; Perez et al., 1996; Hwa et al., 1996, 1997; Porter et al., 1996). Earlier models of constitutive activity have shown that the high affinity binding of agonists is independent of G protein coupling either as assessed by no changes in agonist binding with GTP analogs or in receptor solubilization studies (Samma et al., 1993), where no G protein can interact. CAMs based on these earlier studies, are thought to cause intrinsic changes to the receptor, a G protein-independent conformational change that causes an increase in the J or isomerization constant of the receptor in isomerizing from R to R* (Fig.6A). In addition, CAMs were shown to have little effect on the M constant, a receptor to G protein affinity constant in which an increased value would result in an enhanced coupling of the receptor to the G protein and formation of an active R*G (Fig. 6A), although this has not been empirically tested with purified components. Although computer simulations of a constitutively active β2-AR and the basis for the revision of the ternary complex model (1) (Fig. 6A) indicated a minor contribution was because of altered G protein coupling (the M constant), this was necessary to account for the full extent of the increase in basal signal transduction which is mostly because of changes in J. This revision of the ternary complex model was necessary to account for the existence of an efficacy-related change in affinity for the uncoupled state (G protein independence) of the mutant receptor. In contrast, the EX3 results would suggest that the M constant and not the J constant is being altered in this model because its constitutive activity and high affinity state was dependent on the G protein. However, the revised TCM also makes a large assumption that only the R* form is able to bind and activate the G protein, so that HR*G is the only possible ternary complex formed. Alternately, according to the cubic ternary complex model of Kenakin (1995) (Fig. 6B), the EX3 chimera could have altered equilibrium because of an increased M constant but the commensurate increase in the inactive species, RG1, would ultimately effect the equilibrium leading to a greater amount of the active species, R*G1 or AR*G1. Although our data here cannot distinguish between these two models, the results suggest that the EX3 constitutive activity was because of an altered affinity/coupling for the G protein, which ultimately through mass action influences the amount of active complex that is formed. This is distinct from other CAMs that increase active species directly without alterations in G protein coupling.
Because the location of the replacement in the EX3 chimera is in the extracellular environment, it is not possible to propose a direct interaction of these residues in altering the affinity of the receptor to the G protein. There is precedent, however, in soluble proteins that loop regions can influence the arrangements and packing of α-helices (Munoz et al., 1995). Our observations are most likely indirectly attributable to helical changes/packing in TM 6 and/or 7, which have been implicated as important movements in the activation process of rhodopsin (Altenbach et al., 1996). These changes could effect the conformation of the third intracellular loop, which is implicated in the binding and activation of G proteins. Nevertheless, it is interesting to note that the constitutive activity in EX3 was independent of any subsequent changes in EX1 and/or EX2, suggesting that the conformation adopted by EX3 was fixed and nonreversible as though a constraining factor was broken. The implications of the current study reinforce an emerging concept that the entire structure of G protein-coupled receptors have evolved to constrain the receptor in the inactive state. Altered structure through mutagenesis of the intracellular loops (Samma et al., 1993), transmembrane domains (Perez et al., 1996; Hwa et al., 1996,1997; Porter et al., 1996), and now extracellular loops have all led to constitutive activity, although the mechanism behind these changes may be different.
Footnotes
- Received August 19, 1997.
- Accepted November 21, 1997.
-
Send reprint requests to: Dr. Dianne M. Perez, Department of Molecular Cardiology, FF3-01, The Cleveland Clinic Foundation, 9500 Euclid Ave., Cleveland, Ohio 44195. E-mail:perezd{at}cesmtp.ccf.org
-
↵1 Current affiliation: Division of Hematology, Vanderbilt University, Nashville, TN 37232.
-
This work was supported in part by National Institutes of Health Grant RO1-HL52544 (D.M.P.) and an unrestricted research grant from Glaxo Wellcome. This work was done under the tenure of an Established Investigator Award from the American Heart Association (D.M.P.).
Abbreviations
- AR
- adrenergic receptor
- CAM
- constitutively active mutation
- EX
- extracellular loop
- [125I]CYP
- iodocyanopindolol
- TCM
- ternary complex model
- TM
- transmembrane
- WT
- wild-type
- GTPγS
- guanosine 5′-3-O-(thio)triphosphate
- ANOVA
- analysis of variance
- EGTA
- ethylene glycol bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid
- HEPES
- 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
- The American Society for Pharmacology and Experimental Therapeutics