Modification of the beta 2-adrenergic receptor to engineer a receptor-effector complex for gene therapy.

Depressed G-protein-coupled receptor (GPCR) signaling has been implicated as a component of the pathophysiology of a number of complex diseases including heart failure and asthma, and augmentation or restoration of signaling by various means has been shown to improve organ function. Because some properties of native GPCRs are disadvantageous for ectopic therapeutic expression, we utilized the beta(2)-adrenergic receptor (beta(2)AR) as a scaffold to construct a highly modified therapeutic receptor-effector complex (TREC) suitable for gene therapy. Altogether, 19 modifications were made to the receptor. The ligand-binding site was re-engineered in TM-3 so that a beta-hydroxylmethyl side chain acts as a proton donor for the binding of a novel ligand. In addition, sites critical for agonist-promoted down-regulation in the amino terminus and for phosphorylation by GPCR kinases, and protein kinases A and C, in the third intracellular loop and the carboxyl terminus of the receptor were altered. These modifications of the receptor resulted in depressed agonist-stimulated adenylyl cyclase activity (26.8 +/- 2.1 versus 41.4 +/- 8 pmol/min/mg for wild-type beta(2)AR). This was fully restored by fusing the carboxyl terminus of the modified receptor to G alpha(s) (43.3 +/- 2.7 pmol/min/mg). The fully modified fused receptor was not activated by beta-agonists but rather by a nonbiogenic amine agonist that itself failed to activate the wild-type beta(2)AR. This two-way selectivity thus provides targeted activation based on physiologic status. Furthermore, the TREC did not display tachyphylaxis to prolonged agonist exposure (desensitization was 1 +/- 5% versus 55 +/- 4% for wild-type beta(2)AR). Thus, despite extensive alterations in regions of conformational lability, the beta(2)AR can be tailored to have optimal signaling characteristics for gene therapy. As a general paradigm, TRECs for enhancement of other G-protein signaling appear to be feasible for modification of other pathologic states.

Cell surface receptors that activate signaling via coupling to G proteins comprise a large superfamily consisting of several hundred distinct receptors. G-protein-coupled receptors are widely expressed and serve diverse functions within hormonal, neurotransmission, immune, and growth systems. Receptor signaling in this family is dynamically regulated such that the cell integrates a large number of incoming signals and ulti-mately adapts to short and long term events. Although such regulation can occur at the level of G proteins or subsequent downstream participants in the pathway, highly specific mechanisms acting at the receptor itself are the primary mechanisms by which the signal transduction of a given receptor is selectively modulated (1,2).
Both the ␤ 1 -adrenergic receptor and ␤ 2 -adrenergic receptor (␤ 2 AR) 1 subtypes couple to the stimulatory guanine nucleotide binding protein G s , resulting in activation of adenylyl cyclase and increased intracellular cAMP. Rapid regulation of ␤AR coupling efficiency (1) is mediated by phosphorylation of the receptor by G protein-coupled receptor kinases (GRKs), as well as protein kinase A (PKA) and protein kinase C (PKC). GRKmediated desensitization is evoked by agonist occupancy, whereas PKA-and PKC-mediated phosphorylation are mechanisms of heterologous desensitization. Receptor-mediated responsiveness is also dependent on cell surface ␤AR expression, which is regulated by complex sorting and trafficking mechanisms involving internalization, recycling, and degradation of receptors, as well as transcriptional mechanisms (3)(4)(5). Although regulation of ␤AR function serves important roles in maintenance of homeostasis under normal and compensatory states, in certain pathologic conditions depressed function appears to contribute to the disease state or may limit the therapeutic response to agonist (6). For example, in the lung airway smooth muscle ␤ 2 AR act to relax the muscle resulting in bronchodilation (7). ␤ 2 AR on airway epithelial cells also contribute to the establishment of bronchomotor tone (7,8). In asthma, a disease characterized by bronchial smooth muscle contraction, airway epithelial and smooth muscle ␤ 2 AR function is depressed, thus contributing to airway obstruction (7,9). Furthermore, regular use of ␤-agonists in the treatment of bronchospasm acts to desensitize ␤ 2 AR (termed tachyphylaxis), thus potentially worsening obstruction and limiting the effectiveness of this therapy (10 -12). In the heart, ␤ 1 -adrenergic receptor and ␤ 2 AR expressed on myocytes act to increase the rate and force of contraction (13,14). In human heart failure as well as in many animal models of the syndrome, changes in ␤AR responsiveness are well documented (15,16). Mechanisms of such dysfunction include down-regulation of receptor number, phosphorylation of receptors by the ␤AR kinase (␤ARK) and PKC, and changes in downstream elements of the signal transduction cascade such as G proteins and adenylyl cyclase (14 -19).
Overexpression of ␤ 2 AR or inhibition of ␤ARK has been shown to markedly improve organ function in animal models of asthma and/or heart failure. For example, we have recently shown that transgenic overexpression of ␤ 2 AR in airway smooth muscle results in mice that are resistant to methacholine-induced bronchoconstriction (20). In another set of studies, we showed that ␤ 2 AR overexpression (of as little as ϳ2-fold over background) in airway epithelial cells of transgenic mice decreased ozone-induced bronchial hyperreactivity and the constrictive response to methacholine (8). We have also found that overexpression of ␤ 2 AR in the hearts of transgenic mice increases resting and agonist-stimulated ventricular function (21,22). Lefkowitz and co-workers (23,24) have demonstrated that transgenic expression and ex vivo (25) and in vivo (26,27) delivery of viral constructs to express ␤ 2 AR or a ␤ARK inhibitor can significantly enhance ventricular function. In several animal models with impaired cardiac contractility (28 -31), such intervention results in improved, often fully restored, cardiac function.
The above studies indicate the potential for gene therapy via expression of the ␤ 2 AR in the lung and heart. The wild-type receptor, however, has a number of properties that may limit the effectiveness of this approach, particularly with lower levels of expression, as is likely in human therapy. Thus, under pathologic conditions where ␤ARK activity is increased, phosphorylation and desensitization of the genetically expressed ␤ 2 AR may ultimately limit therapeutic efficacy. Similarly, PKA or PKC activation in the cell, because of homologous or heterologous mechanisms, could alter receptor function. Also, agonist-promoted receptor down-regulation, which is partially due to degradation of receptor protein (5), could decrease the available number of cell surface receptors after initial successful expression. In addition, certain conformational and spatial properties of the receptor-G s interface limit the efficiency of functional coupling of the wild-type ␤ 2 AR (32). Finally, a therapeutically expressed wild-type ␤ 2 AR is activated by the same endogenous catecholamines and synthetic agonists as is the endogenously expressed ␤ 2 AR. Thus, under these circumstances there is no mechanism to selectively activate or deactivate the genetically expressed receptor as may be dictated by the clinical status.
We have considered, however, that the ␤ 2 AR can be utilized as a scaffold upon which to construct a synthetic G-proteincoupled receptor with specific properties to optimize gene therapy. It is not clear, however, whether such a highly modified receptor can maintain functional integrity, because multiple mutations throughout the protein would need to be introduced in known areas of conformational lability. In the current work, we utilize the human ␤ 2 AR in the above fashion to create a protein with markedly different properties from an adrenergic receptor and with characteristics suitable for gene therapy. As a general paradigm the creation of such designer receptors is likely to be amendable for enhancement of other G-protein signaling for the modification of other pathologic states.

EXPERIMENTAL PROCEDURES
Constructs-The human wild-type ␤ 2 AR cDNA cloned into the expression vector pBCBI was used as a template for mutagenesis. From this, an 1862-base pair SacI/SalI fragment encompassing the entire wild-type ␤ 2 AR coding region was isolated and subcloned into M13mp19 for oligonucleotide-directed mutagenesis (33). Sequence en-coding the HA epitope tag (MGYPYDVPDYAS) was incorporated at the amino terminus of the receptor for all subsequent constructs. Amino acid 27 was changed from Gln to Glu, and amino acid 113 was changed from Asp to Ser, to make the receptor refractory to agonist-promoted down-regulation and to alter the ligand binding specificity, respectively. In addition, potential PKA/PKC phosphorylation sites at amino acids 272, 273, 356, and 357 were changed to Ala. Similarly, potential GRK phosphorylation sites at amino acids 366, 367, 371, 375, 395, 404, 407, 412, 418, 419, and 422 were changed to Ala. Another modification included the mutation of the stop codon to alanine and an in-frame ligation at the NcoI site immediately preceding the initiator methionine codon of the entire open reading frame of the rat G␣ s cDNA (long form).
Tissue Culture-Chinese hamster fibroblasts (CHW-1102) were grown in Dulbecco's modified Eagle's medium, 10% fetal calf serum, 100 units/ml penicillin, and 100 g/ml streptomycin. The cells were transiently transfected with a concomitant infection of a "helper" adenovirus using methods as described (34). Briefly, 7 ϫ 10 6 cells were plated in a T-75 flask ϳ16 h prior to transfection. The cells were washed twice with Hank's balanced salt solution and incubated in 0.5 ml of a transfection/infection mixture containing 1 ϫ 10 10 plaque-forming unit purified E1A-deleted adenovirus, 2% fetal calf serum, 80 g/ml DEAEdextran, and 15 g of plasmid DNA in the aforementioned expression vectors, at 37°C and 5% CO 2 with gentle rocking. After 2 h, the solution was aspirated, and the cells were incubated with 10% Me 2 SO in Dulbecco's modified Eagle's medium for 2 min. After a final wash, the cells were incubated in complete medium for 48 h prior to analysis. For long term desensitization studies, the cells were incubated for 24 h in the presence of 10 M albuterol or 100 M of the catechol butanone compound L-158,870 (Merck).
Radioligand Binding and Adenylyl Cyclase Activities-The cells were washed twice with phosphate-buffered saline, detached with a rubber policeman in 5 mM Tris, pH 7.4, 2 mM EDTA buffer, and centrifuged at 30,000 ϫ g for 10 min. To determine receptor density, membrane pellets were resuspended in 75 mM Tris, pH 7.4, 12.5 mM MgCl 2 , 2 mM EDTA, and radioligand binding with 125 I-cyanopindolol was performed for cells expressing the wild-type ␤ 2 AR as described previously (35). To determine adenylyl cyclase activities, the membranes were incubated with 30 mM Tris, pH 7.4, 2 mM MgCl 2 , 0.8 mM EDTA, 120 M ATP, 60 M GTP, 2.8 mM phosphoenolpyruvate, 50 g/ml myokinase, 100 M cAMP, and 1 Ci of [␣-32 P]ATP for 30 min as described (35). The reactions were carried out with water, 100 M forskolin, 10 M epinephrine, 100 M norepinephrine, and varying concentrations of albuterol or L-158,870. [ 32 P]cAMP was separated from ␣-32 P by chromatography over alumina columns. A [ 3 H]cAMP standard included in the stop buffer accounted for individual column recovery.
Western Blots-The Western blots were carried out using immunoprecipitated protein essentially as described (35). Briefly, the cells were lysed in buffer containing 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, and protease inhibitors (10 g/ml benzamidine, 10 g/ml soybean trypsin inhibitor, 10 mg/ml aprotinin, and 5 g/ml leupeptin) in phosphate-buffered saline. Equal amounts of solubilized protein were preincubated with protein G-agarose beads to remove nonspecific binding for 2 h at 4°C, followed by incubation with protein G-agarose beads and 1 g/ml of rat anti-HA antibody (Roche) for 18 h at 4°C. Following immunoprecipitation, the beads were washed three times by centrifugation and resuspension and boiled for 5 min in SDS sample buffer. The proteins in the supernatant were fractionated on a 10% SDS-polyacrylamide gel and transferred to nitrocellulose. Western blots were performed using a mouse monoclonal anti-HA antibody (Babco) at a dilution of 1:200.
Miscellaneous-Protein determination was by the copper bicinchoninic acid method. The results are presented as the means Ϯ S.E. Statistical comparisons were made by paired two-way t tests with significance imparted for p Ͻ 0.05. The response curves were generated using Prism 3.0 software (GraphPad).  Fig. 1 are the modifications imposed on the ␤ 2 AR. To facilitate identification by immunoblotting, all receptors were fused at the amino terminus with the hemagglutinin (HA) epitope tag. The combined effect of mutations that remove sites required for down-regulation and phosphorylation was considered to be unlikely to alter ligand binding, so these were imposed collectively in the initial round of mutagenesis. Ultimately, two modified receptors and the wild-type ␤ 2 AR were studied. One receptor contained the single mutations introduced to change amino acids at the indicated positions to alter the ligand-binding site, receptor down-regulation and receptor phosphorylation as further described below. This receptor is denoted as the "modified receptor," A second receptor contained these mutations in addition to being fused to G␣ s and is denoted as the therapeutic receptor-effector complex (TREC). Receptors were transiently expressed in CHW-1102 cells, and expression was verified by 125 I-cyanopindolol binding (ϳ400 fmol/mg) or Western blots.

Shown in
The agonist-binding site was re-engineered to satisfy several criteria. First, we wanted to render the receptor unresponsive to endogenous catecholamines and other synthetic ␤-agonists. In addition, we aimed to have the mutated receptor respond to a unique agonist. Finally, we preferred that this unique agonist not activate wild-type ␤ 2 AR. The key interactions (36 -38) of the ␤ 2 AR with catecholamines occur in transmembrane domain III at Asp 113 , transmembrane domain VI at Asn 293 , and transmembrane domain V at Ser 204 and Ser 207 . At Asp 113 , it has been proposed that the binding of agonists (and antagonists) involves the formation of an ion pair between the amine group of the ligand and the Asp carboxylate side chain. This interaction can apparently be brought about in mutant forms of the receptor by using different classes of ligands (39). We thus mutated Asp 113 to Ser, which substitutes the ␤-carboxymethyl side chain with a ␤-hydroxylmethyl group. This renders the receptor incapable of activation by biogenic amines. However, compounds that otherwise satisfy requirements for ␤ 2 AR binding and are capable of accepting hydrogen bonds from the engineered Ser in transmembrane domain III, would activate the receptor. Based on work by Strader et al. (39), we chose 1-(3Ј,4Ј-dihydroxyphenyl)-3-methyl-1-butanone as the agonist specific for the modified ␤ 2 AR. This compound, also known as L-158,870, retains the catechol ring but instead of the amine containing alkyl-side chain has a four carbon length moiety, keto-substituted at the ␣-carbon. The functional results of the Asp to Ser substitution are shown in Fig. 2. Because the therapeutic ␤-agonists commonly utilized clinically are partial agonists, we used albuterol as the benchmark agonist. As shown in Fig. 2A, albuterol stimulated adenylyl cyclase by the Asp 113 (the wild-type allele) receptor ϳ50% over basal levels, whereas there was no stimulation by albuterol of the modified mutant (including Ser 113 ) receptor. In contrast, L-158,870 stimulated adenylyl cyclase by the modified receptor ϳ70% above basal but did not activate wild-type ␤ 2 AR (Table I). In additional studies with the modified receptor, neither norepinephrine nor epinephrine stimulated adenylyl cyclase activities over basal levels (5 Ϯ 10% and 10 Ϯ 10%, respectively; n ϭ 6; p Ͼ 0.05 versus basal levels).
One consequence that we observed with the Ser substitution at position 113 of the third transmembrane domain was a decrease in basal adenylyl cyclase activities compared with the wild-type ␤ 2 AR (16.6 Ϯ 1.3 versus 27.8 Ϯ 2.6 pmol/min/mg). So, whereas the two agonists stimulated their respective receptors to the same fold extent over basal, the absolute levels of maximal stimulation of the Ser 113 modified ␤ 2 AR was nonetheless depressed (Fig. 2B). This effect on basal coupling is likely due to conformational effects on the intracellular coupling domains that are imposed by this substitution in transmembrane domain III. Of note, Asp in this position is invariant in all biogenic amine-liganded G-protein-coupled receptors. To attempt to overcome this therapeutically disadvantageous characteristic, the intracellular carboxyl terminus of the modified receptor was fused to the amino terminus of the ␣-subunit of G s . At least with wild-type ␤ 2 AR, such a fusion has been reported to increase basal (non-agonist-stimulated) levels of adenylyl cyclase activity (35). Because the Ser substitution at residue 113 ablated 125 I-cyanopindolol binding, immunoblotting using HA antibody was used for verification of the expression of the fused receptor. As shown in Fig. 3, the wild-type ␤ 2 AR migrated at a molecular mass of ϳ60 kDa, with an additional band representing more highly glycosylated states observed between 80 and 100 kDa. The TREC migrated at molecular masses consistent with the addition of the 42 kDa G␣ s protein, with a band at ϳ100 kDa and another broad band representing the highly glycosylated states at Ն120 kDa.
The results from adenylyl cyclase studies of the TREC are shown in Fig. 4. The fusion with G␣ s indeed increased basal levels from 16.2 Ϯ 1.3 to 23.1 Ϯ 1.9 pmol/min/mg. Agonist stimulation was preserved and in fact increased from ϳ70% over basal in the absence of G␣ s to ϳ90% over basal with the fusion. With these changes, maximal adenylyl cyclase activities were equivalent between wild-type ␤ 2 AR and the fully modified fused receptor in the basal state and with the respective agonist activation (Table I).
We next considered the effects of modifications imposed to limit receptor desensitization. These mutations included substitution of all Ser and Thr in the cytoplasmic tail of the receptor with Ala. A similar mutant ␤ 2 AR, which consisted of Gly and Ala substitutions for Ser and Thr, has been shown to lack sites for GRK (␤ARK)-mediated phosphorylation (40). There are two potential PKA/PKC sites (Arg-Arg-Ser-Ser*) at amino acid positions 273 and 357 that were changed to Ala. Because the Ser immediately adjacent to the second Arg could also serve as a phosphoacceptor, these were mutated to Ala as well. Because long term agonist exposure evokes desensitization via these phosphorylation events as well as down-regulation of receptor number, we also sought to render the receptor resistant to down-regulation. We thus substituted Glu for Gln at amino acid position 27 of the amino terminus (which is a naturally occurring polymorphism in the human gene), which we have previously shown to result in a functional receptor that fails to undergo agonist-promoted destabilization and subsequent degradation (41). Thus, to test whether the above modifications in fact limit desensitization, cells in culture expressing the wild-type ␤ 2 AR or the TREC were incubated with equipotent concentrations of albuterol (10 M) or L-158,870 (100 M) for 24 h. This exposure provides the ultimate verification of a functional knockout of desensitization from all the above mechanisms and also mimics the in vivo setting of repetitive dosing over prolonged periods of time. The cells were then placed on ice and washed, membranes were prepared, and agonist-stimulated adenylyl cyclase activities were determined. The results of such studies are shown in Fig. 5. Wildtype ␤ 2 AR underwent extensive functional desensitization under these conditions that amounted to a loss of 55 Ϯ 4% of the maximal response observed in control cells (Fig. 5A). In marked contrast, the fully modified receptor failed to exhibit any loss of function with long term agonist exposure. As shown in Fig. 5B, the responses are superimposable, with mean desensitization being quantitated at 1 Ϯ 5%. Also, the EC 50 s did not significantly change with the TREC following agonist exposure (6.8 Ϯ 1.2 versus 12 Ϯ 3.9 M). Forskolin-stimulated activities were not altered by agonist treatments (untreated versus treated: 93 ϩ 7 versus 86 ϩ 6 pmol/min/mg for the wild-type ␤ 2 AR and 94 ϩ 6 versus 110 ϩ 7 pmol/min/mg for the TREC), indicating that these responses are receptor-specific.
The above results indicate that constructing synthetic Gprotein-coupled receptors with distinct properties for targeted modification of specific signaling pathways is feasible. Here we utilized the ␤ 2 AR as a backbone for constructing such a highly modified receptor. In this case, the receptor created is no longer an adrenergic receptor, because it is not activated by catecholamines. This is a critical feature of the approach, because it provides for selective activation of the therapeutically expressed receptor via an agonist that does not activate endogenously expressed ␤AR. The need to activate (and to deactivate by withdrawal of agonist) the therapeutically expressed receptor is a mechanism by which the response can be modulated in accordance with physiologic need. We have shown, for example, that high overexpression of wild-type ␤ 2 AR in the hearts of G␣ q overexpressing transgenic mice worsens cardiac hypertrophy and dysfunction and increases mortality, whereas low levels of ␤ 2 AR overexpression in the G␣ q mouse are beneficial (28). So, by selectively activating the therapeutically expressed receptor with appropriate dosing of the agonist, a certain level of enhanced signaling could be obtained for effectiveness without adverse events. There may also be times, however, when no signal enhancement is desirable. Withdrawal of the specifically designed agonist would thus result in cessation of signaling by the genetically expressed receptor despite elevated catecholamines. It should be noted, however, that the ability of the therapeutically expressed receptor to spontaneously activate could still result in some degree of enhanced signaling even in the absence of agonist. As discussed above, our initial mutations resulted in a receptor with significantly depressed basal and agonist-stimulated adenylyl cyclase activities, which was considered therapeutically disadvantageous, so the fusion with G␣ s was undertaken. Basal adenylyl cyclase activities were  3. Western analysis of the expression of ␤ 2 AR and the TREC. Western blot analysis of anti-HA immunoprecipitated protein was performed to identify wild-type ␤ 2 AR and TREC in CHW-1102 cells transiently expressing each receptor as described under "Experimental Procedures." The approximate molecular masses (kDa) are noted. NT, nontransfected. Shown is a representative experiment.

FIG. 4. Functional properties of the nonfused modified receptor and the TREC.
Adenylyl cyclase activities in response to the catechol butanone compound L-158,870 were determined in membranes isolated from CHW-1102 cells transiently transfected with each receptor as described under "Experimental Procedures." Fusion of the modified receptor to G␣ s to create the TREC resulted in increased basal and agonist-stimulated activities that were the same as wild-type ␤ 2 AR (Table I). Shown are the results from four independent experiments. increased with the fused receptor but were no more than basal levels of the wild-type ␤ 2 AR. Of note, we have not tested L-158,870 against other G-protein-coupled receptors, but based on its structure, it is unlikely to activate any other natively expressed receptors.
The TREC also lacks the ability to undergo agonist-promoted phosphorylation and down-regulation by degradation. These modifications were based on multiple reports that ␤ 2 AR function is depressed in asthma and heart failure via these and other mechanisms (see the Introduction). It seemed appropriate, then, to render the modified receptor incapable of such regulation, thus providing therapeutic efficacy particularly with lower levels of expression. Of course, down-regulation of receptor number can also be the result of changes in transcription and message stability (3,4). Some of these mechanisms are based on certain 5Ј-and 3Ј-untranslated region sequences of the ␤ 2 AR, which we purposely did not include in the construct. However, complex regulatory events that alter mRNA levels may be in place in certain pathologic states, and expression of modified G-protein-coupled receptors for gene therapy may nevertheless be subject to such processes.
The potential for gene therapy to augment ␤AR signaling has been studied in a number of animal models of heart and lung disease. In the lung, ␤ 2 AR are expressed in many cell types including airway epithelial cells, airway smooth muscle, and alveolar type II cells. We have overexpressed wild-type ␤ 2 AR in transgenic mice targeting the expression to each of these three cell types. Using the Clara cell secretory protein promoter, we overexpressed by 2-fold ␤ 2 AR in airway epithelium and found that these mice are hyporesponsive to bronchoconstriction from exposures to ozone and methacholine (8). Additional studies indicated that epithelia secrete a smooth muscle relaxation factor (under ␤ 2 AR control), which results in this effect. In airway smooth muscle, the ␣ smooth muscle actin promoter was used to target transgenic expression of the ␤ 2 AR (20). These mice were highly resistant to bronchoconstriction to methacholine in vivo and ex vivo. Indeed, in the absence of agonist they were less responsive than nontransgenic mice treated with agonist, indicating a very high degree of physiologically relevant signaling. However, ϳ75-fold overexpression was required for this effect. The above two studies are relevant for potential gene therapy for asthma and chronic obstructive pulmonary disease, which are characterized by hyperreactive, constricted airways. Epithelial overexpression of ␤AR may also partially correct the defect in chloride transport in cystic fibrosis (42). In alveolar type II cells we utilized the surfactant protein C promoter to overexpress ␤ 2 AR and found that alveolar fluid clearance was enhanced (43), suggesting a potential therapy for pulmonary edema.
In the heart, ␤ 2 AR overexpression in transgenic mice increases inotropy and chronotropy, as has been shown in multiple studies (21)(22)(23). Cardiomyopathic changes are noted, however, with high levels of expression or with moderate levels over a long period of time (22). This emphasizes the need for careful control of receptor signaling, a major feature of our receptor-effector complex. In double transgenic mice, we have recently shown that the hypertrophy and resting cardiac contractile dysfunction due to G␣ q overexpression is partially rescued by ␤ 2 AR overexpression (28). Another method of improving ␤AR function in the heart has been with an inhibitor of the ␤AR kinase. Transgenic overexpression of a peptide inhibitor (␤ARK CT ) rescues ventricular function, and in some cases hypertrophy and other pathologic features, of several mouse models of heart failure. This includes the muscle-specific LIM protein knockout mouse (29), the modified myosin heavy chain overexpressing mouse (31) and a mouse overexpressing calsequestrin (30). The use of a ␤ARK inhibitor may have particular advantages in pathologic states where overstimulation is deleterious, because the inhibitor would act to restore receptor function, avoiding excessive activation of the pathway. However, ␤ARK inhibitors are likely to be most effective when ␤ARK activity is enhanced in the disease. And, such inhibition lacks specificity because many G-protein-coupled receptors are phosphorylated by GRKs. Nevertheless, in the setting of the compromised heart, the in vivo consequences of ␤ARK inhibition are substantial, and appropriately, the approach continues to be pursued.
Our modified receptor effector complex differs by a number of features from a mutated receptor reported by Coward et al. (44). This receptor consisted of a opioid receptor that was mutated such that the affinity for an endogenous peptide agonist (dynorphin) was decreased 200 -2000-fold; yet affinity for some small synthetic molecules was relatively unchanged. Interestingly, the maximal response (inhibition of adenylyl cyclase) by dynorphin was not changed by the mutations. So, with high concentrations as might be seen in certain microenvironments, this receptor could still be activated by endogenous agonist. Furthermore, the small molecule agonists activated wild-type opioid receptors, so "two-way" selectivity was not attained. No modifications were made to alter coupling or regulation.
In summary, the framework of the ␤ 2 AR was utilized to create a G-protein-coupled receptor suitable for therapeutic The membranes were prepared, and the adenylyl cyclase activities were determined as described under "Experimental Procedures." The results are presented as percentages of the maximum activity for the untreated control. Although the wild-type ␤ 2 AR underwent ϳ55% desensitization, the TREC failed to undergo any decrease in function with agonist exposure. The results are from four independent experiments. modulation of signaling in pathologic states. Extensive modifications were undertaken to re-engineer the ligand-binding site, to remove sites for phosphorylation by GRKs and PKA/PKC, to depress agonist-promoted down-regulation, and to increase G s coupling. The resulting receptor-effector complex was unresponsive to catecholamines and the prototypic therapeutic agonist albuterol but was activated by a synthetic agonist that itself did not activate wild-type ␤ 2 AR. Despite decreased coupling imposed by the ligand binding mutation, fusion of the receptor to G␣ s restored basal and maximal agonist-stimulated adenylyl cyclase activities. The receptor-effector complex showed no desensitization during agonist occupancy despite high agonist concentrations and prolonged exposures. Concomitant modifications of all these domains in a single receptor has not been previously reported. The fact that a stable, functional receptor, with such specific therapeutic properties can be made, is promising as a modality for enhancing ␤AR-like signaling in diseases where function is impaired. As a general paradigm, based on these results, it is likely that other "designer" Gprotein-coupled receptor-effector complexes can be constructed to therapeutically modulate other signaling pathways.