Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

-Adrenergic receptors are members of the G protein-coupled receptor superfamily with seven transmembrane helices and are classified into three subtypes: ₁, ₂, and ₃. ₁-, ₂-, and ₃-adrenergic receptors are predominantly expressed in the heart, trachea, and adipose tissue, respectively. Therefore, several ₁-selective agonists are used for the improvement of the function of failing heart, whereas ₂-selective agonists are used for treatment of bronchial asthma. To design a better drug, it is important to determine the binding sites for these agonists. Endogenous catecholamine such as norepinephrine and epinephrine, and synthetic catecholamine such as isoproterenol are full but nonselective agonists of -adrenergic receptors. Ligand binding domains of -adrenergic receptor for these agonists were shown to be located within transmembrane domains (Dixon et al., 1987; Dohlman et al., 1988; Wong et al., 1988; Hockerman et al., 1996). Site-directed mutagenesis studies of ₂-adrenergic receptor have identified several key residues for isoproterenol binding. The amino group of catecholamine was revealed to interact with Asp¹¹³ in transmembrane domain 3 of ₂-adrenergic receptor (Strader et al., 1987, 1988, 1989a). The two catechol hydroxyl groups of agonists were shown to form hydrogen bond with Ser²⁰³ (Sato et al., 1999) as well as Ser²⁰⁴ and Ser²⁰⁷ (Strader et al., 1989b) in transmembrane domain 5. Furthermore, the catechol phenyl group and -hydroxyl group of catecholamine were reported to interact with Phe²⁹⁰ (Strader et al., 1994) and Asn²⁹³ (Wieland et al., 1996), respectively, in transmembrane domain 6. Because these residues in ₂-adrenergic receptor are conserved among three subtypes, these residues are not responsible for the subtype-selective binding.

Several groups have studied the binding sites for subtype-selective ligands, by using the chimeric receptors of ₁- and ₂-adrenergic receptor. Frielle et al. (1988) suggested that transmembrane domain 6 and transmembrane domain 7 play an important role in binding of ₁-selective antagonist betaxolol and ₂-selective antagonist ICI118,551. The binding sites of the ₁- and ₂-selective antagonists were also analyzed by Marullo et al. (1990). They reported that the subtype-selective binding of antagonists cannot be determined by single transmembrane domain. It has been reported that transmembrane domain 4 is responsible for the ₁-selective binding of an endogenous agonist norepinephrine (Frielle et al., 1988; Dixon et al., 1989). However, key amino acids or subdomains of each -adrenergic receptor for subtype-selective agonists have not been investigated in detail.

We have studied the binding domains for ₁- and ₂-selective agonists and have reported that transmembrane domains 2 and 7 of ₁- and ₂-adrenergic receptors form a binding pocket, and that Leu¹¹⁰, Thr¹¹⁷, and Val¹²⁰ in transmembrane domain 2 of ₁-adrenergic receptor or Tyr³⁰⁸ in transmembrane domain 7 of ₂-adrenergic receptor are major determinants for ₁- or ₂-selective agonists, respectively (Isogaya et al., 1998, 1999; Kikkawa et al., 1998). However, ₁-selectivity of T-0509 and denopamine [K_i(₂)/K_i(₁)] used in the previous study was at most 10-fold. Therefore, we could not determine the interaction of each amino acid in transmembrane domain 2 with ₁-adrenergic receptor in an active state.

()-RO363 is one of the ₁-selective agonists with structural similarity to T-0509 and denopamine (Fig. 1). The affinity of ()-RO363 for ₁-adrenergic receptor is about 100- and 3000-fold higher than for ₂- and ₃-adrenergic receptors, respectively (McPherson et al., 1984; Molenaar et al., 1997). Therefore, we used ()-RO363 as a tool to examine the contribution of key amino acids responsible for the ₁-selective binding to the binding in the resting and an active states. We constructed chimeric ₁/₂-adrenergic receptors and Ala-substituted ₁-adrenergic receptors, and found that several key amino acids play an essential role in ₁-selective binding of ()-RO363, by competitive binding assay. To analyze a role of these amino acids in an active state, we constructed double mutants, in which Leu³²³ located at intracellular third loop was changed to Lys to transform the receptor into a constitutively active form, and key amino acids responsible for ₁-selective binding were changed to Ala. Based on these results, we built a three-dimensional model of the binding domain for ()-RO363.

View larger version (11K): [in this window] [in a new window]   Fig. 1.   Structures of ()-RO363, T-0509, and denopamine.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. The plasmids pBC-₁ and -₂ encoding human ₁- and ₂-adrenergic receptors were kindly provided by Dr. R. J. Lefkowitz (Duke University, Durham, NC). The plasmid encoding a constitutively active ₁-adrenergic receptor mutant (L323K) was described by Lattion et al. (1999). (±)-Propranolol, ()-isoproterenol, DEAE-dextran, and GTP were purchased from Sigma-Aldrich (St. Louis, MO). ¹²⁵I-Cyanopindolol was obtained from Amersham Biosciences, Inc. (Piscataway, NJ). Dulbecco's modified Eagle's medium was obtained from Invitrogen (Carlsbad, CA). Fetal bovine serum was from JRH Biosciences (Lenexa, KS). Pfu DNA polymerase was purchased from Stratagene (La Jolla, CA). ()-RO363 was provided by Roche Applied Science (Palo Alto, CA).

Construction of Chimeric Receptors and Mutant Receptors. Chimeric ₁/₂-adrenergic receptors were constructed by PCR, as described by Kikkawa et al. (1998). The structures of these chimeras are shown in Fig. 2. The positions and amino acids of the junctions for the individual chimeras are as follows: chimera 1 (CH1), ₁ Met¹-Ala⁸⁴/₂ Lys⁶⁰-Leu⁴¹³; chimera 2 (CH2), ₂ Met¹-Phe⁷¹/₁ Ile⁹⁷-Cys¹³¹/₂ Glu¹⁰⁷-Leu⁴¹³; chimera 3 (CH3), ₂ Met¹-Val²⁹⁵/₁ Lys³⁴⁷-Pro³⁸¹/₂ Asp³³¹-Leu⁴¹³; chimera 4 (CH4), ₂ Met¹-Phe⁷¹/₁ Ile⁹⁷-Cys¹³¹/₂ Glu¹⁰⁷-Val²⁹⁵/₁ Lys³⁴⁷-Pro³⁸¹/₂ Asp³³¹-Leu⁴¹³; chimera 5 (CH5), ₂ Met¹-Ala⁵⁹/₁ Lys⁸⁵-Val⁴⁷⁷; chimera 6 (CH6), ₁ Met¹-Phe⁹⁶/₂ Ile⁷²-Cys¹⁰⁶/₁ Glu¹³²-Val⁴⁷⁷; chimera 7 (CH7), ₁ Met¹-Val³⁴⁶/₂ His²⁹⁶-Pro³³⁰/₁ Asp³⁸²-Val⁴⁷⁷; and chimera 8 (CH8), ₁ Met¹-Phe⁹⁶/₂ Ile⁷²-Cys¹⁰⁶/₁ Glu¹³²-Met¹-Val³⁴⁶/₂ His²⁹⁶-Pro³³⁰/₁ Asp³⁸²-Val⁴⁷⁷. Ala-substituted mutants of ₁-adrenergic receptor were constructed by PCR with QuickChange site-directed mutagenesis kit (Stratagene). After confirming the mutation, the PCR products were ligated with the rest of ₁-adrenergic receptor sequences to obtain full-length mutated ₁-adrenergic receptors. Ala-substituted mutants of constitutively active ₁-adrenergic receptor were constructed as follows. L323K-mutant of ₁-adrenergic receptor shows constitutive activity (Lattion et al., 1999). The fragment encoding L323K-mutation was digested from this constitutively active mutant and ligated with the rest of ₁-adrenergic receptor sequence, which encodes Ala-substituted sequences, to obtain full-length constitutively active ₁-adrenergic receptor mutants. These cDNAs were cloned into mammalian expression vector pCMV5 or pEF/myc/cyto.

View larger version (37K): [in this window] [in a new window]   Fig. 2.   Structures of chimeric receptors of 1- and 2-adrenergic receptors. The transmembrane domains of 1-adrenergic receptor (1AR) are shown by , and those of 2-adrenergic receptor (2AR) are represented by  The positions of junctions are described under Experimental Procedures.

Transient Expression of Chimeric and Mutant Receptors. COS-7 cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and 10 µg/ml gentamicin. For radioligand binding assay, COS-7 cells were seeded at 1.0 to 1.5 × 10⁶ cells/100-mm dish. The next day, the plasmid constructs were transfected by DEAE-dextran method as described (Isogaya et al., 1998, 1999; Kikkawa et al., 1998). Two days after the transfection, the cells were harvested for preparation of crude membrane fraction.

Radioligand Binding Assay. Two days after the transfection, COS-7 cells were rinsed three times with 10 ml of ice-cold phosphate-buffered saline and mechanically detached in 1 ml of ice-cold lysis buffer containing 10 mM Tris, pH 7.4, 5 mM EDTA, 5 mM EGTA, 1 µg/ml benzamidine, 10 µg/ml soybean trypsin inhibitor (type II-S), and 5 µg/ml leupeptin. The cell lysate was centrifuged at 45,000g for 10 min at 4°C. The pellet containing crude membrane fraction was resuspended in 1 ml of ice-cold lysis buffer with Potter type homogenizer, frozen, and stored at 80°C until use. The concentration of membrane protein was determined by the method of Lowry et al. (1951). Membrane protein (0.1-5 µg) was used for the binding studies. The membrane was incubated with 50 pM ¹²⁵I-cyanopindolol in 75 mM Tris, pH 7.4, 12.5 mM MgCl₂, 2 mM EDTA, and various concentrations of ()-RO363 in the presence of 0.1 mM GTP for 1 h at 37°C. Nonspecific binding was determined in the presence of 5 µM (±)-propranolol. The reaction mixture was filtered over Whatman GF/C filters (Whatman, Maidstone, UK). The filters were washed with ice-cold buffer containing 25 mM Tris, pH 7.4, and 1 mM MgCl₂. The bound ¹²⁵I-cyanopindolol on the filters was measured with a gamma counter.

Data Analysis. The results are expressed as mean ± standard error of n determinations. Equilibrium dissociation constants were determined from saturation isotherms. The competition curves for determination of IC₅₀ and K_i values were analyzed by Prism software (GraphPad Software, San Diego, CA). Statistical significance was evaluated by one-way analysis of variance for multiple comparisons. Analysis of variance post hoc comparisons were assessed with Dunnett's test.

Molecular Modeling. The model of the -carbon template based on the structure of rhodopsin was presented by Baldwin et al. (1997) and was used as described previously (Isogaya et al., 1999). ()-RO363-₁-adrenergic receptor complex was built by the following assumption: Asp¹³⁸ (corresponding to Asp¹¹³ of ₂-adrenergic receptor) with the protonated amine, Ser²²⁹ (Ser²⁰⁴ of ₂-adrenergic receptor) and Ser²³² (Ser²⁰⁷ of ₂-adrenergic receptor) with the catechol hydroxyl groups, and Phe³⁴¹ (Phe²⁹⁰ of ₂-adrenergic receptor) with the catechol phenyl ring. The procedures of docking and energy minimization were described previously (Tanimura et al., 1994).

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Affinities of ()-RO363 for ₁/₂ Chimeric Receptors. The receptor binding analysis with human recombinant ₁- and ₂-adrenergic receptors showed that ()-RO363 has 40-fold higher affinity for ₁-adrenergic receptor than ₂-adrenergic receptor (Table 1). To determine the domain responsible for this selectivity, we constructed eight chimeric receptors of ₁- and ₂-adrenergic receptors (Fig. 2). The binding characteristics of these chimeric receptors for ()-RO363 are summarized in Table 1. No significant differences were found when transmembrane domain 1 of ₁- or ₂-adrenergic receptors was replaced with the corresponding regions of ₂- or ₁-adrenergic receptors (CH5 or CH1). This result indicates that transmembrane domain 1 of ₁-adrenergic receptor does not participate in the subtype-selective binding of ()-RO363. The replacement of transmembrane domain 2 or transmembrane domain 7 of ₁-adrenergic receptor (CH6 or CH7) decreased the affinity of ()-RO363 for the resulting chimeras (33- or 5.5-fold), whereas the replacement of transmembrane domain 2 or transmembrane domain 7 of ₂-adrenergic receptor (CH2 or CH3) increased the affinity for these chimeras (21- or 3.6-fold). Moreover, the replacement of transmembrane domain 2 together with transmembrane domain 7 of ₁- or ₂-adrenergic receptors with the corresponding regions of ₂- or ₁-adrenergic receptors (CH8 or CH4) resulted in the chimera that showed essentially the same affinities as those of wild type ₂- or ₁-adrenergic receptors. The affinity of ()-RO363 for CH4 (₂-adrenergic receptor with transmembrane domains 2 and 7 of ₁-adrenergic receptor) was 58-fold higher than that for wild type ₂-adrenergic receptor and was essentially the same as that of wild-type ₁-adrenergic receptor. In contrast, the affinity of ()-RO363 for CH8 (₁-adrenergic receptor with transmembrane domains 2 and 7 of ₂-adrenergic receptor) was 55-fold lower than that for wild-type ₁-adrenergic receptor, and was almost equivalent to that of wild-type ₂-adrenergic receptor. These data strongly suggest that transmembrane domains 2 and 7 of ₁-adrenergic receptor are responsible for the ₁-selective binding of ()-RO363.

View this table: [in this window] [in a new window]   TABLE 1 Binding affinity of ()-RO363 for wild-type 1-, 2-adrenergic receptors, and chimeric receptors Binding of ()-RO363 to the receptors was assayed by competition with 50 pM 125I-cyanopindolol in the presence of 0.1 mM GTP. The data were analyzed by the nonlinear least-squares regression computer program Prism. The results are shown as mean ± standard error from three to five separate experiments.

Affinities of ()-RO363 for Ala-Substituted ₁-Adrenergic Receptor Mutants. To examine further the binding region of ()-RO363, eight amino acids in transmembrane domain 2 of ₁-adrenergic receptor that are different from those of ₂-adrenergic receptor were individually changed to Ala (Fig. 3). Among these eight mutants, L110A- and T117A-₁-adrenergic receptors showed the significant decrease in the binding affinities of ()-RO363 (6.5- and 2.5-fold, respectively), whereas the affinities of ¹²⁵I-cyanopindolol to these two mutants were unaltered (Table 2; Fig. 4). We also replaced 10 amino acids in transmembrane domain 7, which are different from ₂-adrenergic receptor, individually with Ala (Fig. 3). In this region, only replacement of Phe³⁵⁹ significantly decreased the affinity of ()-RO363 (4.1-fold; Table 3; Fig. 4). The substitution of amino acids other than Phe³⁵⁹ with Ala did not significantly change the affinity of ()-RO363. The substitution of Phe³⁶¹ in transmembrane domain 7 with Ala resulted in low expression of the receptor, and we could not analyze its binding characteristics (data not shown). These data indicated that Leu¹¹⁰, Thr¹¹⁷, and Phe³⁵⁹ contribute to the ₁-selective binding of ()-RO363.

View larger version (39K): [in this window] [in a new window]   Fig. 3.   Alignment of amino acid sequences of transmembrane domains 2 and 7 in 1- and 2-adrenergic receptors. The same amino acids between 1- and 2-adrenergic receptors (1AR and 2AR) are shown by . The eight amino acids in transmembrane domain 2 (TMD 2) and 10 amino acids in transmembrane domain 7 (TMD 7), which are different between two -adrenergic receptors, were individually replaced with Ala (shown by ).

View this table: [in this window] [in a new window]   TABLE 2 Effects of replacement of amino acids in transmembrane domain 2 of the 1-adrenergic receptor Binding of ()-RO363 to the receptors was assayed by competition with 50 pM 125I-cyanopindolol, in the presence of 0.1 mM GTP. The data were analyzed by the nonlinear least-squares regression computer program Prism. The results are shown as mean ± standard error from three to four separate experiments.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Competition binding curves for ()-RO363 in the wild-type and Ala-substituted 1-adrenergic receptors. Binding assays were performed as described under Experimental Procedures by using 125I-cyanopindolol as radioligand. , wild-type 1-adrenergic receptor; , L110A-1-adrenergic receptor; , T117A-1-adrenergic receptor; , F359A-1-adrenergic receptor; and , wild-type 2-adrenergic receptor.

View this table: [in this window] [in a new window]   TABLE 3 Effects of replacement of amino acids in transmembrane domain 7 of the 1-adrenergic receptor Binding of ()-RO363 to the receptors was assayed by competition with 50 pM 125I-cyanopindolol, in the presence of 0.1 mM GTP. The data were analyzed by the nonlinear least-squares regression computer program Prism. The results are shown as mean ± standard error from three to five separate experiments.

Effects of Multiple Ala-Substitution of Key Amino Acids on Affinity of ()-RO363 for ₁-Adrenergic Receptor. Although Leu¹¹⁰, Thr¹¹⁷, and Phe³⁵⁹ were suggested to participate in ₁-selective binding of ()-RO363, the decreases in the affinity of L110A-, T117A-, and F359A-₁-adrenergic receptors were relatively small compared with the difference of K_i values of ()-RO363 for wild-type ₁- and ₂-adrenergic receptors. To investigate the possibility that the key amino acids (Leu¹¹⁰, Thr¹¹⁷, and Phe³⁵⁹) form a binding pocket in a concerted manner, we changed two or three of these amino acids to Ala. L110A/T117A-, L110A/F359A-, and T117A/F359A-₁-adrenergic receptors showed 3.5-, 26-, and 21-fold lower affinity for ()-RO363 than wild-type ₁-adrenergic receptor, respectively (Table 4; Fig. 5). Furthermore, the affinity of ()-RO363 for the triple mutant L110A/T117A/F359A-₁-adrenergic receptor was essentially the same as that for wild-type ₂-adrenergic receptor (Table 4; Fig. 5). Except for L110A/T117A-₁-adrenergic receptor, the affinities for the double or triple mutants were almost multiplicatively decreased, indicating that these amino acids contribute to the free energy of binding in an additive manner. These results suggest that these three amino acids form a binding pocket in a cooperative manner, and that this binding pocket is a major determinant for ₁-selective, high-affinity binding of ()-RO363. To test this hypothesis, we introduced two of these key amino acids (Leu¹¹⁰ and Thr¹¹⁷) into the corresponding positions of ₂-adrenergic receptor (A85L/A92T-₂-adrenergic receptor). However, the level of expression of this mutant was too low to analyze (data not shown).

View this table: [in this window] [in a new window]   TABLE 4 Effects of multiple alanine-substitution of Leu110, Thr117, and Phe359 Binding of ()-RO363 to the receptors was assayed by competition with 50 pM 125I-cyanopindolol, in the presence of 0.1 mM GTP. The data were analyzed by the nonlinear least-squares regression computer program Prism. The results are shown as mean ± standard error from three to six separate experiments.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Competition binding curves for ()-RO363 in the wild-type and double or triple Ala-substituted 1-adrenergic receptors. Binding assays were performed as described under Experimental Procedures by using 125I-cyanopindolol as radioligand. , wild-type 1-adrenergic receptor; , L110A/T117A-1-adrenergic receptor; , L110A/F359A-1-adrenergic receptor; , T117A/F359A-1-adrenergic receptor; , L110A/T117A/F359A-1-adrenergic receptor; , wild-type 2-adrenergic receptor.

Role of Key Amino Acids in Active State of Receptor. As mentioned above, it was suggested that Leu¹¹⁰, Thr¹¹⁷, and Phe³⁵⁹ are key amino acids in ₁-selective binding of ()-RO363. To examine a role of these amino acids in active state of the receptor, we introduced the mutation into a constitutively active ₁-adrenergic receptor (L323K-₁-adrenergic receptor). As reported by Lattion et al. (1999), this mutation confers the constitutive activity on ₁-adrenergic receptor. ()-RO363 showed 3.6-fold higher affinity for L323K-₁-adrenergic receptor than for wild-type ₁-adrenergic receptor (Table 5), consistent with the previous observation that the increase in the affinity for agonists reflect the property of constitutively active receptor. The affinity of ()-RO363 for constitutively active ₁-adrenergic receptor was significantly decreased by mutation of Leu¹¹⁰, Thr¹¹⁷, and Phe³⁵⁹ (4.8-, 5.8-, and 12.6-fold, respectively). Among three mutants, only F359A constitutively active ₁-adrenergic receptor showed a larger decrease in the affinity of ()-RO363 than F359A-₁-adrenergic receptor (Fig. 6).

View this table: [in this window] [in a new window]   TABLE 5 Binding affinity of ()-RO363 for wild-type, constitutively active, and alanine-substituted/constitutively active 1-adrenergic receptors Binding of ()-RO363 to the wild-type, constitutively active, and alanine-substituted/constitutively active 1-adrenergic receptors was assayed by competition with 50 pM 125I-cyanopindolol, in the presence of 0.1 mM GTP. The data were analyzed by the nonlinear least-squares regression computer program Prism. The results are shown as mean ± standard error from three to four separate experiments.

View larger version (11K): [in this window] [in a new window]   Fig. 6.   Decrease in affinity by the replacement of Leu110, Thr117, and Phe359 in wild-type (WT) and constitutively active (CA) 1-adrenergic receptors. The ratios of Ki values of L110A-, T117A-, or F359A-1-adrenergic receptors to wild-type 1-adrenergic receptor () were compared with those of constitutively active 1-adrenergic receptors ().

Three-Dimensional Model of ()-RO363-₁-Adrenergic Receptor Complex. We constructed a three-dimensional model of ()-RO363-₁-adrenergic receptor complex in the resting state to explain the results of binding studies (Fig. 7). The structure of rhodopsin presented by Baldwin et al. (1997) was used as a model of the -carbon template. The amino acids of ₁-adrenergic receptor that interact with catecholamine (Asp¹³⁸ in transmembrane domain 3, Ser²²⁹ and Ser²³² in transmembrane domain 5, and Phe³⁴¹ in transmembrane domain 6) were assumed to interact with the catecholamine-like structure of ()-RO363, except for Asn³⁴⁴. Because Asn³⁴⁴ did not reach -hydroxyl group of ()-RO363, Asn³⁴⁴ may interact with ether oxygen instead of -hydroxyl group of ()-RO363². As previously described by Isogaya et al. (1999), the half of amino acids in transmembrane domains 2 and 7 toward the extracellular space forms binding pocket of ₂-selective agonists. As shown in Fig. 7, Phe³⁵⁹ and Leu¹¹⁰ are located at the top and the bottom of a binding pocket of ()-RO363, respectively. Phe³⁵⁹ seems to participate in ₁-selective, high-affinity binding of ()-RO363 through hydrophobic interaction with the phenyl group of N-substituent of ()-RO363. On the other hand, Thr¹¹⁷ is located near the methoxy group of N-substituent of ()-RO363 and seems to interact through hydrogen bond.

View larger version (36K): [in this window] [in a new window]   Fig. 7.   Three-dimensional model of the ()-RO363-1-adrenergic receptor complex. The model was built based on the predicted structure of rhodopsin. ()-RO363 interacts with 1-adrenergic receptor at the following sites: Asp138 in transmembrane domain 3 with the protonated amine, Ser229 and Ser232 in transmembrane domain 5 with the catechol hydroxyl groups, and Phe341 in transmembrane domain 6 with the catechol phenyl ring. Three amino acids (Leu110, Thr117, and Phe359) to be involved in 1-selectivity of ()-RO363 are shown in figure. ()-RO363-1-adrenergic receptor complex is viewed from top (A) or side (B).

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

In this study, we analyzed the site on ₁-adrenergic receptor conferring the ₁-selective binding. Based on analysis of the binding characteristics of several chimeric ₁/₂-adrenergic receptors, transmembrane domains 2 and 7 were found to be involved in the ₁-selective binding of ()-RO363. This result is consistent with our previous reports that transmembrane domains 2 and 7 of ₁- and ₂-adrenergic receptors form a binding pocket for the ₁- and ₂-selective agonists (Isogaya et al., 1998, 1999; Kikkawa et al., 1998). In the previous report, we examined the contribution of each of amino acids in transmembrane domain 2 of ₁-adrenergic receptor to subtype-selective binding, and found that Leu¹¹⁰, Thr¹¹⁷, and Val¹²⁰ of ₁-adrenergic receptor play an important role in the ₁-selective binding of T-0509 and denopamine, which are ₁-selective agonists with similar structure to ()-RO363 (Isogaya et al., 1999; Fig. 1). Because the effect of the replacement of Val¹²⁰ with Ala was relatively small compared with those of Leu¹¹⁰ and Thr¹¹⁷, it suggested that Leu¹¹⁰ and Thr¹¹⁷ mainly determine ₁-selective binding of T-0509 and denopamine. We also found in the present study that the replacement of Val¹²⁰ did not significantly decrease the affinity of ()-RO363 for ₁-adrenergic receptor. These results indicate that Leu¹¹⁰ and Thr¹¹⁷ in transmembrane domain 2 play a major role in the subtype-selective binding of the ₁-selective agonists containing dimethoxyphenyl group such as T-0509, denopamine, and ()-RO363.

We have previously reported that transmembrane domain 7 of ₂-adrenergic receptor played an essential role in ₂-selective binding irrespective of structure of ₂-selective agonists such as salmeterol, TA-2005, and so on. We demonstrated in the present study that transmembrane domain 7 of ₁-adrenergic receptor is important for the subtype-selective binding. It is interesting to note that the particular amino acid in transmembrane domain 7 located at specific position (Tyr³⁰⁸ of ₂-adrenergic receptor or Phe³⁵⁹ of ₁-adrenergic receptor) plays an essential role in determining subtype-selective binding irrespective of receptor subtypes. We suggested in the previous reports that Tyr³⁰⁸ prevented the N-substituent of ₂-selective agonists from freely moving into extracellular space like a cover. Phe³⁵⁹ of ₁-adrenergic receptor locates at the position corresponding to Tyr³⁰⁸ of ₂-adrenergic receptor. Therefore, Phe³⁵⁹ may work as a cover of the binding pocket formed by transmembrane domains 2 and 7 of ₁-adrenergic receptor like Tyr³⁰⁸ of ₂-adrenergic receptor.

The structures of ₁-selective agonists [denopamine, T-0509, and ()-RO363] used for the previous and present studies are similar to each other. They possess a dimethoxyphenyl group as N-substituent. One of ₁-selective agonists structurally different from these agonists is xamoterol. However, the affinity of xamoterol for ₁-adrenergic receptor is at most 10-fold higher than for ₂-adrenergic receptor as determined by binding experiment with recombinant ₁- and ₂-adrenergic receptors. Therefore, it is necessary to develop a structurally different ₁-selective agonist for demonstrating the importance of Leu¹¹⁰, Thr¹¹⁷, and Phe³⁵⁹ of transmembrane domains 2 and 7, and generalization of the present model. The present model of the binding pocket helps to design new ₁-selective agonists.

One of the chimeras (CH7), in which transmembrane domain 7 of ₁-adrenergic receptor was replaced with that of ₂-adrenergic receptor, did not decrease the affinity of ()-RO363. However, we considered the possibility that the hydrophobic nature of the substituted amino acid may compensate the interaction of the chimera with N-substituent of ()-RO363 because transmembrane domain 7 of ₂-adrenergic receptor plays an essential role in the ₂-selective binding. Therefore, we changed to Ala each of amino acids in transmembrane domain 7 of ₁-adrenergic receptor that is different from ₂-adrenergic receptor. Among them, F359A-₁-adrenergic receptor mutant only showed the significantly decreased affinity of ()-RO363 for ₁-adrenergic receptor. The position of Phe³⁵⁹ of ₁-adrenergic receptor is exactly the same as that of Tyr³⁰⁸ of ₂-adrenergic receptor, which determines the subtype-selective binding of the ₂-selective agonists (Isogaya et al., 1998, 1999; Kikkawa et al., 1998). The homology of amino acids between the upper halves of transmembrane domains 2 and 7 of ₁- and ₂-adrenergic receptors is 54% (Fig. 3). Therefore, it may be feasible to imagine that the region of ₁-adrenergic receptor formed by the amino acids different from ₂-adrenergic receptor such as Leu¹¹⁰, Thr¹¹⁷, and Phe³⁵⁹ determines ₁-selective binding. Although the ₂-selectivity is mainly determined by one amino acid in transmembrane domain 7 (Tyr³⁰⁸), ₁-selectivity for ()-RO363 does not seem to be determined by a single amino acid. To examine whether Leu¹¹⁰, Thr¹¹⁷, and Phe³⁵⁹ work for subtype-selective binding of ()-RO363 in a concerted manner, these three residues were replaced with Ala individually or in combination. L110A/F359A-, T117A/F359A-, and L110A/T117A/F359A-₁-adrenergic receptors showed lower affinity than the mutants with a single substitution. This result supports our conclusion that Leu¹¹⁰, Thr¹¹⁷, and Phe³⁵⁹ form the binding site in a concerted manner (Fig. 5; Table 4). However, the affinity of ()-RO363 for L110A/T117A-₁-adrenergic receptor was essentially the same as those of L110A- or T117A-₁-adrenergic receptor with unknown reason (Fig. 5; Table 4). Molecular modeling of the ()-RO363-₁-adrenergic receptor complex supports the notion that ₁-selective agonist binding site is formed by three amino acids Leu¹¹⁰, Thr¹¹⁷, and Phe³⁵⁹ (Fig. 7). The model shows that N-substituent of ()-RO363 can locate at the space surrounded by these three key amino acids. T-0509 and denopamine, which we used in the previous study, have the same structure (dimethoxyphenyl group) of N-substituent as ()-RO363. However, these two agonists show lower ₁-selectivity than ()-RO363. The model suggests that N-substituent of T-0509 and denopamine interacts with these amino acids, but the interaction is not enough to be tight to produce strong binding. The difference of the affinities between ()-RO363 and T-0509 and denopamine is explained by the fact that the ()-RO363 molecule is longer than T-0509 and denopamine by two atoms. Thus, the pocket formed by Leu¹¹⁰, Thr¹¹⁷, and Phe³⁵⁹ can contribute to the ₁-selective binding of ()-RO363 in a more efficient way.

The proximity between transmembrane domains 2 and 7 has been reported in several GPCRs such as gonadotropin-releasing hormone receptor (Zhou et al., 1994; Ballesteros et al., 1998), tachykinin neurokinin 2 receptor (Donnelly et al., 1999), serotonin 5-hydroxytryptamine_2A receptor (Perlman et al., 1997), and thyrotropin-releasing hormone receptor (Sealfon et al., 1995). Two amino acids (Asp in the middle of transmembrane domain 2, and Asn in the middle of transmembrane domain 7) are highly conserved among GPCR superfamily (Gether and Kobilka, 1998). Interaction of these two amino acids seems to be a driving force to keep transmembrane domains 2 and 7 in a close position. Furthermore, Asn in the middle of transmembrane domain 1, and Asp-Arg-Tyr motif (DRY motif) in the cytoplasmic end of transmembrane domain 3 are also highly conserved (Ballesteros et al., 1998; Gether and Kobilka, 1998). The DRY motif plays an important role in activation of GPCRs, through the interaction with G protein and catalysis of GDP-GTP exchange of G protein -subunit (Acharya and Karnik, 1996). Moreover, highly conserved Asn in transmembrane domain 7 of GPCRs forms NPXXY motif, which is involved in activation and internalization of GPCRs (Abdulaev and Ringe, 1998; Konvincka et al., 1998). Therefore, it is likely that these polar residues form a network responsible for activation of GPCRs induced by agonists. Actually, in _1B-adrenergic receptor (Scheer et al., 1996, 1997) and ₂-adrenergic receptor (Rasmussen et al., 1999), mutations that disrupt the interactions between some of these highly conserved polar residues were found to evoke the agonist-independent activation of these receptors. Furthermore, in rhodopsin (Farrens et al., 1996; Sheikh et al., 1996) and ₂-adrenergic receptor (Gether et al., 1997), it was reported that changes around transmembrane domains 3 and 6 occurred after receptor activation. These conformational changes seem to reflect the movement of transmembrane domains induced by agonist binding. However, it is remains to be determined how agonists induce these conformational changes at molecular level. N-substituent of ()-RO363 seems to fit a binding pocket formed by transmembrane domains 2 and 7 of ₁-adrenergic receptor. If this binding causes disruption of hydrophobic and hydrophilic interactions of these conserved polar amino acids in several transmembrane domains, it is possible that the ₁-selective binding is dependent on the conformation induced by agonist. To examine this possibility, we replaced Leu¹¹⁰, Thr¹¹⁷, or Phe³⁵⁹, which is responsible for the ₁-selectivity of ()-RO363, of a constitutively active mutant of ₁-adrenergic receptor with Ala. If a conformational change occurs around these amino acids, the degree of decrease in the affinity caused by Ala-substitution will be changed by alteration of distance between these amino acids and ()-RO363. Although the replacement of Leu¹¹⁰ or Thr¹¹⁷ of constitutively active ₁-adrenergic receptor decreased the K_i value, the degree of reduction was essentially the same as that of wild-type ₁-adrenergic receptor (Table 5; Fig. 7). Therefore, Leu¹¹⁰ and Thr¹¹⁷ seem to contribute to the initial stage of ()-RO363 binding. However, Ala-substitution of Phe³⁵⁹ of constitutively active ₁-adrenergic receptor caused a larger decrease in the affinity of ()-RO363 than that of wild-type ₁-adrenergic receptor (Table 5; Fig. 7). This result suggests that Phe³⁵⁹ in an active state locates in a closer position than in the resting state.

In conclusion, in ₁AR, Leu¹¹⁰ and Thr¹¹⁷ in transmembrane domain 2, and Phe³⁵⁹ in transmembrane domain 7 are major determinants of ₁-selectivity of ()-RO363. By combining the present result with the previous observations, we propose that Leu¹¹⁰, Thr¹¹⁷, and Phe³⁵⁹ of ₁-adrenergic receptor form in a concerted manner the subtype-selective binding pocket of ₁-selective agonists possessing dimethoxyphenyl group as N-substituent.