Characteristics for a Salt-bridge Switch Mutation of the α1bAdrenergic Receptor

Agonist-dependent activation of the α1-adrenergic receptor is postulated to be initiated by disruption of an interhelical salt-bridge constraint between an aspartic acid (Asp-125) and a lysine residue (Lys-331) in transmembrane domains three and seven, respectively. Single point mutations that disrupt the charges of either of these residues results in constitutive activity. To validate this hypothesis, we used site-directed mutagenesis to switch the position of these amino acids to observe, if possible, regeneration of the salt-bridge reverses that the constitutive activity of the single point mutations. The transiently expressed switch mutant receptor displayed an altered pharmacological profile. The affinity of selective α1b-adrenergic receptor antagonists for the switch mutant (D125K/K331D) was no different from the wild-type α1b-adrenergic receptor, suggesting that both receptors are maintaining similar tertiary structures in the cell membrane. However, there was a significant 4–6-fold decrease in the affinity of protonated amine receptor agonists and a 3–6-fold increase in the affinity of carboxylated catechol derivatives for the switch mutant compared with the wild-type α1b-adrenergic receptor. This pharmacology is consistent with a reversed charge at position 125 in transmembrane domain three. Interestingly, the ability of either a negatively or positively charged agonist to generate soluble inositol phosphates was similar for both types of receptors. Finally, the switch mutant (D125K/K331D) displayed similar basal signaling activity as the wild-type receptor, reversing the constitutive activity of the single point mutations (D125K and K331D). This suggests an ionic constraint has been reformed in the switch mutant analogous to the restraint previously described for the wild-type α1b-adrenergic receptor. These results strongly establish the disruption of an electrostatic interaction as an initial step in the agonist-dependent activation of α1-adrenergic receptors.

* This work was done under the tenureship of an Established Investigator Award from the National American Heart Association (to D. M. P.). This work was also supported in part by National Institutes of Health Grants RO1HL52544 and RO1HL61438, an unrestricted research grant from Glaxo Wellcome (to D. M. P.), and an American Heart Association Postdoctoral Fellowship, Northeast Ohio Affiliate (to J. E. P.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and the transient transfection was performed as described previously using the DEAE-dextran method (3).
Inositol Phosphate Hydrolysis-Total soluble inositol phosphate (IP) production was determined from COS-1 cells that were pre-labeled with [ 3 H]inositol (1 Ci/ml) for 16 -24 h prior to the assay. On the assay day, these cultured cells were prepared as described previously (3). In separate and more sensitive studies, basal measurements for phospholipase C-dependent inositol 1,4,5-trisphosphate (IP 3 ) production were determined by a [ 3 H]IP 3 radioreceptor assay (DuPont) using the same pre-assay condition as described previously (3). Receptor density was determined on separate but parallel plates from the same transfection experiment.
Statistical Analysis-For each individual experiment, the fitted iterative non-linear regression curve that best represented the data was determined using a partial f-test (p Ͻ 0.05). Significance between groups was tested using an unpaired two-tailed Student's t test (p Ͻ 0.05). All values are reported as the mean Ϯ S.E. of N experiments, each performed in duplicate.

RESULTS AND DISCUSSION
Recent studies have described a salt bridge constraint mechanism of activation for the rhodopsin receptor system involving amino acids Glu-113 and Lys-296 in TMDs three and seven of the receptor (2). Constitutively active rhodopsin receptors generated by site-directed mutagenic disruption of the electrostatic interaction between these two amino acids were used to support this activation model. In further experiments, these amino acids were switched in a double mutation (E113K/ K296E) of the rhodopsin receptor to substantiate this ionic constraint hypothesis. Results of these experiments were inconclusive. Regeneration of the ionic pair in the switch mutant reversed the constitutive activity of the single mutant K296E. However, functionality of the double and E113K mutations could not be verified because they could not reconstitute with the chromophore. This is because of the strict structural requirements for 11-cis retinal to covalently attach with Lys-296 suggesting that these mutants were misfolded or just incapable of attaching retinal (2). To validate our ␣ 1b -AR salt-bridge constraint hypothesis, we used sight-directed mutagenesis to generate a double mutation of the wild type (WT) receptor. This mutant ␣ 1b -D125K/K331D-AR switches the amino acids that constitute the previously described receptor salt-bridge constraint (3). However, unlike the rhodopsin receptor system, there is no requirement for the adrenergic ligand to become covalently bound to form a functional ␣ 1 -AR.
This switch ␣ 1b -AR mutation could be transiently expressed on ␣ 1 -AR negative COS-1 cell membranes at a level of 22 Ϯ 4 fmol/mg of protein (data not shown). In addition, there were no differences in binding affinity of the competitive receptor antagonists 125 I-HEAT, prazosin, or phentolamine for the switch mutation (110 Ϯ 31 pM, 299 Ϯ 99 pM, and 29 Ϯ 20 nM, respectively; Table I) when compared with the WT ␣ 1b -AR (83 Ϯ 10 pM, 243 Ϯ 30 pM, and 36 Ϯ 10 nM; Table I). A negative charge at position 125 on TMD three is thought to be important for agonist as well as antagonist affinity at ␤and ␣ 2 -ARs (9). Competitive ␣ 1 -AR antagonists are larger molecules when compared with epinephrine, although they may have common bind-ing contacts with receptor agonists in the ligand binding pocket. Therefore, the Asp-125 binding contact may contribute less to the overall binding affinity of competitive ␣ 1 -AR antagonists than for agonists. Alternately, Asp-125 may not participate in antagonist binding to the ␣ 1 -ARs. For example, an ␣ 1b -D125A-AR mutant does not have a significantly different antagonist affinity value when compared with the WT receptor (3). Previous studies have also characterized amino acid residues on the extracellular loop as ␣ 1 -AR binding contacts for competitive receptor antagonists (10). Molecular modeling suggests that competitive receptor antagonists may contact the ␣ 1 -AR above or near the surface of the agonist binding pocket involving amino acid residues near the top of the TMD. Nevertheless, similarities in antagonist binding between WT and the switch mutant suggest that the overall tertiary structure of these two receptors is comparable.
In contrast, AR agonists such as epinephrine, phenylephrine, and oxymetazoline, which all contain a protonated amine (  Table I). Conversely, carboxylated catechol derivatives such as 3-(2-hydroxyphenyl) proprionic acid and 3-methoxyphenyl acetic acid ( Fig. 1), which replace the protonated amine with an acid group, displayed a significant (p Ͻ 0.05) 3-6-fold higher binding affinity for the switch receptor mutation (0.5 Ϯ 0.4 mM and 1.9 Ϯ 0.4 mM, respectively; Table I) when compared with WT ␣ 1b -AR (2.9 Ϯ 0.9 mM and 5.7 Ϯ 1.4 mM; Table I). Although the carboxylated derivatives display poor overall binding affinity, they are not the exact cogners of epinephrine. Both derivatives have altered distances of the charged head group to the aromatic ring, a crucial feature of agonist binding affinity. However, regardless of the low affinity, the switched binding profile of the double mutant is consistent with a positively charged amino acid now substituted at position 125 in TMD three of the receptor. The degree of decreased affinity of protonated amine agonists (4 -6-fold) for the switch mutant is consistent with agonist affinity losses (3-17-fold) for the D125A ␣ 1b -AR mutation (3).
We also examined the ability of these agonists to activate transiently expressed WT or switch mutant ␣ 1b -ARs. Demonstrating that the switch mutation is not signaling defective, both WT and switch mutant displayed a similar ability to invoke release of soluble IPs in a concentration-response to phenylephrine (Fig. 2). While the maximum IP response of the switch mutant was not significantly different from the WT ␣ 1b -AR, the EC 50 of phenylephrine for the switch mutation was  This lowered potency is consistent with the 6-fold decrease in affinity of phenylephrine for the switch mutation when compared with the WT ␣ 1b -AR. Low expression levels and the poor affinity of the carboxylated catechol derivatives for the switch ␣ 1b -AR mutation did not allow for the accurate generation of concentration-response curves in the presence of these ligands. However, when expression levels of WT ␣ 1b -AR were titrated to equivalent receptor densities as the switch mutation, a single ligand concentration that occupied at least 50% or greater of these transiently expressed receptors was used to increase soluble IPs in COS-1 cells (Fig. 3). Interestingly, either charged agonist was able to significantly increase soluble IPs over basal. Furthermore, the extent of activation was comparable for both the WT and switch mutant receptor. These responses are specific because a blocking concentration of the selective ␣ 1 -AR antagonist prazosin was able to inhibit the increase of soluble IPs for all ligands used in this study (Fig. 3). Although the binding affinity differences are no greater than 6-fold between the WT and switch ␣ 1b -AR mutation for these ligands, their ability to bind and activate both receptors suggests a plasticity of the agonist binding pocket. In consideration of a salt-bridge activation mechanism, these results are consistent in that either type of charge introduced by the ligand should break an ionic interaction in the receptor agonist binding pocket.
We also investigated the basal signaling properties for the switch mutation, the WT ␣ 1b -AR and the individual mutations (D125K or K331D) that make up the double mutant. We have previously described the constitutive properties of the ␣ 1b -D125K-AR and an analogous K331E mutation (3). The amount of agonist-independent production of IP 3 was significantly greater (p Ͻ 0.05) for the D125K mutant (52.2 Ϯ 3.0 pmol/fmol of receptor; Fig. 4) when compared with the WT ␣ 1b -AR (18.8 Ϯ 1.8 pmol/fmol of receptor; Fig. 4). A similar 3-fold increase in basal IP 3 production over the WT receptor was calculated for the newly characterized ␣ 1b -K331D-AR (55.2 Ϯ 3.6 pmol/fmol; Fig. 4). Finally, when both single mutations were generated in the same receptor (switch mutation), the amount of agonistindependent IP 3 produced (19.3 Ϯ 2.1 pmol/fmol of receptor) is not significantly different (p Ͼ 0.05) from the WT ␣ 1b -AR (Fig.  4). Thus, the switch mutant reversed the constitutive activity of the single mutations. This data strongly suggests an interhelical electrostatic interaction has reformed between Lys-125 and Asp-331 that holds the switch mutant in a basal conformation similar to what has been previously described for the WT ␣ 1b -AR (3).
This manuscript characterizes a switch mutation of the ␣ 1b -AR in which the amino acids of a postulated ionic constraint for the WT receptor have been reversed. This switch mutant receptor has comparable characteristics to the WT ␣ 1b -AR (i.e. affinity for receptor antagonists, basal production of IP 3 , and agonist-induced activation), suggesting that the overall tertiary structure and function of these receptors in the membrane bilayer is similar. However, there are distinct differences in the pharmacological properties of the agonist binding pocket between the switch mutant and the WT receptor. The density of expression was 59 Ϯ 7 fmol/mg membrane protein for the WT, 17 Ϯ 6 fmol/mg for D125K, 45 Ϯ 7 fmol/mg for K331D, and 21 Ϯ 4 fmol/mg for the D125K/K331D ␣ 1b -ARs. Data is presented as the mean Ϯ S.E. for n ϭ 3 transfections of ten replicates each.
Interestingly, both WT and the switch receptor mutation are bound to and activated by either catecholamines or carboxylated catechol derivatives, suggesting some flexibility in the docking of agonists in ␣ 1 -AR binding pocket. This is in contrast to the more stringent structural requirements of the opsin receptor ligand binding pocket for reacting with 11-cis-retinal to generate rhodopsin although alterations in the chromophores length and ability to form a Schiff's base has been shown to generate viable rhodopsin receptors with limited ability to activate transducin (11,12). Furthermore, the ability to activate the switch mutant or WT ␣ 1b -AR should be dependent on the effectiveness of a ligand to compete for the appropriate counterion that forms the ionic constraint of the receptor. Therefore, it is not unexpected to observe activation of the WT or switch mutant ␣ 1b -AR by phenylephrine or carboxylated catechol ligands, even though the orientation of the charged ligand to the salt-bridge constraint on the receptor may not be optimal in some combinations. This is evident in the reduced potency of phenylephrine to generate soluble IPs mediated by the switch ␣ 1b -AR mutation. If we were able to accurately generate a concentration-response curve for the carboxylated catechol derivatives, predicted changes in potency that correlated to binding affinity differences would likely have become apparent.
Although Asp-125 is a highly conserved amino acid in biogenic amine receptors, evidence for the conservation of this mechanism to other family members is still unresolved. Previous mutation of Asp-113 in the ␤ 2 -AR resulted in a dramatic loss of agonist and antagonist affinity (13). The receptor was still capable of signaling although at reduced potency. The constitutive behavior of this mutation was not explored at that time because of the field's lack of knowledge for active state receptors. Analysis is also complicated because of the poor expression of these mutations and the endogenous expression of ␤-ARs in most cell types. Mutations of Asp-113 in the ␣ 2a -AR also resulted in a similar phenotype to the ␤ 2 -AR (14). However, recent ␤ 2 -AR work in which a histidine replaced Asp-113 in TMD three and a cysteine replaced Asn-312 in TMD seven, followed by chelation with zinc, resulted in a constitutive phenotype (15). The altered distance between these two helices introduced by the zinc is thought to be responsible for the active state stabilization of this mutant ␤ 2 -AR. This suggests that these two helices and/or residues are linked in the WT ␤ 2 -AR activation process. Analogous amino acids in the ␦ opioid receptor have also generated constitutively active receptors with mutations at Asp-128 in TMD three and Tyr-308 in TMD seven (16). It has been proposed that these two residues participate in hydrogen bonding forming a constraint similar to our salt-bridge hypothesis for the ␣ 1b -AR. This suggests that similar constraining paradigms may be part of a universal activation mechanism among GPCRs.
To summarize, disruption of this ionic constraint has been suggested to be an initial but partial event that is required for full activation of the WT ␣ 1 -AR (4). In addition, pH-dependent binding studies resulting in pK a shifts for the WT and Lys-331 mutations have implied that Asp-125 and Lys-331 are in close proximity to each other in the unbound WT ␣ 1b -AR. 2 Finally, reversing the constitutive properties of D125K and K331D by combining these mutations in a single receptor and activation of this switch mutant by either type of charged agonist strongly supports our original hypothesis for a salt-bridge constraint holding the WT ␣ 1b -AR in a basal conformation (3).