A single mutation in the 5-HT4 receptor (5-HT4-R D100(3.32)A) generates a Gs-coupled receptor activated exclusively by synthetic ligands (RASSL).

To better understand G-protein-coupled receptor (GPCRs) signaling, cellular and animal physiology, as well as gene therapy, a new tool has recently been proposed. It consists of GPCR mutants that are insensitive to endogenous ligands but sensitive to synthetic ligands. These GPCRs are called receptor activated solely by synthetic ligands (RASSL). Only two examples of such engineered receptors have been described so far: one G(i)-coupled (opioid receptors) and one G(s)-coupled (beta(2)-adrenergic receptors). Here, we describe the first RASSL related to serotonin receptors (D100(3.32)A G(s)-coupled 5-HT(4) receptor or 5-HT(4)-RASSL). 5-HT(4)-RASSL is generated by a single mutation, is totally insensitive to serotonin (5-HT), and still responds to synthetic ligands. These ligands have affinities in the range of nanomolar concentrations for the mutant receptor and exhibit full efficacy. More interestingly, two synthetic ligands behave as antagonists on the wild type but as agonists on the 5-HT(4)-RASSL.

changes lead either to a loss or a gain of function, such as blindness, diabetes insipidus, and hypo-or hyperthyroidism (3). Depressed GPCR signal transduction is related to numerous complex diseases. Heart failure and asthma are associated with a decrease in the G s -signaling pathway, whereas an increase in G i signaling is a potential cause of dilated cardiomyopathy).
The concept that modified GPCRs could be used as tools to better understand GPCR-controlled signal transduction pathways in a given cell or organ or for gene therapy has recently been proposed (5)(6)(7). Several types of modified GPCRs have been developed to prepare such tools. The first, engineered by Conklin and collaborators (8), were named RASSL for "receptor activated solely by synthetic ligands." The idea is to engineer receptors that would be insensitive to their endogenous ligand(s) but can be fully activated by synthetic ligands. Among the GPCRs, only the opioid receptors have been modified so far by Conklin's group (8) to produce two opioid receptor-RASSLs, Ro1 and Ro2. Ro1 was constructed by substituting the second extracellular loop of the opioid receptor with the corresponding portion of the human ␦ opioid receptor. This substitution induced a lower affinity for the endogenous peptide ligands (including dynorphin) without significantly reducing the response to synthetic ligands, like spiradoline (8). The specificity of Ro1 for the synthetic ligands was further enhanced in Ro2 by substituting glutamine for Glu 297 in Ro1.
More recently, a second type of mutant GPCR the "therapeutic receptor-effector complex" (or "TREC") was proposed to be a biotechnological tool to study GPCR signal transduction (9). The ␤ 2 -adrenergic receptor mutated in 19 positions and fused with G␣ s was not activated by ␤-adrenergic agonists, but only by a non-biogenic amine agonist, L156870, although with relatively low potency.
In this report, we describe a third example of RASSL type receptors: a mutant serotonin receptor, the 5-HT 4 receptor. 5-HT 4 receptors are G s -coupled receptors expressed in the gastrointestinal tract, human and pig atria, urinary bladder, adrenal medulla, and central nervous system including limbic areas (olfactory tubercles, limbic system, basal ganglia) (10). Over the past 10 years, the pharmacology and structure of 5-HT 4 receptors have been extensively studied (10 -12). One of the interesting characteristics of this receptor is its ability to be activated by a wide range of compounds from very different chemical classes (13). Some are used in clinic to treat gastroparesis, dyspepsia gastro-esophageal reflux, or irritable bowel syndromes. The 5-HT 4 agonists are related to tryptamines like 5-HT, to carbazimidamides (HTF-919 or Zelmac), benzamides (metoclopramide or Primperan TM , cisapride or "Prepulsid TM ," benzoates (SL 10302), benzimidazolones (BIMU8), or aryl ketones (for reviews, see Refs. 10 and 13).
The demonstration that tryptamines and benzamides have different pharmacophores (14) prompted us to generate mutant 5-HT 4 receptors with the aim of disrupting the 5-HT recognition site, keeping the recognition site for synthetic 5-HT 4 agonists. 4(a) Receptor cDNA-The mutant was generated by exchanging the endogenous residue Asp 100 to Ala in the m5-HT 4(a) R cDNA sequence with the QuikChange site-directed mutagenesis kit (Stratagene). The sense primer used was: D100A, 5Ј-ACC TCT CTG GCT GTC CTA CTC ACC-3Ј.

Construction of Mutant m5-HT
Cell Culture and Transfection-The cDNAs, subcloned into the pRK5 * 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.
Determination of Cyclic AMP (cAMP) Production in Intact Cells-Six hours after transfection, the surrounding cell medium was exchanged for DMEM without dFBS with 2 Ci of [ 3 H]adenine/ml to label the ATP pool and incubated overnight (16 h). cAMP accumulation was measured, as described previously (15).
Membrane Preparation and Radioligand Binding Assay-Membranes were prepared from transiently transfected cells plated on 15-cm dishes and grown in DMEM with 10% dFBS for 6 h, followed by incubation for 20 h in DMEM without dFBS. The cells were washed twice in PBS, scraped with a rubber policeman, harvested in PBS, and centrifuged at 4°C (200 ϫ g for 4 min). The pellet was resuspended in buffer containing 10 mM HEPES, pH 7.4, 5 mM EGTA, 1 mM EDTA, and 0.32 M sucrose and homogenized 10 times with a glass-Teflon potter at 4°C. The homogenate was centrifuged at 20,000 ϫ g for 20 min. The membrane pellet was resuspended in 50 mM HEPES, pH 7.4 (5 mg of protein in 1 ml of solution) and stored at Ϫ80°C until use. Saturation experiments were performed using the specific 5-HT 4 receptor radioligand [ 3 H]GR 113808 at height concentrations ranging from 0.048 to 0.51 nM. The 5-HT 4 receptor binding site density was estimated with [ 3 H]GR 113808 at a saturating concentration (0.5 nM), as described previously (16). 5-HT (5 ϫ 10 Ϫ5 M) or RS 100235 (10 Ϫ7 M), a 5-HT 4 receptor antagonist, was used to determine nonspecific binding. Protein concentration in the samples was determined with the Bio-Rad protein assay.
Data Analysis-Competition and saturation experiments were analyzed by non-linear regression curves using the computer program LIGAND (17). Saturation experiments were also analyzed according to Scatchard. IC 50 values required to displace 50% of [ 3 H]GR 113808 binding sites were converted to K D values, according to the equation K D ϭ IC 50 /1 ϩ S/K DS (18), where S is [ 3 H]GR 113808 concentration, and K DS is the equilibrium constant of [ 3 H]GR 113808.
Using Kaleidagraph software, the dose-response curves were fitted according to the following equation.
Y ϭ ͑͑y max Ϫ y min ͒/1 ϩ ͑x/EC 50 ) nH )) ϩ y min (Eq. 1) where EC 50 (or EC 50inv ) is the concentration of agonist (or inverse agonist) that evokes a half-maximal response, y max and y min correspond to the maximal and minimal responses, respectively, and nH is the Hill coefficient. Data were compared using the Stat-View Student program (Abacus Concepts, Berkeley, CA) with t tests.
Drugs-The following drugs were used:

RESULTS AND DISCUSSION
First, we wanted to render the receptor unresponsive to 5-HT, but responsive to synthetic 5-HT 4 agonists. Based on Strader's work on ␤-adrenergic receptors and on many other biogenic amine receptors (19), the key interaction between 5-HT 4 receptors and 5-HT should obviously occur at the highly conserved Asp residue found in TM-III within the biogenic amine receptor family. In the 5-HT 4 receptor this Asp is at position 100 in TM-III (Asp 100 (3.32)). This Asp residue is believed to interact with the positively charged nitrogen of the neurotransmitter protonated amine.
Indeed, the mutation of Asp 100 (3.32) to alanine in the 5-HT 4 receptor totally abolished the 5-HT stimulation of cAMP accumulation (11,12), as well as the stimulation by other tryptamines (5-CT or 5-MeOT) or by a substituted indole carbazimidamide (HTF-919) (Fig. 1, A and B) (Table I). Compared with 5-HT, HTF-919 has a guanidine function instead of a protonated amine in the indole side chain (20). Surprisingly, when we mutated other well conserved residues within the biogenic amine receptor family and likely to be involved in the 5-HT binding pocket (Ser 197 (5.43), Trp 272 (6.48), and Phe 275 (6.51)), none were able to totally suppress the 5-HT or HTF-919 responses. 5-HT has a 50 -500-fold reduced affinity
Interestingly, and in contrast to other biogenic amine receptors (9), the mutation of the conserved Asp 100 (3.32) only moderately affected the specific 5-HT 4 R radioligand, [ 3 H]GR 113808 (K D ϭ 0.17 Ϯ 0.08 nM and 0.38 Ϯ 0.09 nM at WT and D100(3.32)A receptors, respectively) (Fig. 1D). Using [ 3 H]GR 113808 binding, we verified that 5-HT was unable to bind the 5-HT 4 D100(3.32)A receptor (Fig 1C). The binding of [ 3 H]GR 113808 was the first indication that ligands, which include the basic nitrogen of the aromatic ring side chain, in a structured ring, associated with an increase in the distance between this basic nitrogen and the main aromatic ring of the compound, suppressed the requirement of the Asp 100 (3.32) carboxylic group for ligand binding. This was confirmed when we screened most of the non-tryptamine 5-HT 4 receptor agonists for their ability to activate the 5-HT 4 D100(3.32)A receptor.
BIMU8, a specific 5-HT 4 agonist structurally very different   4 receptor Efficiency was determined in cAMP accumulation experiments in COS-7 cells transiently expressing either 5-HT 4(a) and mutant D100(3.32)A receptors at the same level of receptor density (about 2500 fmol/mg proteins). The results are expressed as percentages of 5-HTinduced cAMP accumulation via the 5-HT 4 WT receptors and BIMU8induced cAMP accumulation via the 5-HT 4 -R D100(3.32)A mutant receptors. Data are the means Ϯ S.E. values of three experiments performed in triplicate (see "Experimental Procedures"). The chemical structures of some of the 5-HT 4 receptor drugs tested in this study are shown. 4 Receptors for Gene Therapy 701 from 5-HT, belonging to the azabicycloalyl benzymidazolone class (21), was found to be a potent agonist. As shown in Fig. 2, this compound remained fully active and even showed higher efficacy and potency on the D100A mutant than on the WT (EC 50 values for cAMP stimulation were 4 Ϯ 1.5 and 1 Ϯ 0.5 nM for WT and D100A, respectively) ( Fig. 2A). The affinity of BIMU8 for the 5-HT 4 receptor was also slightly better on the D100A mutant than on the WT. K D values for BIMU8, measured by competition with the [ 3 H]GR 113808 radioligand, were 30 Ϯ 11 nM and 6.5 Ϯ 3 nM for WT and D100A, respectively (Fig. 2B). Similarly, the benzamides bearing the 2-methoxy-4-amino-5-chloro substitution (renzapride, S-zacopride, or cisapride) were equi-effective on the WT and the D100(3.32)A mutant (Table I) with nanomolar affinity. A slight decrease in their potency on the mutant D100A receptor was observed. These data suggest that the D100(3.32)A mutant could still be activated as long as the agonist could bind the receptor. Furthermore, this mutation had no effect on the receptor expression level. Two drugs were antagonists on WT and agonists at the D100(3.32)A mutant.
The structure-activity relationships and the structural analyses of the 5-HT 4 receptor ligands used in this study are consistent with a recent report on comparative receptor mapping of 5-HT 4 binding sites (26). The authors proposed structural insights to assist the design of selective 5-HT 4 receptor ligands. The structural features that define the 5-HT 4 ligands are an aromatic moiety, a coplanar carbonyl, carboxyl or ketone function, and a basic nitrogen atom. The substitute of the basic nitrogen must be voluminous (as in all the active drugs acting on the 5-HT 4 RASSL) to interact in an hydrophobic pocket of the receptor. The size of the substitute improves selectivity and potency. One of our hypotheses is that the basic nitrogen of these drugs interacts within the hydrophobic pocket, with a negative charge, but not with Asp(3.32). In contrast, this Asp is necessary for the binding of the protonated primary amine of 5-HT derivatives. The affinity of the voluminous substituted ligands for the 5-HT 4 RASSL (structures in Table I) would depend of the interactions between their hydrophobic rings and the hydrophobic pocket, defined by highly conserved aromatic residues in TMVI and VII (Trp(6.48), Phe(6.51), and Tyr(7.43)) (11,27,28).
The above results indicate that the D100(3.32)A G s -coupled 5-HT 4 receptor described here is the first RASSL related to serotonin receptors and that it has unique properties not found in two previously described RASSL receptors (5, 7).
RASSL is generated by a single mutation and is completely insensitive to its endogenous agonist: 5-HT. RASSL can be stimulated by numerous synthetic agonists from different chemical classes.
Many of these synthetic ligands have a nanomolar affinity for the 5-HT 4 RASSL, and some of them can be administered orally such as cisapride or metoclopramide. Cisapride has been used for years to treat dyspepsia and gastro-esophageal reflux and has only been removed from the market because of one second effect (arrhythmia probably caused by action on K ϩ channels) (29,30). However, other benzamides are being developed and will certainly be used in humans. Interestingly, RS 67333 was found to be another potent agonist of 5-HT 4 RASSL. This compound possesses a great ability to cross the blood-brain barrier.
Antagonist compounds (ML 10375 and GR 113808), which are highly specific on the native receptor, exhibit agonist properties on the 5-HT 4 RASSL. GR 113808 is also able to cross the blood-brain barrier. ML 10375 and GR 113808 will be particularly interesting for in vivo stimulation of 5-HT 4 RASSL. Indeed, with these compounds, native 5-HT4 receptors can be kept silent, while generating a G s signal by activating 5-HT 4 RASSL, expressed in a given tissue at a given developmental time, by bioengineering in animals and possibly in humans.