Chimeric Receptor Analysis of the Ketanserin Binding Site in the Human 5-Hydroxytryptamine 1 D Receptor : Importance of the Second Extracellular Loop and Fifth Transmembrane Domain in Antagonist Binding

The 5-hydroxytryptamine (5-HT)1B/1D receptor subtypes are involved in the regulation of 5-HT release and have gained particular interest because of their apparent role in migraine. Although selective antagonists for both receptor subtypes recently have been developed, the receptor domains involved in the pharmacological specificity of these antagonists are defined poorly. This was investigated with a chimeric 5-HT1B/1D receptor analysis and using ketanserin as a selective antagonist of h5-HT1D (h5-HT1D) Ki 5 24–27 nM) as opposed to h5-HT1B (Ki 5 2193–2902 nM) receptors. A domain of the h5-HT1D receptor encompassing the second extracellular loop and the fifth transmembrane domain is necessary and sufficient to promote higher affinity binding (Ki 5 65–115 nM) for ketanserin to the h5-HT1B receptor. The same domain of the h5-HT1B receptor, when exchanged in the h5-HT1D receptor, abolished high affinity binding of ketanserin (Ki 5 364-1265 nM). A similar observation was made with the antagonist ritanserin and seems specific because besides the unmodified binding affinities for 5-HT and zolmitriptan, only minor modifications (2–4-fold) were observed for the agonists L 694247 and sumatriptan and the antagonists GR 127935 and SB 224289. Generating point mutations of divergent amino acids compared with the h5-HT1B receptor did not demonstrate a smaller peptide region related to a significant modification of ketanserin binding. The antagonists ketanserin and ritanserin are likely to bind the h5-HT1D receptor by its second extracellular loop, near the exofacial surface of the fifth transmembrane domain, or both. Dysregulation of serotonin receptor function may contribute to various peripheral and central nervous system disorders (Glennon and Westkaemper, 1993). The 5-HT1B/1D receptor subtypes are, among other 5-HT receptors, involved in the regulation of 5-HT release and have gained particular interest because of their potential role in migraine, depression, and diseases involving the basal ganglia (Hoyer et al., 1994). These receptors may serve a presynaptic autoreceptor function inasmuch as their activation acts to inhibit 5-HT release (Middlemiss et al., 1988; Hamblin et al., 1992). They also seem to function as heteroreceptors as indicated by studies of nonserotonergic neurons in which 5-HT inhibits the release of acetylcholine, glutamate, dopamine, norepinephrine, and g-aminobutyric acid (Hen, 1992). Similarly, 5-HT1B/1D receptors have been suggested to inhibit peptide release from trigeminal nerve endings in the dura mater (Buzzi et al., 1991). The precise function of each receptor subtype is still controversial. Available data favor the view that vasoconstriction is mediated primarily by 5-HT1B receptors, whereas neuroinhibition in the trigeminovascular system involves predominantly the 5-HT1D receptor subtype (Longmore et al., 1997). It can be put forward that the expression of the 5-HT1D receptor is less abundant than the 5-HT1B receptor. Molderings et al. (1996) suggested norepinephrine release to be inhibited by 5-HT1D receptors located on the noradrenergic axon terminals in human atrial appendages. Regulation of [H]5-HT release in raphé nuclei of 5-HT1B receptor gene knockout mice seems to be mediated by a 5-HT1D-like receptor (Piñeyro et al., 1995). In these mice, 5-HT1B, but not 5-HT1D, autoreceptors inhibit 5-HT release at nerve terminals located in the frontal cortex and ventral hippocampus (Trillat et al., 1997). The h5-HT1B and h5-HT1D receptor subtypes show a relatively low (63%) overall amino acid identity with 77% identity within the TMDs (Weinshank et al., 1992). TMD I is most divergent (59% identity), whereas the six other TMDs share between 71% (TMD V) to 96% (TMD III) amino acid identity. Notwithstanding this low homology, the h5-HT1D and h5HT1B receptor subtypes first were reported to display similar binding profiles (Weinshank et al., 1992). Both receptor subABBREVIATIONS: 5-HT, 5-hydroxytryptamine (serotonin); 5-CT, 5-carboxamidotryptamine; TMD, transmembrane domain; ECL, extracellular loop; PCR, polymerase chain reaction; WT, wild-type; h, human. 0026-895X/98/061088-09$3.00/0 Copyright © by The American Society for Pharmacology and Experimental Therapeutics All rights of reproduction in any form reserved. MOLECULAR PHARMACOLOGY, 54:1088–1096 (1998). 1088 at A PE T Jornals on A uust 7, 2017 m oharm .aspeurnals.org D ow nladed from types can be pharmacologically differentiated using the 5-HT2 receptor antagonists ketanserin and ritanserin (Peroutka, 1994; Pauwels et al., 1996). Both compounds show potent binding affinity for and are silent antagonists at cloned 5-HT1D receptors of human, rat, and guinea pig (Wurch et al., 1997b). Surprisingly, they show micromolar affinity for canine 5-HT1D receptors (Zgombick et al., 1997). A series of benzanilides, such as GR 125743 (N-[4-methoxy-3(4-methyl-piperazin-1-yl)phenyl]3-methyl-4-(4-pyridyl)benzamide) and GR 127935 (N-[4-methoxy-3-(4-methyl-1piperazinyl)phenyl]-29-methyl-49-(5methyl-1,2,4-oxadiazol3-yl)[1,19-biphenyl]-4-carboxamide), have been reported as examples of mixed 5-HT1B/1D receptor antagonists (Clitherow et al., 1994). However, they also display agonist properties both in vitro and in vivo (Pauwels, 1997). Recently, some antagonists have been communicated as selective for 5-HT1B receptors [SB 216641 (N-[3-(2-dimethylamino) ethoxy-4-methoxyphenyl]29-methyl-49-(5-methyl-1,2,4oxadiazol-3-yl)-(1,19-biphenyl)-4-carboxamide) (Price et al., 1997) and SB 224289 (19-methyl-5-(29-methyl-49-(5-methyl1,2,4-oxadiazol-3-yl)biphenyl-4-carbonyl)-2,3,6,7-tetrahydrospiro [furo[2,3f]indole-3,49-piperidine) (Roberts et al., 1997)] and for 5-HT1D receptors [BRL-15572 (3-[4-(3-chlorophenyl)piperazin-1-yl]1,1-diphenyl-2-propanol) (Price et al., 1997)]. Hence, several ligands are available that distinguish between 5-HT1B and 5-HT1D receptor subtypes. We are interested in the ligand binding divergence between h5-HT1D and h5-HT1B receptors and the molecular correlates that underlie their pharmacological specificity. Therefore, a study was undertaken to identify the domain or domains of 5-HT1B and 5-HT1D receptors that determine ligand binding specificity. This was performed using a chimeric receptor approach by combining different parts of h5HT1D and h5-HT1B receptors. The binding profile of the various chimeric receptors was determined on transient expression in Cos-7 cells with two different radioligands (i.e., the agonist [H]5-CT and the putative antagonist [H]GR 125743). The current report summarizes our findings on the selective interaction of ketanserin and ritanserin with a 5-HT1D receptor domain restricted to the second ECL and the fifth TMD. Experimental Procedures Construction of chimeric 5-HT1D/1B receptors and pointmutated 5-HT1D receptors. Chimeric 5-HT1D/1B receptors and point-mutated 5-HT1D receptors were constructed by a modified PCR-based overlap extension technique. It allows the construction of chimeric receptors without generating a frame-shift or insertion or deletion of amino acids. Briefly, each chimeric receptor is realized by a three-step PCR-based method that allows the fusion of two or three PCR fragments corresponding to the respective segments that represent the chimeric receptor. The first PCR step corresponds to the amplification of the different fragments of the chimeric receptor, which will be fused together in a second PCR step. A third PCR step amplifies the obtained fusion product. The corresponding PCR-primers were designed according to the reported h5-HT1B and h5-HT1D receptor gene sequences (Weinshank et al., 1992) such that they possess a 59 extension that is complementary to the adjacent PCR fragment that has to be fused. These primer sequences are listed in Table 1. For each series of primers, the first PCR mixture (50 ml each) contained 10 ng of purified full-length receptor gene fragment, 25 mM concentration of each dNTP, 400 nM concentration of each primer, and 1.25 units of Taq DNA polymerase in PCR buffer (50 mM KCl, 1.5 mM MgCl2, 10 mM TriszHCl, pH 8.3). The PCR program consisted of 10 repetitive cycles with a strand separation step at 94° for 30 sec, an annealing step at 58° for 1, min and an elongation step for 1 min at 72°. The PCR products were separated by 2% agarose gel electrophoresis and purified using a Geneclean II kit. For the second PCR step, an equimolecular amount of each fragment (;10 nmol