A library of robust ghrelin receptor mutants with single substitutions at 22 positions in the main ligand-binding pocket was employed to map binding sites for six different agonists: two peptides (the 28-amino-acid octanoylated endogenous ligand ghrelin and the hexapeptide growth hormone secretagogue GHRP-6) plus four nonpeptide agonists-the original benzolactam L-692,429 [3-amino-3-methyl-N-(2,3,4,5-The strongest mutational effect with respect to decrease in potency for stimulation of inositol phosphate turnover was for all six agonists the GluIII:09-to-Gln substitution in the extracellular segment of TM-III. Likewise, all six agonists were affected by substitutions of PheVI:16, ArgVI:20, and PheVI:23 on the opposing face of transmembrane domain (TM) VI. Each of the agonists was also affected selectively by specific mutations. The mutational map of the ability of L-692,429 and GHRP-6 to act as allosteric modulators by increasing ghrelin's maximal efficacy overlapped with the common mutational map for agonism but it was not identical with the map for the agonist property of these small-molecule ligands. In molecular models, built over the inactive conformation of rhodopsin, low energy conformations of the nonpeptide agonists could be docked to satisfy many of their mutational hits. It is concluded that although each of the ligands in addition exploits other parts of the receptor, a large, common binding site for both small-molecule agonists-including ago-allosteric modulators-and the endogenous agonist is found on the opposing faces of TM-III and -VI of the ghrelin receptor.The gastrointestinal peptide hormone ghrelin is an important regulator of appetite, energy expenditure, and acute growth hormone secretion through interaction with the ghrelin receptors located mainly in the central nervous system (Tschöp et al., 2000). From a drug discovery point of view, these functions of ghrelin qualify ghrelin receptor agonists as anabolic compounds with potentials for treatment of, for example, cachexia.The initial agonists synthesized for the ghrelin receptor were peptides, sharing common structural features, including a central hydrophobic motif and a positive charge in the amidated C-terminal end. Despite relatively poor bioavailability and low therapeutic index/window, these peptides efficiently induced GH secretion in vitro as well as in vivo both in animal and human models (Bowers et al., 1984). GHRP-6 is a prototype for these peptides (Fig. 1). In an attempt to increase oral bioavailability, a series of nonpeptide compounds was subsequently developed. They were