Residues in transmembrane domain (TM)-III, TM-V, TM-VI, and TM-VII believed to be facing the deep part of the presumed main ligand-binding pocket of the NK1 receptor were probed by alanine substitution and introduction of residues with larger and/or chemically distinct side chains. Unaltered or even improved binding affinity for four peptide agonists, substance P, substance P-O-methyl ester, eledoisin, and neurokinin A, as well as normal EC50 values for substance P in stimulating phosphatidylinositol turnover indicated that these mutations did not alter the overall functional integrity of the receptor. The alanine substitutions in general had only minor effects on nonpeptide antagonist binding. However, the introduction of the larger and polar aspartic acid and histidine residues at positions corresponding to the monoamine binding aspartic acid in TM-III of the beta 2-adrenoceptor (ProIII:08, Pro112 in the NK1 receptor) and to the presumed monoamine interacting "two serines" in TM-V (ThrV:09, Thr201; and IleV:12, Ile204) impaired by > 100-fold the binding of a group of nonpeptide antagonists, including CP96,345, CP99,994, RP67,580, RPR100,893, and CAM4092. In contrast, another group of nonpeptide antagonists, LY303,870, FK888, and SR140,333, were little or not at all affected by the space-filling substitutions. Two of these compounds, FK888 and LY303,870, were those most seriously affected (75-89-fold) by alanine substitution of PheVI:20 located in the upper part of the main ligand-binding crevice. Surprisingly, substitution of AlaIII:11 (Ala115), which is located in the middle of TM-III, conceivably pointing toward TM-VII, with a larger valine residue increased the affinity for all 13 ligands tested, presumably by creating a closer interhelical packing. It is concluded that the introduction of larger side chains at positions at which molecular models indicate that this is structurally allowed can be a powerful method of locating ligand-binding sites due to the considerable difference between positive and negative results. Such steric hindrance mutagenesis strongly indicates that one population of nonpeptide antagonists bind in the deep pocket of the main ligand-binding crevice of the NK1 receptor, whereas another group of nonpeptide antagonists, especially SR140,333, was surprisingly resistant to mutational mapping in this pocket.