Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The mammalian tachykinins substance P (SP), neurokinin A (NKA), and neurokinin B (NKB) form a family of peptides that share the common C-terminal sequence Phe-X-Gly-Leu-MetNH₂. Tachykinins are found in many regions of the central and peripheral nervous system (Maggio, 1988; Maggi et al., 1993; Otsuka and Yoshioka, 1993) and produce their biological effects by stimulating three distinct G protein-coupled receptors termed tachykinin NK₁, NK₂, and NK₃ receptors (Maggi, 1995, for review).

The tachykinin NK₁ receptor has been the object of intense investigations to assess the structural determinants of its interaction with ligands of both peptide and nonpeptide nature, agonists and antagonists, respectively (Schwartz et al., 1995, for review). The site-directed mutagenesis approach has been extensively used to identify residues putatively involved in the binding of agonists and antagonists to the tachykinin NK₁ receptor. From these studies, it has been proposed that residues forming the binding site of natural agonists on the tachykinin NK₁ receptor are quite distinct from those that form the binding site of antagonists, especially those of antagonists of nonpeptide nature (Fong et al., 1992; Gether et al., 1993a; Huang et al., 1994).

Less information is available on the structural determinants of agonist and antagonist binding at the tachykinin NK₂ receptor. Studies using chimeric tachykinin NK₁/NK₂ receptors have shown that the binding site of the nonpeptide tachykinin NK₂ receptor antagonist SR 48968 is not fully coincident with that of NKA, the natural tachykinin with preferential binding affinity for tachykinin NK₂ receptors (Gether et al., 1993b). A few site-directed mutagenesis studies have also been performed on the human tachykinin NK₂ receptor (hNK₂R; Bhogal et al., 1994; Huang et al., 1995). The results of these studies have identified some aromatic residues in transmembrane segments (TM) V to VII of the hNK₂R as being involved in the binding of NKA and other natural tachykinins, as well as in the binding of the nonpeptide tachykinin NK₂ receptor antagonists SR 48968 (Emonds-Alt et al., 1992) and GR 159897 (Beresford et al., 1995; Fig. 1). In particular, Tyr²⁸⁹ (TM VII), along with Tyr²⁶⁶ and Phe²⁷⁰ (TM VI), were proposed to form part of the SR 48968 binding pocket (Huang et al., 1995). Discrepant results were reported with regard to the role played by His¹⁹⁸ (TM V) in determining the binding affinity of SR 48968. Bhogal et al. (1994) and Huang et al. (1995) reported a dramatic fall in the binding affinity of NKA at the H¹⁹⁸A mutant tachykinin NK₂ receptor. However, although Bhogal et al. (1994) reported that this residue is crucial for preserving the binding affinity of SR 48968, Huang et al. (1995) failed to detect any significant change of the affinity with the H¹⁹⁸A mutation, although they reported reduced binding affinity for another nonpeptide antagonist, GR 159897. The reasons for this discrepancy have not been found.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Chemical structures of the tachykinin NK2 receptor antagonists MEN 11420, SR 48968, and GR 159897.

In this study, we used Chinese hamster ovary (CHO) cell lines stably transfected with the wild-type and mutant tachykinin NK₂ receptors to more precisely assess the role played by aromatic residues in TM V (His¹⁹⁸), TM VI (Tyr²⁶⁶ and Phe²⁷⁰), and TM VII (Tyr²⁸⁹; Fig. 2) in the binding of natural agonists and peptide (including both linear and cyclic structures) and nonpeptide antagonists. We performed saturation and competition binding experiments with CHO cell membranes by using [¹²⁵I]NKA, [³H]SR 48968, and [³H]MEN 11420, a new tachykinin NK₂ receptor-selective ligand (Catalioto et al., 1998, Renzetti et al., 1998; Fig. 1) as radioligands. In particular, we aimed at verifying the hypothesis of the highly variable, flexible mode of interaction of the receptor with the different classes of ligands that could be inferred from the previous work of Bhogal et al. (1994) and Huang et al. (1995).

View larger version (35K): [in this window] [in a new window]   Fig. 2.   Schematic of the tachykinin NK2 receptor showing the putative localization of the residues subjected to mutation.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals

Penicillin, streptomycin, and trypsin were purchased from GIBCO (Grand Island, NY). Dialyzed fetal calf serum (FCS) was obtained from BioWhittaker (Walkersville, MD). Minimum essential Eagle's medium -modification (-MEM) and GTP were purchased from Sigma Chemical Co. (St. Louis, MO). Flasks and Petri dishes were purchased from Falcon (Becton Dickinson, Milan, Italy). Protein binding dye was from Bio-Rad (Richmond, CA). NKA, NKB, and SP were obtained from Neosystem Laboratoire (Strasbourg, France). SR 48968 [(S)-N-methyl-N[4-(4-acetyl-amino-4-phenylpiperidino)-2-(3,4-dichlorophenyl)butyl]benzamide] was kindly provided by Drs. X. Emonds-Alt and G. Le Fur (Sanofi Recherche, Montpellier, France). MEN 11420 (c{[(-D-GlcNAc)Asn-Asp-Trp-Phe-Dpr-Leu]c(2-5)}), MEN 10376 ([Tyr⁵,D-Trp^6,8,9,Arg¹⁰]NKA (4-10), and GR 159897 [(R)-1-[2-(5-fluoro-1H-indol-3-yl)ethyl]-4-methoxy-4[(phenylsulfinyl)-methyl]piperidine] were synthesized at the Chemistry Department of Menarini Ricerche (Firenze, Italy). R396 (Ac-Leu-Asp-Gln-Trp-Phe-Gly-NH₂) was a gift from Prof. D. Regoli.

[³H]MEN 11420 (specific activity, 62 Ci/mmol) was synthesized by SibTech Inc. (Elmsford, NY). [¹²⁵I]NKA (specific activity, 2000 Ci/mmol) was purchased from Amersham International (Buckinghamshire, UK). [³H]SR 48968 (specific activity, 25.7 Ci/mmol) was purchased from Dupont-NEN Life Sciences (Boston, MA). All other reagents, available from commercial sources, were of analytical grade.

Site-Directed Mutagenesis of hNK₂R cDNA. Plasmid pBS/hNK₂R (kindly provided by Dr. J. E. Krause, Washington University, St. Louis, MO) contained a 1.2-kb cDNA for the ileum hNK₂R cloned in the SmaI site of the pBlueScript II SK() phagemid. Site-directed mutagenesis of the tachykinin NK₂ receptor cDNA was performed according to the phosphorothioate technique of Eckstein (Taylor et al., 1985; Nakamaye and Eckstein, 1986) using single-stranded DNA of the tachykinin NK₂ receptor cDNA in pBlueScript II SK() and an in vitro mutagenesis kit (Sculptor; Amersham International) according to the manufacturer's instructions. Wild-type cDNA was isolated from HindIII/XbaI-digested pBS/hNK₂R and cloned in pmCMVSV1dhfr, partially digested with HindIII, between the HindIII and XbaI sites of the polylinker region, under the transcriptional control of the murine cytomegalovirus major immediate-early promoter. The resulting vector was designated as pmCMVSV1dhfr-hNK₂R. Expression vectors carrying the mutated cDNAs (H¹⁹⁸A, Y²⁶⁶A, Y²⁶⁶F, F²⁷⁰A, Y²⁸⁹A, Y²⁸⁹F) were constructed by exchanging the 1.2-kb EcoRV/Xbal wild-tachykinin NK₂ receptor cDNA in pmCMVSV1dhfr-hNK₂R with the EcoRV/Xbal mutated cDNAs excised from pBS/hNK₂R. pmCMVSV1 was constructed by removing G-CSF cDNA sequences from XbaI-digested pmCMVG-CSFSV1dhfr (Rotondaro et al., 1997). The structure of all constructs was confirmed by restriction analysis. The complete coding sequence of wild-type and mutated tachykinin NK₂ receptor cDNAs cloned into the pmCMVSV1dhfr expression vector was confirmed by DNA sequencing.

Receptor Expression in CHO Cells. Large-scale preparation of vector DNA for transfection experiments was carried out using a Qiagen maxipreparation column (Qiagen, Hilden, Germany). Wild-type and (H¹⁹⁸A, Y²⁶⁶A, Y²⁶⁶F, F²⁷⁰A, Y²⁸⁹A, or Y²⁸⁹F) mutated tachykinin NK₂ receptor cDNAs in pmCMVSV1dhfr were introduced by lipofection as described previously (Rotondaro et al., 1997) into dihydrofolate reductase (DHFR)-deficient CHO DUKX-B11 cells (Urlaub and Chasin, 1980; referred to as CHOdhfr). Stable DHFR+ transformants were selected in nucleoside-free -MEM containing 5% dialyzed FCS; 12 to 14 days after transfection, more than 100 individual DHFR+ clones were pooled, grown to mass culture, and used for ligand-binding studies. CHOdhfr cells and stable CHO transfectants were grown in -MEM containing ribonucleosides and deoxyribonucleosides and supplemented with 5% FCS, in a humidified atmosphere of 5% CO₂/95% air at 37°C until slightly confluent.

Binding Assays. Confluent cells were harvested in PBS, pelleted by centrifugation at 200g (4°C), and homogenized using a Polytron PT3000 (Kinematica, Lucerne, CH) at 13,000 rpm for 15 s in 20 ml of 50 mM Tris, pH 7.4, containing bacitracin (100 µg/ml), chymostatin (10 µg/ml), leupeptin (5 µg/ml), and 10 µM thiorphan (buffer A). The homogenate was centrifuged for 1 h at 25,000g (4°C), and the pellet was resuspended in the binding buffer (pH 7.4) composed of buffer A supplemented with 150 mM NaCl, 5 mM MnCl₂, and 0.1% BSA at a protein concentration of about 0.6 mg/ml. The membranes (50-90 µg protein/assay) were incubated for 30 min ([¹²⁵I]NKA and [³H]SR 48968) or 60 min ([³H]MEN 11420) at 20°C. In saturation experiments, either increasing concentrations (0.09-15.0 nM) of [³H]SR 48968 or [³H]MEN 11420 (hot experiments) or a fixed concentration (150 pM) of [¹²⁵I]NKA with various concentrations of unlabeled NKA (0.01 nM to 1 µM; cold experiment) was used. For competition experiments, 0.4 to 1.0 nM [³H]MEN 11420, 0.6 nM [³H]SR 48968, or 150 pM [¹²⁵I]NKA was used, with or without varying concentrations (0.001 nM to 10 µM) of the competing compounds in a final volume of 0.5 ml. Then, 1 µM unlabeled MEN 11420, SR 48968, or NKA was used to define nonspecific binding, depending on the radioligand. The reaction was terminated by the addition of 4 ml of ice-cold Tris (50 mM, pH 7.4) followed by rapid filtration through Whatman GF/B filter sheets (presoaked in 0.3% polyethylenimine for [³H]MEN 11420 or 0.5% BSA for [³H]SR 48968 and [¹²⁵I]NKA, for at least 3 h using a Brandel (Montreal, Quebec, Canada) cell harvester. Filters were washed three times with 4 ml of ice-cold 50 mM Tris buffer, pH 7.4. The trapped radioactivity was determined by liquid scintillation using a -scintillation counter (2200 CA; Packard, Milan, Italy) or using a gamma counter (Cobra, Packard).

Analysis of Binding Data

Saturation and competition data were processed according to Munson and Rodbard (1980) by the step-wise application of EBDA and LIGAND programs in the KELL software for Macintosh (Biosoft, Cambridge, UK). The maximal binding (B_max), equilibrium dissociation constant (K_d), and equilibrium inhibition constant (K_i) values were calculated. They are reported as mean values with approximate S.E.s as estimated by LIGAND. The goodness of fit, according to the single- or multiple-site binding models, was evaluated by the F ratio.

Statistical comparisons were made by using one-way ANOVA followed by Tukey's test where appropriate.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Saturation Experiments

The specific binding of [¹²⁵I]NKA, [³H]SR 48968, and [³H]MEN 11420 was directly proportional to the membrane concentration (data not shown). At protein concentrations between 50 and 90 µg/assay, as used in competition experiments, the specific binding represented approximately 80 to 90% of the total binding for each radioligand, respectively. Scatchard transformation of the saturation experiments data showed a monophasic interaction with the wild-type NK₂ receptor with K_d values of 2.4 ± 0.43, 0.3 ± 0.08, and 4.0 ± 0.44 nM and B_max values of 206 ± 28, 3274 ± 1489, and 1664 ± 416 fmol/mg protein for [¹²⁵I]NKA, [³H]SR 48968, and [³H]MEN 11420, respectively (n = 5-6; Table 1).

Y²⁶⁶A and Y²⁸⁹A mutant receptors did not bind any of the radioligands (Table 1). For the other mutant receptors, the specific binding of [¹²⁵I]NKA, [³H]SR 48968, and [³H]MEN 11420 was directly proportional to the membrane concentration (data not shown). Saturation experiments always produced data congruent with monophasic ligand-receptor interaction. Neither K_d nor B_max values for [¹²⁵I]NKA changed with the mutant receptors, which maintained the ability to interact with the agonist ligand (Table 1). The B_max values for [³H]SR 48968 and [³H]MEN 11420, considerably higher than that measured for [¹²⁵I]NKA, were not statistically different from each other. They, too, despite an apparent variability, did not change in a significant manner throughout the responsive receptors; the only exception was the H¹⁹⁸A mutant, which showed both lower expression and reduced affinity for the antagonist ligands (Table 1).

Competition Experiments

Wild-Type. Among the tachykinins, only NKA competed with high affinity for the binding of the three radioligands to the wild-type hNK₂R. When using either [³H]SR 48968 or [³H]MEN 11420 as tracers, the competition of unlabeled NKA was biphasic, probably reflecting its binding to the high-affinity (i.e., G protein-coupled) and low-affinity (i.e., G protein-uncoupled) states of the tachykinin NK₂ receptor (Fig. 3). The rank order of affinity of the receptor antagonists, which was independent from the radioligand used to label tachykinin NK₂ receptors, was as follows: SR 48968 > GR 159897 > MEN 11420 > MEN 10376  R396 (Table 2). The curves describing the displacement of [³H]SR 48968 and [³H]MEN 11420 by the prototype antagonists SR 48968, GR 159897, and MEN 11420 are shown in Figs. 4A and 5A.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Homologous and heterologous displacements of [125I]NKA (), [3H]SR 48968 (), and [3H]MEN 11420 () by unlabeled NKA in membrane preparations of CHO cells transfected with wild-type nNK2R. Data shown represent the mean of at least three separate determinations in duplicate. Curves were fitted by the simultaneous analysis of all data sets to a one-site (for [125I]NKA) or two-site (for [3H]SR 48968 and [3H]MEN 11420) binding model by GraphPAD Prism.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Displacement curves of [3H]SR 48968 binding to CHO cell membranes transfected with wild-type (A), H198A (B), Y266F (C), and F270A (D) hNK2R by GR 159897 (), MEN 11420 (), and SR 48968 (). Data shown represent the mean of at least three separate determinations in duplicate. Curves were fitted by the simultaneous analysis of all data sets to a one-site binding model by GraphPAD Prism.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Displacement curves of [3H]MEN 11420 binding to CHO cell membranes transfected with wild-type (A) and Y289F (B) hNK2R by GR 159897 (), MEN 11420 (), and SR 48968 (). Data shown represent the mean of at least three separate determinations in duplicate. Curves were fitted by the simultaneous analysis of all data sets to a one-site binding model by GraphPAD Prism.

H¹⁹⁸A. This mutation completely abrogated the binding of [¹²⁵I]NKA to the hNK₂R and reduced by about 16-fold the binding affinity of [³H]SR 48968, whereas binding affinity of [³H]MEN 11420 was reduced by about 3-fold (Table 1). Because a sizable binding affinity was measured with [³H]SR 48968 and [³H]MEN 11420, these tracers were used in competition experiments with natural tachykinins and antagonists of both peptide and nonpeptide nature (Table 2, Fig. 4B).

Y²⁶⁶F. This mutation did not affect the binding of [¹²⁵I]NKA, slightly affected (4-fold decrease) the binding of [³H]SR 48968, and abrogated the binding of [³H]MEN 11420 to the hNK₂R (Table 1). The binding affinity of unlabeled MEN 11420 at the mutant receptor was decreased by about 12-fold when using either [¹²⁵I]NKA or [³H]SR 48968 as tracers (Table 2, Fig. 4C). The binding affinity of NKB, MEN 10376, and R396 was not substantially altered by the Y²⁶⁶F mutation (Table 2), whereas the binding affinity of GR 159897 was markedly decreased, by about 185- and 149-fold when using [¹²⁵I]NKA or [³H]SR 48968 as tracers, respectively (Table 2, Fig. 4C).

F²⁷⁰A. This mutation completely abrogated the binding of [¹²⁵I]NKA and [³H]MEN 11420 with only a minor effect on the binding of [³H]SR 48968 (Table 1). When using [³H]SR 48968 as a tracer, a 66-fold decrease in the binding affinity of MEN 11420 was estimated (Table 2, Fig. 4D), whereas no competition was exerted by NKA up to 10 µM (Table 2). Likewise, no binding affinity of NKB and SP could be estimated using [³H]SR 48968 as a tracer. The binding affinity of GR 159897 was, if any, slightly increased (5-fold) by the F²⁷⁰A mutation (Table 2, Fig. 4D), whereas the binding affinity of the linear peptide antagonists was decreased (MEN 10376; 17-fold) or abolished (R396; Table 2).

Y²⁸⁹F. This mutation completely abrogated the binding of [³H]SR 48968 without affecting the binding of [¹²⁵I]NKA and [³H]MEN 11420 (Table 1). In competition experiments, a 2741- and 3088-fold decrease in binding affinity of SR 48968 was estimated when using [¹²⁵I]NKA and [³H]MEN 11420 as tracers, respectively (Table 2, Fig. 5B). The binding affinity of NKB and SP was unaffected by the Y²⁸⁹F mutation (Table 2). The binding affinity of GR 159897 was totally abrogated (Table 2, Fig. 5B), whereas that of MEN 10376 was unaffected when using either [¹²⁵I]NKA or [³H]MEN 11420 as tracers (Table 2). Interestingly, the binding affinity of the linear antagonist R396 was consistently increased, by about 18-fold, by the Y²⁸⁹F mutation, when using either radioligand (Table 2).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the attempt to overcome problems that may be related to transient transfection and variable levels of expression of the receptor, we prepared cell lines stably transfected with wild-type and mutant hNK₂Rs. This approach enabled us to produce a systematic pharmacological analysis of the mutant receptors by using three different radioligands and a panel of agonists and antagonists.

The even B_max values found with the agonist [¹²⁵I]NKA generally represent only a fraction of the receptor density estimated with the antagonist radioligands, which are not able to discriminate between G protein-coupled and -uncoupled states of the NK₂ receptor. This is likely due to the fact that transfected receptors are expressed in excess relative to the G protein-coupling capability of the CHO. In the case of the H¹⁹⁸A mutant, a significant reduction in the B_max value accompanied the decrease in affinity for the recognized antagonists.

It is interesting to note that three of the four aromatic residues mutated in this study (His¹⁹⁸, Phe²⁷⁰, and Tyr²⁸⁹ in the sequence of the hNK₂R) are fully conserved at equivalent positions in the sequence of the various mammalian tachykinin receptors cloned so far; the fourth residue (Tyr²⁶⁶ in the sequence of the hNK₂R) is conserved in the mammalian tachykinin NK₂ and NK₃ receptors and is replaced by Phe at the equivalent position of the mammalian tachykinin NK₁ receptor. Due to their lipophilic nature, these conserved residues are likely to play a structural role in the correct positioning of the TMs. From the present data, we have evidence that each one of the H¹⁹⁸A, Y²⁶⁶F, F²⁷⁰A, and Y²⁸⁹F mutant receptors maintained high-affinity binding for at least one of the three radioligands used. This can be considered as a kind of positive control (i.e., it allows us to exclude the possibility that these mutations produced a marked alteration in the gross structure of the receptor or that the mutant receptor protein was synthesized but not expressed at the cell membrane level). Therefore, the observed changes in the affinity of a given ligand for each of these mutant receptors can be confidently assumed as indications that the corresponding residues represent a true point of drug-receptor interaction or, alternatively, that the mutation has modified the geometry of the ligand-binding pocket, thus indirectly influencing the affinity of the ligand.

With regard to the binding of natural tachykinins, our data demonstrate that both His¹⁹⁸ and Phe²⁷⁰ are crucial for maintaining the high-affinity binding of NKA (see Huang et al., 1995) and that these residues are also of relevance for the binding of NKB and SP to the wild-type tachykinin NK₂ receptor.

His¹⁹⁸, putatively located on TM V of the hNK₂R, corresponds to His¹⁹⁷ in the sequence of the hNK₁R. In the latter, the PheAla replacement abolished the binding of several nonpeptide receptor antagonists without affecting the binding of SP (Fong et al., 1993). On the other hand, SP displayed a weak but sizable binding to the wild-type tachykinin NK₂ receptor (K_i = 1.8 µM against [¹²⁵I]NKA), which was lost in the H¹⁹⁸A mutant (K_i > 10 µM). This observation, along with the data of Fong et al. (1993), implies that the mode of interaction of SP with the tachykinin NK₁ and NK₂ receptors are different (i.e., the same tachykinin interacts with different regions of the tachykinin NK₁ and NK₂ receptors).

Our data also indicate that the ---OH function of Tyr²⁶⁶ and Tyr²⁸⁹ is not involved in determining the binding affinity of natural tachykinins to the hNK₂R. On the other hand, the actual role of the aromatic function of these two Tyr residues in the binding of tachykinins could not be definitely assessed; indeed, Ala replacement of Tyr²⁶⁶ and Tyr²⁸⁹ completely abolished the binding affinity for each of the three radioligands tested. Unfortunately, this also means that no positive control (as defined above) was any more available, which hampered any speculation on the relevance of these data.

Our findings provide a clear demonstration that the binding of peptide antagonists to the hNK₂R is partially overlapping but not coincident to that of nonpeptide antagonists SR 48968 and GR 159897. Moreover, differences in binding behavior also emerged between linear and cyclic peptide antagonists. In general, the observed changes in the binding affinity of linear peptide antagonists MEN 10376 (Maggi et al., 1991) and R396 (Dion et al., 1990) for the mutant tachykinin NK₂ receptors follow quite closely those observed for NKA and agonists. In fact, Ala substitution of His¹⁹⁸ and Phe²⁷⁰ almost completely abrogated the binding of MEN 10376 and R396, whereas the TyrPhe replacement at position 266 was irrelevant for both antagonists and, for MEN 10376, the TyrPhe replacement at position 289 had no effect on binding affinity. Our data for MEN 11420, a glycosylated bicyclic peptide (Catalioto et al., 1998), indicate that in common with NKA, Phe²⁷⁰ is relevant for determining the high-affinity binding of this antagonist to the hNK₂R. On the other hand, although His¹⁹⁸ is clearly relevant for the binding of NKA, MEN 10376, and R396, this residue is apparently less important for the binding of MEN 11420.

Huang et al. (1995) described some pharmacological properties of Tyr²⁶⁶ mutants of the hNK₂R. They reported that although the Y²⁶⁶F mutant maintains a full binding affinity for both NKA and SR 48968, the binding affinity was largely lost when Tyr²⁶⁶ was mutated to Ser or Ala (findings confirmed by the present results). Because a functional response (phosphoinositol formation) to NKA was preserved in the Y²⁶⁶S mutant, the derived affinity estimates for NKA and SR 48968 could be computed and found to be obviously low (0.3 and 3.4 µM, respectively; Huang et al., 1995). From this, they concluded that the aromatic group of Tyr²⁶⁶ is involved, directly or indirectly, in the binding of NKA and SR 48968 to the hNK₂R. In this study, we observed that it is sufficient to remove the ---OH function of Tyr²⁶⁶ to abrogate the high-affinity binding of [³H]MEN 11420 to the hNK₂R. This indicates that the binding site of MEN 11420 on the hNK₂R is partly different from that of both NKA and SR 48968. The drop in affinity of MEN 11420, as estimated from competition experiments against [¹²⁵I]NKA or [³H]SR 48968, is relatively small (about 10-fold) but consistent and evidently sufficient to prevent the detection of a specific binding at the low concentrations of the radioligand used in saturation experiments. This same mutation did not affect the binding affinity of the linear peptides MEN 10376 and R396, nor that of SR 48968, but decreased the binding affinity of GR 159897. This may suggest that the ---OH function of Tyr²⁶⁶ directly interacts with some moiety present in both MEN 11420 and GR 159897 but not in SR 48968. An hydrogen-bond donor/acceptor interplay between the ---OH function of Tyr²⁶⁶ and the indole amino group present in both MEN 11420 and GR 159897 could be postulated (Fig. 6).

View larger version (12K): [in this window] [in a new window]   Fig. 6.   Molecular models of the binding of the tachykinin NK2 receptor antagonists MEN 11420 (top) and SR 48968 (bottom) to the tachykinin NK2 receptor. Presented view is from outside the cell membrane, along the -helix axis. TM I and TM II have been omitted for clarity.

With regard to the binding of SR 48968 at the H¹⁹⁸A mutant receptor, we found a sizable but limited decrease in affinity (about 16-fold): in this respect, our data differ from those of Huang et al. (1995), who reported that the binding of SR 48968 is unaffected by the H¹⁹⁸A mutation, and from those of Bhogal et al. (1994), who reported that binding of SR 48968 is abrogated by the H¹⁹⁸A mutation. Because both previous studies were performed on transiently transfected cell lines, it is possible that variable levels of receptor expression had influenced the results. With regard to the other residues, our data confirm the observation (Huang et al., 1995) that the removal of the ---OH function of Tyr²⁸⁹ almost totally eliminated the binding affinity of SR 48968. On the other hand, the F²⁷⁰A mutation, which impaired the binding of most other ligands tested (with the exception of GR 159897), also had little effect on the binding of SR 48968 to the hNK₂R. The equivalent residue (Phe²⁶⁸) in the sequence of the hNK₁R has been recently proposed to be partially involved, directly or indirectly, in the binding of both SP and various nonpeptide antagonists to the tachykinin NK₁ receptor, with the exception of SR 140333 (Holst et al., 1998).

On the whole, the two nonpeptide antagonists appear to have quite different sensitivities to residue mutation: the binding affinity of GR 159897 was heavily compromised by mutations at sites (His¹⁹⁸ and Tyr²⁶⁶), which have a weaker effect on the binding of SR 48968. This strongly suggests that similar to the tachykinin NK₁ receptor (Holst et al., 1998), nonpeptide ligands of different chemical classes possess different modes of interaction with the tachykinin NK₂ receptor.

An interesting further observation is that relative to the ligand-dependent variation of the affinity of GR 159897 for the H¹⁹⁸A mutant receptor, the affinity of GR 159897 was about 4-fold higher if measured in competition with [³H]MEN 11420 than with [³H]SR 48968 (K_i = 58 and 214 nM, respectively). A similar (5-fold) difference was observed when SR 48968 was tested in homologous (saturation) or heterologous (competition versus [³H]MEN 11420) binding assays (K_d = 5.2 nM, K_i = 1.0 nM, respectively). These observations suggest that [³H]MEN 11420 and [³H]SR 48968 may label different conformers of the tachykinin NK₂ receptor, both unable to signal second messenger activation, and that the His¹⁹⁸ residue is relatively less important in the binding energy balance for the ligand-receptor complex formed by [³H]MEN 11420 with the selected receptor conformer.

In conclusion, the present findings provide a detailed pharmacological analysis of the role of some aromatic residues in TM V to VII for binding of different tachykinin receptor agonists and antagonists to the hNK₂R. With regard to the residues examined, it appears that the binding of linear peptide antagonists is affected in a similar manner to that of agonists, whereas the chemical modification introduced in the cyclic backbone of MEN 11420 partially changed the regions of the receptor that provide high-affinity binding for this ligand. The binding of agonists and antagonists (of both peptide and nonpeptide nature) to the hNK₂R only partly overlaps.