Cannabinoid-1 receptor inverse agonists: current understanding of mechanism of action and unanswered questions.

Rimonabant and taranabant are two extensively studied cannabinoid-1 receptor (CB1R) inverse agonists. Their effects on in vivo peripheral tissue metabolism are generally well replicated. The central nervous system site of action of taranabant or rimonabant is firmly established based on brain receptor occupancy studies. At the whole-body level, the mechanism of action of CB1R inverse agonists includes a reduction in food intake and an increase in energy expenditure. At the tissue level, fat mass reduction, liver lipid reduction and improved insulin sensitivity have been shown. These effects on tissue metabolism are readily explained by CB1R inverse agonist acting on brain CB1R and indirectly influencing the tissue metabolism through the autonomic nervous system. It has also been hypothesized that rimonabant acts directly on adipocytes, hepatocytes, pancreatic islets or skeletal muscle in addition to acting on brain CB1R, although strong support for the contribution of peripherally located CB1R to in vivo efficacy is still lacking. This review will carefully examine the published literature and provide a perspective on what new tools and studies are required to address the peripheral site of action hypothesis.

Plasma agouti-related protein level: a possible correlation with fasted and fed states in humans and rats.

We measured plasma concentrations of agouti-related protein (AGRP) in humans and rats and determined whether these were affected by ingestion of a meal after fasting. In 17 healthy human subjects, the mean plasma concentration of AGRP was lower in the fed state than in the fasted state. Two hours after a breakfast meal, AGRP levels dropped by 39%. By contrast, a continued fast for 2 h increased the average AGRP concentration by 73%. In rats with diet-induced obesity, refeeding resulted in a 50% decrease in plasma AGRP concentrations following a fasting-refeeding protocol. Our results support the notion that plasma AGRP may serve as a biomarker for the transition from a fasted to the satiated state.

Molecular determinants of ligand binding to the human melanocortin-4 receptor.

To elucidate the molecular basis for the interaction of ligands with the human melanocortin-4 receptor (hMC4R), agonist structure-activity studies and receptor point mutagenesis were performed. Structure-activity studies of [Nle(4), D-Phe(7)]-alpha-melanocyte stimulating hormone (NDP-MSH) identified D-Phe7-Arg8-Trp9 as the minimal NDP-MSH fragment that possesses full agonist efficacy at the hMC4R. In an effort to identify receptor residues that might interact with amino acids in this tripeptide sequence 24 hMC4R transmembrane (TM) residues were mutated (the rationale for choosing specific receptor residues for mutation is outlined in the Results section). Mutation of TM3 residues D122 and D126 and TM6 residues F261 and H264 decreased the binding affinity of NDP-MSH 5-fold or greater, thereby identifying these receptor residues as sites potentially involved in the sought after ligand-receptor interactions. By examination of the binding affinities and potencies of substituted NDP-MSH peptides at receptor mutants, evidence was found that core melanocortin peptide residue Arg8 interacts at a molecular level with hMC4R TM3 residue D122. TM3 mutations were also observed to decrease the binding of hMC4R antagonists. Notably, mutation of TM3 residue D126 to alanine decreased the binding affinity of AGRP (87-132), a C-terminal derivative of the endogenous melanocortin antagonist, 8-fold, and simultaneous mutations D122A/D126A completely abolished AGRP (87-132) binding. In addition, mutation of TM3 residue D122 or D126 decreased the binding affinity of hMC4R antagonist SHU 9119. These results provide further insight into the molecular determinants of hMC4R ligand binding.

A melanocortin agonist reduces neuronal firing rate in rat hypothalamic slices.

Bath application of alpha-melanocyte stimulating hormone (alpha-MSH) to rat hypothalamic slices inhibited the spontaneous firing rate of continuously firing neurons in the ventromedial hypothalamic nucleus or paraventricular nucleus. This inhibitory effect is most likely direct and independent of synaptic transmission. The alpha-MSH-responsive neurons tested did not respond to neuropeptide Y (NPY) application. On the other hand, alpha-MSH did not inhibit the intraburst firing rate of phasic bursting neurons, although these bursting neurons were highly responsive to a serotonin 5HT2a/2b/2c agonist with a change of firing pattern to continuous firing and an increase in firing rate which was reversed by NPY. These results suggest that a change of neuronal firing rate may represent a neural correlate of satiety induced by anorexic agents.

Species-dependent pharmacological properties of the melanocortin-5 receptor.

The genes encoding the melanocortin-3 receptor and melanocortin-5 receptor have been cloned from rhesus monkey. Heterologous expression in CHO cells indicated species dependent in vitro pharmacological properties for the human and rhesus melanocortin-5 receptors. Several peptides including NDP-alpha-MSH, alpha-MSH, MT-II and ACTH1-24 are more potent at the rhesus melanocortin-5 receptor than the human melanocortin-5 receptor by more than 10-fold. In contrast, we found no species difference in pharmacological properties between the human and rhesus melanocortin-3 receptors. Such a species-dependent pharmacological difference for melanocortin-5 receptor appears to be an exception compared to other G protein-coupled receptors from human and rhesus monkey.

Use of bioluminescent aequorin for the pharmacological characterization of 5HT receptors.

A convenient functional assay for 5HT2a and 5HT2c receptors is reported utilizing the bioluminescent aequorin to detect intracellular calcium changes. Using this assay, the pharmacological properties of many 5HT ligands can be determined in a 96-well format. The data indicate that the aequorin detection method is superior to the inositol phosphate assay with regard to speed and scope. This system is also appropriate for kinetic studies of receptor desensitization. We showed that the human 5HT2c receptor desensitizes in a biphasic manner, with a fast desensitization of approximately 90% of the total response occurring within 15 minutes while the remaining 10% response remains for at least 3 hours.

Molecular interaction of Agouti protein and Agouti-related protein with human melanocortin receptors.

Agouti protein and the Agouti-related protein (AGRP) are antagonists of the melanocortin-3 receptor and melanocortin-4 receptor. Both proteins contain 10 cysteines in the C-terminal domain arranged in five disulfide bonds. One possible arrangement of the disulfide bonds predicts an octapeptide loop, and the chemical properties of four residues within this loop (residues 111-114 in human AGRP) bear striking resemblance to those of several melanocortin peptides, including alpha-MSH, MT-II, and SHU-9119. We showed that cyclic synthetic octapeptides based on the sequence of this loop from Agouti protein or human AGRP are functional antagonists of the human melanocortin-4 receptor. All peptides had a lower affinity for the melanocortin-3 receptor than for the melanocortin-4 receptor. Substitution of serines for cysteines resulted in linear peptides which had reduced binding affinities for both receptors. Mutational analysis of human AGRP indicated that its C-terminal domain is functionally equivalent to the intact human AGRP. The RFF111-113 triplet appears to be the most critical portion of AGRP in determining the binding affinity for both melanocortin-3 and melanocortin-4 receptors. These data strongly suggest that the loop defined by Cys-110 and Cys-117 is critical in determining the antagonist activity of human AGRP. Our data provide indirect evidence for the suggestion that the Cys-110 to Cys-117 octapeptide loop of human AGRP mimics the conformation of alpha-MSH, MT-II, and SHU-9119.

Localization of leptin binding domain in the leptin receptor.

The leptin receptor is a member of the class I cytokine receptor family and is involved in the control of appetite and body weight. The predicted amino acid sequence of the extracellular region of the cloned leptin receptor differs from that of many other cytokine receptors in that it contains two homologous segments representing potential ligand binding sites. After the analysis of various deletion and substitution mutants of the leptin receptor, we found that the first potential binding motif is not required for leptin binding and receptor activation, whereas modification of the second potential binding motif can lead to inactive receptor mutants. Further deletion analysis generated a minimal binding domain that retains high affinity leptin binding. The leptin binding domain thus has been localized to residues 323-640, which contain the second segment of cytokine receptor domain/fibronectin type 3 domain (residues 428-635). Coexpression of the active isoform of leptin receptor (OB-Rb) with an inactive mutant lacking high affinity leptin binding site led to suppression of the activity mediated by OB-Rb, suggesting that the leptin receptor may exist as a multimeric complex in the absence of leptin.

ART (protein product of agouti-related transcript) as an antagonist of MC-3 and MC-4 receptors.

The mRNA encoding an agouti related protein (ART) of unknown biochemical function was previously reported to be up-regulated in the hypothalamus of two genetically obese mouse strains. We have expressed human ART as a secreted protein in COS-7 cells, and show that recombinant ART is functionally active in inhibiting the binding of a radiolabeled alpha-melanocyte stimulating hormone (alpha-MSH) analog to the human melanocortin-3 (MC-3) and melanocortin-4 (MC-4) receptors, while it is not a potent inhibitor of the human melanocortin-5 (MC-5) receptor. ART is an antagonist of the human MC-3 and MC-4 receptors as determined in functional assay. ART appears to be approximately 100-fold more potent than agouti with reference to the MC-3R and MC-4R binding affinity. These data suggest that ART may be a physiological regulator of feeding behavior.

Molecular cloning of the dog beta 1 and beta 2 adrenergic receptors.

We report the isolation of the genes encoding the beta 1 and beta 2 adrenergic receptors from dog genomic DNA. Sequence analysis of both genes revealed intronless open reading frames of 473 and 415 amino acid residues, receptively. Heterologous expression of both receptors in CHO cells indicated that both receptors are functionally similar to the human homologs. Comparing the dog beta 1 and beta 2 adrenergic receptors, the beta 1 receptor appears to bind to G proteins more tightly than the beta 2 receptor. Heterologously expressed receptors provide a convenient system for evaluating novel receptor agonists and antagonists.

Characterization of the binding and activity of a high affinity, pseudoirreversible morpholino tachykinin NK1 receptor antagonist.

2(S)-((3,5-Bis(trifluoromethyl)benzyl)-oxy)-3(S)-phenyl-4-((3-oxo-1,2,4- triazol-5-yl)methyl)morpholine (L-742,694) is a selective morpholino tachykinin NK1 receptor antagonist that inhibits the binding of 125I-substance P to the human tachykinin NK1 receptor with a Kd = 37 pM. Increasing concentrations of L-742,694 added to cells 15 min prior to agonist progressively increase the apparent EC50 of substance P for inducing the synthesis of inositol phosphate in Chinese hamster ovary (CHO) cells expressing human tachykinin NK1 receptor and decrease the maximal level of stimulation observed. In contrast, addition of substance P and L-742,694 to the cells at the same time results in an increase in the EC50 for substance P with no decrease in the maximal level of stimulation. The compound also decreases the apparent number of binding sites for 125I-substance P observed by Scatchard analysis. Analysis of the binding of [3H]L-742,694 to the tachykinin NK1 receptor shows that it associates with the receptor with k(a) = 3.98 x 10(8) M(-1) min(-1), and dissociates with k(d) = 0.026 min(-1) and t1/2 = 27 min at 22 degrees C. The slow rate of dissociation of L-742,694 from the tachykinin NK1 receptor and the observation that altering the order of addition of antagonist and substance P attenuates the effect of the antagonist on the maximal activation suggest that L-742,694 is a competitive antagonist that can behave as a pseudoirreversible antagonist under some experimental conditions. L-742,694 has reduced affinity for tachykinin NK1 receptors in which alanine has been substituted for Gln165, His197 or His265 in transmembrane helices 4, 5 and 6, respectively. These three residues have previously been shown to be present in the binding site of tachykinin NK1 receptor antagonists of several structural classes. In addition, L-742,694 inhibits binding of the quinuclidine antagonist (2S,3S)-cis-2-(diphenyl methyl)-N-[(2-iodophenyl)-methyl]-1-azabicyclo[2.2.2]octane 3-amine ([125I]L-703,606) with the same affinity as it inhibits binding of 125I-substance P. These data indicate that L-742,694 binds to the same site within the transmembrane domain of the receptor as previously described competitive antagonists.

Relative contribution of polar interactions and conformational compatibility to the binding of neurokinin-1 receptor antagonists.

Based on single residue substitutions, previous studies suggested that Gln165, His197, and His265 of the neurokinin-1 receptor interact directly with many nonpeptide antagonists, including CP-96,345. To further test this model, all three residues have been substituted simultaneously with alanine. The Q165A-H197A-H265A triple mutant bound CP-96,345 and eight analogs with similar affinity (2-20 microM), even though the same series of compounds bound to the wild-type receptor with affinities over a range of 1000-fold. These observations correspond exactly to the prediction of the binding site model. The micromolar binding affinity of all tested CP-96,345 analogs for the triple mutant seems to reflect solely van der Waals interactions, which suggests a significant contribution of conformational compatibility (or shape complementarity) to binding affinity. The primary role of conformational compatibility in ligand binding was consistent with the observation that simply transferring the residues involved in polar interactions with beta2-agonists into the neurokinin-1 receptor did not lead to increased binding affinity for the beta2-agonists. Taken together, these results support a general principle of ligand-receptor binding in which specific polar interactions can take place only if the overall ligand conformation is compatible with the stereochemistry of the binding pocket. In addition, double-residue and triple-residue substitutions, in combination with single-residue substitutions, can provide an alternative route to reveal multiple interactions that may not be detectable by single-residue substitutions and represent a novel approach to examine ligand-receptor interactions in the absence of high-resolution structural data.

Functional STAT 1 and 3 signaling by the leptin receptor (OB-R); reduced expression of the rat fatty leptin receptor in transfected cells.

The leptin receptor (OB-R) bears homology to members of the class I cytokine receptor family. We demonstrate that leptin binding to OB-R stimulates formation of STAT-1 and STAT-3 complexes, thereby defining transcriptional motifs for genes that are under leptin control. Transfected fa OB-R bound leptin with equal affinity to that of wild type OB-R. fa OB-R abundance was about 7 fold reduced compared to control cells. Surprisingly, the low level of fa OB-R is fully capable of activating the STAT signal transduction pathway. We discuss plausible explanations for the obese phenotype in Zucker fatty rats.

Purification and reconstitution of a recombinant human neurokinin-1 receptor.

Recombinant human neurokinin-1 receptors expressed in insect cells have been purified to near homogeneity by sequential metal-chelating chromatography and gel filtration chromatography. The purified receptor consists of a single polypeptide with an apparent molecular weight of 50 kD as revealed by SDS gel electrophoresis, and exhibits a specific activity of 19 nmol of L-703,606 bound per mg of protein. Immunoblot experiments further confirm the identity of the stained protein band. The purified receptor binds the antagonist L-703,606 with an affinity similar to that of native human neurokinin-1 receptor, and binds the agonist substance P with an affinity similar to that of the low affinity state of uncoupled native receptor. The purified receptor can be reconstituted with membranes from uninfected insect cells, and the reconstitution results in an increased affinity for substance P, consistent with the reappearance of the high affinity state of the receptor for agonist in the presence of endogenous G proteins. These data indicate that the purified neurokinin-1 receptor is functional with respect to agonist and antagonist binding and G protein coupling.

Mechanistic hypotheses for the activation of G-protein-coupled receptors.

G-protein-coupled receptors (GPCR) play an important role in the functioning of the nervous and endocrine systems. Intensive efforts devoted to mutational analyses of GPCR have greatly enhanced our understanding of the receptor-G-protein signalling pathway. Nonetheless, the lack of receptor structural data mandates that the mechanism of receptor activation can only be probed indirectly through hypothesis testing. Mechanistic models at two different levels, a thermodynamic model and a structural model, are reviewed here. A general thermodynamic model can be used as a framework for analyzing macroscopic experimental data, while a probability function based on a helical-rotation model can potentially provide a link between microscopic conformational changes and macroscopic rate constants and equilibrium constants. Conformational induction and conformational selection are two inseparable consequences of the receptor-binding system. Mechanistic models based on the integration of mutational data, spectroscopic data, and model simulation will form a foundation upon which further experiments can be designed, and consideration of both thermodynamic and structural principles will provide coherent insights into the receptor-activation process.

Conformational compatibility as a basis of differential affinities of tachykinins for the neurokinin-1 receptor.

The neurokinin-1 receptor is characterized by sub-nanomolar affinity for substance P and 30-100 nM affinity for other substance P-related peptides, including neurokinin B and septide. We have characterized a neurokinin-1 receptor mutant containing a Y216A substitution in the fifth transmembrane segment. This receptor mutant binds substances P with sub-nanomolar affinity but loses much of its peptide discrimination capability, exhibiting 1-2 nM binding affinity for other tachykinins. Kinetic measurements of ligand binding indicate that the increased binding affinity of neurokinin B and septide for the Y216A mutant compared to the wild-type receptor is due to a 100-fold increase in the association rate constant without appreciable change in the dissociation rate constant. The substantially increased association rate constant for the Y216A mutant suggests that the mutant receptor is probably more flexible in accommodating the approaching peptide molecule. It is proposed that a major determinant of peptide specificity for the wild-type neurokinin-1 receptor is the overall conformational compatibility between the receptor and the ligand, rather than residue-specific interactions with the divergent N-terminal residues of different peptides. Furthermore, the highly conserved nature of Tyr-216 in the G protein coupled receptor family suggests that this residue may also play an important role in the receptor activation process in general.

Identification of a series of 3-(benzyloxy)-1-azabicyclo[2.2.2]octane human NK1 antagonists.

The synthesis and in vitro and in vivo evaluation of a series of 3-(benzyloxy)-1-azabicyclo-[2.2.2]octane NK1 antagonists are described. While a number of 3,5-disubstituted benzyl ethers afford high affinity, the 3,5-bis(trifluoromethyl)benzyl was found to combine high in vitro affinity with good oral activity. Detailed structure-activity relationship studies in conjunction with data from molecular modeling and mutagenesis work have allowed the construction of a model of the pharmacophore. Specific interactions that have been identified include an interaction between His-197 and one of the rings of the benzhydryl, a lipophilic pocket containing His-265 that the benzyl ether occupies, and a possible hydrogen bond between Gln-165 and the oxygen of the benzyl ether.

Molecular characterization of a common binding site for small molecules within the transmembrane domain of G-protein coupled receptors.

The cloning of over 100 members of the superfamily of G-protein coupled receptors has resulted in the identification and characterization of novel targets for therapeutic intervention. In addition, mutagenesis studies aimed at understanding the nature of the molecular interactions of these receptors with peptides and small molecules have led to advancements in the ability to synthesize novel therapeutic agents with high affinity and specificity for these receptors. These experiments have shown that there is a common binding site for small molecules within the transmembrane domain of G-protein coupled receptors regardless of the nature of the endogenous, physiologically relevant receptor agonist. Finally, the demonstration that the binding sites for peptide agonists and nonpeptide competitive antagonists are not necessarily identical has provided insights regarding the mechanism of competitive antagonism in these receptors.

Identification of residues involved in ligand binding to the neurokinin-2 receptor.

Several residues of the human neurokinin-2 receptor have been identified to be critical for the binding of peptide agonists and non-peptide antagonists. Amino acid substitutions in the first and second extracellular segments and the second transmembrane segment led to substantial reduction in peptide affinity without affecting the affinity of antagonist SR48968. These effects are identical to those observed for homologous residues in the neurokinin-1 receptor, suggesting that these three regions are involved in high-affinity peptide binding to both receptor subtypes. On the other hand, some conserved residues in the fourth to seventh transmembrane segments are required for peptide binding to only one receptor subtype but not both. The conserved nature and location of these receptor residues suggest that the distance between bound peptide and helices 4-7 varies depending on the receptor subtype. It is likely that the conformational compatibility between a ligand and a given receptor determines the magnitude of binding affinity, and thus receptor subtype selectivity. While many single-residue substitutions did not affect the binding affinity of the antagonist SR48968, two double mutants in the sixth and seventh transmembrane segments were found to reduce its affinity substantially. Therefore, receptor residues participate cooperatively in the binding of SR48968. These results demonstrate the usefulness of combining single-residue substitutions in studying and confirming the role of receptor residues in ligand binding. Finally, the overlapping nature of agonist and antagonist binding sites is consistent with the observation that substitutions of some residues modify the binding affinities of both peptide agonists and non-peptide antagonists.