Homodimerization of Drosophila Class A neuropeptide GPCRs: Evidence for conservation of GPCR dimerization throughout metazoan evolution.

While many instances of GPCR dimerization have been reported for vertebrate receptors, invertebrate GPCR dimerization remains poorly investigated, with few invertebrate GPCRs having been shown to assemble as dimers. To date, no Drosophila GPCRs have been shown to assemble as dimers. To explore the evolutionary conservation of GPCR dimerization, we employed an acceptor-photobleaching FRET methodology to evaluate whether multiple subclasses of Drosophila GPCRs assembled as homodimers when heterologously expressed in HEK-293 T cells. We C-terminally tagged multiple Drosophila neuropeptide GPCRs that exhibited structural homology with a vertebrate GPCR family member previously shown to assemble as a dimer with CFP and YFP fluorophores and visualized these receptors through confocal microscopy. FRET responses were determined based on the increase in CFP emission intensity following YFP photobleaching for each receptor pair tested. A significant FRET response was observed for each receptor expressed as a homodimer pair, while non-significant FRET responses were displayed by both cytosolic CFP and YFP expressed alone, and a heterodimeric pair of receptors from unrelated families. These findings suggest that receptors exhibiting positive FRET responses assemble as homodimers at the plasma membrane and are the first to suggest that Drosophila GPCRs assemble as homodimeric complexes. We propose that GPCR dimerization arose early in metazoan evolution and likely plays an important and underappreciated role in the cellular signaling of all animals.

Coordinated RNA-Seq and peptidomics identify neuropeptides and G-protein coupled receptors (GPCRs) in the large pine weevil Hylobius abietis, a major forestry pest.

Hylobius abietis (Linnaeus), or large pine weevil (Coleoptera, Curculionidae), is a pest of European coniferous forests. In order to gain understanding of the functional physiology of this species, we have assembled a de novo transcriptome of H. abietis, from sequence data obtained by Next Generation Sequencing. In particular, we have identified genes encoding neuropeptides, peptide hormones and their putative G-protein coupled receptors (GPCRs) to gain insights into neuropeptide-modulated processes. The transcriptome was assembled de novo from pooled paired-end, sequence reads obtained from RNA from whole adults, gut and central nervous system tissue samples. Data analysis was performed on the transcripts obtained from the assembly including, annotation, gene ontology and functional assignment as well as transcriptome completeness assessment and KEGG pathway analysis. Pipelines were created using Bioinformatics tools and techniques for prediction and identification of neuropeptides and neuropeptide receptors. Peptidomic analysis was also carried out using a combination of MALDI-TOF as well as Q-Exactive Orbitrap mass spectrometry to confirm the identified neuropeptide. 41 putative neuropeptide families were identified in H. abietis, including Adipokinetic hormone (AKH), CAPA and DH31. Neuropeptide F, which has not been yet identified in the model beetle T. castaneum, was identified. Additionally, 24 putative neuropeptide and 9 leucine-rich repeat containing G protein coupled receptor-encoding transcripts were determined using both alignment as well as non-alignment methods. This information, submitted to the NCBI sequence read archive repository (SRA accession: SRP133355), can now be used to inform understanding of neuropeptide-modulated physiology and behaviour in H. abietis; and to develop specific neuropeptide-based tools for H. abietis control.

Urbilaterian origin of paralogous GnRH and corazonin neuropeptide signalling pathways.

Gonadotropin-releasing hormone (GnRH) is a key regulator of reproductive maturation in humans and other vertebrates. Homologs of GnRH and its cognate receptor have been identified in invertebrates-for example, the adipokinetic hormone (AKH) and corazonin (CRZ) neuropeptide pathways in arthropods. However, the precise evolutionary relationships and origins of these signalling systems remain unknown. Here we have addressed this issue with the first identification of both GnRH-type and CRZ-type signalling systems in a deuterostome-the echinoderm (starfish) Asterias rubens. We have identified a GnRH-like neuropeptide (pQIHYKNPGWGPG-NH2) that specifically activates an A. rubens GnRH-type receptor and a novel neuropeptide (HNTFTMGGQNRWKAG-NH2) that specifically activates an A. rubens CRZ-type receptor. With the discovery of these ligand-receptor pairs, we demonstrate that the vertebrate/deuterostomian GnRH-type and the protostomian AKH systems are orthologous and the origin of a paralogous CRZ-type signalling system can be traced to the common ancestor of the Bilateria (Urbilateria).

Structural and functional divergence of growth hormone-releasing hormone receptors in early sarcopterygians: lungfish and Xenopus.

The evolutionary trajectories of growth hormone-releasing hormone (GHRH) receptor remain enigmatic since the discovery of physiologically functional GHRH-GHRH receptor (GHRHR) in non-mammalian vertebrates in 2007. Interestingly, subsequent studies have described the identification of a GHRHR(2) in chicken in addition to the GHRHR and the closely related paralogous receptor, PACAP-related peptide (PRP) receptor (PRPR). In this article, we provide information, for the first time, on the GHRHR in sarcopterygian fish and amphibians by the cloning and characterization of GHRHRs from lungfish (P. dolloi) and X. laevis. Sequence alignment and phylogenetic analyses demonstrated structural resemblance of lungfish GHRHR to their mammalian orthologs, while the X. laevis GHRHR showed the highest homology to GHRHR(2) in zebrafish and chicken. Functionally, lungfish GHRHR displayed high affinity towards GHRH in triggering intracellular cAMP and calcium accumulation, while X. laevis GHRHR(2) was able to react with both endogenous GHRH and PRP. Tissue distribution analyses showed that both lungfish GHRHR and X. laevis GHRHR(2) had the highest expression in brain, and interestingly, X. laevis(GHRHR2) also had high abundance in the reproductive organs. These findings, together with previous reports, suggest that early in the Sarcopterygii lineage, GHRHR and PRPR have already established diverged and specific affinities towards their cognate ligands. GHRHR(2), which has only been found in xenopus, zebrafish and chicken hitherto, accommodates both GHRH and PRP.

International Union of Basic and Clinical Pharmacology. LXXXVI. Orexin receptor function, nomenclature and pharmacology.

Orexin signaling is essential for normal regulation of arousal and behavioral state control and represents an attractive target for therapeutics combating insomnia. Alternatively termed hypocretins, these neuropeptides were named to reflect sequence similarity to incretins and their potential to promote feeding. Current nomenclature reflects these molecular and biochemical discovery approaches in which HCRT, HCRTR1, and HCRTR2 genes encode prepro-orexin, the orexin 1 receptor (OX(1)) and the orexin 2 receptor (OX(2))-gene names designated by the Human Genome Organization and receptor names designated by the International Union of Basic and Clinical Pharmacology. Orexinergic neurons are most active during wakefulness and fall silent during inactive periods, a prolonged disruption in signaling most profoundly resulting in hypersomnia and narcolepsy. Hcrtr2 mutations underlie the etiology of canine narcolepsy, deficiencies in orexin-producing neurons are observed in the human disorder, and ablation of mouse orexin neurons or the Hcrt gene results in a narcolepsy-cataplexy phenotype. The development of orexin receptor antagonists and genetic models targeting components of the orexin pathway have elucidated the OX(2) receptor-specific role in histamine-mediated arousal and the contribution of both receptors in brainstem pathways involved in vigilance state gating. Orexin receptor antagonists of varying specificity uncovered additional roles beyond sleep and feeding that include addiction, depression, anxiety, and potential influences on peripheral physiology. Combined genetic and pharmacological approaches indicate that orexin signaling may represent a confluence of sleep, feeding, and reward pathways. Selective orexin receptor antagonism takes advantage of these properties toward the development of novel insomnia therapeutics.

Allatostatin-C receptors in mosquitoes.

In the present work we describe the functional and molecular characterization of two Aedes aegypti allatostatin-C receptor paralogs (AeAS-CrA and AeAS-CrB) and provide a detailed quantitative study of the expression of the AS-C receptor genes in an adult insect. The tissue distribution of the two AS-C receptors differed significantly; the mRNA levels of AeAS-CrB in the Malpighian tubules were the highest detected, while transcripts for AeAS-CrA were relatively low in this tissue. In addition, the transcript levels of both receptors were different in the thoracic and abdominal ganglia, corpora allata (CA) and the testis of the male. In the CA, the AeAS-CrB mRNA levels were constant from 0 to 72 h after female emergence, while the AeAS-CrA levels increased at 72 h. To complement the receptor expression studies, we analyzed the tissue specificity for allatostatin-C mRNA in female mosquitoes. Expression was high in abdominal ganglia and brain. Transcript levels of allatostatin-C in the head of females were elevated at eclosion and there were no major changes during the first week of adult life or after blood feeding. Fluorometric Imaging Plate Reader (FLIPR) recordings of calcium transients in HEK293T cells transiently expressing both putative receptors showed that they both responded selectively to allatostatin-C stimulation in the nanomolar concentration range. However, the peptide showed slightly greater affinity for AeAS-CrB than AeAS-CrA. Our studies suggest that some of the pleiotropic effects of allatostatin-C in mosquitoes could be mediated by the different receptor paralogs. Transcriptional regulation of the AS-C receptors may not have a critical role in the changes of CA responsiveness to the peptide that we previously described.

G protein-coupled receptor genes in the FANTOM2 database.

G protein-coupled receptors (GPCRs) comprise the largest family of receptor proteins in mammals and play important roles in many physiological and pathological processes. Gene expression of GPCRs is temporally and spatially regulated, and many splicing variants are also described. In many instances, different expression profiles of GPCR gene are accountable for the changes of its biological function. Therefore, it is intriguing to assess the complexity of the transcriptome of GPCRs in various mammalian organs. In this study, we took advantage of the FANTOM2 (Functional Annotation Meeting of Mouse cDNA 2) project, which aimed to collect full-length cDNAs inclusively from mouse tissues, and found 410 candidate GPCR cDNAs. Clustering of these clones into transcriptional units (TUs) reduced this number to 213. Out of these, 165 genes were represented within the known 308 GPCRs in the Mouse Genome Informatics (MGI) resource. The remaining 48 genes were new to mouse, and 14 of them had no clear mammalian ortholog. To dissect the detailed characteristics of each transcript, tissue distribution pattern and alternative splicing were also ascertained. We found many splicing variants of GPCRs that may have a relevance to disease occurrence. In addition, the difficulty in cloning tissue-specific and infrequently transcribed GPCRs is discussed further.

Galanin and galanin receptors in human gliomas.

Galanin-like immunoreactivity (GAL-LI) and specific GAL binding sites have been shown to be widely distributed in the central nervous system (CNS) and in CNS tumors. GAL and its receptors have also been shown to be present in glial cells, but to date it is still unknown whether human gliomas produce GAL and express GAL receptors. In this study 20 brain tumors consisting of 15 glioblastomas, 4 meningiomas and 1 gliosarcoma were investigated for the presence of GAL-LI and GAL receptors. Immunofluorescence analysis revealed a dense network of GAL-LI positive cellular processes and cell bodies in 18 of the 20 tumors. In contrast, in vitro (125)I-labeled GAL receptor autoradiography showed substantial GAL binding in only 6 glioblastoma tissues. Reverse transcription-PCR analysis detected mRNA of all three known galanin receptors in the tumor tissues, with most tumors expressing multiple receptor subtypes. Pharmacological analysis of tumor membrane homogenates with GAL and the specific GAL receptor GalR2 agonist, AR-M1896, revealed that the GAL receptor GalR1 is most likely the receptor responsible for the observed GAL binding in the glioblastomas. No correlation could be found between GAL-LI, the level of GAL binding and proliferative activity as determined by immunostaining with the cell proliferation marker Ki-67.

Systematic G-protein-coupled receptor analysis in Drosophila melanogaster identifies a leucokinin receptor with novel roles.

Leucokinins are insect neuropeptides that stimulate hindgut motility and renal fluid secretion. Drosophila has a single leucokinin gene, pp, encoding the longest known leucokinin, Drosokinin. To identify its receptor, a genome-wide scan for G-protein-coupled receptors was performed in silico and candidate receptors identified by similarity to known tachykinin receptors. The deduced peptides were expressed, with a transgene for the calcium reporter aequorin, in S2 cells and only one gene (CG10626) encoded a protein that responded to Drosokinin. The properties of the heterologously expressed receptor (action through intracellular calcium with an EC(50) of 4 x 10(-11) m and a t(1/2) <1 s) match closely those reported for the action of Drosokinin on Malpighian (renal) tubules. Antibodies raised against the receptor identified known sites of leucokinin action: stellate cells of the Malpighian tubule, two triplets of cells in the pars intercerebralis of the adult central nervous system, and additional cells in larval central nervous system. Western blots and reverse transcription-PCR confirmed these locations, but also identified expression in male and female gonads. These tissues also displayed elevated calcium in response to Drosokinin, demonstrating novel roles for leucokinin. A functional genomic approach has thus yielded the first complete characterization of a leucokinin receptor in an insect.

Discrimination of galanin receptor subtypes in RINm5F cells by structurally different galanin radioligands.

Galanin (GAL) is a biologically active peptide that is involved in a variety of physiological functions. The purpose of this study was to evaluate whether porcine and rat galanin radioligands could be used as probes to discriminate GAL receptors (GALR) subtypes using a cell line, RINm5F, that express multiple GALR subtypes. Data from parallel equilibrium binding experiments using the same RINm5F membrane homogenates reveal that [125I]pGAL labels 20% more GALRs with a 2-fold lower affinity than those values identified when using [125I]rGAL. Competition studies using various GAL peptides showed different rank order of potencies depending on the radioligand used. Preincubation of RINm5F membranes with GppNHp, a non-hydrolizable GTP analog, prior to radioligand labeling suggests that a portion of GALRs is precoupled to G proteins. In addition, receptors labeled by [125I]rGAL appear more sensitive to GppNHp-induced uncoupling of G proteins than those labeled by [125I]pGAL. In conclusion, our data suggest that pGAL and rGAL radioligands define different pharmacological profiles of GALRs, and hence, these ligands can be used as pharmacological tools to discriminate GALR subtypes. Additionally, our data suggests that GALRs exist in a precoupled state with their respective G-proteins prior to interaction with the agonist.

Neuropeptides and neuropeptide receptors in the Drosophila melanogaster genome.

Recent genetic analyses in worms, flies, and mammals illustrate the importance of bioactive peptides in controlling numerous complex behaviors, such as feeding and circadian locomotion. To pursue a comprehensive genetic analysis of bioactive peptide signaling, we have scanned the recently completed Drosophila genome sequence for G protein-coupled receptors sensitive to bioactive peptides (peptide GPCRs). Here we describe 44 genes that represent the vast majority, and perhaps all, of the peptide GPCRs encoded in the fly genome. We also scanned for genes encoding potential ligands and describe 22 bioactive peptide precursors. At least 32 Drosophila peptide receptors appear to have evolved from common ancestors of 15 monophyletic vertebrate GPCR subgroups (e.g., the ancestral gastrin/cholecystokinin receptor). Six pairs of receptors are paralogs, representing recent gene duplications. Together, these findings shed light on the evolutionary history of peptide GPCRs, and they provide a template for physiological and genetic analyses of peptide signaling in Drosophila.

Galanin receptor subtypes.

The neuropeptide galanin, which is widely expressed in brain and peripheral tissues, exerts a broad range of physiological effects. Pharmacological studies using peptide analogues have led to speculation about multiple galanin receptor subtypes. Since 1994, a total of three G-protein-coupled receptor (GPCR) subtypes for galanin have been cloned (GAL1, gal2 and gal3). Potent, selective antagonists are yet to be found for any of the cloned receptors. Major challenges in this field include linking the receptor clones with each of the known physiological actions of galanin and evaluating the evidence for additional galanin receptor subtypes.

The GalR2 galanin receptor mediates galanin-induced jejunal contraction, but not feeding behavior, in the rat: differentiation of central and peripheral effects of receptor subtype activation.

The neuropeptide galanin mediates a diverse array of physiological functions through activation of specific receptors. Roles of the three recently cloned galanin receptors (GalRs) in rat intestinal contraction and food intake were examined using GalR-selective ligands and the results were compared with the pharmacological profiles of defined GalRs. The action profile of these ligands in jejunal contraction resembled only that of GalR2 and only a high level of GalR2 mRNA was detected in the tissue, supporting GalR2 as the receptor mediating jejunal contraction. The action profile for food intake in rats excluded GalR2, GalR3 and the putative pituitary galanin receptor as the 'feeding receptor', suggesting that either GalR1 or an unidentified GalR is responsible for mediating this function.

Cloned human and rat galanin GALR3 receptors. Pharmacology and activation of G-protein inwardly rectifying K+ channels.

The neuropeptide galanin has been implicated in the regulation of processes such as nociception, cognition, feeding behavior, and hormone secretion. Multiple galanin receptors are predicted to mediate its effects, but only two functionally coupled receptors have been reported. We now report the cloning of a third galanin receptor distinct from GALR1 and GALR2. The receptor, termed GALR3, was isolated from a rat hypothalamus cDNA library by both expression and homology cloning approaches. The rat GALR3 receptor cDNA can encode a protein of 370 amino acids with 35% and 52% identity to GALR1 and GALR2, respectively. Localization of mRNA by solution hybridization/RNase protection demonstrates that the GALR3 transcript is widely distributed, but expressed at low abundance, with the highest levels in the hypothalamus and pituitary. We also isolated the gene encoding the human homologue of GALR3. The human GALR3 receptor is 90% identical to rat GALR3 and contains 368 amino acids. Binding of porcine 125I-galanin to stably expressed rat and human GALR3 receptors is saturable (rat KD = 0.98 nM and human KD = 2.23 nM) and displaceable by galanin peptides and analogues in the following rank order: rat galanin, porcine galanin approximately M32, M35 approximately porcine galanin-(-7 to +29), galantide, human galanin > M40, galanin-(1-16) > [D-Trp2]galanin-(1-29), galanin-(3-29). This profile resembles that of the rat GALR1 and GALR2 receptors with the notable exception that human galanin, galanin-(1-16), and M40 show lower affinity at GALR3. In Xenopus oocytes, activation of rat and human GALR3 receptors co-expressed with potassium channel subunits GIRK1 and GIRK4 resulted in inward K+ currents characteristic of Gi/Go-coupled receptors. These data confirm the functional efficacy of GALR3 receptors and further suggest that GALR3 signaling pathways resemble those of GALR1 in that both can activate potassium channels linked to the regulation of neurotransmitter release.

Growth hormone-releasing factor mobilizes cytosolic free calcium through different mechanisms in two somatotrope subpopulations from porcine pituitary.

Porcine somatotropes can be separated by Percoll density gradient centrifugation into low (LD) and high density (HD) subpopulations that differ ultrastructurally and functionally. Here, we report the effects of growth hormone-releasing factor (GRF) on the cytosolic free calcium concentration ([Ca2+]i) of single LD and HD somatotropes. Resting [Ca2+]i in LD somatotropes was 2-fold higher than in HD cells. GRF induced [Ca2+]i increases in a similar percentage of somatotropes from both subsets. However, amplitude and kinetics of the responses were markedly different. In all responsive LD somatotropes, GRF evoked a rapid initial peak followed by a sustained plateau (plateau-type response). Blockade of extracellular Ca2+ entry by 3 mM EDTA, 2 mM CoCl2, or 100 microM verapamil completely abolished the plateau phase without affecting the initial Ca2+ spike. Conversely, only the plateau phase was preserved in thapsigargin (TG)-treated LD cells. The vast majority of GRF-responsive HD somatotropes exhibited a transient [Ca2+]i peak that returned gradually to baseline (transient-type response). This response was completely blocked by removal of extracellular Ca2+, whereas TG treatment had no effect. Taken together, our results indicate that the response of LD somatotropes to GRF depends on mobilization of Ca2+ of both extra- and intracellular origin, whereas that of HD somatotropes seems to be exclusively dependent on extracellular Ca2+ entry through L-type voltage sensitive Ca2+ channels (VSCC). These findings are the first to demonstrate a differential effect of GRF on Ca2+ mobilization in two somatotrope subpopulations, and suggest the existence of differences in the GRF receptor(s) expressed in each subpopulation and/or in the intracellular signalling pathways activated upon GRF binding.

Identification of a pituitary growth hormone-releasing peptide (GHRP) receptor subtype by photoaffinity labeling.

Hexarelin, an analogue of GHRP-6, in which D-Tryptophan has been replaced by its 2-methyl derivative, is known to release growth hormone (GH) in vivo and in vitro by direct action on receptors present in anterior pituitary cells. Measurement of second messengers (c-AMP, Ca++, IP3) upon somatotrophs stimulation, suggests the existence of more than one GHRP receptor subtype. In order to document such an hypothesis, we have used a new photoactivatable derivative of Hexarelin, Tyr-Bpa-Ala-Hexarelin. This derivative was shown to be fully active in the release of GH in vivo with neonate rats. Using this photoactivatable ligand, we have specifically labeled a protein with an apparent Mr of 57,000 in human, bovine and porcine anterior pituitary membranes. Hexarelin and the spiroindoline sulfonamide MK-0677 displaced the Mr-57,000 photolabeled band with an apparent ED50 of 6x10(-7) M and 2x10(-5) M respectively. Taking into account the high efficiency (>60%) of covalent incorporation of the Bpa residue, this photoactivatable Hexarelin derivative has allowed the identification of a pituitary GHRP receptor subtype, which is apparently distinct from the recently cloned GH secretagogue receptor.

Pharmacological studies of proctolin receptors on foregut and hindgut of Blaberus craniifer.

Proctolin (Arg-Tyr-Leu-Pro-Thr) and proctolin analogs modified at position 1, 2, or 5 caused dose dependent contractions of Blaberus fore- and hindgut. The varying contractile effects between both tissues revealed the possible presence of receptor subtypes as identified by [GABA1]-proctolin. A single population of binding sites (Kd approximately 100 nM) was deduced from Scatchard analysis. In addition, nanomolar concentrations of proctolin induced a dose-dependent hydrolysis of phosphoinositides (PIns) augmented by GTPgammaS (1 microM) on foregut membranes but no accumulation of cAMP. Proctolin induced contractions are likely mediated via a phospholipase C linked to a heptahelical receptor bound to heterotrimeric G-proteins.

Binding sites for growth hormone-releasing peptide.

Growth hormone-releasing peptides (GHRPs) are known to release growth hormone (GH) in vivo and in vitro by a direct action on receptors in anterior pituitary cells. Measurement of second messengers released following somatotroph stimulation suggests the existence of more than one GHRP receptor subtype in the hypothalamic-pituitary system. Furthermore, hexarelin, a hexapeptide of the GHRP family and a potent GH secretagogue, is reported to increase left ventricular ejection fraction, suggesting the expression of specific myocardial GHRP binding sites. In order to confirm such a hypothesis, a photoactivatable derivative of hexarelin, Tyr-p-benzoyl phenylalanine-Ala-hexarelin, was developed. A putative GHRP receptor with an apparent relative molecular mass of 57,000 was specifically labelled and characterized in human, bovine and porcine anterior pituitary membranes using this hexarelin derivative. The existence of myocardial binding sites was also demonstrated using the same approach. The differential binding affinity of GHRP analogues to cardiac tissue raises the possibility of the existence of distinct GHRP receptor subtypes in the pituitary and the cardiovascular system, for which physiological roles have yet to be determined.

Heterogeneity of motilin receptors in the gastrointestinal tract of the rabbit.

Motilin, a 22-amino acid peptide synthesized in endocrine cells of intestinal mucosa, stimulates GI smooth muscle contractility. To elucidate the mode of action of motilin, we attempted to determine whether motilin receptors are localized on nerve cells or on smooth muscle cells of the GI tract. Mucosa-free tissues from rabbit antrum and duodenum were homogenized separately with a Polytron prior to differential centrifugation to obtain synaptosome or plasma membrane-enriched fractions, as determined by the distribution of [3H]saxitoxin (SAX) binding (neural membranes) and 5' nucleotidase (5'N) activity (smooth muscle plasma membranes). Motilin binding was evaluated by the displacement of [125I]motilin by motilin (1-22) on the various membrane fractions. In the antrum, motilin binding was highly correlated with SAX binding (r = 0.81, p < 0.0005), and also significantly with 5'N activity (r = 0.54, p < 0.05). In the duodenum, motilin binding correlated significantly with 5'N activity (r = 0.67, p < 0.005), but not with SAX binding (r = -0.11, NS). Receptor affinity, for the motilin antagonist MOT(1-12)[CH2NH]10-11, for motilin(1-22), and for the motilin agonist erythromycin lactobionate was significantly (p < 0.001, p < 0.001, and p < 0.05, respectively) higher in SAX-enriched fractions from the antrum than in 5'N-enriched fractions from the duodenum. Therefore, in the rabbit: 1) motilin receptors appear to be predominantly located on nerve tissues in the antrum and restricted to smooth muscle cells in the duodenum, and 2) antral receptors and duodenal receptors displayed different pharmacological characteristics, probably corresponding to two specific and heterogeneous motilin receptor subtypes.