Identification of Peptide Ligand-binding Domains within the Human Motilin Receptor Using Photoaffinity Labeling*

The cDNA encoding the human motilin receptor was recently cloned and found to represent a G protein-coupled receptor that is structurally related to the growth hormone secretagogue receptors. Together, these represent a new Class I receptor family. Our aim in the present work is to gain insight into the molecular basis of binding of motilin to its receptor using photoaffinity labeling. To achieve this, we developed a Chinese hamster ovary cell line that overexpressed functional motilin receptor (CHO-MtlR; 175,000 sites per cell, with K i = 2.3 ± 0.4 nmmotilin and EC50 = 0.3 ± 0.1 nm motilin) and a radioiodinatable peptide analogue of human motilin that incorporated a photolabilep-benzoyl-l-phenylalanine (Bpa) residue into its pharmacophoric domain. This probe, [Bpa1,Ile13]motilin, was a full agonist at the motilin receptor that increased intracellular calcium in a concentration-dependent manner (EC50 = 1.5 ± 0.4 nm). This photolabile ligand bound specifically and with high affinity to the motilin receptor (K i = 12.4 ± 1.0 nm), and covalently labeled that molecule within its M r = 45,000 deglycosylated core. Cyanogen bromide cleavage demonstrated its covalent attachment to fragments of the receptor having apparent M r = 6,000 and M r = 31,000. These were demonstrated to represent fragments that included both the first and the large second extracellular loop domains, with the latter representing a unique structural feature of this receptor. The spatial approximation of the pharmacophoric domain of motilin with these receptor domains support their functional importance as well.

Motilin has long been recognized as an important endogenous peptide regulator of gastrointestinal motor function (1,2). The receptor for this hormone has also been demonstrated to represent a clinically useful pharmacological target, activated by erythromycin (3,4). While there have been extensive physiological and pharmacological studies of motilin action, including extensive analysis of peptide structure-activity relationships (5)(6)(7), the motilin receptor (MtlR) 1 has been difficult to isolate and biochemically characterize.
Using a unique high throughput screen of compounds that can interact with cloned orphan receptors, the cDNA encoding the human motilin receptor (originally isolated as orphan clone GPR38) was finally identified in 1999 (8). Sequence analysis demonstrated that this represents a member of a new subgroup within the Class I family of G protein-coupled receptors that is defined by structural homology with a series of growth hormone secretagogue receptors (9,10). The motilin receptor is 52% identical to the human growth hormone secretagogue receptor, with 86% sequence identity in the predicted transmembrane domains (8,10). Because this group of receptors has only been recognized for a short time, no specific receptor domains of importance for binding or activation have yet been identified. Better understanding of the molecular basis of activation of this receptor by its natural ligand may provide important insights for drug development.
Many existing structure-activity studies have focused on the relationship between the structure of motilin, the natural peptide ligand, and its ability to bind to or to stimulate contractile activity of natural motilin receptor-bearing gastrointestinal smooth muscle tissue (5-7, 11, 12). These studies have demonstrated that the amino-terminal portion of motilin, including residues 1-9, is devoid of any activity, while extension of this domain beyond the first nine residues restores binding and biological activity (6,8). Such observations have established that the pharmacophoric domain of this hormone represents its amino-terminal decapeptide. The carboxyl-terminal region of motilin forms an ␣-helix that is thought to stabilize the interaction of the critical amino-terminal residues at the active site of the receptor (7).
In this work, we have attempted to develop the tools necessary to begin to gain an understanding of the molecular basis of motilin binding to its receptor. This included the establishment of a cell line that expresses a high density of functional motilin receptors and the development of a radioiodinatable motilin analogue that incorporates a photolabile residue within this critical pharmacophoric domain of the natural peptide ligand. In the present work, we have developed and fully characterized these reagents and have applied them to the photoaffinity labeling of the motilin receptor and to the initial identification of distinct subdomains within that molecule. The [Bpa 1 ,Ile 13 ]motilin probe was shown to covalently label the motilin receptor within two cyanogen bromide fragments, with one including its large second extracellular loop, a structurally unique and functionally important region.

MATERIALS AND METHODS
Reagents-The solid phase oxidant, N-chlorobenzenesulfonamide (Iodo-beads), was from Pierce Chemical Co. Endoglycosidase F was prepared in our laboratory, as we have reported (13). Fura-2AM was from Molecular Probes (Eugene, OR). Other reagents were analytical grade. The wild type human motilin receptor cDNA was kindly provided by Merck Research Laboratories, Rahway, NJ.
Receptor Preparations-We established a Chinese hamster ovary (CHO) cell line stably expressing the wild type human motilin receptor (CHO-MtlR) by transfecting CHO-K1 cells with the motilin receptor cDNA construct (8). A clonal cell line was selected by repeated cycles of limiting dilution, screening for motilin radioligand binding (as described below). This cell line was used as a source of receptor for the current study. Cells were cultured at 37°C on tissue culture plasticware in Ham's F-12 medium supplemented with 5% Fetal Clone-2 (Hyclone Laboratories, Logan, UT). Cells were passaged twice a week and were lifted mechanically after incubation in cell dissociation medium before membrane preparation or study of intact cells. Enriched plasma membranes from these cell lines were prepared as previously described (14).
Mutant motilin receptor constructs were prepared using an oligonucleotide-directed approach (QuikChange site-directed mutagenesis system, Stratagene). Two such constructs were prepared for this work, representing replacement of motilin receptor residues Leu 98 and Ile 185 with Met residues (L98M and I185M motilin receptors). These residues were in positions predicted to reside in the second transmembrane segment and in the second extracellular loop domain. Sequences of these constructs were confirmed by direct DNA sequencing (15). These receptor constructs were expressed transiently on COS cells after transfection using a modification of the DEAE-dextran method (16). Sixty hours after transfection, cells were lifted mechanically, and enriched plasma membranes were prepared using the method previously described (14).
Synthesis of Peptides-Human motilin-(1-22) (motilin), [Ile 13 ]motilin, and [Bpa 1 ,Ile 13 ]motilin peptides ( Fig. 1) were synthesized by manual solid-phase techniques using Boc-protected amino acids, hydroxybenzotriazol, and Pam resin. D,L-Bpa was prepared as described (17,18). The synthetic products were purified to homogeneity by sequential reversedphase HPLC separations (19). Identities of the peptides were determined by quantitative amino acid analysis and mass spectrometry. L-and D-Bpa stereoisomers were identified using the technique of Miller and Kaiser (20), involving the full cleavage of the peptide containing the L form and limited or absent cleavage of the peptide containing the D form with amino peptidase M. The L-Bpa peptide was used in the current studies. 13 ]motilin and [Bpa 1 ,Ile 13 ]motilin were radioiodinated oxidatively on Tyr residues in position 7 of each peptide using Na 125 I and a 15 s exposure to the solid phase oxidant, N-chlorobenzene-sulfonamide (Iodo-beads). The products were purified to homogeneity by reversed-phase HPLC to yield specific radioactivities of 2,000 Ci/mmol (19).

Radioiodination of Probes-[Ile
Biological Activity Determination-The agonist activities of motilin and [Bpa 1 ,Ile 13 ]motilin were studied using an assay for stimulation of intracellular calcium activity in the CHO-MtlR cell line or COS cells expressing motilin receptor mutants. In this assay, ϳ2 million cells were loaded with 5 M Fura-2AM (Molecular Probes, Eugene, OR) in Krebs-Ringer-HEPES medium (KRH) containing 25 mM HEPES, pH 7.4, 104 mM NaCl, 5 mM KCl, 1.2 mM MgSO 4 , 2 mM CaCl 2 , 1 mM KH 2 PO 4 , 0.2% bovine serum albumin, and 0.01% soybean trypsin inhibitor. They were incubated for 20 min at 37°C. After washing, cells were stimulated with variable concentrations of motilin or [Bpa 1 ,Ile 13 ]motilin at 37°C. Fluorescence was quantified in a Perkin-Elmer LS50B luminescence spectrometer (Norwalk, CT). Excitation was performed at 340 and 380 nm, and emissions were determined at 520 nm, with calcium concentration calculated from the ratios, as described by Grynkiewicz et al. (21). The peak intracellular calcium transient was utilized to determine the agonist concentration dependence of the biological responses. All assays were performed in duplicate and repeated at least three times in independent experiments.
Receptor Binding Studies-Radioligand binding assays utilized intact motilin receptor-bearing cells or an enriched plasma membrane fraction of these cells (1-10 g per tube), 125 I-[Ile 13 ]motilin (3-5 pM radioligand) and KRH medium. Incubations were for 60 min at 25°C. The intact cell binding assay was performed in a 24-well tissue culture plates, while the membrane binding assay utilized separation of bound and free radioligand using the Skatron apparatus (Molecular Devices, Sunnyvale, CA) and receptor-binding filtermats. Nonspecific binding was determined in the presence of 1 M motilin and represented less than 20% of total binding.
Photoaffinity Labeling of the Motilin Receptor-Photoaffinity labeling was performed using enriched plasma membrane preparations (100 g per tube) and the radioiodinated photolabile analogue of motilin described above (100 pM). Incubations were performed for 60 min in the dark at 25°C. After binding, membranes were exposed to photolysis for 30 min at 4°C using a Rayonet Photochemical Reactor (Southern New England Ultraviolet, Hamden, CT) equipped with 3500 Å lamps. For selected experiments, the affinity-labeled motilin receptor and its relevant fragments were deglycosylated with endoglycosidase F under the conditions we previously reported (22). Membranes were then washed, solubilized, and separated by gel electrophoresis using either Laemmli conditions with 10% SDS-polyacrylamide gels or 10% NuPAGE gels (InVitrogen, Carlsbad, CA). Labeled products were visualized by autoradiography.
Chemical Cleavage of the Labeled Motilin Receptor-Gel-purified, affinity-labeled native and deglycosylated motilin receptor were digested with cyanogen bromide in 70% formic acid, according to the procedure previously described (22). The products of cleavage were separated on 10% NuPAGE gels using MES running buffer, with labeled products visualized by autoradiography. The apparent molecular masses of radiolabeled receptor fragments were determined by interpolation on a plot of the mobility of Multimark™ protein standards (In-Vitrogen) versus the log values of their apparent masses.
Statistical Analysis-All observations were repeated at least three times in independent experiments and are expressed as means Ϯ S.E. Binding curves were analyzed using the LIGAND program of Munson and Rodbard (23) and were plotted using the nonlinear regression analysis routine for radioligand binding in the Prism software package (GraphPad Software, San Diego, CA).

RESULTS
Binding and Biological Activity Studies with Natural Ligand-A clonal cell line stably expressing motilin receptors was established (CHO-MtlR). The binding of motilin to the human motilin receptors expressed on these cells was specific, saturable, and high affinity, with a K i of 2.3 Ϯ 0.4 nM motilin (Fig. 2). The calculated receptor density on these cells was 175,000 sites per cell.
Characterization of the Probe-[Bpa 1 ,Ile 13 ]motilin contains a photolabile Bpa in the position of Phe 1 as a site for covalent attachment to the receptor and a Tyr residue that is naturally present in position 7 as a site for radioiodination. An Ile residue was incorporated in the position of Met 13 to eliminate a site for potential oxidative damage during radiolabeling. This motilin analogue was synthesized by manual solid-phase techniques, purified by reversed-phase HPLC, and characterized chemically by mass spectrometry. The L-and D-Bpa stereoisomers of this probe were separated and identified (20), with the L-stereoisomer used in the described studies.
The binding of [Bpa 1 ,Ile 13 ]motilin to the human motilin receptor was saturable and specific, but displayed a slightly lower affinity (K i ϭ 12.4 Ϯ 1.0 nM) than natural motilin (Fig. 3). This analogue was a full agonist, inducing a rise in intracellular calcium concentration in CHO-MtlR cells in a concentrationdependent manner with an EC 50 ϭ 1.5 Ϯ 0.2 nM (Fig. 3). Photoaffinity Labeling of the Motilin Receptor-Membranes prepared from CHO-MtlR were used as source of receptor for affinity labeling. The 125 I-[Bpa 1 ,Ile 13 ]motilin analogue covalently saturably labeled two membrane proteins from these cells that migrated on an SDS-polyacrylamide gel at M r ϭ 78,000 and at M r ϭ 58,000, with the former representing greater than 90% of labeling (Fig. 4). Photoaffinity labeling of each of these bands was inhibited in a concentration-dependent manner with increasing concentrations of competing unlabeled motilin (IC 50 between 1 and 10 nM motilin). Each band had similar apparent affinity for this probe, raising the possibility that they represented distinct glycoforms of a single protein, representing the motilin receptor. Deglycosylation of each of the glycoprotein bands with endoglycosidase F resulted in bands that migrated similarly at approximate M r ϭ 45,000 (Fig. 5). This corresponds with the expected mass of the core receptor protein. As control, labeled bands of this size were absent in affinity-labeled CHO cell membranes lacking the human motilin receptor.
Active Site Identification- Fig. 6 shows the ten fragments that would theoretically be obtained after cyanogen bromide cleavage of the human motilin receptor, ranging in mass from less than 1 to 12 kDa. Conditions for quantitative cleavage of this receptor with cyanogen bromide were established and applied. The spectrum of molecular masses and the presence of glycosylation at distinct sites within the amino terminus and second extracellular loop may, therefore, provide suggestive evidence for the identification of the domain being labeled using this single manipulation.
Cyanogen bromide cleavage of the M r ϭ 78,000 band covalently labeled by 125 I-[Bpa 1 ,Ile 13 ]motilin resulted in two distinct labeled bands (Fig. 6). These had apparent migrations of M r ϭ 6,000 and M r ϭ 31,000. The migration of the M r ϭ 6,000 fragment did not shift further after deglycosylation with endoglycosidase F. Similar treatment of the M r ϭ 31,000 fragment resulted in its shifting to apparent M r ϭ 21,000. Considering the mass of the covalently bound ligand probe (2,909 Da) and the glycosylation status of the labeled bands, the most probable identifications of these bands were the cyanogen bromide fragments that included the first and second extracellular loop domains.
To further establish these identities, two motilin receptor mutants were constructed that introduced additional sites of cyanogen bromide cleavage at new Met residues within each of these candidate domains. These included L98M and I185M motilin receptor constructs. Fig. 7 shows that motilin bound with normal affinity (K i values of 3.4 Ϯ 2.5 nM motilin at the L98M receptor and 7.0 Ϯ 0.8 nM motilin at the I185M receptor) and stimulated a normal intracellular calcium concentration response in both of these receptor constructs (EC 50 values of 1.8 Ϯ 0.6 nM motilin at the L98M receptor and 3.7 Ϯ 0.6 nM motilin at the I185M receptor).
The 125 I-[Bpa 1 ,Ile 13 ]motilin probe efficiently and saturably labeled the wild type and mutant motilin receptor constructs expressed on the COS cells (Fig. 8). Glycosylation of the motilin receptor was different in the COS cells from that in the CHO cell line, with the labeled glycosylated cyanogen bromide fragment in the COS cells migrating as a broad band between M r ϭ 42,000 -52,000 rather than at M r ϭ 31,000 (observed in the CHO cell line). Each of the non-glycosylated fragments from both cell types migrated similarly. Fig. 8 shows the effects of cyanogen bromide cleavage of the labeled receptor constructs. As predicted, the labeled M r ϭ 6,000 band from the wild type receptor shifted to approximately M r ϭ 5,000 in the L98M receptor. Also as expected, the labeled glycosylated M r ϭ 42,000 band from the wild type receptor shifted to a nonglycosylated M r ϭ 12,000 band in the I185M receptor construct. These results definitively confirmed the identities of the labeled cyanogen bromide fragments of the motilin receptor and refined the labeled domains. The domains labeled include the region between receptor residues 99 and 129 and that between residues 130 and 185. DISCUSSION The G protein-coupled receptor superfamily is remarkable for the diversity of structures of its natural agonist ligands, ranging from extremely small photons, odorants, and biogenic showing the labeling of intact receptors, shows the saturable labeling of these molecules and the differential migration of the glycoforms in CHO and COS cells. In the bottom panel, cyanogen bromide cleavage of the labeled wild type receptor yielded a non-glycosylated band migrating at approximate M r ϭ 6,000 and a glycosylated band migrating at approximate M r ϭ 42,000. The former band shifted to approximate M r ϭ 5,000 in the L98M receptor, confirming the region between receptor residues 99 and 129 as being the labeled domain. The labeled glycosylated M r ϭ 42,000 band from the wild type receptor shifted to a non-glycosylated band migrating at M r ϭ 12,000 in the I185M receptor (although the cleavage was not fully complete). This confirmed that the second region of labeling was between residues 130 and 185 of the receptor. amines to larger peptides, glycoproteins, and even viral particles. It is well recognized that these receptors can be clustered based on sequence homology, with structurally related groups often representing targets of structurally related ligands (24). It has been assumed that themes of ligand binding are analogous in closely related receptors within this superfamily.
There is substantial structural homology between the motilin receptor and the previously cloned growth hormone secretagogue receptors (8 -10). This is most marked in the predicted transmembrane domains, where these receptors share 86% sequence identity. With the recent identification of natural ligands for one of the growth hormone secretagogue receptors (a 28-amino acid peptide, ghrelin, and its splice variant representing des-Gln 14 -ghrelin) (25,26), a structural relationship between these ligands and motilin also became clear. Alignment of motilin with ghrelin revealed that 8 of its 22 amino acid residues were identical (motilin residues Phe 5 , Glu 9 , Gln 11 , Arg 12 , Gln 14 , Lys 16 , Glu 17 , and Lys 20 ). Of interest, these were distributed throughout motilin, with several of these residues present within the amino-terminal pharmacophoric domain of this hormone.
In this work, we have established a CHO-MtlR cell line that overexpresses the fully functional recombinant human motilin receptor, and we have developed a radioiodinatable analogue of motilin that incorporates a photolabile residue within the pharmacophoric domain, in position 1 of this peptide. This probe represents a full agonist that binds to the motilin receptor with high affinity and that efficiently covalently labels two glycoforms of this molecule that are expressed on these cells. It efficiently labeled two cyanogen bromide fragments of the receptor that include the first and second extracellular loop domains. This is consistent with the pharmacophoric domain of motilin being held in close proximity to these regions of the receptor. These data provide the first direct evidence for molecular approximations between a residue within a motilin-like agonist and this receptor, and begin the definition of important domains within this receptor.
It is of particular interest that alignment of the motilin receptor with the growth hormone secretagogue receptors reveals that this large second extracellular loop domain that is covalently labeled in this study is absent in the growth hormone secretagogue receptors. There is a 40-residue insertion in the motilin receptor relative to those receptors in this region. An analogous region can, therefore, not be involved in the binding of ghrelin ligands to those receptors.
It is noteworthy that ghrelin and des-Gln 14 -ghrelin both have a unique post-translational modification, representing N-octanoylation of Ser 3 , that has been shown to be necessary for their biological activity, the pulsatile secretion of growth hormone (25,26). No such modification has been described for motilin. It is quite possible that this biologically essential fatty acid acylation of the natural ligands for the growth hormone secretagogue receptors establishes a distinct mode of binding from that observed for the receptor that normally interacts with non-acylated motilin. The incorporation of a modification that is so hydrophobic may bring these ligands closer to the lipid bilayer for their sites of recognition than for a non-modified peptide ligand.
The active sites of G protein-coupled receptors are of substantial interest for rational drug design. The identification of the ligand-binding domain of the human motilin receptor may lead to the development of more selective and potent agonists acting at this important receptor. Particularly in view of the well described gastrointestinal prokinetic properties of motilin and of the established clinical efficacy for the non-peptidyl motilin receptor ligand, erythromycin (4), new compounds may be of potential benefit to patients with diabetic gastroparesis (3) or other disturbances of normal food propulsion.