Characterization of deletion and truncation mutants of the rat glucagon receptor. Seven transmembrane segments are necessary for receptor transport to the plasma membrane and glucagon binding.

Glucagon receptor mutants were characterized with the aim of elucidating minimal structural requirements for proper biosynthesis, ligand binding, and adenylyl cyclase coupling. One N-terminal deletion mutant and five truncation mutants with progressively shorter C termini were expressed in transiently transfected monkey kidney (COS-1) cells. Each truncation mutant was designed so that the truncated C-terminal tail would remain on the cytoplasmic surface of the receptor. In order to characterize the cellular location of the expressed receptor mutants, a highly specific, high affinity antipeptide antibody was prepared against the extracellular, N-terminal tail of the receptor. Immunoblot analysis and immunofluorescence microscopy showed that the presence of all seven putative transmembrane segments, but not not an intact N-terminal tail, was required for cell surface expression of the receptor. Membranes from cells expressing receptor mutants lacking a large portion of the N-terminal tail or any of the seven putative transmembrane segments failed to bind glucagon. Membranes from cells expressing the C-terminal tail truncation mutants, which retained all seven transmembrane segments, bound glucagon with affinities similar to that of the native receptor and activated cellular adenylyl cyclase in response to glucagon. These results indicate that all seven helices are necessary for the proper folding and processing of the glucagon receptor. Glycosylation is not required for the receptor to reach the cell surface, and it may not be required for ligand binding. However, the N-terminal extracellular portion of the receptor is required for ligand binding. Most of the distal C-terminal tail is not necessary for ligand binding, and the absence of the tail may increase slightly the receptor binding affinity for glucagon. The C-terminal tail is also not necessary for adenylyl cyclase coupling and therefore does not play a direct role in G protein (GS) activation by the glucagon receptor.

The receptors for the glycoprotein hormones luteinizing hormone, follicle-stimulating hormone, thyroid-stimulating hormone, and chorionic gonadotropin contain very large extracellular N-terminal domains of over 300 amino acids (Salesse et al., 1991). The specificity of gonadotropin receptors is determined by high affinity hormone binding to N-terminal leucinerich repeats. Activation of the receptor may then proceed after the bound ligand interacts with the transmembrane segments of the receptor (Braun et al., 1991). Thus, the glycoprotein hormone receptors use the N-terminal portion of the receptor as well as other extracellular loops and the transmembrane helices to bind their ligands. In contrast, the photopigment rhodopsin (Zhukovsky et al., 1991) and the ␤-adrenergic receptors (Strader et al., 1989) use only the transmembrane helices to bind chromophore or agonist ligands, respectively. These receptors have considerably smaller N-terminal extensions compared with those of the glycoprotein hormone receptors (Probst et al., 1992).
According to tentative structural models, the hormone-binding site of the glucagon receptor probably consists of a contribution from the large extracellular domain of the receptor (Fig.  1), which includes the N-terminal tail and loops connecting transmembrane helices. However, transmembrane signaling must involve ligand-mediated communication between the extracellular domain and the intracellular domain where heterotrimeric G proteins are activated. To investigate the molecular mechanism of hormone-receptor interaction and of receptor activation, we previously designed and synthesized a gene for the rat glucagon receptor. COS cell membranes expressing the synthetic receptor gene bound glucagon with high affinity and displayed the appropriate peptide hormone specificity. The transfected COS cells also showed increased intracellular cAMP levels in response to glucagon (Carruthers et al., 1994).
In order to evaluate the minimal structural requirements for glucagon binding and for adenylyl cyclase activation, a series of six site-directed glucagon receptor mutants was prepared (Fig.  1). Mutants T7b and T7a had portions of the C-terminal tail removed. Mutants T5, T3, and T1 consisted of the N terminus followed by five, three, or one transmembrane helix, respectively. In addition to the truncated receptors, mutant D1 contained a 96-residue deletion from the N terminus, such that the seven helices were intact, but none of the N-linked glycosylation sites remained. The mutant receptor genes were expressed in COS cells and studied by immunoblot analysis, glycosidase treatment, immunofluorescence microscopy, competitive displacement ligand-binding assays, and adenylyl cyclase activation assays. The results showed that to reach the plasma membrane, the receptor did not need to be glycosylated but did require all seven helices. The N terminus was shown to be important for ligand binding, but the detailed role of the transmembrane helices for ligand binding remains to be elucidated. In contrast to the N terminus, most of the C terminus was not necessary either for the binding of glucagon or for activating adenylyl cyclase.

Construction of the Glucagon Receptor Mutant Genes-Construction
of the synthetic rat glucagon receptor gene (Gen Bank /EMBL Data Bank accession number U14012) has been reported (Carruthers et al., 1994(Carruthers et al., , 1995. Site-directed mutagenesis of the gene was performed using restriction fragment replacement "cassette mutagenesis" (Lo et al., 1984) in a modified pGEM2 (Invitrogen) cloning vector (Carruthers et al., 1994). Oligonucleotide synthesis was carried out on an Applied Biosystems model 392 synthesizer. Purification and characterization of synthetic DNA was performed essentially as described (Ferretti et al., 1986;Sakmar and Khorana, 1988). General methods for the cloning of synthetic duplexes to prepare mutant receptor genes were carried out as described (Carruthers et al., 1994(Carruthers et al., , 1995. Mutant receptors are shown schematically in Fig. 1. The N-terminal deletion mutant D1 was prepared by replacing a 366-bp EcoRI-MluI restriction fragment with a synthetic duplex that deleted the nucleotides corresponding to amino acid residues 20 -115. The truncation mutant T1 was prepared by replacing a 77-bp BssHII-BstEII restriction fragment with a synthetic duplex that substituted sequence coding for amino acid residues 172-178 with two successive stop codons. The truncation mutant T3 was prepared by replacing a 48-bp BspDI-NsiI restriction fragment with a synthetic duplex that substituted sequence coding for amino acid residues 258 -268 with a sequence coding for the tetrapeptide RKLH followed by two successive stop codons. The tetrapeptide, which is found as the cytoplasmic border of the first transmembrane domain, was introduced because the cytoplasmic boundary of the third transmembrane segment was not well defined. The truncation mutant T5 was prepared by replacing a 238-bp SacI-BspE1 restriction fragment with a synthetic duplex that replaced residues 341-412 with two stop codons. The truncation mutant T7a was prepared by replacing a 65-bp BspEI-NcoI restriction fragment with a duplex substituting sequence coding for amino acid residues 416 -432 with two stop codons. To prepare truncation mutant T7b, a 168-bp NcoI-NotI restriction fragment encoding amino acid residues 443-485 was replaced with a duplex containing two stop codons. All cloned synthetic sequences were confirmed by dideoxy sequencing on double-stranded plasmid DNA (Sequenase, U.S. Biochemical Corp.) (Sanger et al., 1977). A clone for each mutant with the correct DNA sequence was transferred into the eukaryotic expression vector, pMT3, as an EcoRI-NotI restriction fragment (Franke et al., 1988).
Peptide Synthesis and Preparation of Anti-glucagon Receptor Antibody-A rabbit polyclonal anti-peptide antibody was prepared against peptide DK-12, which corresponds to a dodecapeptide sequence in the extracellular N-terminal tail of the receptor (Fig. 1). Peptide synthesis, conjugation, rabbit immunization, and affinity purification of antiserum were carried out as described previously (Goldsmith et al., 1987;Carruthers et al., 1994). A complete description of anti-peptide, antireceptor antibodies is presented elsewhere (Unson et al., in press).

Expression of Glucagon Receptor and Mutant
Receptor Genes in COS-1 Cells-Receptor genes were expressed transiently in COS-1 cells according to the DEAE-dextran procedure previously reported for the expression of rhodopsin (Oprian et al., 1987;Sakmar et al., 1989).
Preparation of Membranes from Transfected COS-1 Cells-Cells from six 100-mm culture plates were washed in phosphate-buffered saline, followed by the addition of a hypotonic buffer (1 mM Tris-HCl, pH 6.8, 0.1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 5 g/ml leupeptin, 0.7 g/ml pepstatin, 10 mM EDTA) at 4°C. The cells were collected in a 15-ml Falcon tube and spun for 1 min in a clinical centrifuge. The cell pellet was resuspended in 750 l of hypotonic buffer and forced through a 26-gauge needle three times. The lysate was layered onto 750 l of a 37.7% (w/v) sucrose solution in Buffer A (20 mM Tris-HCl, pH 6.8, 150 mM NaCl, 1 mM MgCl 2 , 1 mM CaCl 2 , 10 mM EDTA) in an 11 ϫ 34-mm TLS-55 centrifuge tube and spun at 22,000 rpm for 20 min at 4°C. The interface band containing the membrane fraction was collected into a 1-ml syringe with a 25-gauge needle, and the volume was brought to 3.5 ml with Buffer A supplemented with 0.1 mM phenylmethylsulfonyl fluoride in a 13 ϫ 51-mm TLA-100.3 centrifuge tube. The membrane fraction was spun at 60,000 rpm for 30 min at 4°C, and the pellet was resuspended in 3.5 ml of Buffer A. The membrane pellet was resuspended and spun a second time, and the washed pellet was resuspended in 0.5 ml of Buffer A, frozen in liquid nitrogen, and stored in 0.1-ml aliquots at Ϫ80°C. The protein content of each sample was determined by the modified Bradford method (Bio-Rad).
Treatment of Expressed Glucagon Receptor and Mutants with Glycosidases-For cleavage with N-glycosidase F, membrane preparations in Buffer A (18 l) were mixed with 2 l of 10% dodecyl maltoside detergent (w/v in double distilled H 2 0) (Anatrace, Inc.). Samples were incubated with 0.3 units of N-glycosidase F (Boehringer Mannheim) for 2 h at 37°C. For cleavage with endoglycosidase H (Endo H), membrane preparations were pelleted and resuspended in 20 l of 50 mM sodium acetate, pH 5.0. Samples were incubated with 1.5 milliunits of Endo H (Boehringer Mannheim) for 2 h at 37°C.
Immunoblot Analysis of Expressed Glucagon Receptor and Receptor Mutants-Immunoblot analysis was carried out essentially as described (Carruthers et al., 1994). Briefly, membrane preparations were loaded without boiling onto a 1.5-mm-thick 4% stacking, 10% separating SDS-polyacrylamide gel and electrophoretically separated along with prestained protein markers, broad range (New England Biolabs). Proteins were transferred electrophoretically to Immobilon-P transfer membrane (Millipore) with 15 V for 10 min at room temperature. The membrane was then blocked, incubated with 3.5 g of the DK-12 anti-peptide antibody, washed, incubated with a 1:10,000 dilution of goat anti-rabbit IgG peroxidase (Boehringer Mannheim), and washed again. The immunoreactive bands were visualized by ECL (Amersham Corp.) and exposure to X-OMAT AR film (Eastman Kodak).
Immunofluorescence Microscopy of Transfected COS-1 Cells Expressing Glucagon Receptor and Receptor Mutants-The design for the imaging experiments was adapted from a previously described procedure (Moore et al., 1987). Circular microscope coverglasses (12-mm diameter) were placed in tissue culture dishes and sterilized. COS-1 cells were split into the plates for transient transfection. After transfection, the cells were maintained in serum-free HyQ-CCM5 media (HyClone). About 64 h after transfection, the medium was aspirated, and the cells were washed twice with a phosphate-buffered solution (137 mM NaCl, 8.0 mM Na 2 HPO 4 , 2.7 mM KCl, 1.5 mM KH 2 PO 4 ). The coverglasses were incubated in 20 ml of 3% formaldehyde in Buffer B (150 mM NaCl, 10 mM Na 2 HPO 4 , 2 mM MgCl 2 , pH 7.4) for 40 min, washed twice in Buffer B for 5 min, and transferred to 24-well plates, where all subsequent incubations were performed. Cell permeabilization, when indicated, was carried out by treatment with 1 ml of 0.2% Triton X-100 in Buffer B for 15 min at 4°C. After washing the cells twice with Buffer B for 5 min, nonspecific binding sites were blocked with 1% bovine serum albumin in Buffer B for 40 min at room temperature, followed by incubation at room temperature for 1 h with 0.7 g of DK-12 antibody in 0.1% bovine serum albumin in Buffer B. Three washes in 1 ml of Buffer B for 10 min each were followed by incubation at room temperature with 0.7 g of lissamine rhodamine B sulfonyl chloride-conjugated goat anti-rabbit IgG (Jackson ImmunoResearch Laboratories) diluted in 0.1% bovine serum albumin in Buffer B. All incubations and washes were performed subsequently in the dark. After washing three times in Buffer B for 10 min, coverglasses were blotted on Kimwipes and mounted on slides with Gel/Mount (Biomeda). All cells were viewed with a Nikon Microphot-SA epifluorescence microscope. Images were photographed at 1000-fold magnification with a motorized FX-35DX dark box connected to a Nikon Microflex UFX-DX photomicrographic attachment using Kodak Ektachrome ISO 400 35-mm film. The exposure time for each image was determined automatically by the photomicrographic attachment, except for the nonpermeabilized mutants where no positive staining was seen. For these images, the shutter was opened manually for the same duration as that determined automatically for the corresponding permeabilized sample.
Binding of 125 I-Glucagon to Transfected Cell Membrane Preparations-Competitive binding studies were performed essentially as described (Carruthers et al., 1994), with the following modifications: aliquots of membrane samples containing 15 g of protein were used, and the radioactivity retained on the nitrocellulose filters was quantified with a Packard 5221 Autogamma Scintillation Spectrometer (Packard).
Intracellular cAMP Assay-Adenylyl cyclase activity was determined by measuring cAMP levels in transfected COS-1 cells in response to increasing concentrations of glucagon as described (Carruthers et al., 1994).
Data Analysis for Competition Binding and cAMP Assays-In Figs. 5 and 6, symbols represent the mean of duplicate determinations. For the binding studies, experiments were performed at least four times on independent samples to verify reproducibility. For an individual experiment, the data at each concentration of glucagon were fit to a fourparameter logistic function of the form The IC 50 values were calculated from the inflection point (c) of the best-fit curve (Sigmaplot, Jandel Scientific Software), with the values for parameters a and d fixed at 100 and 0, respectively. For the cAMP assays, experiments were performed at least two times on independent samples. For an individual experiment, the data at each concentration of glucagon were fit to a four-parameter sigmoid function of the form The EC 50 values were calculated from the inflection point (c) of the best fit curve (Tablecurves, Jandel Scientific Software).

Preparation of Anti-peptide Antibody Directed against the N-terminal Tail of the Rat Glucagon
Receptor-A rabbit antibody (DK-12) against a synthetic peptide was prepared and purified. The peptide sequence was derived from the amino acid residues 126 -137 near the first transmembrane segment of the rat glucagon receptor (Fig. 1). The antibody detected the presence of an immunoreactive band of the appropriate molecular weight in rat hepatocyte preparations (not shown). The specificity of the antibody is demonstrated by immunoblot analysis of membrane preparations of transfected cells (Fig. 2) as described below.
Preparation and Expression of Mutant Glucagon Receptor Genes-The glucagon receptor mutant genes D1, T1, T3, T5, T7a, and T7b were prepared by site-directed mutagenesis of the synthetic gene for the rat glucagon receptor as described under "Experimental Procedures." The sites of all truncations are shown schematically in Fig. 1. In addition, mutant T3 included the tetrapeptide RKLH derived from the first intracellular loop after Thr 257 because the intracellular boundary of transmembrane helix 3 was not well defined. The genes were expressed in COS-1 cells by transient transfection.
Immunoblot analysis of the synthetic rat glucagon receptor mutants expressed in transfected COS cells is shown in Fig. 2. Membrane preparations of cells transfected with the expression vector (pMT3) without the glucagon receptor gene did not react with the antibody upon immunoblot analysis. Membrane preparations of cells transfected with vector containing the synthetic glucagon receptor gene (pMT5) showed a faint broad FIG. 1. Schematic representation of the rat glucagon receptor primary and secondary structure. Seven putative transmembrane helices (helix A through helix G) based on previous models of G protein-coupled receptors are shown. The N terminus and extracellular surface is toward the top, and the C terminus and cytoplasmic surface is toward the bottom of the figure. The four sites of potential N-linked glycosylation on the N terminus are labeled with asterisks. The 12 amino acids in the N-terminal tail, which were used to design a peptide DK-12 for antibody production, are boxed. Asp 64 , which was previously studied by site-specific mutagenesis, is numbered and labeled with an arrow (Carruthers et al., 1994). Mutant D1 contained a deletion of 96 amino acid residues from the N-terminal tail as indicated by arrows. Mutants T1, T3, T5, T7a, and T7b were truncated as shown. In addition, mutant T3 included the tetrapeptide RKLH derived from the first intracellular loop after Thr 257 because the intracellular boundary of transmembrane helix 3 was not well defined.
band migrating with an apparent molecular mass of 55-75 kDa. A potential receptor dimer band migrated at about 100 kDa. An additional weak band migrating at about 35 kDa not seen in the pMT3 lane was also apparent as previously discussed (Carruthers et al., 1994).
Membrane preparations from cells transfected with the Nterminal deletion mutant (D1) and the five truncation mutants (T1, T3, T5, T7a, and T7b) were immunoreactive with the DK-12 antibody since the peptide epitope was not disrupted by the deletion or the truncations (Fig. 2). Each of the mutant receptors was present in membrane preparations of transfected cells. The immunoblot band patterns and levels of expression of the individual mutants are described below. The first step in evaluating whether the mutant receptors were capable of glucagon binding and signal transduction was to demonstrate that they had been properly inserted into the plasma membrane of the cell. Membrane localization was measured by a series of deglycosylation experiments and by immunofluorescence microscopy described below.
Deglycosylation of the Expressed Glucagon Receptor Mutants with N-Glycosidase F-The enzyme N-glycosidase F cleaves N-linked carbohydrates from glycoproteins and leaves an aspartic acid residue at the position originally occupied by asparagine. After transient transfection, cell membranes were treated with N-glycosidase F. The immunoblot analysis in Fig.   2 shows that each mutant was present in the cell membrane preparation.
The pattern of N-linked glycosylation differed among the truncated mutant receptors and fell into two groups. The first group consisted of mutants T1, T3, and T5, where each displayed a doublet band pattern. Upon deglycosylation, each of the doublet band patterns collapsed to a single band with lower apparent molecular weight. Mutant receptors T3 and especially T5 also displayed bands corresponding to receptor dimer and higher order multimers as described previously for the native receptor (Carruthers et al., 1994). Dimer bands, and multimer bands in the case of mutant T5, were also noted after deglycosylation.
The second group of mutant receptors consisted of T7a and T7b. These two mutants displayed behavior similar to that of the native receptor. Each of these receptors was visualized as a faint broad band corresponding to the receptor monomer. Deglycosylation yielded one band for these receptors, again with a lower apparent molecular weight. With each of these receptors, the DK-12 anti-peptide antibody seemed to bind more avidly to the N-glycosidase F-treated form compared with the untreated form. In addition to the major bands, mutant T7b and native pMT5 each displayed a faint band migrating faster than the principal bands. The positions of these bands did not change upon N-glycosidase F treatment. The significance of this 35-kDa band in the case of pMT5 was been discussed previously (Carruthers et al., 1994). Mutant T7a had a consistently lower expression level compared with the other receptors as judged by immunoblot analysis.
Deletion mutant D1 showed no difference in immunoblot band pattern between untreated and treated membranes. This result was anticipated since mutant D1 contained none of the four potential N-linked glycosylation sites on the N-terminal tail (Fig. 1). Also, the immunoreactive band intensity was also not affected by the N-glycosidase F treatment of cells transfected with mutant D1.
Endo H Treatment of Membrane Preparations from Cells Transfected with Mutant Receptor Genes-Endo H sensitivity has been used as a marker to determine the cellular location of membrane proteins during biosynthesis (Kornfeld and Kornfeld, 1985;Robbins et al., 1977). The sensitivities of the two groups of glycosylated mutant receptors to Endo H cleavage is shown in Fig. 3. The native glucagon receptor (pMT5), as well as the mutants T7a and T7b were all resistant to Endo H cleavage while remaining sensitive to N-glycosidase F cleavage, demonstrating that these three receptors were glycosylated with complex oligosaccharides. In contrast, mutant receptors T1, T3, and T5 all were sensitive to both N-glycosidase F and Endo H, showing that these three receptor mutants were glycosylated with high mannose carbohydrate moieties. The significance of these results with respect to cellular localization is discussed below.
Cellular Localization of Mutant Receptors Using Immunofluorescence Microscopy-The results of the glycosidase digestion of the mutant receptors were confirmed using immunofluorescence microscopy. Fig. 4 shows that after permeabilizing transiently transfected COS-1 cells to permit entry of the primary and secondary antibodies (ϩ panels), each of the expressed receptors gave a positive reaction. All staining was seen within the internal membrane network of the cells. However, when cells were not permeabilized (Ϫ panels), only pMT5, T7a, T7b, and D1 gave positive signals. Since under nonpermeabilizing conditions the N-terminal antibody DK-12 should bind only to receptors with an extracellular epitope (i.e. receptors in the plasma membrane), these results confirm that pMT5, T7a, T7b, and D1 were processed correctly in the endoplasmic reticulum

FIG. 2. Immunoblot analysis of the rat glucagon receptor mutants expressed in transiently transfected COS-1 cells.
Membrane preparations were separated by 10% SDS-polyacrylamide gel electrophoresis, transferred to membranes, and probed with DK-12 anti-peptide glucagon receptor antibody. Immunoreactive bands were visualized by chemiluminescence (ECL). Control cells were transfected with vector alone (pMT3), or with vector containing the synthetic glucagon receptor gene (pMT5). Samples were divided, and one fraction was treated with N-glycosidase F to remove N-linked carbohydrates. Lanes labeled (ϩ) were treated with N-glycosidase F and lanes labeled (Ϫ) were untreated. Molecular mass (kDa) indicators are shown to the right of each immunoblot. Each lane contains 10 g of total protein. A, mutant receptors T1, T3, D1, and controls undigested (Ϫ) and digested (ϩ) with N-glycosidase F. B, mutant receptors T5, T7a, T7b, and controls undigested (Ϫ) and digested (ϩ) with N-glycosidase F. and Golgi and were displayed in the proper orientation on the extracellular surface. In addition, the observation that the fluorescence signal for mutant D1 was generally stronger relative to the other receptors corroborates the immunoblot finding that DK-12 bound more avidly to deglycosylated receptors. Fig. 5. Membranes from COS cells transiently transfected with the synthetic glucagon receptor gene (pMT5) or mutant receptor genes were incubated with radiolabeled glucagon and increasing concentrations of unlabeled glucagon. The levels of receptor in the membrane preparation assayed were essentially identical in each case. Of the three mutant receptors capable of reaching the cell surface (D1, T7a, and T7b), only mutants T7a and T7b were seen to bind glucagon with affinities comparable with that of native receptor ( Fig.  5a and Table I). The binding curves were well approximated by a four-parameter logistic function characteristic of ligand-receptor equilibrium binding. The concentrations of unlabeled glucagon required to displace 50% of receptor-bound 125 I-glucagon (IC 50 values) determined from the binding curves are presented in Table I. In contrast, Fig. 5b shows that all other mutant receptors did not bind glucagon, including deletion mutant D1, which was detected on cell surface, and the three mutants that did not reach the cell surface (T1, T3, and T5).

Binding of 125 I-Glucagon to Transfected Cell Membrane Preparations-Competition for 125 I-glucagon binding to COScell membranes expressing native or mutant glucagon receptor genes is shown in
Glucagon-dependent Stimulation of Adenylyl Cyclase in Transfected COS Cells-Adenylyl cyclase activity of COS cells expressing the synthetic glucagon receptor gene (pMT5) and the two mutant receptors (T7a and T7b) that bound glucagon is shown in Fig. 6. The increase in intracellular cAMP level was determined when cells were incubated with increasing concentrations of glucagon. cAMP was quantitated using an assay method that measured the ability of cAMP in each sample to displace [8-3 H]cAMP from a cAMP-binding protein. Both of the mutant receptors that demonstrated glucagon binding (T7a FIG. 3. Endo H sensitivity of glucagon receptor mutants in membrane preparations. Membrane preparations were separated by 10% SDS-polyacrylamide gel electrophoresis, transferred to a membrane, and probed with DK-12 anti-peptide glucagon receptor antibody. Each lane contains 10 g of total protein. Immunoreactive bands were visualized by chemiluminescence (ECL). Samples (T1, T3, T5, T7a, and T7b) were divided, and one fraction was untreated (Ϫ), one fraction was treated with N-glycosidase F (N) to remove N-linked carbohydrates, and one fraction was treated with Endo H (E) to remove only high mannose carbohydrates. Mutant receptor D1 was not tested since it lacks glycosylation sites. The native receptor pMT5 (not shown) displayed the same pattern as that of mutants T7a and T7b. It was resistant to Endo H cleavage, but sensitive to N-glycosidase F cleavage (see Fig. 2).

FIG. 4. Immunofluorescence microscopy of COS cells transfected with mutant glucagon receptor genes.
Transfected cells were prepared as described under "Experimental Procedures." For each mutant receptor (T1, T3, T5, T7a, T7b, and D1) and control (pMT3 and pMT5), permeabilized cells (ϩ) are compared with nonpermeabilized cells (Ϫ). Anti-peptide receptor antibody DK-12, which recognizes the extracellular N-terminal tail of the receptor, was used as the primary antibody. A rhodamine-conjugated antibody (lissamine rhodamine B sulfonyl chlorideconjugated goat anti-rabbit IgG) was used as the secondary antibody. and T7b) also were able to increase intracellular cAMP concentrations in response to glucagon. Fig. 6 shows that for each of the three receptors (pMT5, T7a, and T7b), there was a sigmoidal increase in cAMP accumulation that peaked at a glucagon concentration of approximately 2.9 ϫ 10 Ϫ7 M. At higher concentrations of glucagon, the total accumulated cAMP was lower, as described previously (Carruthers et al., 1994). The effective concentrations at 50% stimulation of adenylyl cyclase (EC 50 values) for receptors T7a, T7b, and pMT5 were in the nanomolar range and consistently lower than each IC 50 value (Table I) as expected (Carruthers et al., 1994). DISCUSSION A series of six glucagon receptor mutants was prepared to attempt to define the minimum size required for proper biocentration, where each symbol represents the mean of triplicate measurements. Each of the four mutations resulted in the complete inability of the mutant receptor to bind glucagon.

FIG. 5. Competition for 125 I-glucagon binding to COS cell membranes expressing native or mutant glucagon receptor genes.
A, competitive displacement of 125 I-labeled glucagon bound to transfected cell membranes was determined by incubation with 125 I-glucagon alone and with the indicated concentrations of unlabeled glucagon. Data are presented as percentage of total binding of the radiolabeled hormone versus the log of glucagon concentration. Maximum binding (100% on the y axis) was less than 10% of total added radioactivity. Each symbol represents the mean of duplicate determinations and was curve-fitted where appropriate based on a single ligand-binding site model as described under "Experimental Procedures." Cells transfected with control vector pMT3 showed insignificant binding of 125 I-glucagon that was considered to be nonspecific. The concentration of unlabeled glucagon required to displace 50% of receptor-bound 125 I-glucagon, the IC 50 value, was calculated to be 10.4 nM for the native receptor (pMT5), 4.7 nM for mutant receptor T7a, and 4.0 nM for mutant receptor T7b. B, membranes from COS cells transiently transfected with the glucagon receptor gene (pMT5), T1, T3, T5, or D1 were incubated with radiolabeled glucagon and increasing concentrations of unlabeled glucagon. Total radioactivity bound (cpm) is plotted versus log of glucagon con- a The concentration of unlabeled glucagon required to displace 50% of receptor-bound 125 I-glucagon. The average of four independent competition binding curve determinations in which each point was measured in duplicate is given.
b The effective glucagon concentration at 50% stimulation of adenylyl cyclase. The average of two independent competition binding curve determinations in which each point was measured in triplicate is given.
FIG. 6. Adenylyl cyclase activity of COS cells expressing the glucagon receptor gene and C-terminal mutants. COS cells were transfected with vector containing the synthetic glucagon receptor gene (pMT5), mutant receptor T7a, or mutant receptor T7b. The increase in intracellular cAMP level when cells were incubated with increasing concentrations of glucagon was determined. cAMP was quantitated using an assay method that measured the ability of cAMP in each sample to displace [8-3 H]cAMP from a cAMP-binding protein. Each symbol represents the mean of duplicate determinations and is plotted as the percentage of total cAMP accumulation versus the log of glucagon concentration. The effective concentrations at 50% stimulation of adenylyl cyclase (EC 50 values) for cells expressing the glucagon receptor or the two mutants are given in Table I. synthesis and ligand binding. A schematic representation of the rat glucagon receptor primary and secondary structure is shown in Fig. 1. Seven putative transmembrane helices (helix A through helix G) based on previous models of G proteincoupled receptors are shown (Dratz and Hargrave, 1983;Sakmar et al., 1989;Baldwin, 1993). Four sites of potential Nlinked glycosylation on the N terminus are labeled with asterisks (*). The N-terminal amino acid residues 126 -137 were used to design a peptide, DK-12, for anti-peptide antibody production. A highly specific, high affinity antibody was obtained, which was used for immunoblot analysis of the expression of the glucagon receptor gene and site-directed mutant genes. The immunoblots in Fig. 2 show that the antibody reacted with the products of expression of the vector containing the synthetic glucagon receptor gene. The antibody showed affinity for the monomeric and oligomeric forms of the receptor.
Three of the six mutant receptors, T1, T3, and T5, were not expressed on the cell surface of transfected COS cells. The Endo H susceptibility analysis (Fig. 3) combined with immunofluorescence microscopy (Fig. 4) showed that these receptors could have been transported only as far as the medial Golgi (Kornfeld and Kornfeld, 1985). However, most of the population of these mutant receptors probably were located in the endoplasmic reticulum (ER). Abnormally processed membrane proteins often are specifically retained in the rough ER (Lodish, 1988). Figs. 2 and 3 show that mutants T1, T3, and T5 all were glycosylated to a high mannose form, which is the degree of glycosylation expected to occur in the ER. However, mutants T1 and T3 both lack the highly conserved pair of Cys 225 -Cys 295 , which are likely to form an extracellular disulfide bond analogous to the essential Cys 110 -Cys 187 linkage in rhodopsin (Karnik et al., 1988;Ridge et al., 1995). Therefore, there is at least one key structural element lacking in mutants T1 and T3 that might prevent proper folding. Mutant T5 has the cysteine residues for the putative disulfide bond, but lacks the last two transmembrane helices. It is likely that these two helices are essential for assuming a native-like conformation in the ER membrane. In the case of rhodopsin, it was shown that a structure involving all transmembrane helices, the N-terminal tail, and the three intradiscal loops were all essential for the correct folding and assembly of a functional receptor (Anukanth and Khorana, 1994). Thus, each of these three mutants lacks at least one key structural element.
In spite of the demonstrated importance of glycosylation for proper folding for some receptors, the glucagon receptor clearly does not require glycosylation to reach the plasma membrane. The immunofluorescence data in Fig. 4 show surface expression of mutant D1, which lacks all four potential N-linked glycosylation sites. A lack of mutant D1 receptor glycosylation is shown experimentally by a failure of electrophoretic mobility change following N-glycosidase F treatment ( Fig. 2A). Glycosylation cannot be a required signal for intracellular trafficking and targeting of this receptor to the plasma membrane in COS cells. In addition, a considerable portion of the N terminus can be deleted without affecting the global structure responsible for ER and Golgi processing. The glucagon receptor may therefore differ from other membrane glycoproteins, for which it has been argued that glycosylation is required for protein folding and for routing to the cell membrane (Kusui et al., 1994). In addition, receptor glycosylation also may not be necessary for glucagon binding, since preliminary results show that treating pMT5 membranes with N-glycosidase F does not alter glucagon-binding affinity (data not shown).
A tentative model for receptor-hormone interaction suggests the existence of at least three distinct functional regions: Region 1, the proximal N-terminal loop, which contains all four putative N-linked glycosylation sites; Region 2, the distal portion of the N-terminal loop, the seven transmembrane helices, and the three extracellular loops; and Region 3, the intracellular region consisting of the three interhelical loops and the C-terminal tail. Region 1 is likely to contain structural elements necessary for proper receptor structure, as suggested by the finding that substitution of a conserved aspartic acid residue (Asp 64 ) resulted in a receptor mutant that failed to bind glucagon (Carruthers et al., 1994). In another recent analysis, a series of chimeras in which various domains of the human glucagon receptor were replaced by homologous regions from the GLP-1 receptor was generated (Buggy et al., 1995). These two receptors share 47% amino acid identity, and glucagon binds to the GLP-1 receptor with an affinity 1000-fold less than for its own receptor. With respect to glucagon binding, the chimeras were classified as being more similar to either the glucagon receptor or to the GLP-1 receptor. For those receptors that did not bind well to glucagon, the cell surface expression was confirmed by immunofluorescence. The dichotomous nature of binding specificities was used to identify certain segments of Region 2 as important for glucagon binding: the distal portion of the N-terminal domain, the first extracellular loop, and the third, fourth, and sixth transmembrane domains (Buggy et al., 1995). The importance for biological activity of the N-terminal domain and the first extracellular loop was also demonstrated in the secretin-vasoactive intestinal peptide receptor family (Holtmann et al., 1995). These findings are generally consistent with the present results. Mutants T7a and T7b both lack part or most of the intracellular C-terminal tail, and yet both bind glucagon with an affinity comparable with that of the native receptor (Fig. 5, Table I). All other truncation mutants and the deletion mutant D1 lack at least one of the domains implicated to be important for glucagon binding.
A direct role for Region 1 in ligand binding cannot be ruled out. The chimeric receptor analysis was inherently incapable of distinguishing a structural from a functional role for Region 1, since the proximal N-terminal region of the GLP-1 receptor shares many of the same residues as the glucagon receptor, including Asp 64 (Buggy et al., 1995). Four glucagon receptor mutants (D64E, D64G, D64K, and D64N), were previously shown to be incapable of binding glucagon (Carruthers et al., 1994). Each of these mutant receptors is resistant to Endo H cleavage (data not shown) and is present on the cell surface. The strict conservation of this residue throughout the glucagon receptor family suggests a generalized role for Asp 64 . Whether this Asp is involved directly in ligand-binding is still open to question.
The intracellular portions of the receptor, Region 3, are directly responsible for signal transduction. Previous studies have shown the second and third intracellular loops of G protein-coupled receptors to mediate G protein activation (Ernst et al., 1995). The role for the C-terminal tail, however, is less well defined. The present work demonstrates that most of the tail is not required for binding glucagon. In fact, mutants T7a and T7b have affinities for the hormone that are slightly stronger than normal (Table I). This improvement may be explained by a receptor-ligand interaction where the receptor is in equilibrium between an active and inactive state. An agonist for the receptor, such as glucagon, would stabilize the active state, and lead to activation of the signal transduction pathways (Bond et al., 1995;Samama et al., 1993). The presence of the native C-terminal tail destabilizes the active state and pushes the equilibrium toward an inactive state that has less affinity for a pure agonist such as glucagon. Such a response already has been observed with the rat PTH/PTH-related peptide receptor, where the removal of wild-type C-terminal sequences increased the mutant receptor affinity for agonist and increased the efficacy with which the receptor interacted with G s (Iida-Klein et al., 1995).
In an analogous study, mutant avian ␤-adrenergic receptors with progressively truncated C-terminal tails were prepared (Parker and Ross, 1991). Membranes from cells that expressed the truncated receptors displayed elevated basal and agoniststimulated adenylyl cyclase activities. Counterintuitively, the truncation mutants also demonstrated adenylyl cyclase activation upon exposure to alprenolol and propranolol, molecules classically considered as pure ␤-adrenergic receptor antagonists. It would appear that the C-terminal tail does modulate the receptor's affinity for various ligands. While it is not directly required for G s activation, the C-terminal tail does affect the overall coupling efficiency between receptor and G protein.
Accumulation of cAMP was measured as an assessment of the ability of a receptor to couple to G s . A large body of work has demonstrated a clear connection between glucagon binding to its receptor and the generation of cAMP via G s (Iyengar et al., 1988). However, it is possible that this ligand-receptor complex can activate additional effector pathways. Glucagon-induced elevations in inositol 1,4,5-trisphosphate (Wakelam et al., 1986) and [Ca 2ϩ ] i have been reported (Jelinek et al., 1993). The key question remains as to whether this change occurs through G q , (Christophe, 1995) as has been postulated for the glucagon receptor. G q has been implicated as the G protein for other peptide hormone receptors such as that for thyroid-stimulating hormone (Allgeier et al., 1994) and PTH/PTH-related peptide, (Iida-Klein et al., 1995) but recent evidence has shown that G ␤␥ (Sternweis, 1994) and cAMP (Cooper et al., 1995) also can elevate intracellular calcium concentrations. However, the source for calcium may not necessarily be the ER stores released by inositol 1,4,5-trisphosphate (Berridge and Irvine, 1985). Rather, a G ␤␥ -stimulated adenylyl cyclase or cAMP directly may open channels at the plasma membrane to increase calcium influx (Cooper et al., 1995). Investigations are under way to distinguish the possible roles for each of these components.
In conclusion, the present research provides biochemical and cell biological information about the glucagon receptor, which may be relevant to related G protein-coupled peptide hormone receptors. It is clear that all seven helices are necessary for the proper folding, processing, and cell surface expression of the glucagon receptor. Truncated receptors may be glycosylated, but they are not transported to the Golgi and subsequently to the plasma membrane. At least for the glucagon receptor, glycosylation is not required to reach the cell surface, and it may not be required for ligand binding. The N-terminal extracellular portion of the receptor is required for ligand binding, but it is unclear from the present study which particular transmembrane helices interact directly with the ligand glucagon. In contrast, most of the C-terminal region of the receptor is not necessary for ligand binding, and the presence of an intracellular tail may in fact decrease the receptor's binding affinity for glucagon. Finally, the distal C-terminal region was shown to be unnecessary for coupling to adenylyl cyclase. We are currently investigating which portions of the intracellular region are responsible for activating adenylyl cyclase, increasing intracellular calcium concentrations, and regulating receptor activity at the cell surface.