INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

Pituitary adenylate cyclase-activating polypeptide (PACAP)¹ was first discovered in 1989 as a novel hypothalamic hormone that increases adenylate cyclase activity in pituitary cells (1). PACAP exists in two carboxyl-terminal-amidated forms: PACAP38 with 38 amino acid residues (1) and PACAP27 with the same amino-terminal 27 residues (2). Molecular cloning studies revealed that the structure of PACAP is highly conserved among rat, sheep, and humans (3, 4). PACAP has structural similarity to peptides in the secretin/glucagon peptide family, especially to vasoactive intestinal polypeptide (VIP) (1). PACAP is distributed in the central nervous system and various peripheral organs (5) and elicits a wide variety of biological functions, such as neuroprotective action against gp120-induced cell death (6), protection of cerebellar granule neurons from apoptosis (7), secretion of pituitary hormones (1, 8), secretion of interleukin-6 from astrocytes or folliculo-stellate cells (9, 10), secretion of catecholamines from chromaffin cells (11) or adrenal glands (12), and insulin release (13). The biological actions of PACAP are mediated by a PACAP-specific receptor (type I receptor) and a PACAP/VIP-nonselective receptor (type II receptor). The type I PACAP receptor includes the PACAP₁ receptor (14-21) and a novel variant PACAPR-TM4 (22). There are two alternatively spliced exons, rat hip and hop (20) or human SV-1 and SV-2 (21), in the PACAP₁ receptor gene, resulting in the possible existence of five splicing variants in the PACAP₁ receptor (20, 21). All of these receptors belong to the G protein-coupled receptor superfamily and are subdivided structurally into the secretin/glucagon receptor family (23) that is distinguished from rhodopsin-type receptors.

All G protein-coupled receptors have seven hydrophobic segments that probably form transmembrane -helices. Direct evidence for the arrangement of transmembrane domains was obtained from the two-dimensional crystallography of rhodopsin (24), providing valuable information for molecular modeling of other G protein-coupled receptors. More precise modeling requires elucidating the structures of another receptors. In particular, receptors in the secretin/glucagon receptor family are predicted to have a different arrangement in the transmembrane domains (25). On the other hand, structural biology directly clarifying the three-dimensional structure of a G protein-coupled receptor has been hindered by several difficulties in the purification and crystallization of the receptor protein. Most G protein-coupled receptors exist at very low level in tissue membranes. Thus, it is essential to develop an expression system that can produce a large amount of the recombinant receptor. Parker et al. (26) first described that a baculovirus expression system was beneficial for the expression of -adrenergic and muscarinic receptors at high levels (5-30 pmol/mg). Recombinant -adrenergic receptors purified from the baculovirus-infected insect cells were functionally active as were the -adrenergic receptors from turkey erythrocytes (26). The expression system was also used to produce various G protein-coupled receptors, however, only a few reports (27) succeeded in providing a practical amount of purified receptor for further biochemical or structural studies. This is probably due to difficulty in the solubilization and purification of G protein-coupled receptors.

We previously described successful purification of the PACAP₁ receptor from bovine brain membranes in a high affinity state (28). In the present study we conducted large-scale purification of the recombinant PACAP₁ receptor by combining the previously described purification procedures and the baculovirus expression system. The recombinant PACAP₁ receptor purified in a digitonin-solubilized form retained high affinity for PACAP and was functionally active when reconstituted with G_s protein in lipid vesicles. The purified receptor will likely contribute to the understanding of the regulatory mechanisms and structure of G protein-coupled receptors.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Materials-- PACAP27, PACAP38, PACAP(6-38), VIP, leupeptin, pepstatin, and E-64 were obtained from the Peptide Institute (Osaka, Japan). GTP, GDP, and ATP were from Yamasa (Tokyo, Japan). Guanosine 5'-O-3-thiotriphosphate (GTPS) was obtained from Boehringer Mannheim GmbH (Mannheim, Germany). Bovine serum albumin (BSA), GMP, Nonidet P-40, and brain extract type VII were from Sigma. Flavobacterium menigosepticum peptide-N⁴-(N-acetyl--glucosaminyl) asparagine amidase (N-glycanase, recombinant), F. menigosepticum endo--N-acetylglucosaminidase F₂ (endoglycosidase F₂), Streptomyces plicatus endo--N-acetylglucosaminidase (endoglycosidase H, recombinant), and Streptococcus sp. sialidase (neuraminidase) were from Genzyme (Cambridge, MA). Xanthomonas manihotis -N-acetylglucosaminidase and X. manihotis 1-2,3-mannosidase were from New England Biolabs, Inc. (Beverly, MA). SDS, digitonin, hydroxyapatite, and phenylmethylsulfonyl fluoride were from Wako Pure Chemicals (Osaka, Japan). Digitonin was dissolved in water at 80-90 °C, cooled, and ultracentrifuged for removal of insoluble materials. BIGCHAP, CHAPS, and 5-[5-(N-succinimidyloxycarbonyl)penthylamido]hexyl D-biotinamide (Biotin-(AC₅)₂-OSu) were obtained from Dojindo Laboratories (Kumamoto, Japan). Avidin-Affi-Gel 10 was prepared by immobilizing avidin (Wako Pure Chemicals, Osaka, Japan) to Affi-Gel 10 (Bio-Rad) following the manufacturer's instructions. Lentil lectin-Sepharose 4B was obtained from Pharmacia Biotech (Uppsala, Sweden). ¹²⁵I-PACAP27 was prepared by the previously described method (29). [³⁵S]GTPS was obtained from NEN Life Science Products (Boston, MA).

Expression in Sf9 Insect Cells-- The null variant lacking the SV-1 and SV-2 insertion sequences (19, 21) was expressed in Sf9 insect cells as described previously (32). The human PACAP receptor cDNA fragment (nucleotides 1-1664) was excised from pTS847 plasmid (19) by EcoRI digestion and cloned in pcDNAI/Amp (Invitrogen). An EcoRV fragment was excised from the resulting plasmid, ligated with the Sse8387I linker, digested with BamHI and Sse8387I, and cloned in the transfer vector pBlueBacIII (Invitrogen). The resultant transfer vector (pHPR-7) and Autographa californica nuclear polyhedrosis virus genomic DNA were co-transfected to Sf9 cells to generate recombinant baculovirus. The recombinant virus producing the highest amount of the PACAP receptor was selected. The Sf9 cells (2 × 10⁸ cells) were cultured with 200 ml of Grace's insect cell culture medium (Life Technologies, Inc., Grand Island, NY) containing 0.1% Pluronic F-68 (Life Technologies, Inc., Grand Island, NY), 10% fetal calf serum, and 20 µg/ml gentamicin in a 1-liter spinner flask at 27 °C for 25 h, infected with the recombinant virus at a multiplicity of infection of 3-5, and cultured at 27 °C for 4 days. The cells were harvested, washed with phosphate-buffered saline containing 2.7 mM EDTA, and stored at 70 °C until used.

Stable Expression in Chinese Hamster Ovary Cells-- The PACAP receptor cDNA fragment (nucleotides 245-1652) was obtained by polymerase chain reactions using pTS847 plasmid as a template (19) and cloned at SalI site in a pAKKO1.11 expression vector (33) containing SR promoter and mouse dihydrofolate reductase gene as a selective marker. The resulting plasmid was transfected to Chinese hamster ovary cells deficient in dihydrofolate reductase (CHO/dhfr cells) by the calcium phosphate-coprecipitation method. A clonal cell line expressing the maximum level of the PACAP receptor (PACR19 clone) was obtained by selection in Dulbecco's modified Eagle's medium containing 10% dialyzed fetal calf serum, 100 units/ml penicillin, and 100 µg/ml streptomycin. The PACR19 cells were grown with Dulbecco's modified Eagle's medium containing 10% fetal calf serum, 100 units/ml penicillin, and 100 µg/ml streptomycin in Nunc Cell Factories (Nunc A/S, Roskilde, Denmark). The cells at 70-80% confluency were harvested by washing with phosphate-buffered saline containing 2.7 mM EDTA, and stored at 70 °C until used.

Preparation of Biotinylated PACAP38-- PACAP38 (2.2 µmol, 10 mg) was reacted with a 1.4 mol equivalent (3.0 µmol, 1.7 mg) of biotinylating reagent, biotin-(AC₅)₂-Osu in dimethyl sulfoxide (5 ml) containing a 10 mol equivalent of triethylamine (21 µmol, 3 µl) at room temperature for 2 h. An aliquot (1 ml) of the reaction mixture was diluted with 0.05% trifluoroacetic acid and injected into a reversed phase high performance liquid chromatography column (7.8 mm × 30 cm, ODS80TM, Tosoh, Tokyo, Japan) equilibrated with 0.05% trifluoroacetic acid. The biotinylated PACAP38 was eluted with a linear gradient of acetonitrile from 20 to 40% for 60 min at a flow rate of 2 ml/min. Several peaks of biotinylated ligands eluted behind the peak of unbiotinylated PACAP38 were collected, lyophilized, and dissolved in 0.05% CHAPS.

Preparation of the Membrane Fraction-- The infected Sf9 cells were homogenized with HOM buffer (10 mM NaHCO₃, 5 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, 20 µg/ml leupeptin, 10 µg/ml pepstatin, and 8 µg/ml E-64, pH 7.3) using a Polytron homogenizer (Kinematica GmbH, Littau, Switzerland) and centrifuged at 700 × g for 10 min (32, 33). The pellet was subjected twice to a repeated cycle of homogenization and centrifugation. The final pellet was suspended in the HOM buffer (P₀ membranes). The supernatant fractions were combined and ultracentrifuged at 100,000 × g for 60 min. The resultant membrane pellet was suspended in the HOM buffer (P₁₊₂₊₃ membranes). The P₁₊₂₊₃ membranes from the transformed CHO cells (PACR19 clone) were prepared as above.

Solubilization and Purification of the PACAP Receptor-- The membrane protein was solubilized with 1% digitonin at a protein concentration of 2 mg/ml for 16-18 h. The clear solubilized protein fraction was obtained after ultracentrifugation at 100,000 × g for 60 min. The solubilized membrane protein was mixed with a 4-fold equivalent of biotinylated PACAP38 and avidin-Affi-Gel 10 (1 ml/10 nmol of biotinylated ligand). The mixture was gently agitated on a rotary shaker for 2 days at 4 °C. The gel was packed in a glass column and washed with the HOM buffer containing 1 M NaCl and 0.2% digitonin. The PACAP receptor was eluted with 20 mM magnesium acetate buffer (pH 4.0) including 0.2% digitonin, 1 M NaCl, and 10% glycerol.

Reconstitution of the Purified Receptor with G Protein-- Reconstitution of the purified receptor was performed as follows. Bovine brain crude lipid (brain extract type VII) was dissolved in REC buffer (20 mM Tris, 1 mM EDTA, 3 mM MgCl₂, and 160 mM NaCl, pH 7.4) containing 17% CHAPS at 40 mg/ml and stored at 70 °C. The lipid solution was diluted 8-fold with the REC buffer before use. The diluted lipid solution (60 µg/12 µl), purified PACAP receptor (20 pmol/4.5 µl), purified G_s (80 pmol/8 µl), and purified brain G (160 pmol/25 µl) were mixed and dialyzed at 4 °C for 24-36 h in a dialyzing apparatus (Microdialysis system, Life Technologies, Inc., Gaithersburg, MD) equipped with dialysis membranes (Spectra/Por 2, MWCO:12-14,000, Spectrum Medical Industries, Inc., Houston, TX) against the REC buffer supplied at a flow rate of 15 ml/h. The reconstituted receptor was used in 2-3 days.

Receptor Binding Experiments-- Receptor binding experiments were performed by the previously described method (28) in DG-BSA/TED buffer (0.05% DG, 0.1% BSA, 20 mM Tris, and 1 mM EDTA, pH 7.4), BSA/TED buffer (1% BSA, 20 mM Tris, and 1 mM EDTA, pH 7.4), and BSA/TED-Mg buffer (1% BSA, 20 mM Tris, 1 mM EDTA, and 5 mM MgCl₂, pH 7.4). BSA concentration was increased in the BSA/TED and BSA/TED-Mg buffer to avoid severe sticking of ¹²⁵I-PACAP27 on test tubes. Every binding reaction mixture contained 0.005% CHAPS that was derived from the vehicle of ¹²⁵I-PACAP27.

GTPS Binding Experiments-- The reconstituted receptor diluted 200-fold with the BSA/TED-Mg buffer (10 µl), PACAP, or a related ligand dissolved in the BSA/TED buffer supplemented with 0.05% CHAPS (1 µl), and 0.5 nM [³⁵S]GTPS diluted with the BSA/TED-Mg buffer (100 µl) were mixed and incubated at 25 °C for 1 h. The reaction mixture was diluted with 1.5 ml of chilled TEM buffer (0.05% CHAPS, 0.1% BSA, 5 mM MgCl₂, 1 mM EDTA, and 50 mM Tris, pH 7.4) and filtered through a pre-wetted GF/F glass fiber filter (Whatman). The filter was washed with 1.5 ml of the TEM buffer, dried, and subjected to liquid scintillation counting. The experiment with the membrane fraction was performed in the same manner but in the BSA/TED-Mg buffer supplemented with 1 µM GDP and 150 mM NaCl.

Glycosidase Digestion-- Purified PACAP receptor (0.5 mg/ml, 10 µl) was mixed with 1% SDS (10 µl), denatured at 100 °C for 3 min and then diluted with 10% Nonidet P-40 (10 µl) and distilled water (10 µl). An aliquot (4 µl) of the denatured receptor was digested with N-glycanase (250 milliunits) in 0.1 M Tris (pH 7.6), endoglycosidase F₂ (0.2 milliunits) in 0.2 M sodium acetate (pH 4.75), endoglycosidase H (2 milliunits) in 50 mM sodium citrate (pH 6.0), Streptococcus sp. sialidase (5 milliunits) in 50 mM sodium citrate (pH 6.0), X. manihotis 1-2,3-mannosidase (1 units) in 50 mM sodium citrate (pH 6.0), or X. manihotis -N-acetylglucosaminidase (1 units) in 50 mM sodium citrate (pH 4.5) at 37 °C for 18 h. The reaction mixture was diluted with a sample buffer for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (34), boiled at 100 °C for 3 min, and analyzed by SDS-PAGE (34). Protein bands were visualized using a silver staining kit, 2D-Silver StainII (Daiichi Pure Chemicals, Tokyo, Japan).

Miscellaneous Methods-- Protein determination was carried out by the method described by Schaffner and Weissman (35). Protein sequencing was performed as described previously (28).

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Low Affinity PACAP Receptor in Sf9 Cell Membranes-- The predominant splicing variant (null variant) of the human PACAP₁ receptor, that lacks SV-1 and SV-2 insertion sequences in the third intracellular loop (19, 21), was overexpressed in Sf9 insect cells under the control of the baculovirus polyhedrin promoter and was stably expressed in CHO cell transformants under the control of SR promoter. Saturation receptor binding experiments performed in the absence of digitonin (in BSA/TED buffer) followed by Scatchard plot analysis demonstrated that membranes from the baculovirus-infected Sf9 cells (Sf9 P₁₊₂₊₃ membranes) contained a single class of binding sites having low affinity for ¹²⁵I-PACAP27 (K_d = 155.3 ± 16.7 pM, Fig. 1A). In contrast, membranes from the transformed CHO cells (CHO P₁₊₂₊₃ membranes) contained high affinity binding sites (K_d = 44.4 ± 2.2 pM, Fig. 1B). The maximum receptor binding (B_max) to the Sf9 P₁₊₂₊₃ membranes was 82.6 ± 3.8 pmol/mg of protein (Fig. 1A) and that to the CHO P₁₊₂₊₃ membranes was 24.1 ± 1.5 pmol/mg (Fig. 1B).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Scatchard plot analysis for saturation binding experiments with the membranes. Equilibrium binding of increasing concentration of 125I-PACAP27 to: A, the P1+2+3 membranes from the infected Sf9 cells (0.37 µg/ml); or B, that from the transformed CHO cells (PACR19 clone) (0.94 µg/ml) was determined in the BSA/TED or DG-BSA/TED buffer. Nonspecific binding was determined in the presence of 0.1 µM PACAP27. Binding data was plotted versus total ligand (inset) and analyzed by Scatchard plot.

View this table: [in this window] [in a new window]   Table I Effect of GTPS on 125I-PACAP binding to the membranes Equilibrium binding of 100 pM 125I-PACAP27 to the Sf9 and CHO P1+2+3 membranes was determined in various buffers containing 20 µM GTPS. Specific binding was derived by subtracting nonspecific binding from total binding. Results were shown in percent of control specific binding determined in the absence of GTPS.

Lack of Functional Coupling to G Protein in Sf9 Cell Membranes-- Agonist-dependent stimulation of [³⁵S]GTPS binding to the P₁₊₂₊₃ membranes was determined in BSA/TED-Mg buffer supplemented with 1 µM GDP and 150 mM NaCl. These additives were required to decrease the basal [³⁵S]GTPS binding occurring in the absence of the agonist. The binding of [³⁵S]GTPS to the CHO P₁₊₂₊₃ membranes increased markedly in the presence of 1 µM PACAP27 (Table II). The increase in [³⁵S]GTPS binding to the CHO membranes was dependent on the agonist concentration and was specific to PACAP (Fig. 2). The EC₅₀ values were 581 ± 43 pM for PACAP27 and 107 ± 4.4 pM for PACAP38 (Fig. 2). PACAP did not increase [³⁵S]GTPS binding to the membranes of mock transfectants (data not shown). On the other hand, the Sf9 P₁₊₂₊₃ membranes did not exhibit a significant increase in [³⁵S]GTPS binding in response to PACAP27 stimulation (Table II). This result further demonstrates that the PACAP receptor is functionally coupled to G protein in the CHO cell membranes but not in the Sf9 cell membranes.

View this table: [in this window] [in a new window]   Table II Agonist stimulation of [35S]GTPS binding to the membranes Binding of [35S]GTPS to the Sf9 and CHO P1+2+3 membranes in the presence and absence of 1 µM PACAP27 was determined in the BSA/TED-Mg buffer containing 1 µM GDP and 150 mM NaCl and in the BSA/TED-Mg buffer containing 1 µM GDP.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Agonist concentration dependence of [35S]GTPS binding to the membranes. Binding of [35S]GTPS to the P1+2+3 membranes from the transformed CHO cells (13.6 µg/ml) was determined in the BSA/TED-Mg buffer containing 1 µM GDP, 150 mM NaCl, and increasing concentration of indicated peptide. Increase in GTPS binding was obtained by subtracting basal binding (16 pM) from total binding.

The PACAP Receptor in a Digitonin-induced High Affinity State-- Saturation receptor binding experiments were further performed in the presence of 0.05% digitonin (in DG-BSA/TED buffer). In contrast with the experiments in the absence of digitonin, both PACAP receptors in the Sf9 and CHO P₁₊₂₊₃ membranes had high affinity for ¹²⁵I-PACAP27. The K_d values were 44.6 ± 2.1 pM (Fig. 1A) and 39.3 ± 2.3 pM (Fig. 1B), respectively. The B_max values, 146 ± 7.1 pmol/mg for Sf9 P₁₊₂₊₃ membranes (Fig. 1A) and 54.2 ± 0.5 pmol/mg for CHO P₁₊₂₊₃ membranes (Fig. 1B), were two times higher than the respective values determined in the absence of digitonin. Neither PACAP receptor was sensitive to GTPS in the DG-BSA/TED buffer (Table I).

Solubilization and Purification of the Recombinant PACAP Receptor-- The combined Sf9 membranes were subjected to solubilization with digitonin. The solubilized protein contained a single class of the PACAP receptor with high affinity (K_d = 20-40 pM) at a 3-fold higher concentration than the membranes (Table III). Digitonin worked best in that it solubilized the receptor in the high affinity state and did not destroy the receptor activity even at higher concentrations as shown for the bovine brain PACAP receptor (28, 36).

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Purification of the PACAP receptor by biotinylated PACAP38/avidin-Affi-Gel 10 and lentil lectin-Sepharose 4B chromatography. A, recombinant PACAP receptor (144 nmol) solubilized from the combined Sf9 P1+2+3 and P0 membranes (8-liter culture) was adsorbed onto avidin-Affi-Gel 10 (80 ml) via biotinylated PACAP38 (600-700 nmol), eluted with 0.2% digitonin, 1 M NaCl, 10% glycerol, magnesium acetate buffer (pH 4.0) and neutralized with 1 M Tris/HCl (pH 7.5). Ligand (100 pM 125I-PACAP27) binding activity in each fraction (10 ml) was assayed at 33,000-fold dilution. B, the pooled fractions (3.5 mg of protein) from biotinylated PACAP38/avidin-Affi-Gel 10 chromatography were loaded onto a lentil lectin-Sepharose 4B column (9 ml). After washing the column with 0.2% digitonin, 0.5 M NaCl, 20 mM Tris buffer (pH 7.5), the receptor was eluted with the same buffer including 0.5 M -methylmannoside. Ligand binding activity in each fraction (2.5 ml) was assayed at 33,000- and 3,300-fold dilution.

Ligand Binding Properties of the Purified PACAP Receptor-- Saturation receptor binding experiments in the DG-BSA/TED buffer indicated that the purified PACAP receptor from the insect cells had a single class of high affinity binding sites with a K_d value of 17.3 ± 1.3 pM (Fig. 4A). The affinity of the purified receptor was slightly higher than the membranous PACAP receptor determined in the DG-BSA/TED buffer. The specific activity was 23.9 ± 1.5 nmol/mg of protein (Table III). Competitive binding experiments demonstrated that the purified receptor retained selectivity for PACAP27 (IC₅₀ = 0.21 ± 0.02 nM) and PACAP38 (IC₅₀ = 0.086 ± 0.005 nM) against VIP (IC₅₀ = 0.37 ± 0.04 µM) (Fig. 4B). These binding properties were very similar to those of the purified PACAP receptor from the CHO cells (K_d = 14.7 ± 1.1 pM, Fig. 4A; IC₅₀ = 0.20 ± 0.02 nM for PACAP27, 0.12 ± 0.008 nM for PACAP38, and 0.27 ± 0.01 µM for VIP, data not shown). Similar results have been obtained for the PACAP receptor purified from bovine brain (28).

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Binding experiment with the purified PACAP receptor. A, equilibrium binding of increasing concentration of 125I-PACAP27 to the purified PACAP receptor from the infected Sf9 cells (1.8 ng/ml) or from the transformed CHO cells (1.4 ng/ml) was determined in the DG-BSA/TED buffer. Nonspecific binding was determined in the presence of 0.1 µM PACAP27. Binding data was plotted versus total ligand (inset) and was analyzed using Scatchard plot. B, equilibrium binding of 125I-PACAP27 (100 pM) to the purified PACAP receptor (2.6 ng/ml) from the infected Sf9 cells was determined in the DG-BSA/TED buffer including increasing concentration of indicated peptide.

Biochemical Properties of the Purified PACAP Receptor-- The PACAP receptor purified from Sf9 cells showed a broad silver-stained band with M_r = 48,000 (Fig. 5A) in SDS-PAGE analysis, while the PACAP receptor purified from the CHO cells showed a major silver-stained band at M_r = 58,000 (Fig. 5A). Both receptor bands presented the same amino-terminal amino acid sequence of Met-His-Ser-Asp-(unidentified)-Ile-Phe-Lys-Lys-Glu-Gln-. This sequence corresponds to the previously reported amino-terminal amino acid sequence of purified bovine brain PACAP receptor (28). Therefore, insect cells as well as mammalian cells recognize and cleave the signal sequence of PACAP receptor correctly.

View larger version (38K): [in this window] [in a new window]   Fig. 5.   SDS-PAGE of the purified receptor and deglycosylated receptor. The purified PACAP receptor (0.5 µg) from the infected Sf9 insect cells and from the transformed CHO cells were digested with: A, N-glycanase, endoglycosidase F2, endoglycosidase H; B, X. manihotis 1-2,3-mannosidase, X. manihotis -N-acetylglucosaminidase, and Streptococcus sp. sialidase. The digested and undigested PACAP receptors were analyzed using SDS-PAGE in the absence (A) and presence (A and B) of dithiothreitol.

Ligand Binding Properties of the Reconstituted PACAP Receptor-- The purified receptor from insect cells was reconstituted with and without recombinant G_s/bovine brain G at the molar ratio of 1:4:8 (receptor:G_s:G) in crude brain lipid vesicles. Saturation binding experiments performed in the BSA/TED buffer demonstrated that the purified PACAP receptor required G_s/G for expressing high affinity for PACAP27 when it was reconstituted into lipid vesicles (Fig. 6A). The dissociation constant of the reconstituted receptor without G_s/G (K_d = 182.6 ± 26 pM, Fig. 6A) was similar to that of the membranous PACAP receptor in the Sf9 P₁₊₂₊₃ membranes (K_d = 155.3 ± 16.7 pM, Fig. 1A). In contrast, the reconstituted receptor with G_s/G had a single class of high affinity binding sites with a K_d value of 40.2 ± 4.2 pM in the BSA/TED buffer (Fig. 6A). The dissociation constant was comparable to the value of the membranous PACAP receptor in the CHO P₁₊₂₊₃ membranes (K_d = 44.4 ± 2.2 pM, Fig. 1B) but two times larger than the value of the purified receptor in a digitonin-solubilized form (K_d = 17.3 ± 1.3 pM, Fig. 4A).

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Scatchard plot analysis for saturation binding experiments with reconstituted receptor. The purified PACAP receptor (20 pmol) from the Sf9 cells was reconstituted with or without recombinant Gs (80 pmol)/bovine brain G (160 pmol) in lipid vesicles. Equilibrium binding of increasing concentration of 125I-PACAP27 to the reconstitute (2400-fold dilution) with or without Gs/G was determined in the BSA/TED buffer (A: , ), in the BSA/TED buffer containing 100 µM GTPS (A: ), or in the DG-BSA/TED buffer (B). Binding data was plotted versus total ligand (inset) and analyzed using Scatchard plot.

View larger version (21K): [in this window] [in a new window]   Fig. 7.   Effect of nucleotide on the 125I-PACAP27 binding to the reconstituted PACAP receptor. Equilibrium binding of 100 pM 125I-PACAP27 to the purified PACAP receptor reconstituted with Gs/G (prepared as in Fig. 6, 2400-fold dilution) was determined in the BSA/TED buffer (open symbols) or in the DG-BSA/TED buffer (closed) including increasing concentrations of the indicated nucleotide. Nonspecific binding was determined in the presence of 0.1 µM PACAP27. Percent of control specific binding was plotted versus nucleotide concentration.

G Protein Activation by the Reconstituted PACAP Receptor-- Agonist-dependent stimulation of [³⁵S]GTPS binding to the reconstitute was determined to investigate functional coupling of the recombinant PACAP receptor to G protein. Spontaneous [³⁵S]GTPS binding to the reconstitute was very slow in the absence of PACAP27; however, it was greatly enhanced in the presence of 1 µM PACAP27 (Fig. 8). The reagents GDP and NaCl, which decrease the basal [³⁵S]GTPS binding level, were not added to the reaction mixture because the addition of higher concentrations of GDP diminished the agonist-dependent stimulation of [³⁵S]GTPS binding (Fig. 9).

View larger version (15K): [in this window] [in a new window]   Fig. 8.   Time course of [35S]GTPS binding to the reconstitute in the presence and absence of PACAP27. Binding of [35S]GTPS to the reconstitute with Gs/G (prepared as in Fig. 6, 2200-fold dilution) for the indicated incubation period was determined in the BSA/TED-Mg buffer with or without 1 µM PACAP27.

View larger version (17K): [in this window] [in a new window]   Fig. 9.   Effect of GDP on [35S]GTPS binding to the reconstitute. Binding of [35S]GTPS to the reconstitute with Gs/G (prepared as in Fig. 6, 2200-fold dilution) in the presence of the indicated concentration of GDP was determined in the BSA/TED-Mg buffer with or without 1 µM PACAP27.

View larger version (29K): [in this window] [in a new window]   Fig. 10.   Agonist concentration dependence of [35S]GTPS binding to the reconstitute. The purified PACAP receptor from the Sf9 cells (20 pmol, open symbols) or the purified receptor from the CHO cells (20 pmol, closed) was reconstituted with Gs (80 pmol)/G (160 pmol) in lipid vesicles. Binding of [35S]GTPS to the reconstitute (2200-fold dilution) was determined in the BSA/TED-Mg buffer with increasing concentration of the indicated peptide. A similar experiment was performed using the reconstituted Gs (80 pmol)/G (160 pmol) without the PACAP receptor (inset). Increase in GTPS binding was obtained by subtracting basal binding (20 pM) from total binding.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

In the present study, the human PACAP receptor was overexpressed in Sf9 insect cells under the control of a strong polyhedrin promoter. The expression level (50-150 pmol/mg) was higher than those of other G protein-coupled receptors so far reported (ranging from 0.5 to 100 pmol/mg) (26, 33, 38-42). The presence of a signal sequence and many potential glycosylation sites in the PACAP receptor, favoring the translocation of the receptor to cell membranes, presumably contributes to the high expression level. Furthermore, trimming of the 5'- and 3'-noncoding regions along with the use of pAKKO vector having SR promoter might increase the expression level of the receptor in CHO/dhfr cells compared with the previous expression study using CHO-K1 cells and pRc/CMV expression vector (19). Butkerait et al. (38) suggested that the shortening of the 5'-untranslated sequence to 11 authentic bases might be the reason for their high expression level of the 5-HT_1A receptors in insect cells (5-34 pmol/mg) compared with the values reported by others (0.15 and 3 pmol/mg).

The native PACAP receptor is believed to be coupled primarily to G_s proteins for activating adenylate cyclase. Additional coupling to another G protein species is suggested by the pleiotropic cellular response to PACAP stimulation such as increases in phosphoinositide turnover (20, 21, 43) and intracellular calcium ion concentration (11, 44). The recombinant receptor in the Sf9 cell membranes, however, was not coupled to endogenous G proteins as evidenced by its low affinity and GTPS insensitivity in agonist binding. This is most probably due to a deficiency in the appropriate G proteins. The concentrations of immunoreactive G_i2 and G_o proteins in High 5 insect cells were estimated as 10.7 and 0.5 pmol/mg, respectively, while those in Sf9 insect cells were below the detection limit (45). This finding strongly suggests that the expression level of G_s-like protein in Sf9 insect cells is also very low. Butkerait et al. (38) suggested that the coupling of recombinant receptors to endogenous G proteins in insect cells is related to the receptor expression level. Receptors expressed at a low level, such as the D₄ dopamine receptor (5 pmol/mg) (39), serotonin 5-HT_1B receptor (0.5 pmol/mg) (40), and 5-HT_1A receptor (0.15 pmol/mg) (41), were sensitive to guanine nucleotides in agonist binding, indicating coupling to G protein. On the other hand, the 5HT_1A receptor (5-34 pmol/mg) (38), the formyl peptide receptor (27 pmol/mg) (42), the -adrenergic receptor and the muscarinic receptor (30 pmol/mg) (26), did not exhibit GTPS sensitivity in agonist binding. Therefore, it is not surprising that most of the PACAP receptors overexpressed in insect cells were not coupled to endogenous G proteins.

The recombinant PACAP receptor in the Sf9 cell membranes was purified in a digitonin-solubilized form. The purified receptor had a protein core similar to that purified from CHO cells but with different N-linked sugar chains. The apparent molecular weight of the N-glycanase digested band (M_r = 43,000) was smaller than the calculated molecular weight of the protein core (M_r = 51,354) (19). A possible reason for this discrepancy is proteolytic degradation of the carboxyl terminus region. The carboxyl-terminal amino acid sequence could not, however, be determined in the present study. Another possibility is that strong hydrophobic interactions between membrane spanning -helices restricted complete unfolding, resulting in a higher mobility than soluble proteins with a similar molecular weight.

Glycosidase digestion studies suggested structural differences in the N-linked sugar chains of the PACAP receptors. The PACAP receptor from CHO cells was resistant to endoglycosidase H and endoglycosidase F₂. Endoglycosidase H digests high mannose-type and hybrid-type sugar chains (46). Endoglycosidase F₂ cleaves biantennary complex-type sugar chains preferentially, but also cleaves high mannose-type sugar chains at a slower rate (46). Taken together with the result from digestion by exoglycosidases, it is suggested that the receptor from CHO cells has sialylated tri- or tetraantennary complex-type sugar chains. In contrast, the PACAP receptor from Sf9 cells was digested by endoglycosidase F₂, indicating that it has biantennary complex-type and/or high mannose-type sugar chains. The presence of biantennary complex-type sugar chains, however, is somewhat controversial, because exoglycosidase digestion studies indicate that the PACAP receptor from Sf9 cells has only mannosyl residues at non-reducing terminal but does not have NeuAc and GlcNAc residues. The presence of high mannose-type sugar chains is compatible with the result from mannosidase digestion but inconsistent with the resistance to endoglycosidase H. Considering that the PACAP receptor from Sf9 cells has truncated N-linked sugar chains (paucimannosidic N-linked sugar chains) (47) rather than high mannose-type sugar chains, this discrepancy could be explained by possible difference between substrate specificity of endoglycosidase H and endoglycosidase F₂. Endoglycosidase H scarcely digests some kinds of paucimannosidic N-linked sugar chains such as Man13(Man16)Man14GlcNAc₂ (46), while the reactivity of endoglycosidase F₂ on these sugar chains is not revealed. Direct evidences are required to clarify the structure of N-linked sugar chains in the PACAP receptors.

The purified PACAP receptor presented several oligomer bands upon SDS-PAGE. Similar results were found in various G protein-coupled receptors such as rhodopsin (48) and the olfactory receptor (49). Pharmacological evidence also suggests that the m₂-muscarinic receptor forms a dimeric structure (50). On the other hand, bacteriorhodpsin has been shown to form a trimeric structure in orthorhombic two-dimensional crystals (51). The physiological significance of receptor oligomerization observed in SDS-PAGE is still unclear, because high temperatures used during SDS-PAGE sample preparation might promote artificial oligomerization via hydrophobic interactions as described by Sagné et al. (52). In fact, it was reported that the olfactory receptor forms higher oligomers after prolonged boiling (49). Receptor oligomerization in physiological conditions should be further examined.

The purified receptor had high affinity for PACAP27 and PACAP38 but low affinity for VIP. These ligand binding properties were similar to those of the receptor expressed in CHO cells. It has been proposed that digitonin might stabilize the PACAP receptor in the high affinity state based on the observations that purified PACAP receptor has a high affinity by itself (28). This hypothesis was further substantiated by reconstituting the purified PACAP receptor into lipid vesicles. The receptor reconstituted in lipid vesicles alone had low affinity in the absence of digitonin but high affinity in the presence of digitonin. The molecular mechanism of digitonin action, however, is not yet clear. For example, it is not clear whether solubilization into a lipid/digitonin mixed micelle is required for stabilizing the receptor or if the insertion of a small amount of digitonin into membrane bilayers is sufficient. Also, it is not clear whether digitonin acts directly on the receptor protein or acts indirectly by changing the milieu of the membrane or micelle. The small difference found between K_d values for the purified receptor in a digitonin-solubilized form and the reconstituted receptor in the DG-BSA/TED buffer suggests that complete solubilization is required for the full effect of digitonin. Studies on possible conformational changes in the PACAP receptor induced by digitonin may aid in understanding the G protein-dependent regulation of ligand binding affinity.

It should be also noted that the effect of digitonin is different from receptor to receptor. Some G protein-coupled receptors, such as -adrenergic (53) and neuropeptide Y (54) receptors, were solubilized successfully using digitonin. The corticotropin-releasing factor receptor solubilized with digitonin had high affinity for its ligand and no GTPS sensitivity (55), as observed in the present study. In contrast, the use of digitonin failed in the solubilization of gonadotropin-releasing hormone (56) and B₂ bradykinin (57) receptors. Receptors solubilized with digitonin in high affinity states are easier targets for receptor purification.

The PACAP receptor reconstituted with G_s/G in lipid vesicles had high affinity for PACAP27, in contrast to that without G_s/G. GTPS, known to inhibit receptor/G protein coupling by destabilizing the G protein trimer, almost completely neutralized the effect of G_s/G, demonstrating that the purified PACAP receptor reconstituted in lipid vesicles requires G_s/G for expressing high affinity for PACAP27. The B_max value also decreased when the PACAP receptor was uncoupled from G protein. The result leads to the hypothesis that there are at least two states in G protein-uncoupled PACAP receptors, a low affinity state (K_d = 100-200 pM) and a very low affinity state without detectable affinity for ¹²⁵I-PACAP27. Difference between the B_max values in the presence and absence of digitonin represents G protein-uncoupled receptor in a very low affinity state. It should thus be interpreted that the CHO membranes contain G protein-uncoupled spare PACAP receptor as much as G protein-coupled PACAP receptor. The number of the PACAP receptors at the high affinity state (approximately 20-25 pmol/mg) in the CHO membranes reflects the maximum amount of G protein available for coupling to the PACAP receptor.

GDP as well as GTPS or GTP decreased the specific binding of ¹²⁵I-PACAP27 to the reconstituted receptor. The molecular mechanism of GDP action is thought to decrease the dissociation rate of GDP from the G protein -subunit. Thus, this indicates that high affinity binding of PACAP27 requires the dissociation of GDP, whereas the dissociation of GDP is believed to be promoted by agonist stimulation. In other words, the PACAP receptor has high affinity for PACAP27 when it is interacting with nucleotide-free G protein. Therefore, PACAP binding to the receptor shifts the guanine nucleotide binding equilibrium toward the nucleotide free state, leading to the cooperative PACAP binding and GDP dissociation. A similar result with GDP has been observed in other G protein-coupled receptors such as the muscarinic receptor (58). The mathematical solution for the multiequilibrium system explains the observed effect of GDP on ligand binding (59).

The expression of a homogeneous high affinity was attained with a smaller amount of G protein (PACAP receptor:G_s:G = 1:4:8) compared with other reconstitution studies. Florio and Sternweis (60) described that approximately a 1000-fold excess of G_o protein versus the muscarinic receptor is required for complete expression of high affinity. On the other hand, reconstitution of the ₂-adrenergic receptor (61) or muscarinic receptor (58) at a smaller receptor:G protein ratio (1:1-1:20) converted only a portion of the reconstituted receptors into the high affinity state. In our preliminary experiments, the PACAP receptor reconstituted with recombinant G_i/bovine brain G (PACAP receptor G_i:G = 1:4:8) did not have a high affinity (data not shown). Therefore, the present result implies that the recombinant PACAP receptor from insect cells couples to G_s/G efficiently and preferentially. Further reconstitution studies with various kinds of G proteins at different molar ratios will help us to understand the specificity of G protein coupling and the signal transduction mechanism in the PACAP receptor.

Functional coupling between the reconstituted PACAP receptor and G_s/G was further studied by determining the agonist-dependent increase in [³⁵S]GTPS binding. Spontaneous [³⁵S]GTPS binding occurring in the absence of agonist stimulation was very slow, probably due to the slow dissociation rate of bound GDP from the G_s subunit. The addition of GDP strongly diminished the PACAP-dependent increase in [³⁵S]GTPS binding to G_s/G. Wieland and Jakobs (62) described that agonist activation of the -adrenergic receptor interacting with G_s proteins induces a slight increase in [³⁵S]GTPS binding to erythrocyte membranes in the absence of GDP, but not in the presence of high concentrations of GDP. Taken together, an agonist-dependent increase in [³⁵S]GTPS binding to G_s protein should be determined in the absence of GDP, although this makes it difficult to distinguish the agonist-dependent signal from the high basal [³⁵S]GTPS binding usually encountered in crude membrane fractions. It is thus presumed that the PACAP-dependent increase in [³⁵S]GTPS binding to the CHO cell membranes observed in the presence of GDP might not reflect primary coupling to G_s proteins but secondary coupling to other G proteins. This interpretation may account for PACAP's low potency (EC₅₀ values of 580 pM) and low efficacy in the CHO membranes (the accumulated [³⁵S]GTPS binding for 60 min being only about 50 pM at a receptor concentration of 700 pM).

On the other hand, [³⁵S]GTPS binding to the reconstituted G_s/G was potently enhanced by PACAP27 and PACAP38 in the absence of GDP. The EC₅₀ value of PACAP27 was comparable to its K_d value from saturation binding experiments. An increase in [³⁵S]GTPS binding reached a high level, which was compatible with the receptor concentration. An antagonist peptide PACAP(6-38) (37) had no effect on [³⁵S]GTPS binding. These results indicated that the G protein activating machinery was properly reconstituted. Inverse agonist activity, decreasing [³⁵S]GTPS binding less than control levels, was not observed for PACAP(6-38). This finding is related to the lack of G protein activation by an agonist-vacant PACAP receptor. These are all attributed to very slow GDP dissociation from the reconstituted G_s/G. Furthermore, the G protein activation observed with the recombinant PACAP receptor from insect cells was similar to that observed with the recombinant receptor from CHO cells. Thus, the PACAP receptor purified from insect cells is as functionally active as that from CHO cells.

In conclusion, the human recombinant PACAP receptor was purified from the infected Sf9 insect cells on a large scale, yielding 1 to 2 mg. The purified PACAP receptor was comparable to the PACAP receptor purified from CHO cells in ligand binding and G protein activating properties, although the N-linked sugar chains were different. The purified PACAP receptor provides a good model for studying the structure, function, and regulatory mechanisms of G protein-coupled receptors.