Purification and characterization of the bombesin/gastrin-releasing peptide receptor from Swiss 3T3 cells.

The bombesin/gastrin-releasing peptide (GRP) receptor was solubilized from Swiss mouse 3T3 cell membranes in an active form and was purified about 90,000-fold to near homogeneity by a combination of wheat germ agglutinin-agarose and ligand affinity chromatography. The purified receptor displayed a single diffuse band with a Mr of 75,000-100,000 on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. After treatment of the receptor with N-glycanase, removing N-linked oligosaccharide moieties, the protein yielded a Mr = 38,000 band. These results agree with the Mr value estimated for the GRP receptor that was labeled on Swiss 3T3 cells by cross-linking to 125I-GRP1-27. GRP1-27 bound to the purified receptor with a Kd of 0.038 +/- 0.019 nM. By comparison, the soluble receptor in unfractionated extracts and intact membranes displayed a Kd for GRP1-27 of 0.036 +/- 0.003 nM and 0.13 +/- 0.04 nM, respectively. The relative potencies of a series of GRP analogs for the soluble receptor and intact membranes indicated that the extraction procedure did not significantly alter the receptor's ligand binding specificity. However coupling of the receptor to its guanyl nucleotide regulatory protein was not maintained in the soluble extract, and a G-protein did not co-purify with the receptor. Physiological concentrations of NaCl greatly inhibited the binding of some GRP analogs to the receptor, while the binding of other analogs was not affected. A domain on the GRP molecule involving Lys-13 or Arg-17 was identified which promoted binding to the GRP receptor under conditions of low ionic strength. These findings aided the development of an effective ligand affinity resin for the purification of the GRP receptor.

is homologous to the amphibian peptide bombesin (1). Although the GRP molecule contains 27 amino acids, the biologically active region of the molecule has been localized to the C-terminal 8 amino acid residues (2). GRP has been found to elicit a wide range of biological responses such as the stimulation of gastrointestinal hormone release (3, 4), induction of smooth muscle contraction (5), and, when administered to the central nervous system, alteration of homeostasis and thermoregulation (6, 7). Furthermore, GRP has been shown to be a potent mitogen in Swiss 3T3 cells (8-lo), human small cell lung cancer cells (SCLC) (11,12), and bronchial epithelial cells (13). Potentiation of the growth of some SCLC cell lines by GRP has been found to occur in an autocrine manner (12). The possibility of controlling the growth of SCLC by GRP has stimulated the development of potent antagonists of the cell surface GRP receptor (14,15).
Radiolabeled GRP binds to a number of target tissues such as a clonal line of pituitary cells (16), pancreatic. acinar cells (17), central nervous system cells (18), and Swiss 3T3 cells (10). In addition, neuromedin B receptors in rat esophageal muscle have been characterized which also bind GRP and may be highly homologous to the GRP receptor (19). The GRP receptor and its signal transduction mechanism have been studied most extensively in Swiss 3T3 fibroblasts. GRP binding to Swiss 3T3 cells promotes the breakdown of inositol phospholipids, resulting in the release of Ca*+ from intracellular stores, the activation of protein kinase C, and the induction of cellular oncogenes c-fos and c-myc (20-25). The synergistic effect of forskolin and bombesin on CAMP levels, and the stimulation of cell division by GRP, suggest that several signal transduction pathways may be involved in eliciting these complex cellular responses (25,26). Experiments crosslinking lZ51-  to high affinity binding sites on Swiss 3T3 cells indicate that the GRP receptor is a glycoprotein (A4, = 75,000-85,000) that has a protein core with a M, of 45,000 (27) or 43,000 (28, 29).
Fischer and Schonbrunn have shown that the ligand affinity of the GRP receptor is reduced by guanyl nucleotides in membranes from GH4CI pituitary cells and, to a lesser extent, from HIT islet cells and Swiss 3T3 fibroblast cells (30). Such regulation is considered a hallmark of the coupling of a receptor to guanyl nucleotide-binding regulatory proteins (Gproteins) (31,32). In addition, the GDP analog GDPpS was shown to block the stimulation of protein kinase C activity by bombesin in permeabilized Swiss 3T3 cells (33). These results suggest that enzymes which hydrolyze inositol phospholipids can be regulated by G-proteins coupled to the GRP receptor in a manner that is analogous to the coupling of adenylate cyclase to receptors via G-proteins, which has been characterized in detail (31, 32). To identify the components of the GRP receptor system and study their structure and function, it is necessary to purify the receptor. In this paper, we report the solubilization (20 PM) and was terminated after 60 min at 37 "C by the addition of 1.0 ml of ice-cold 50 mM HEPES, pH 7.6, 7 mM MgCl,, 2 mM EDTA, 10 mg/ml bovine serum albumin, and 30 rg/ml bacitracin.
The membranes were then pelleted at 39,000 X g for 10 min at 4 "C. The supernatants were aspirated, and the radioactivity in the pellets was determined.
For the other experiments presented in this paper, the binding reactions were conducted as described above, except that the binding mixture did not contain MgCl, and the final volume was 500 ~1. The resin was then washed with 100 mM NaP04, pH 7.0, followed by alternating washes with a low pH buffer (100 mM NaAc, pH 4.0, 0.5 M NaCI) and a high pH buffer (100 mM Tris, pH 8.0, 0.5 M NaCl).
Over 95% of the peptide was coupled to the resin. The resin was stored in 100 mM NaPO+ pH 7.0, and 0.04% sodium azide at 4 "C.
Purification of the GRP Receptor-Crude membranes were prepared from 2 x 10' cells cultured in 200 roller bottles (1300 cm') as described above. Before extracting the GRP receptor, the membranes were washed twice in high salt buffer (50 mM HEPES, pH 7.5, 2 mM EDTA, 1.0 M NaCl, 50 pg/ml leupeptin, 2.5 rg/ml pepstatin, 10 rg/ ml aprotinin, and 0.5 mM PMSF), washed once in the same buffer, Other Methods-N-Glycanase treatments of labeled receptor were carried out for 19 h at 37 "C in 0.2 M sodium phosphate. PH 8.6, 10 mM l,lO-o-phenanthroline, 1.3% Nonidet P-46, 0.15% SCS, and'0.7 unit of N-glycanase. X g for 60 min at 4 "C.
The soluble extract was then fractionated by polyethylene glycol (PEG) precipitation.
Ice-cold PEG-8,000 (50% w/v in HzO) was added to the soluble extract, giving a final concentration of 20% (w/v). After thorough mixing, the precipitate was collected by centrifugation at 100,000 X g for 10 min. The pellet was suspended in 25 mM HEPES, 25 mM Tris, pH 7.5, 2 mM EDTA, 0.075% CHAPS, 0.0075% CHS, 5 fig/ml leupeptin, and 10 Kg/ml bacitracin in one-fourth of the volume of the extract before fractionation with the aid of a Potter-Elvehjem homogenizer.
Some protein remained insoluble and was removed by centrifugation at 69,000 x g for 10 min at 4 "C. was found to solubilize lz51-GRPl-27 binding activity from Swiss 3T3 membranes (Fig. 1). Extracts were prepared by incubating membranes with various concentrations of detergent followed by centrifugation at 100,000 x g for 60 min to remove insoluble material. Before extracts were assayed for lz51-GRPl-27 binding activity, the concentration of detergent was normalized to 0.1% CHAPS and 0.02% CHS. Therefore, the results reflected the number of receptor molecules extracted and not the effect of detergent on the ligand binding properties of the receptor. The best yield of solubilized receptor was achieved at 0.75% CHAPS, a concentration of detergent that extracted 40% of the total membrane protein. binding activity was observed in the complete absence of CHS. However, receptors solubilized without CHS regained lZ51-GRPl-27 binding activity when assayed in a medium that contained the cholesterol ester. The amount of binding activity seen in this case was about 30% of the level observed in control experiments where CHS was present during the solubilization step (data not shown). These results indicate that CHS can promote a change in the receptor structure which enables it to bind GRP with high affinity. It may also play a limited role in protecting the receptor from irreversible denaturation by CHAPS. Other detergents, including Triton X-100, Nonidet P-40, and digitonin were not able to solubilize lz51-GRPl-27 binding activity from Swiss 3T3 membranes.
However, the possibility that active GRP receptors could be extracted from membranes by a combination of CHS with these detergents was not addressed.
The concentration dependence of CHAPS on lz51-GRPl-27 binding to the GRP receptor is shown in Fig. 2. CHS was present in the binding medium at a level equivalent to onefifth of the concentration of CHAPS. The specific lZ51-GRPl-27 binding activity of the soluble GRP receptor was maximal at 0.075 to 0.15% CHAPS and fell to negligible levels at concentrations of CHAPS above 0.4%. As also evident from Fig. 2, the level of nonspecifically bound lz51-GRPl-27 increased significantly at concentrations of CHAPS greater than about 0.2%, possibly due to the formation of detergent aggregates that trapped ""I-GRPl-27 on PEI-treated filters used to separate the receptor-ligand complexes from unbound ligand. The inhibition of lz51-GRPl-27 binding caused by CHAPS was completely reversed by reducing the concentration of detergent by either dilution or dialysis. The effect of various concentrations of CHS on the binding activity of the soluble receptor at a constant CHAPS concentration (0.075%) was also investigated. Maximal binding activity was observed at a CHAPS to CHS ratio of 10 to 1 (data not shown).
Ligand Specificity of the GRP Receptor in Soluble Extracts and Intact Membranes-The ligand specificity of the soluble GRP receptor was investigated by analyzing the ability of various peptides and GRP analogs to compete with lz51-GRP binding (Table I)  The ligand specificity of the GRP receptor in isolated membranes was also examined using conditions that were comparable to those employed above ( Table I). The relative affinities of each of the peptides tested for the soluble and the membrane-bound forms of the GRP receptor were similar, indicating that the binding specificity of the receptor was not significantly altered by the solubilization procedure.

Inhibition of GRP Receptor Binding by NaCl-Binding of '*'I-GRPl-27
to the soluble GRP receptor and intact Swiss 3T3 membranes was inhibited by NaCl with an I& of 150 mM and 30 mM, respectively (Fig. 3). Other salts such as KC1 and NaAc also inhibited the binding of lz51-GRPl-27 to membranes (data not shown). 150 mM NaCl increased the & of membranes for GRPl-27 by a factor of about 40 (Table II)  without altering the total number of binding sites present (data not shown). NaCl produced analogous effects on "'I-  binding to the receptor in intact Swiss 3T3 cells. However, the binding of N-acetyl-GRP20-27 to the receptor was unaffected by salt (Table II). Accordingly, GRPl-27 exhibited a 50-fold higher affinity than N-acetyl-GRP20-27 under low salt conditions, but had about the same affinity in the presence of 150 mM NaCl. The experiments discussed The specific binding of "'1-GRPl-27 to membranes (o---O) or soluble receptors (a--O) was performed as described under "Experimental Procedures," except that the concentrations of NaCl indicated in the figure were added to the binding medium. For the determination of ""I-GRPl-27 binding to soluble receptors, the CHAPS and CHS concentrations were 0.075% and 0.0075%, respectively. lz5 I-GRPl-27 binding is shown as a percentage of control values with no NaCl added. GRPl-27 4.5 + 1 0.12 + 0.02 3.1 f 2 0.26 f 0.2 N-AC-GRP20-27 5.7 + 1 6.6 + 1 ND ND ' ND, not determined. above indicate that GRPl-27 has a receptor binding determinant involving lysine 13 and/or arginine 17 under low ionic strength conditions that is not present on the N-acetyl-GRP20-27 molecule. Therefore, the inhibition of GRPI-27 binding to the receptor by salt probably results from localized effects of salt on the interaction of these residues with the receptor.
G-protein Coupling-Fischer and Schonbrunn (30) reported that ligand binding to the GRP receptor in Swiss 3T3 fibroblast membranes was reduced by about 25% in the presence of Gpp(NH)p and Mg+. To extend these findings, the effects of various nucleotides on ligand binding to Swiss 3T3 membranes were examined (Fig. 4). Gpp(NH)p, GTP, and GDP (5 PM) caused a 60% reduction in the binding of lZ5-I-GRPl-27 to the membranes. In contrast, ATP and App(NH)p had no effect on binding. The I& of Gpp(NH)p was found to be 0.58 PM, and, at a concentration of 100 pM, the nucleotide reduced the binding of 12"1-GRPl-27 by over 80% (data not shown). Additional experiments showed that Gpp(NH)p reduces receptor affinity without affecting the number of binding sites (data not shown). These results clearly demonstrate that the GRP receptor in Swiss 3T3 fibroblasts is coupled to a GTP-binding regulatory protein. However, as also shown in Fig. 4, 12"1-GRPl-27 binding to the soluble receptor was unaffected by guanyl nucleotides indicating that the soluble receptor does not maintain coupling to its G-protein. The binding of Y-GRPl-27 to Swiss 3T3 membranes (A-F) was performed as described under "Experimental Procedures," except that 7 mM MgCl, was included in the binding medium. In addition, the following nucleotides (5 PM) were included: A, none; B, ATP; C, App(NH)p; D, GTP; E, Gpp(NH)p; F, GDP. Assays of ligand binding to the soluble receptor (G-J) were performed as described under "Experimental Procedures," using a detergent concentration of 0.075% CHAPS and 0.0075% CHS, except that 7 mM MgC12 was included in the binding medium. In addition, nucleotides (10 +M) were included as follows: G, none; H, App(NH)p; Z, GTP; J, GPPOWP. '251-GRPl-27 bound in the experiments is shown as the percent of the total labeled ligand present.
Purification of the GRP Receptor-Out of a total of 10 independent receptor purifications performed, the data from a representative preparation are summarized in Table III. Crude membranes were washed with high salt buffer to remove extrinsic membrane proteins from the preparation and were solubilized with CHAPS and CHS. The extract was further fractionated by precipitation with PEG, giving a 2fold purification of the receptor with negligible loss of GRP receptor binding activity. PEG precipitation and suspension of the pellet in a smaller volume of buffer were also convenient methods of concentrating the extract. In addition, the detergent concentration could be reduced lo-fold, which significantly enhanced the stability of lZ51-GRP binding activity.
After precipitation by PEG, the soluble extract was chromatographed on a WGA-agarose column. The column bound greater than 95% of the lz51-GRP binding activity in the extract, and elution of the column with N,N',N"-triacetylchitotriose (5 mM) produced a large peak of glycoproteins which generally included over 60% of the total lZ51-GRP binding activity applied to the column. In contrast, all attempts to elute the column with N-acetyl-o-glucosamine resulted in unacceptably poor recovery of '2"I-GRP1-27 binding activity from the column.
The eluate from the WGA-agarose column was then chromatographed on a GRP affinity column. The ligand used for the column was [Nle'*~"']GRP13-27, which had similar affinity for the soluble GRP receptor as GRPl-27 (Table I). [Nle '427]GRP13-27 was coupled to an activated aldehydeagarose resin (Actigel Superflow) via a reduced Schiff base linkage to either the t-amino group of lysine 13, or the N terminus of the peptide.
Application of WGA-agarose purified receptor to the ligand affinity column resulted in the binding of about 80% of the lZ51-GRP binding activity in the extract to the column (Fig.  5). After extensive washing, the column was eluted with column buffer + 0.5 M NaCl. Yields of '*"I-GRP binding activity recovered from the column ranged from 35 to 50% of the total activity loaded. The ability to elute the GRP receptor from the column by salt is consistent with the finding that Procedures," is shown by the dashed line. The value for the binding activity in the flow-through peak (fractions 1-30) was determined from an assay of a pool of fractions across the peak. The column was eluted by the addition of 500 mM NaCl to the column buffer at the position indicated by the arrow. Each fraction contained 2 ml.  binding to the receptor is reversibly inhibited by NaCl (EC& = 150 mM) (Fig. 3). The total amount of protein eluted from the column in fractions containing '*"I-GRP binding activity was estimated from the trace of 280 nm absorbance monitored during the run (Fig. 5), which indicated that the specific activity of the receptor was increased about 300-fold. This figure should be taken as a lower limit of the purification achieved since a significant portion of the absorbance at 280 nm detected in the elution peak came from a small amount of CHS that bound to the resin, and was eluted by salt along with the receptor. SDS-PAGE analysis of the pooled fractions containing ""I-GRP binding activity showed that numerous contaminants remained in the preparation (Fig. 6, lane C). However, the performance of the ligand affinity column was difficult to improve since the protein that bound nonspecifically to the affinity resin could not be eluted by a salt wash of the column without also eluting the GRP receptor.
A number of GRP analogs with lower net charge than [Nle'4,*7 ]GRP13-27 were used as the affinity column ligand in an attempt to reduce the level of protein that bound nonspecifically to the column. The peptides tested were GRP19-27, GRP18-27 (neuromedin C), [Lys3]bombesin, and [ N1e14*27, Leui7]GRP14-27. However, as discussed earlier, analogs of GRP lacking the positively charged residues lysine 13 and arginine 17 displayed affinities for the receptor that were 1 to 2 orders of magnitude lower than [Nle'4,27]GRP13-27 (Table I). None of the affinity resins produced from these reduced charge peptides could efficiently bind the GRP receptor (data not shown).
Since the GRP receptor eluted from the [Nle'4~27]GRP13-27-agarose column maintained its ability to bind GRP, it was possible to further purify the protein by performing a second round of ligand affinity chromatography.
The eluate from the first [Nle'4,27]GRP13-27-agarose column was desalted by a regimen of dilution with a low salt buffer and concentration by ultrafiltration.
The recoveries of "'I-GRP binding activity in this step varied from 50-100%. The desalted sample was applied to a smaller version of the first [Nle'4~27]GRP13-27. agarose column and was again eluted with column buffer + 0.5 M salt. The recovery of GRP receptor binding activity from the second affinity column was generally better than that obtained with the first ligand affinity column step, ranging from 50-80%.
The fractions containing 1251-GRP binding activity were pooled, concentrated by ultrafiltration, and analyzed by SDS-PAGE. The gels showed presence of a single diffuse band of silver-stained material, characteristic of a glycoprotein, with a M, of 75,000-100,000 (Fig. 6, lanes D and E). The protein appeared to be free of significant contaminants. Cross-linking of '251-GRP to the GRP receptor in unfractionated soluble extracts (Fig. 6, lane F) or intact Swiss 3T3 cells (Fig. 8 and Refs. 27-29), labeled a glycoprotein with a SDS-PAGE mobility corresponding to that of the purified receptor protein. In both cases, the production of cross-linked receptor was competed by unlabeled GRPl-27 (Fig. 6, lane G, data not shown for intact cells), indicating that the '251-GRP1-27 was cross-linked to high affinity GRP receptor binding sites. The eluate from the [Nle14.27]GRP13-27 affinity column was chromatographed on a Superose 6 gel filtration column (Fig. 7), which gave essentially quantitative recovery of lz51-GRPl-27 binding activity. This step desalted the sample and removed some of the detergent accumulated when the ligand affinity column eluate was concentrated by ultrafiltration. The apparent molecular weight of the soluble receptor was estimated from the column to be 250,000, which was several times larger than the size of the receptor determined by SDS-PAGE. Elution of the purified receptor from the Superose 6 Superose 6 chromatography and the assay of '251-GRP1-27 binding were performed as described under "Rxnerimental Procedures." Each fraction contained 0.5 ml. column could be followed by the protein's absorbance at 280 nm. The receptor emerged from the column on the leading edge of a larger peak of detergent aggregates. Gel analysis of individual fractions of the Superose 6 column indicated that the level of the M, = 75,000-100,000 receptor band paralleled the Azso of the receptor peak and the level of iz51-GRPl-27 binding activity present in each fraction (data not shown). The gel results also showed that the peak of detergent aggregates did not contain detectable amounts of protein.
The total yield of GRP receptor isolated was estimated by quantitating its 280 nm absorbance from the Superose 6 chromatogram and assuming that the receptor displayed the same extinction coefficient as bovine serum albumin and chicken ovalbumin standards. By this method of analysis, the preparation summarized in Table III yielded 0.73 rg of purified receptor protein. This value is essentially the same as the amount of receptor determined from lz51-GRPl-27 binding data (Table III), using 38,000 g/mol as the molecular weight of the deglycoslyated receptor, which was determined from the analysis described below. The data indicate that a large fraction of the purified receptor remained in an active conformation.
N-Glycanase Treatment-The presence of N-linked carbohydrate on the purified GRP receptor was investigated by treating the protein with N-glycanase and analyzing shifts in mobility of the protein on SDS-PAGE. N-Glycanase hydrolyzes the glycosylamine linkages between the oligosaccharide chains and asparagine residues on a protein without any apparent specificity toward the structure of the carbohydrate moiety present. To enhance the detection of receptor bands and avoid interference from the N-glycanase protein present in the sample (M, = 35,000), the experiment was performed with purified receptor protein that was lz51-labeled. As shown in Fig. 8 (right panel), treatment of lz51-GRP receptor with N-glycanase changed its mobility on the gel from M, = 75,000-100,000 to M, = 38,000. The same result was obtained when unlabeled purified receptor was used in the experiment, and deglycosylated products were visualized by silver staining of the gel (data not shown). In addition, affinity-labeled GRP receptor, produced by cross-linking lz51-GRP to the GRP receptor in intact Swiss 3T3 cells, was also subjected to treatment with N-glycanase (Fig. 8, left panel). The deglycosylated form of iz51-GRP affinity-labeled receptor also exhibited a M, of 38,000, strongly indicating that the GRP receptor labeled by cross-linking and the purified GRP receptor are the same protein.  (Table II, low salt).
The binding of '251-GRP to soluble receptors in an unfractionated extract was also analyzed by the method of Scatchard (Fig. 9B). The crude soluble receptor exhibited a Kd for GRPl-27 of 0.036 + 0.003 nM indicating that the ligand binding properties of the soluble receptor were not altered by purifying the protein to homogeneity. The number of receptor sites (B,.,) present in the crude extract was extrapolated to be 0.60 pmol/mg of membrane protein. The reaction products were run on a 12.5% gel and visualized by fluorography. In order for the GRP receptor to bind Y-GRPl-27 after solubilization by CHAPS, it was necessary to include CHS in the medium. CHS was found to promote formation of the active conformation of the receptor, possibly by directly interacting with hydrophobic domains on the protein. CHS was also found to be crucial for the solubilization of the neurotensin receptor from rat brain in an active form (42) and promoted the stability of the dithiothreitol-reduced form of the P-adrenergic receptor after reconstitution into phospholipid vesicles (43).
On SDS-PAGE gels, the purified receptor runs as a single diffuse band with a M, of 75,000 to 100,000. These results agree with estimates of GRP receptor size made by crosslinking '251-GRP to the receptor in intact Swiss 3T3 cells (Fig.  8) (27-29), or soluble membrane extracts (Fig. 6) It has also been observed that N-glycanase treatment of the cross-linked receptor generates a single deglycosylation intermediate, suggesting that at least two N-linked oligosaccharide chains are present on the receptor molecule (29). The purification of the GRP receptor should make it possible to perform a more comprehensive carbohydrate analysis of the protein.
The apparent molecular weight of the native receptor determined by gel filtration was 250,000, or 2 to 3 times higher than that determined by the SDS-PAGE analysis discussed above. The same value for the size of the GRP receptor was estimated from chromatographic runs of unpurified receptor (data not shown). However, with crude extracts, it was necessary to maintain a relatively high CHAPS concentration (0.25%) during the chromatography to prevent aggregation of the receptor with other proteins in the preparation. The apparent size of the receptor estimated from gel filtration data may be larger than that determined from SDS-PAGE because enough detergent binds to the receptor to greatly increase its hydrodynamic radius. Alternatively, the native receptor may exist as a dimer in membranes or in solution.
However, the fact that M, = 250,000 dimers are not produced in ligand-receptor cross-linking experiments argues against this possibility, since the homobifunctional cross-linking reagent used would be expected to also cross-link receptor subunits. In addition, the mobility of the purified receptor on SDS-PAGE was the same, whether or not the protein was run in a reduced form (data not shown). Therefore, the receptor is not a disulfide-linked homodimer and did not form such a species by oxidation in the course of purification. The amount of protein obtained in purified preparations of the GRP receptor was estimated from lz51-GRP binding data, assuming that each receptor molecule has a single ligand binding site and a molecular weight of 38,000 g/mol. The yield of receptor derived in this manner was approximately the same as that estimated from the absorbance of the receptor at 280 nm. Both these methods of receptor quantitation are subject to relatively large errors, but the close agreement of the values obtained, coupled with data from the crosslinking, SDS-PAGE, and carbohydrate analysis described above, strongly indicate that the receptor preparation is free of major contaminants.
The data also indicate that a large fraction of the purified receptor maintains its native conformation.
As observed preciously (30), treatment of Swiss 3T3 membranes with guanyl nucleotides converts the GRP receptor from a high affinity to a low affinity form, demonstrating that the receptor is coupled to a guanyl nucleotide regulatory protein. This G-protein is likely involved in coupling the receptor to the breakdown of inositol phospholipids (33). Kinetic analysis of ligand dissociation from the Swiss 3T3 membranes in the presence and absence of Gpp(NH)p showed that disruption of G-protein coupling decreased the affinity of the receptor for lZ51-GRPl-27 about lo-fold (30).* However, the affinity of the soluble receptor was not affected by guanyl nucleotides and thus the receptor was no longer coupled to its G-protein.
It was not surprising, therefore, that the GRP receptor did not co-purify with an associated G-protein.
The fact that the purified receptor displays a marginally higher affinity for ligand than the G-protein-coupled membrane form of the receptor is consistent with the notion that the receptor adopts its high affinity conformation when extracted from the 'J. Wu and R. Feldman, unpublished observations.
Purified BombesinlGRP Receptor membrane bilayer. However, it is more likely that other factors acting on the receptor in solution alter its affinity and more than compensate for the loss of affinity due to the disruption of G-protein coupling. A number of different G-proteins appear to be able to regulate inositol phospholipid metabolism, since the stimulation of protein kinase C in some, but not all, systems studied is inhibited by pertussis toxin (44)(45)(46)(47). Despite reports that pertussis toxin inhibits the stimulation of DNA synthesis in Swiss 3T3 cells (48), other work indicates that it does not significantly alter bombesin's ability to stimulate phosphoinositide turnover (49) and does not affect the ligand affinity of the GRP receptor (30). It has also been reported that the p21 gene product of the N-ras proto-oncogene, when overexpressed in NIH 3T3 cells, can couple the GRP receptor to the regulation of protein kinase C (50); however, the role of p21 in normal cells is not yet clear. Future work is warranted to develop conditions that maintain G-protein coupling to the GRP receptor in solution or that allow the functional reconstitution of the receptor and its G-protein(s) into phospholipid vesicles. Such methods may make it possible to identify specific members of the G-protein family that are important in GRP receptor function in Go.
Sodium chloride, at a concentration of 150 mM, strongly inhibited the binding of GRPl-27 to the GRP receptor. In contrast, the binding of truncated analogs, such as N-acetyl-GRP20-27 was not affected by salt. N-Acetyl-GRP20-27 was shown previously to contain all of the determinants for receptor binding present on its parent molecule, GRPl-27, under physiological conditions (2). However, the data described in this paper indicate that additional determinants are used to bind GRPl-27 to the receptor in media containing low ionic strength. This enhanced binding affinity, likely involving lysine 13 or arginine 17 of GRPl-27, was eliminated in the presence of 150 mM NaCl. An understanding of the binding properties of different GRP analogs in relationship to salt concentration was useful in developing an effective ligand affinity chromatography method to purify the GRP receptor. The complex effects of salt noted here underscore the importance of maintaining standard assay conditions that are functionally equivalent to physiological conditions before making conclusions regarding the relative potency of different receptor ligands in Go.
The purification of the GRP receptor in an active form, described in this paper, is a significant step toward the resolution of other components involved in the signal transduction pathway of the receptor. This work will also facilitate obtaining an amino acid sequence of the receptor that could be used to clone the GRP receptor gene.