Biochemistry 1999, 38, 6043-6055
6043
Direct Identification of the Agonist Binding Site in the Human Brain CholecystokininB Receptor† Jonas Anders,‡,§ Martin Blu¨ggel,| Helmut E. Meyer,| Ronald Ku¨hne,⊥ Antonius M. ter Laak,⊥ Elzbieta Kojro,‡ and Falk Fahrenholz*,‡ Institut fu¨ r Biochemie, Johannes Gutenberg UniVersita¨ t Mainz, Becherweg 30, D-55099 Mainz, Germany, Institut fu¨ r Immunologie, Medizinisches Proteom Center, Ruhr-UniVersita¨ t Bochum, UniVersita¨ tsstrasse 150, D-44780 Bochum, Germany, and Forschungsinstitut fu¨ r Molekulare Pharmakologie, Alfred Kowalke Strasse 4, D-10315 Berlin, Germany ReceiVed February 4, 1999; ReVised Manuscript ReceiVed March 9, 1999
ABSTRACT: In investigating the agonist binding site of the human brain cholecystokininB receptor (CCKBR), we employed the direct protein chemical approach using a photoreactive tritiated analogue of sulfated cholecystokinin octapeptide, which contains the p-benzoylbenzoyl moiety at the N-terminus, followed by purification of the affinity-labeled receptor to homogeneity. This probe bound specifically, saturably, and with high affinity (KD ) 1.2 nM) to the CCKBR and has full agonistic activity. As the starting material for receptor purification, we used stably transfected HEK 293 cells overexpressing functional CCKBR. Covalent labeling of the WGA-lectin-enriched receptor revealed a 70-80 kDa glycoprotein with a protein core of about 50 kDa. Identification of the agonist binding site was achieved by the application of subsequent chemical and enzymatical cleavage to the purified receptor. A radiolabeled peptide was identified by Edman degradation amino acid sequence analysis combined with MALDI-TOF mass spectrometry. The position of the radioactive probe within the identified peptide was determined using combined tandem electrospray mass spectrometry and peptide mapping. The probe was covalently attached within the sequence L52ELAIRITLY61 that represents the transition between the N-terminal domain and predicted transmembrane domain 1. Using this interaction as a constraint to orientate the ligand within the putative receptor binding site, a model of the CCK-8s-occupied CCKBR was constructed. The hormone was found to be placed in a binding pocket built from both extracellular and transmembrane domains of CCKBR with its N-terminus mainly interacting with residues Arg57 and Tyr61.
Cholecystokinin (CCK)1 plays an important physiological role both in the gastrointestinal tract and in the central nervous system. So far, two CCK receptor subtypes have been described that could be distinguished by their affinities for gastric peptides and for different forms of CCK. The CCKA receptor (CCKAR) exhibits high-affinity binding only for sulfated forms of CCK and weak affinity for gastric peptides and nonsulfated CCK. In contrast, the CCKB receptor (CCKBR) has less extensive structural requirements † This work was supported by the Fonds der Chemischen Industrie and by the Naturwissenschaftlich-Medizinisches Forschungszentrum, Johannes Gutenberg Universita¨t Mainz. * To whom correspondence should be addressed. E-mail:
[email protected]. Fax: +49-6131-395348. Phone: +496131-395833. ‡ Johannes Gutenberg Universita ¨ t Mainz. § Present address: Division of Molecular Neurobiology, Department of Neuroscience, Karolinska Institute, Doktorsringen 12 B Plan 3, S-17177 Stockholm, Sweden. | Ruhr-Universita¨t Bochum. ⊥ Forschungsinstitut fu¨r Molekulare Pharmakologie. 1 Abbreviations: CCK, cholecystokinin; CCK-8s, sulfated cholecystokinin octapeptide; CCKAR, cholecystokininA receptor; CCKBR, cholecystokininB receptor; EDTA, ethylenediaminetetraacetic acid; EGTA, ethylene glycol bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid; FPLC, fast protein liquid chromatography; HPLC, highperformance liquid chromatography; MALDI-TOF mass spectrometry, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry; SDS, sodium dodecyl sulfate; TFA, trifluoroacetic acid.
for its ligands, since sulfated and nonsulfated forms of CCK and gastric peptides are bound with comparable high affinities (reviewed in ref 1). The CCKBR is a particularly important drug target because of its widespread expression throughout the central nervous system and in certain peripheral tissues, including stomach, pancreas, and smooth muscle. In the central nervous system, the major form of CCK is the C-terminally sulfated cholecystokinin octapeptide (CCK-8s) and the CCKBR is the predominant subtype with the highest receptor densities found in the cortex regions and nucleus caudatus (2). Binding of CCK to the CCKBR in the central nervous system increases the rate of release of dopamine and causes interference with the action of opioids where CCK is suggested to be a specific opiate antagonist in the analgesiamediating system implicated in the perception of pain (35). CCK is colocalized with the neurotransmitters serotonin and γ-aminobutyric acid, and accordingly, it mediates anxiety and panicogenic effects (6, 7). Both CCKAR and CCKBR have been cloned (8-10) and share approximately 50% amino acid identity. Despite their different agonist selectivities, the two receptors use the same signal transduction mechanism: activation of phospholipase C with release of inositol trisphosphate and diacylglycerol ultimately leading to an increase in the intracellular Ca2+ concentration (10, 11). Both CCK receptor subtypes are
10.1021/bi990269z CCC: $18.00 © 1999 American Chemical Society Published on Web 04/16/1999
6044 Biochemistry, Vol. 38, No. 19, 1999 members of the family of rhodopsin-like seven-transmembrane domain G-protein-coupled receptors (GPCRs). Within this broad group of GPCRs, it is known that residues belonging to the transmembrane domains of biogenic amine receptors are the primary determinants of agonist binding (12, 13). For protein and peptide ligands, both transmembrane and extracellular domains of the receptors are reported to contribute to the ligand binding site (14-17). Understanding the molecular interactions involved in agonist binding provides insights that could be helpful in targeted drug design and in understanding the basic mechanisms of GPCR activation. The emphasis is on the question about the molecular basis for two receptor subtypes with different agonist binding selectivities that induce the same signal transduction mechanisms. The current insights into the agonist binding domains of the CCKBR were obtained exclusively from mutagenesis studies, and multiple contributions of both transmembrane domains (TMs) and extracellular loops (ECLs) have been reported. Residues responsible for the high-affinity binding of gastric peptides to the CCKBR were found to be located in the second extracellular loop (18), and critical residues for high-affinity CCK-8s binding were identified in ECL 1 and on top of TM 1 (19) while mutations of transmembrane domain residues interfere only weakly with agonist binding (20-23). However, the data currently available from mutagenesis experiments represent mainly contributions of receptor regions to agonist selectivity or affinity; thus, these studies lack direct structure information about the agonistreceptor complex. In this report, we present the protein chemical approach to gaining additional and direct structure information about the interaction between the CCKB receptor and its agonists. The agonist binding site of CCKBR was probed using a tritiated photoreactive CCK analogue [3H]-p-benzoylbenzoylOrn(propionyl)-CCK-8s, which represents a full agonist on the recombinant CCKB receptor. The sequence L52ELAIRITLY61 in CCKBR that forms the transition between the N-terminal domain and the predicted first transmembrane domain was clearly shown to be the contact site between the N-terminus of the probe and the CCKBR. Using this interaction as a constraint to orientate the ligand within the putative receptor binding site, a model of the CCK-8soccupied CCKBR was constructed, including additional information from site-directed mutagenesis studies. MATERIALS AND METHODS Reagents The reagents used were from the following suppliers: fetal calf serum, PAA Laboratories; dodecyl β-D-maltopyranoside, Anatrace Inc.; N-hydroxysuccinimidyl [3H]propionate, Amersham Buchler; n-octyl glucoside, Calbiochem; cholecystokinin octapeptide, Bachem; and Fura-2AM and cyanogen bromide, Fluka. All proteinases were from Boehringer Mannheim, and the photoreactive tritiated CCK analogue [3H]-p-benzoylbenzoyl-Orn(propionyl)-CCK-8s was synthesized and purified as described previously (24).
Anders et al. recently (25). In brief, the receptor cDNA was fused with an additional c-myc tag to allow immunodetection, cloned into the vector pCDNA3, and used for transfection of cells. Selection of stable transfectants was performed using 1 mg/ mL G418 as a selecting antibiotic until a single clone could be isolated which was designated as HEK 293-CCKBR. Cells were maintained in complete Dulbecco’s modified Eagle’s medium (DMEM) supplemented with penicillin, streptomycin, and 10% (w/v) fetal calf serum. Membrane Preparation Stably transfected HEK 293 cells expressing the c-myctagged human brain CCKB receptor (25) were grown to confluence and then washed with ice-cold PBS/10 mM EDTA. After incubation for 5 min in PBS/10 mM EDTA, the cells were lifted mechanically and collected by sedimentation (800g at room temperature for 10 min). The cell pellet was washed (2 × 10 mL of PBS/10 mM EDTA) and finally resuspended at a cell density of 5 × 107 cells/mL in lysis buffer [10 mM Hepes/NaOH (pH 7.4)] and protease inhibitors 10 mg/mL benzamidine, 10 µg/mL leupeptin, 0.1 mg/ mL Pefabloc, 10 µg/mL 1,10-phenanthroline, 1 mM EGTA, 0.1 mg/mL bacitracin, 10 µg/mL SBTI type I, and 5 µg/mL pepstatin A (Pefabloc from Boehringer Mannheim, all others from Sigma). After the mixture had been shaken for 15 min at 4 °C, cell homogenization was performed by a polytron ultrathorax tissue homogenizer at the maximum setting with three pulses of 15 s at 4 °C. Membranes were collected by centrifugation (32000g for 30 min at 4 °C), resuspended in binding buffer [10 mM Hepes/NaOH (pH 7.4), 120 mM NaCl, 5 mM MgCl2, and 1 mM EGTA], and stored at -70 °C. Receptor Binding Assays For saturation studies, membranes were incubated with increasing concentrations of radioligand in a total volume of 100 µL of binding buffer for 30 min at 30 °C. Then the membranes were washed three times with 3 mL of each filtration buffer [10 mM Hepes (pH 7.4)]. The filtration buffer for the solubilized fraction also contained 7% (w/v) polyethylene glycol (PEG) 8000. Prior to filtration, the solubilized proteins were precipitated by adding 0.075% (w/ v) bovine γ-globulin and 7% (w/v) polyethylene glycol 8000 for 15 min on ice. The amount of bound radioactivity was separated from the amount of free radioactivity by rapid filtration over Whatman GF/C filters using a Brandel cell harvester. Filters were washed twice with the corresponding filtration buffers, placed in scintillation vials, and made transparent with 1.5 mL of ethylene glycol monomethyl ether. After 15 min, 10 mL of scintillation cocktail (Rotiszint eco plus) was added. The amount of radioactivity was measured in a Packard Tri-Carb 2100 TR liquid scintillation counter. Nonspecific binding was assessed in the presence of a 200-fold excess of unlabeled cholecystokinin octapeptide. Data analysis for saturation experiments was carried out using the LIGAND program (26).
Generation of Stably Transfected HEK 293 Cells
Probing the Biological ActiVity of [3H]-p-Benzoylbenzoyl-Orn(propionyl)-CCK-8s
A stably transfected HEK 293 cell line expressing the human brain CCKB receptor was constructed as described
HEK 293 cells stably transfected with the c-myc-tagged CCKBR (25) were grown to 70% confluence and loaded with
CCKBR Agonist Binding Site 2 µM Fura-2AM (Fluka) for 40 min at 37 °C. Then the cells were washed with HBS [10 mM Hepes/NaOH (pH 7.4), 140 mM NaCl, 5 mM KCl, 0.5 mM MgCl2, and 2 mM CaCl2] and were finally suspended in HBS with additional 10 mM glucose. Samples (3.0 mL) of the cell suspension equal to about 4 × 105 cells were placed in a stirred thermostated quartz cuvette in a PTI 810 spectrophotometer at 37 °C, and increasing concentrations (from 10-15 to 10-6 M) of either CCK-8s (Bachem) or p-benzoylbenzoyl-Orn-CCK-8s, the nonradioactive precursor of [3H]-p-benzoylbenzoyl-Orn(propionyl)-CCK-8s, were added. Fluorescence was measured by ratiometry at 500 nm after excitation at 340 and 380 nm. The intracellular Ca2+ concentration was calculated as described (27) after calibration using 0.1% (w/v) Triton X-100 to permeabilize cells to determine the maximum fluorescence ratio. The minimum fluorescence ratio was determined after additionally adding 25 mM EGTA (pH 8.0) and 0.005 N NaOH. All data were corrected against the autofluorescence of unloaded cells. Photoaffinity Labeling
Biochemistry, Vol. 38, No. 19, 1999 6045 0.01% (w/v) Serva Blue G]. Samples were subjected to 9% SDS-PAGE according to the method of Laemmli (28) in the case of the receptor and 16.5% T/6% C Tris/Tricine/ urea SDS-PAGE (30) in the case of cyanogen bromide fragments. After electrophoresis, gels were sliced and the amount of radioactivity was determined as described above. Cyanogen Bromide CleaVage Proteins after gel filtration chromatography or HPEC were collected by the chloroform/methanol procedure (29), and the dried pellet was dissolved in 50 µL of 70% (v/v) formic acid and 5% (v/v) β-mercaptoethanol. After the addition of 5 mg of crystalline cyanogen bromide, the sample was incubated at 20 °C for 20 h under a nitrogen atmosphere in the dark. The labeled peptide was identified by subjecting the cleavage fragments to 16.5% T/6% C Tris/Tricine/urea SDS-PAGE (30) with or without prior enzymatical deglycosylation. After electrophoresis, the gels were cut into 2 mm slices which were treated with 0.5 mL of lumasolve (Packard) each. The amount of radioactivity was determined by liquid scintillation counting using lipoluma (Packard) as the scintillator.
The tritiated photoaffinity labeling probe [3H]-p-benzoylbenzoyl-Orn(propionyl)-CCK-8s was synthesized, and photoaffinity labeling of the CCKBR was performed as described previously (24). In brief, the probe was bound to equilibrium (40 min, 30 °C) to the proteins in binding buffer containing the following additional protease inhibitors: 10 mg/mL benzamidine, 10 µg/mL leupeptin, 0.1 mg/mL Pefabloc, 10 µg/mL 1,10-phenanthroline, 1 mM EGTA, 0.1 mg/mL bacitracin, 10 µg/mL SBTI type I, 5 µg/mL pepstatin A. Then the sample was photolyzed using a 200 W mercury lamp (λ > 320 nm) in a thermostated cuvette at 4 °C for 30 min. Nonspecific labeling was assessed in the presence of 1 µM CCK-8s. To subject the samples to analytical SDS-PAGE according to the method of Laemmli (28), proteins from photoaffinity labeling experiments were precipitated using the chloroform/methanol method (29). After electrophoresis, the gels were cut into 2 mm slices which were treated with 0.5 mL of lumasolve (Packard) each. The amount of radioactivity was determined by liquid scintillation counting using 4 mL of lipoluma (Packard) as the scintillator.
For peptide mapping, the radioactive photoaffinity-labeled CCKB receptor after gel filtration chromatography was subjected to cyanogen bromide cleavage as described above, and the radiolabeled peptide was purified by HPLC (see below) and lyophilized and the residue dissolved in 100 mM (NH4)2CO3 and 1.0% (w/v) n-octyl glucoside. Analytical cleavage by V8 protease was performed by diluting portions of this sample 40-fold with 100 mM (NH4)2CO3 to decrease the detergent concentration to 0.025% (w/v) and finally adding V8 protease (sequencing grade, Boehringer Mannheim) at a concentration of 100 ng/µL. Samples were then incubated for 24 h at 37 °C. Enzymatical reactions were stopped by adding the same volume of SDS sample buffer and subjecting the fragments to analytical Tris/Tricine/urea SDS-PAGE, gel slicing, and β-scintillation counting as described above.
Enzymatical Deglycosylation
Purification of the Photoaffinity-Labeled CCKB Receptor
In the case of receptor deglycosylation, a total of 50 µg of WGA-lectin-enriched and photoaffinity-labeled membrane protein was collected by precipitation using the chloroform/methanol method (29). Pellets were dissolved in 10 µL of 10% (w/v) SDS by incubating them for 15 min at room temperature and finally diluted to a volume of 200 µL with deglycosylation buffer containing 50 mM sodium phosphate (pH 7.5), 1% (w/v) NP-40, and 1% (v/v) β-mercaptoethanol. In the case of cyanogen bromide, cleavage product samples after purification by reversed-phase HPLC were at first lyophilized to evaporate the excess of cyanogen bromide, and the residue was dissolved in deglycosylation buffer. The reaction was started by adding the indicated amounts of N-glycosidase F, and the mixture was incubated at 25 °C for 18 h. Then the samples were mixed with equal volumes of 2× SDS-polyacrylamide gel electrophoresis (PAGE) sample buffer [8% (w/v) SDS, 24% (v/v) glycerol, 50 mM Tris-HCl (pH 6.7), 4% (v/v) β-mercaptoethanol, and
Prior to purification, the CCKBR was solubilized using the nonionic detergent dodecyl β-D-maltopyranoside as described previously (31). Starting with portions of 400 mg protein in membranes of stably transfected HEK 293 cells (25), we obtained 150 mg of the solubilized protein. Lectin Affinity Chromatography. The detergent soluble membrane proteins were at first separated by wheat germ agglutinin (WGA) affinity chromatography. A column (18 cm × 1.5 cm) was filled with WGA-agarose (Pharmacia) at hydrostatic pressure and was equilibrated with starting buffer [10 mM Hepes (pH 7.4), 120 mM NaCl, 5 mM MgCl2, 10% (v/v) glycerol, and 0.1% (w/v) dodecyl β-D-maltopyranoside]. Solubilized membranes (150 mg of protein in 80 mL of buffer) were diluted 6-fold with starting buffer to obtain a detergent concentration of 0.1% (w/v) and were then subjected to the column at 4 °C, and a linear flow rate of 0.25 mL/min was applied. After extensive washing with starting buffer, receptor-containing fractions were eluted
V8 Protease Peptide Mapping
6046 Biochemistry, Vol. 38, No. 19, 1999 using 0.6 M N-acetylglucosamine in starting buffer as a competitor. Peak fractions containing radioligand binding activity were pooled and subsequently subjected to photoaffinity labeling with [3H]-p-benzoylbenzoyl-Orn(propionyl)CCK-8s as described above. Gel Filtration Chromatography. To separate the noncovalently bound excess of the probe from the photoaffinitylabeled CCKBR, gel filtration chromatography was employed as a second purification step. A Superdex 200 column (2 cm × 60 cm) connected to an FPLC system (Pharmacia) was equilibrated with buffer [25 mM Tris-HCl (pH 7.0), 150 mM NaCl, 1 mM EGTA, and 2% (w/v) SDS] at room temperature, and proteins were eluted with 400 mL of buffer at a linear flow rate of 1 mL/min at room temperature. The column eluate was fractionated (5 mL fractions), and 50 µL portions were taken for scintillation counting. Peak fractions containing high-molecular mass radioactivity were pooled, and the photoaffinity-labeled CCKBR was identified by subsequent analytical SDS-PAGE, gel slicing, and β-scintillation counting as described above. The procedure starting with the solubilization of a portion of 400 mg of membrane protein in cell membranes was repeated twice, and both eluates after gel filtration chromatography were pooled prior to further purification of the photoaffinity-labeled CCKB receptor. High-Performance Electrophoresis Chromatography (HPEC). HPEC was performed on an Applied Biosystems model 230A device. For each HPEC run, the protein (500 µg) from 5 mL of a protein solution obtained after gel filtration chromatography was collected by the chloroform/ methanol method (31). The protein pellet was incubated for 30 min with 10 µL of 10% (w/v) SDS, and then 140 µL of sample buffer [15 mM Tris-phosphate (pH 7.5), 2% (w/v) SDS, 0.4% (v/v) β-mercaptoethanol, and 30% (v/v) glycerol] was added and the sample incubated for another 30 min at room temperature. Insoluble material was separated by centrifugation at 10000g for 10 min. One hundred fifty microliters of the supernatant from this centrifugation was applied on a calibrated acrylamide tube gel in the HPEC system. Protein separation was performed with a Trisphosphate/SDS buffer system at pH 7.5 according to the protocol of Applied Biosystems; continuous 6% polyacrylamide tube gels (3.5 mm × 100 mm) were used. The gels were prerun for 60 min with a constant current of 0.5 mA and then for 240 min with a current of 2.5 mA. Calibrating runs with standard proteins and protein separation runs were performed for 30 min with a constant current of 0.5 mA and then for 600 min with a constant current of 2.5 mA. All electrophoresis runs were performed at 8 °C. Proteins were eluted from the gel at a flow rate of 10 µL/min and detected at 220 nm. Protein fractions were collected each 5 min at 4 °C; 1 µL portions were taken both for scintillation counting and for SDS-PAGE analysis for analysis of the protein composition (minigels and silver staining). Relevant peak fractions were stored at -20 °C and prepared for chemical and enzymatical cleavage. Purification of the Agonist Binding Domain of the CCKB Receptor The purified affinity-labeled receptor from nine HPEC runs was collected by the chloroform/methanol procedure (31)
Anders et al. and subjected to cyanogen bromide cleavage as described above. Cyanogen bromide fragments were separated by reversed-phase high-performance liquid chromatography prior to further tryptic cleavage. This was achieved using a Varian system equipped with a C4 reversed-phase analytical column (Vydac 215TP5415, 4.6 mm × 250 mm) using a linear flow rate of 1.5 mL/min at room temperature and a gradient of increasing concentrations of 50% (v/v) acetonitrile in 2-propanol (buffer B) with a background of 0.1% (v/v) trifluoroacetic acid. Buffer A was water containing 0.1% (v/ v) trifluoroacetic acid. The gradient consisted of the following: 20% buffer B over the course of 10 min, advancing to 100% buffer B over the course of 60 min, and holding at 100% buffer B for an additional 10 min. Peak radioactive fractions were collected, and the identity of the isolated peptide was again examined by enzymatical deglycosylation and SDS-PAGE as described above. Pooled eluate fractions containing radioactivity were dried in vacuo and then dissolved in 50 µL of 100 mM (NH4)CO3 containing 1% (w/v) n-octyl glucoside. Trypsin (from bovine pancreas, sequencing grade, Boehringer Mannheim) was added to a final concentration of 100 ng/µL, and the sample was incubated for 5 h at 37 °C. Separation of tryptic peptides was achieved by reversed-phase HPLC. One microliter of TFA was added to the digest prior to application on a C4 reversed-phase microbore column (Vydac 214TP215, 2.1 mm × 150 mm). Tryptic peptides were separated using a linear flow rate of 200 µL/min at 25 °C and a gradient of increasing concentrations of 2-propanol (buffer B) with a background of 0.1% (v/v) trifluoroacetic acid. Buffer A was water containing 0.1% (v/v) trifluoroacetic acid. The gradient consisted of the following: 20% buffer B over the course of 10 min, advancing to 30% buffer B over the course of 20 min and then to 100% buffer B over the course of 45 min, and holding at 100% buffer B for an additional 10 min. The relevant peak fractions were dried in vacuo, stored at -20 °C, and prepared for N-terminal Edman sequencing and mass spectrometry. Edman Degradation Amino Acid Sequence Analysis The purified radioactive peptide (0.6 pmol of radioactivity) was dissolved in 5 µL of 70% (v/v) HCOOH/20% (v/v) 2-propanol/10% (v/v) water and was applied to a Biobrenecoated filter. Sequencing was performed in an Applied Biosystems Procise 494 CLC Sequencer in pulsed-liquid mode according to the protocol from Applied Biosystems. MALDI-TOF Mass Analysis The purified radioactive peptide (0.2 pmol of radioactivity) was dissolved in 1 µL of 80% (v/v) HCOOH, and 5 µL of a saturated solution of R-cyano-4-hydroxycinnamic acid in 0.05% (v/v) TFA/50% (v/v) acetonitrile/49.5% (v/v) water was added. The solution was applied to the target of the mass spectrometer and dried. Mass analysis was performed using a Bruker Reflex III instrument in reflectron mode. The acceleration voltage was 20 kV; the energy of the nitrogen laser (λ ) 337 nm) was attenuated by 79%. All signals with an m/z of
