Expression, Characterization, and Purification of C-terminally Hexahistidine-tagged Thromboxane A2 Receptors*

Thromboxane A2(TxA2) receptors belong to the class of G-protein-coupled receptors. Knowledge of the relationship of structure to function for TxA2 receptors is limited because of their low levels of expression, lengthy purification procedures and poor recoveries. A C-terminal hexahistidine-tag (C-His) was ligated to the α-isoform of TxA2 receptors and expressed in COS-7 and Chinese hamster ovary cells. The C-His-TxA2 receptors bound the radioligands125I-7-[(1R,2S,3S,5R)-6,6-dimethyl-3-(4-benzenesulfonylamino)bicyclo[3.1.1]hept-2-yl]-5(Z)-heptenoic acid, an antagonist, and125I-[1S-1α,2β(5Z),3α(1E,3S*),4α]-7-[3[(3-hydroxy-4-(4′-phenoxy)-1butenyl)-7-oxabicyclo-[2.2.1]heptan-2-yl]-5-heptanoic acid, an agonist, with affinities not significantly different from those of the wild type (wt)-TxA2 receptors. LipofectAMINE transfection of the cDNAs resulted in high levels of expression (B max = 95 ± 6 pmol/mg) of the C-His-TxA2 receptors. In competition binding studies the IC50 values of five different ligands were not significantly different between C-His-TxA2 and wt-TxA2 receptors. Agonist-induced stimulation of cAMP and total inositol phosphate formation were not significantly different between the two receptors. Purification on a Ni2+-NTA column resulted in a rapid (within 4 h) purification with a 36 ± 2% recovery and a 30 ± 6-fold purification (n = 5). The partially purified receptors were resolved on SDS-polyacrylamide gel electrophoresis, transferred to a nitrocellulose membrane, dissolved in acetone/trifluoroacetic acid/hexafluoroisopropanol/sinapinic acid, and successfully subjected to matrix-assisted laser desorption ionization-time of flight mass spectrometry analysis. The results suggest that the combination of a high level of expression of C-His-TxA2 receptors and a rapid purification procedure followed by SDS- polyacrylamide gel electrophoresis may provide a useful approach for mass-spectrometry based structure-function and other studies of TxA2receptors.

Thromboxane A 2 (TxA 2 ) 1 and its immediate precursor pros-taglandin H 2 are metabolites of arachidonic acid that are produced through the cyclooxygenase pathway (1) and stimulate TxA 2 receptors, to induce platelet shape change and aggregation (2), and smooth muscle contraction and proliferation (3). Increased synthesis of TxA 2 has been linked to many cardiovascular, renal, and respiratory diseases (4).
TxA 2 receptors are members of the G-protein-coupled class of receptors (5) and exhibit seven transmembrane domains, which are presumed to contain the ligand binding domain. The cDNAs for human TxA 2 receptors have been cloned from placenta (5), K652 cells (6), endothelial cells (7), and human erythroleukemia cells (HEL, a megakaryocyte-like cell line) (8). Thus the amino acid sequence has been deduced. Information concerning the ligand binding site and other functional domains is currently limited and the result of site-directed mutagenesis studies (9,10). Because, like most G-protein-coupled receptors, TxA 2 receptors are natively expressed at low levels, studies employing affinity labeling techniques to determine the ligand binding domain, which require hundreds of picomoles of receptors, have not been conducted.
The methods currently developed for purification of TxA 2 receptors use human platelets as the source (11). These procedures suffer from low yields and are time-consuming, requiring as many as 5 days, with small quantities of receptor typically being purified per week.
Hexahistidine tagging of proteins for purification by means of immobilized metal ion chromatography has been used for a number of proteins expressed in bacteria (12,13), Sf9 cells (14,15), and mammalian cell lines (16). The hexahistidine tag allows for a fast and efficient one-step purification of the receptors using a nickel ion affinity column (17). Such a small tag does not modify the structure of proteins in a significant way (18). Bovine rhodopsin (14), muscarinic m2 receptors (15), adenosine receptors (16), and ␤-adrenergic receptors (19) have * This work was supported in part by National Institutes of Health Grant HL36838. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ** To whom correspondence and reprint requests should be addressed: Dept. of Cell and Molecular Pharmacology and Experimental Therapeutics, Div. of Clinical Pharmacology, Medical University of South Carolina, 171 Ashley Ave., Charleston, SC 29425. Tel.: 803-792-5477; Fax: 803-792-0816; E-mail: halushpv@musc.edu. 1 The abbreviations used are: TxA 2 , thromboxane A 2 ; CHAPS, 3-[(3cholamidopropyl)dimethylammonio]-1-propanesulfonic acid; ECL, enhanced chemiluminescence; ONO11113, 9,11-epithio-11,12-methano- been purified using this approach.
Of the various forms of mass spectrometry, matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS) has proven to be very useful for the study of proteins and peptides (20). MALDI-MS is highly sensitive and accurate and has a high tolerance for impurities. However, MALDI-MS of hydrophobic membrane proteins has proven to be challenging requiring strong organic solvents (21).
In this study, the ␣-isoform of TxA 2 receptors were expressed as C-terminal hexahistidine-tagged recombinant fusion proteins in COS-7 cells and CHO cells, characterized by radioligand binding assays and functional studies, and partially purified using nickel ion affinity chromatography. The receptors were purified to homogeneity by SDS-PAGE, transferred to a nitrocellulose membrane, dissolved in an acetone-based solvent system, and subjected to MALDI-MS.

EXPERIMENTAL PROCEDURES
Construction of the cDNA for Expression of C-His-TxA 2 Receptors-An oligonucleotide encoding the hexahistidine residue and stop codon was synthesized, with a PstI overhang on the 5Ј end and an XhoI overhang on the 3Ј end. The sequences were as follows: forward strand (P814), 5Ј-GCATCATCATCATCATCATTAGC-3Ј; reverse strand (P815), 3Ј-ACGTCGTAGTAGTAGTAGTAGTAATCGAGCT-5Ј.
The TxA 2 receptor cDNA was originally cloned from HEL cells (8) and inserted into the BamHI/EcoRV site of the mammalian expression vector pcDNA3 (Invitrogen, Carlsbad, CA). The recombinant vector pcDNA3-TxA 2 receptor was digested with BamHI and PstI. PstI excises the cDNA immediately before the stop codon. The fragment was gelpurified and ligated to the oligonucleotide. The resulting construct was ligated into pcDNA3 that had been digested with BamHI and XhoI. DNA sequencing confirmed the ligation.
Transfection of COS-7 Cells Using DEAE-Dextran Method-COS-7 cells were transfected with the recombinant plasmid pcDNA3-C-His-TxA 2 receptor or pcDNA3-wt-TxA 2 receptor using the DEAE-dextran technique. Briefly, cells were grown to approximately 70% confluence in 100 mm dishes, washed twice with PBS, and overlaid with 6 ml of DMEM with 10% FBS containing DEAE-dextran (250 g/ml), Tris (100 mM, pH 7.3), and plasmid (0.8 g/ml). Four hours later, the cells were washed twice with PBS, treated with 10% Me 2 SO in PBS for 3 min, and then incubated in DMEM with 10% FBS containing 0.1 mM chloroquine for 2 h. The cells were washed with PBS and incubated in 10 ml of DMEM with 10% FBS and harvested after 60 -72 h.
Transfection of COS-7 Cells and CHO Cells Using LipofectAMINE-Cells were grown to 80% confluence in 100-mm dishes. cDNA (10 g) and LipofectAMINE (30 l) were separately diluted in serum-free DMEM to 200-l volumes, mixed, and incubated at room temperature for 1 h. Cells were washed twice with PBS and overlaid with 5 ml of serum free DMEM to which the DNA-lipid complexes were added. The cells were incubated for 5 h at 37°C, the medium replaced with DMEM-10% FBS, and harvested after 60 -72 h.
Tunicamycin Treatment-Tunicamycin (2 g/ml) was added to the culture medium 24 h after transfection. Cells were harvested after another 48 h.
Western Blot Analysis-Membrane proteins were resolved using SDS-PAGE. After transfer to a nitrocellulose membrane, the proteins were probed using four different anti-TxA 2 receptor peptide antibodies raised against different regions of the TxA 2 receptor. The antibodies and the sequences they recognized were "PH3" (QHAALFEWHAC; amino acids 88 -97), "PH4" (TLCHVYHGQEC; amino acids 221-231), "767" (QPRLSTRPRS; amino acids 320 -329), and "794" (RPAVASQ-RRA; amino acids 141-150). In addition, an anti C-His antibody (Invitrogen) was also used. All Western blots were performed according to standard protocols (Amersham Pharmacia Biotech). The membranes were blocked for 1 h in TBST (0.1% Tween 20) containing 7.5% milk, incubated for 60 min at room temperature in the antisera diluted 1:5000 in TBST, and washed 2 ϫ 5 min with large volumes of TBST. The secondary antibodies were anti-rabbit (Amersham Pharmacia Biotech) or anti-mouse (Bio-Rad), diluted 1:5000 in TBST. After incubation with the second antibody for 60 min at room temperature, membranes were washed 4 ϫ 5 min with TBST. Detection was performed using the ECL method (Amersham Pharmacia Biotech).
Radioligand Binding Studies-COS-7 cells were harvested 72 h after transfection. Crude membranes were prepared by Dounce homogenization followed by a centrifugation at 100,000 ϫ g. Membranes were resuspended in 25 mM HEPES, pH 7.4. Protein concentrations were determined using a Bio-Rad protein assay. Membrane protein (2-5 g) was incubated with 50,000 -100,000 cpm/tube of [ 125 I]I-BOP (22) or [ 125 I]I-SAP (23) for 30 min at 30°C in the presence of increasing concentrations of I-BOP and I-SAP in a buffer containing HEPES (25 mM), NaCl (150 mM), MgCl 2 (10 mM), KCl (5 mM), pH 6.5. Separation of bound from free ligand was done by rapid filtration through glass fiber filters using a Brandel Harvester (Gaithersburg, MD). Filters were washed three times, and the radioactivity associated with the filters was counted using an LKB ␥ counter. L-657925 (100 M), a TxA 2 receptor antagonist, was used to determine specific binding. Nonspecific binding was 1-2% of the total bound counts for [ 125 I]I-BOP and 10 -15% for [ 125 I]I-SAP.
Stimulation of cAMP and Total Inositol Phosphates-TxA 2 receptormediated signaling was studied by measuring I-BOP stimulation of cAMP and inositol phosphate (IP) formation in transiently transfected CHO cells and COS cells (cAMP only), using a method modified from Kosugi et al. (24). COS-7 and CHO cells in 100-mm dishes were grown to 80% confluence and transfected with wt-TxA 2 receptor or C-His-TxA 2 receptor as described above. After 24 h, the cells were passaged into 12 well plates and grown in inositol-free DMEM supplemented with 10% FBS and 2.5 Ci of myo-2-[ 3 H]inositol (NEN Life Science Products). After 48 h, cells were washed twice with PBS and incubated for 30 min with 0.5 ml of PBS containing 10 mM LiCl, 0.5 mM isobutylmethylxanthine, and 0 -100 nM I-BOP Ϯ 10 or 50 M SQ29548. The reaction was stopped by adding 0.5 ml of ice-cold 10% trichloroacetic acid. The cells were scraped and centrifuged and the supernatant was divided into two aliquots: 100 l for cAMP assay and 900 l for IP assay. cAMP was assayed using a standard radioimmunoassay, modified from Lima (25). For measurement of IPs, the supernatants were extracted with ether three times. The upper layer containing ether was discarded and the pH of the lower organic layer was adjusted to between 6.0 and 7.0 with 50 mM Tris base. The samples were loaded on AG1-X8 anion exchange columns, and total IPs were eluted using 1.0 M ammonium formate in 0.1 M formic acid. Samples were counted in a liquid scintillation counter (Beckman).
Solubilization of COS-7 Cell Crude Membranes-COS-7 membranes were thawed on ice and resuspended in HEPES (25 mM; pH 7.4) buffer at a protein concentration of 4 mg/ml. Glycerol was added at 0.25 ml/ml of resuspended membranes. Protease inhibitors (see above) were added followed by CHAPS (to a final concentration of 10 mM). The membranes were kept on ice for 4 h with occasional mixing. The membranes were centrifuged at 250,000 ϫ g for 60 min. The supernatant was transferred to a 15-ml conical tube and 0.25 volumes of dilution buffer (25 mM HEPES, 10 mM CHAPS, 20% glycerol, 2.5 M KCl, 1 mg/ml asolectin) were added. The solubilized membranes were kept on ice until used.
Purification of C-His-TxA 2 Receptors Using Ni 2ϩ -NTA Column Chromatography-Ni 2ϩ -NTA resin (0.5 ml) was equilibrated by washing with double-distilled H 2 O followed by the equilibration buffer (25 mM HEPES, 10 mM CHAPS, 20% glycerol, 0.5 M KCl, 0.2 mg/ml asolectin; pH 7.4). Solubilized membranes were added to the resin (0.5 ml) in a 15-ml conical tube and nutated at 4°C for 4 hr. The mixture was transferred to a 10-ml column, and the void was collected. The column was washed with 20 column volumes of equilibration buffer containing 25 mM imidazole. The receptor was then eluted in 0.5-ml fractions in equilibration buffer without the glycerol containing 200 mM imidazole at pH 6.5.
Photoaffinity Labeling of Partially Purified C-His-TxA 2 Receptors-Partially purified receptors (25 l; ϳ2 pmol) were added to 62.5 mM MES, 10 mM CHAPS buffer, pH 6.5 containing 2-4 ϫ 10 6 cpm of [ 125 I]I-SAP-N 3 in a total volume of 100 l. The samples were incubated in the dark for 30 min at 30°C. Cross-linking was achieved by photolysis with UV light for 10 s. The reaction was stopped by adding SDS sample buffer and heating to 70°C for 30 s. The samples were electrophoresed on an SDS gel and autoradiographed.
MALDI-TOF MS Analysis of C-His-TxA 2 Receptors-The procedure for MALDI-TOF MS analysis was modified from Liang et al. (26). Partially purified receptors (ϳ100 pmol) were electrophoresed on an 11% SDS-polyacrylamide gel and electroblotted to nitrocellulose. The nitrocellulose membrane was stained with Ponceau S (0.5% in 1% acetic acid) for 2 min and destained using distilled water. The band representing C-His-TxA 2 receptors was excised with a sharp blade, dried, and dissolved in acetone/trifluoroacetic acid (5%) saturated with sinapinic acid. Hexafluoroisopropanol was added to a final concentration of 12.5%. After thorough mixing, 1 l of the sample was placed on the sample plates, dried and subjected to MALDI-TOF MS analysis. Mass spectra were obtained using a Voyager-DE (PerSeptive Biosystems) time of flight mass spectrometer equipped with delayed extraction technology for improved mass resolution. External calibration was performed using insulin and myoglobin as standards.
Protein Assays-Protein assays were performed using a Bio-Rad protein assay kit. For very dilute samples, Amido Black protein assay (27) was used.
Other Materials-The following were generous gifts: ONO11113 from Ono Pharmaceutical Co., Osaka, Japan; U46619, Upjohn Co., Kalamazoo, MI; L657925 and L657926, Merck Frosst Canada Inc., Point Claire-Dorval, Quebec, Canada; SQ29548, Dr. Martin Ogletree, Squibb Institute for Medical Research, Princeton, NJ. Restriction enzymes and T4 ligase were purchased from New England Biolabs, Beverly, MA. Acrylamide, protein molecular weight standards, and other reagents used for SDS-polyacrylamide gel electrophoresis were purchased from Bio-Rad. Unless stated, all other reagents were from Sigma.
Data Analysis-Radioligand binding data were subjected to Scatchard analysis using the programs EBDA and LIGAND for the Macintosh (28). Statistical analyses (means and standard errors, Student's t test, analysis of variance) were performed using the software package Statview.

RESULTS
Radioligand Binding Assays of C-His-TxA 2 and wt-TxA 2 Receptors Expressed in COS-7 Cells-Membranes prepared from COS-7 cells transfected using the DEAE-dextran method were assayed for ligand binding to determine the affinities and the level of expression of C-His-TxA 2 and wt-TxA 2 receptors. Saturation-equilibrium binding assays were performed with [ 125 I]I-BOP. Scatchard analysis of the binding data revealed a K d of 1.8 Ϯ 0.2 nM and a B max of 24 Ϯ 1.8 pmol/mg protein for C-His-TxA 2 receptors (n ϭ 26), which are not significantly different from the values obtained for wt-TxA 2 receptors (n ϭ 3) ( Table I).
As there have been reports of differences between agonist and antagonist binding to TxA 2 receptors under certain conditions (29), saturation-equilibrium binding assays were also performed using the TxA 2 receptor antagonist, [ 125 I]I-SAP. The K d values of [ 125 I]I-SAP for the C-His-TxA 2 receptor and wt-TxA 2 receptor were not significantly different (Table I).
LipofectAMINE has been reported to give high levels of transfection efficiency. Thus, the LipofectAMINE method was also used to transfect COS-7 cells and the expression levels were determined by radioligand binding assays (Table I). Competition Binding Studies of TxA 2 Receptors Expressed in COS-7 Cells-To determine if the C-terminal hexahistidine tag interferes with agonist or antagonist binding, competition binding studies were performed. Two agonists (ONO11113 and U46619) and three antagonists (SQ29548, L657925, and L657926) were studied. The latter two, L657925 and L657926, are stereoisomers that exhibit a 100 -150-fold difference in their affinity for TxA 2 receptors in platelets. Neither the IC 50 values nor the rank order of potency for the compounds were significantly different for the C-His-TxA 2 receptors compared with the wt-TxA 2 receptors (Table II).
Stimulation of cAMP and Total IP Formation-Agonist stimulation of wt-TxA 2 receptors has been shown to result in increases in cAMP (30) and IPs (31). As the C-terminal domain of the TxA 2 receptor may be involved in signaling (32), it was important to determine if the hexahistidine tag significantly altered agonist induced generation of second messengers. Thus, I-BOP stimulation of cAMP and IP formation was measured in transiently transfected CHO cells because COS-7 cells were found to express an endogenous TxA 2 receptor at a low level of 0.5 pmol/mg protein (33). I-BOP stimulation of C-His-TxA 2 receptors and wt-TxA 2 receptors resulted in concentrationdependent increases in cAMP formation that were not significantly different (Fig. 1A). There were also no significant differences in the maximal increase in cAMP levels. I-BOP stimulation of C-His-TxA 2 receptors and wt-TxA 2 receptors resulted in concentration-dependent stimulations of IP formation that did not significantly differ from each other (Fig. 1B). There was no significant difference in the maximal increase in IP formation between the two receptors.
As the receptors were transfected into COS-7 cells for the purposes of expression and purification, their ability to couple to the G-proteins was also assessed in this cell line (Fig. 1C). Stimulation with I-BOP produced significantly (p Ͻ 0.05) greater maximal cAMP responses (6 -9-fold stimulation) in COS-7 cells transfected with either of the two TxA 2 receptors compared with the untransfected cells (3-fold stimulation (NS)). A significantly (p Ͻ 0.05) greater maximal cAMP response with wt-TxA 2 receptors transfected using Lipo-fectAMINE compared with DEAE-dextran was observed. There were no significant differences between the maximal cAMP responses with C-His-TxA 2 receptors transfected using DEAEdextran compared with LipofectAMINE or between maximal cAMP response with C-His-TxA 2 receptors compared with wt-TxA 2 receptors for the same transfection method.
Western Blot Analysis of TxA 2 Receptors Expressed in COS-7 Cells-Expression of the TxA 2 receptors was confirmed by a Western blot analysis using the anti-peptide antibody PH-4 or with anti-C-His antibody (Invitrogen). The lanes containing solubilized membranes prepared from COS-7 cells transfected with C-His-TxA 2 and wt-TxA 2 receptors were probed with PH-4. There was a band of similar intensity at approximately 55 kDa, consistent with the mass of TxA 2 receptors (Fig. 2A). These results are in concert with the B max values obtained for the two receptors. Although COS-7 cells express an endogenous TxA 2 receptor, the amount of protein loaded on the gel is insufficient to see it. When probed with the anti-C-His antibody (Fig. 2B), the band was seen only in the lane containing membranes from the C-His-TxA 2 receptors and not the wt-TxA 2 receptors, verifying the expression of the hexahistidine peptide fused to the TxA 2 receptor.
To verify that the C-His-TxA 2 receptors were glycosylated as has been reported for wt-TxA 2 receptors, the transfected COS-7 cells were incubated with tunicamycin which blocks glycosylation of proteins (Fig. 2C). There was an attenuation in the intensity of the band at 55 kDa, with the simultaneous appearance of a band at 40 kDa, which is the predicted mass of nonglycosylated C-His-TxA 2 receptors.
Purification of C-His-TxA 2 Receptors-Purification of the TxA 2 receptors was performed using Ni 2ϩ -NTA affinity chromatography. Fig. 3 shows a representative elution profile for C-His TxA 2 receptors eluted from the Ni 2ϩ -NTA column. Based on the low level of binding activity (13 Ϯ 3%) in the void fraction, it appears that there is a high efficiency of uptake for the C-His-TxA 2 receptors. There was also 74 Ϯ 4% of the original protein in the void. The recovery of partially purified receptors (n ϭ 4) was 36 Ϯ 2%. The -fold purification was 30 Ϯ 6. The B max obtained in the purified receptor was 2600 Ϯ 800 pmol/mg protein (Table III). The theoretical B max of the receptors purified to homogeneity is approximately 15 nmol/mg protein.
The purification was also followed by SDS-PAGE/silver stain and immunoblot analysis (Fig. 4). The silver stain shows an enrichment of a protein band at about 56 kDa, corresponding to the molecular mass of TxA 2 receptors. The samples were probed with the anti-C-His antibody (Invitrogen). The antibody recognized the band that is also enriched on the silver stain. To verify that the protein was the partially purified C-His-TxA 2 receptor immunoblots with four different antipeptide antibodies were conducted (Fig. 5). All the antibodies recognized a protein band at approximately 56 kDa. A protein band at approximately 56 kDa was also seen when the partially purified receptor was probed with the Ni 2ϩ -NTA conjugate (Qiagen) (data not shown).
Photoaffinity Labeling of Partially Purified C-His-TxA 2 Receptors-As an additional assessment of the ligand binding properties of C-His-TxA 2 receptors, photoaffinity labeling with [ 125 I]I-SAP-N 3 was performed. There was specific labeling of a band at approximately 56 kDa, which corresponds to the molecular weight of the TxA 2 receptor (Fig. 6).

MALDI-MS Analysis of C-His-TxA 2 Receptors-MALDI-MS
was performed to demonstrate the feasibility of performing structure-function studies of C-His-TxA 2 receptors. The mass of the C-His-TxA 2 receptor was 54,667 Ϯ 47 Da (n ϭ 16) (Fig.  7). This agrees well with the migration on SDS-PAGE (approximately 55 kDa). Doubly charged species were seen at 27,427 Ϯ 27 mass units (n ϭ 16).

DISCUSSION
A C-terminal hexahistidine-tagged TxA 2 receptor has been constructed and expressed at high levels. Its ligand binding properties (in COS-7 cells) (Table II) and signaling properties (in COS cells and CHO cells) (Fig. 1) are not altered compared with wt-TxA 2 receptors. It can be rapidly and partially purified using Ni 2ϩ -NTA affinity chromatography without alteration in its ligand recognition properties.  1. A and B To verify that the ligand binding properties were not altered, two different approaches were used. A series of competition binding studies with [ 125 I]I-BOP were performed to determine the ligand binding properties of the C-His-TxA 2 receptor compared with the wt-TxA 2 receptor. The results showed that there were no significant differences in the IC 50 values of the two agonists and the three antagonists for the two receptors. Collectively, these results support the notion that the C-terminal hexahistidine tag does not interfere with ligand binding.
The endothelial cell TxA 2 receptor is a splice variant that differs from the platelet TxA 2 receptor because it has a longer C-terminal tail and, although it shares some signaling properties with the platelet TxA 2 receptor (32), e.g. increases in cAMP and inositol phosphate production, it also differs in that at low concentrations of agonist, it inhibits cAMP formation. Because we modified the C terminus of the wt-TxA 2 receptor by adding a hexahistidine peptide, we determined if the major signaling pathways of this modified receptor were still present. Agonist stimulation of the C-His-TxA 2 and the wt-TxA 2 receptors with I-BOP revealed that the concentration-response curves were not significantly different in both CHO and COS-7 cells, demonstrating that the C-terminal hexahistidine tag does not alter signaling. In COS-7 cells, we saw 3-fold stimulation of cAMP formation by I-BOP in untransfected cells. This is because COS-7 cells express an endogenous TxA 2 receptor (33). In transfected cells, we had a 6 -9-fold increase in cAMP formation, which is 2-3 times higher than in untransfected cells. The lack of even greater cAMP response in the transfected COS-7 cells, reflecting the increase in receptor (200-fold for Lipo-fectAMINE transfection) compared with untransfected COS-7 cells expressing the endogenous TxA 2 receptor, could be due to the fact that the limiting factor is the subtype of membranebound adenylyl cyclase that is coupled to TxA 2 receptors. Similar results were obtained by Gao et al. (34), who reported that increased expression of type VI adenylyl cyclase, but not ␤adrenergic receptor or G-protein, led to proportionately in-  (37)(38)(39)(40). Allan et al. (8) expressed the TxA 2 receptor in COS-7 cells and found that the expression level was higher than the levels reported in the literature for other receptors. In this study, it was found that LipofectAMINE transfection resulted in a B max of approximately 100 pmol/mg protein. This resulted in approximately 1 nmol of receptor being harvested by transfection of 10 100-mm dishes of COS-7 cells followed by passaging into 150-mm dishes. We have found that the current procedure yields sufficient quantities of C-His-TxA 2 receptors for MALDI-MS (see below). Thus, this level of expression should provide quantities sufficient to conduct structure-function studies.
The recovery of C-His-TxA 2 receptors from the Ni 2ϩ -NTA affinity column was approximately 36%. Other G-protein-coupled receptors have been purified using metal ion affinity chromatography. Janssen et al. (14) purified hexahistidine-tagged bovine rhodopsin and obtained a higher recovery of 80%, as did Kobilka (19), purifying the ␤-adrenergic receptor. Their results differ from the present method because they used larger amounts of starting materials. Kobilka used 28 nmol of ␤-adrenergic receptor and 4 ml of Ni 2ϩ -NTA resin. Optimization of the purification procedure requires that the amount of tagged protein is closely matched to the capacity of the resin used (41). As these were feasibility studies, we loaded approximately 0.7 nmol of C-His-TxA 2 receptors onto 0.5 ml of Ni 2ϩ -NTA resin, which is far below the capacity of the resin (30 -50 nmol/ml). Thus the recovery should be improved by starting with a larger amount of solubilized membranes containing C-His-TxA 2 receptors. Hayashi and Haga (15) obtained a 38% recovery by loading 1.63 nmol of hexahistidine-tagged mAChR onto 0.5 ml of chelating Sepharose containing Co 2ϩ ions. They also reported that they had a poorer recovery of 15% when using Ni 2ϩ ions. Thus different metal ions and resins may be optimal for different receptors.
TxA 2 receptors obtained from platelets have been purified using S-145 affinity chromatography followed by wheat germ agglutinin chromatography (11). This approach has limited utility because it is time-consuming, requiring 5 days, and associated with low recoveries of about 20%. As access to the ligand binding domain is required for the purification, affinitylabeled receptors cannot be purified using this method. The Ni 2ϩ -NTA column chromatography circumvents these problems. It has higher recoveries, is very rapid, and because the procedure does not require that the receptor be in active conformation, can be used to purify denatured, mutant, or affinitylabeled receptors. The S-145/WGA results in a higher -fold purification, with a B max value of the partially purified receptor of approximately 5 nmol/mg protein. The Ni 2ϩ -NTA column chromatography resulted in a B max of the partially purified receptors of 2.6 Ϯ 0.8 nmol/mg protein. This B max is based on the receptor maintaining its affinity for the ligand. The actual amount of receptor protein may be greater than this estimate, as some of it may have lost its binding affinity but still eluted in fractions containing the high concentrations of imidazole.
Studies of the domains involved with ligand binding have relied on site-directed mutagenesis and affinity labeling (42). These are complimentary approaches that together allow for a greater understanding of the structure of the ligand binding  Fig. 3 for description) were separated on a 10% SDS-PAGE. A, silver staining shows the enrichment of a protein band in the 200 mM imidazole-eluted fractions at around 56 kDa, corresponding to the mass of the TxA 2 receptor, which is not prominent in the solubilized membranes or in the 25 mM imidazole wash fractions. The highest intensity of the band is in fraction B, corresponding to the fraction with the highest [ 125 I]I-SAP binding activity (Fig. 3). B, proteins from a second gel with the same samples were transferred to a nitrocellulose membrane and probed with anti-C-His antibody (Invitrogen). The Western blot shows a protein band at approximately 56 kDa, which is seen in the solubilized COS-7 membranes, disappears in the void fraction and the 25 mM imidazole washes, and is seen in the 200 mM imidazole fractions, peaking in fraction B, corresponding to the fraction with the highest [ 125 I]I-SAP binding activity (Fig. 3). domain. Photoaffinity labeling followed by limited proteolysis and sequencing of the labeled fragment has required nanomole quantities of receptors in the starting material (43), and thus has had limited use. Recently, there have been reports of low picomole amounts (1-25 pmol) of proteins being separated on SDS-polyacrylamide gels, transferred to nitrocellulose or polyvinylidene difluoride membranes, and subjected to MALDI-MS (26,44,45). Most of the approaches were reported for water-soluble proteins. We decided to adapt the method of Liang et al. (26), who showed that dissolving the nitrocellulose membrane in acetone was more efficient than elution of the protein from the membrane. However, initial experiments following their method were not successful. We were able to optimize the solvent for hydrophobic proteins by having 5% trifluoroacetic acid rather than 1%, and hexafluoroisopropanol. Additionally, sinapinic acid was found to generate stronger signals than the ␣-cyano-4hydroxycinnamic acid employed by Liang et al. (26). The molecular mass for the C-His-TxA 2 receptor derived from MALDI-MS (54,667 Ϯ 47) agrees well with the migration on SDS-PAGE (approximately 55 kDa).
We were able to photoaffinity-label the partially purified C-His-TxA 2 receptor, demonstrating that it maintained its li-gand binding domain. We have also previously labeled the TxA 2 receptor in crude membranes. 2 Thus it may also be possible to label the receptors in the crude COS-7 cell membranes and then subject them to a denaturing purification, potentially giving higher yields of purified photoaffinity-labeled receptors for peptide mapping studies. The partially purified labeled receptor can be purified to homogeneity by SDS-PAGE and subjected to limited proteolysis (46) to determine the site of attachment of the probe by either repeat SDS-PAGE or by MALDI-MS. In preliminary experiments, we were able to digest the C-His-TxA 2 receptor on the nitrocellulose membranes with CNBr and obtain mass spectra corresponding to some of the predicted peptide fragments of the receptor. 2 We have shown that C-His-TxA 2 receptors can be expressed at high levels and can be rapidly and partially purified. The receptor can then be purified to homogeneity by SDS-PAGE, transferred to nitrocellulose, and subjected to MALDI-TOF MS analysis. The availability of these methods will provide a valuable approach to structure-function and other studies documenting post-translational modifications of TxA 2 receptors, and should be applicable to other G-protein-coupled receptors as well. FIG. 5. Western blot analysis of partially purified C-His-TxA 2 receptors using antibodies directed to different regions of the receptor. Partially purified receptors were resolved on a 10% SDS gel and transferred to nitrocellulose membranes. The membranes were probed with four different antibodies, 767, 794, PH-3, and PH-4. All four antibodies recognized a protein with a molecular mass of approximately 56 kDa, corresponding to the mass of C-His-TxA 2 receptors.
FIG. 6. Photoaffinity labeling of the partially purified C-His-TxA 2 receptor. An aliquot (25 l; approximately 2 pmol) of the partially purified receptor was incubated in the dark with 4 ϫ 10 6 cpm of [ 125 I]I-SAP-N 3 in a buffer containing MES (62.5 mM) and CHAPS (10 mM), pH 6.5. Labeling was achieved by exposure to UV light for 10 s, and the reaction was stopped by adding SDS sample buffer. Preincubation for 15 min with L657925 (10 M) was used to determine specificity of labeling. The samples were resolved on a 10% SDS-PAGE gel, dried, and autoradiography performed. The results show that a 56 kDa protein band, corresponding to the mass of C-His-TxA 2 receptors, is labeled in the absence of L657925, and that this labeling is attenuated in the presence of L657925, suggesting that the band represents C-His-TxA 2 receptors.

FIG. 7.
Representative MALDI-TOF mass spectrum of C-His-TxA 2 receptors. Approximately 100 pmol of receptors were resolved on SDS-PAGE and transferred to nitrocellulose. After Ponceau S staining, the strip of nitrocellulose containing the receptor was dissolved in 40 l of acetone containing 5% trifluoroacetic acid and saturating amounts of sinapinic acid. Hexafluoroisopropanol was added to a final concentration of 12.5%. After thorough mixing, 1 l of the sample was placed on the probe. The results show a peak at 54825.5 corresponding to the mass of the C-His-TxA 2 receptors. Doubly charged ions of the receptor are seen at 27391.7.