Affinity chromatography of the muscarinic acetylcholine receptor.

A novel compound, 3-(2'-aminobenzhydryloxy)-tropane (ABT), and an ABT-agarose gel were synthesized and used for the purification of solubilized muscarinic receptors. ABT had a high affinity with an apparent dissociation constant (Kd) of 7 nM for the muscarinic receptors solubilized from the porcine brain by digitonin. An ABT-agarose gel was prepared by coupling ABT with epoxy-activated Sepharose 6B, and the degree of substitution to the gel was determined to be 4-5 mumol/ml of the gel by UV absorption spectrum. During affinity chromatography using 10 ml of the ABT-agarose gel and 100 ml of the digitonin-solubilized preparation, 70% of muscarinic receptors were adsorbed to the gel, in marked contrast with the adsorption of only 2% of proteins. Approximately 25% of muscarinic receptors applied to the gel were eluted biospecifically with 1 mM muscarinic ligands. The purified fraction showed a high affinity for [3H]quinuclidinyl benzylate with a Kd of 0.4 nM and similar specificity for muscarinic ligands to that of unpurified soluble receptors. The protein concentration of the purified fraction was too low to be determined accurately, but very approximately a purification of 10(3)-fold was indicated.

receptors have been shown to be solubilized in their active forms by several detergent systems (Refs. 3-7 and see references of Ref. 3), and their molecular weight has also been estimated by different methods (8-12). However, little progress has been made in the purification of the receptor mainly because of the minute quantities of muscarinic receptors. The high affinity and great specificity of ligand-receptor interactions make affinity chromatography an attractive method for purification. Most muscarinic ligands do not contain a functional group that is necessary for the coupling to gels and contain the ester bond that is not stable under some conditions required for coupling. Therefore, we have synthesized several muscarinic ligands with an amino group and an ether bond instead of an ester bond. In the present paper, we describe an application of affinity chromatography to one of those, ABT, for the purification of muscarinic receptors solubilized by digitonin from porcine cerebral membranes.
Synthesis of Muscarinic Ligands 2-Aminobenzhydrol-TO a solution of 2-aminobenzophenone (30 g, 152 mmol), methanol (300 ml), and benzene (100 ml) in a 500-ml round bottom flask which was chilled in an ice bath, NaBH, (5.7 g, 150 mmol) was added with stirring in 0.5-g portions. After addition was completed, the reaction mixture was stirred for 2 h at room temperature with the flask loosely stoppered. Then the solvent was removed in U~C U O and the oily pellet was poured into ice-cold water.
The white solid which was formed was collected by filtration and washed repeatedly with water. The white solid (2-aminobenzhydrol) was dried in uacuo over PzO,.
ABT-A 100-ml two-necked round bottom flask containing 10 ml of concentrated sulfuric acid was equipped with a sealed stirrer and chilled in an ice-salt bath. To the flask was added 14.1 g (100 mmol) of powdered tropine. After the addition was completed, the mixture was vigorously homogenized at room temperature. To the resulting mixture in an ice bath, 4 g (20 mmol) of 2-aminobenzhydrol were gradually added, and the mixture was stirred to homogeneity. The mixture was then warmed in an oil bath at 60-70 "C with stirring for 30 min. After cooling, it was poured onto a slurry of ice and water. The pH of the mixture was adjusted to 7.0 with 5 M NaOH, and the resulting insoluble yellowish solid was extracted with benzene. This fraction contained 2-aminobenzhydrol and by-products with a little ABT and was discarded. The aqueous layer was separated, washed with benzene, and made alkaline with an excess of 5 M NaOH in an ice bath. The resulting insoluble yellowish oil was extracted with benzene. The benzene extract was washed with water until the water phase became neutral, dried over anhydrous sodium sulfate, and evaporated under reduced pressure. The oily residue (3.75 g) soon became crystalline (yield 58%). It was recrystallized from methanol. m.p. 107-108 "C. The final product showed a single spot in TLC (chloroform:methanol, 9:l) which absorbed UV and reacted with iodine vapor. One of the starting materials (2-aminobenzhydrol) absorbed UV but did not react with iodine, and the other (tropine) reacted with iodine but did not absorb UV.

Preparation of Affinity Gels
Epoxy-activated Sepharose 6B (10 g, dry weight) was washed with 1 liter of water and then mixed with 2 mmol of ABT in a mixture of dioxane (24 ml) and aqueous 0.1 M K2C03 (36 ml). The pH of the suspension was adjusted to above 10.5 by adding 9 ml of 1 M KOH and then 30 ml of dioxane were added to the suspension to prevent precipitation of ABT. The suspension was shaken for 16 h a t 40 "C and then poured on a sintered glass funnel. The gel on the funnel was washed with 2 liters of 50% aqueous dioxane and then 2 liters of water. After suspension in 50 ml of 1 M ethanolamine overnight, the gel was washed successively with 1 liter of water, 0.5 liter of 0.1 M acetate buffer (pH 4), 1 liter of water, 0.5 liter of 0.1 M potassium carbonate buffer (pH 8), and finally with 1 liter of 20 mM Tris-HC1 buffer (pH 7.5). The gel was stored a t 4 "C and washed with 10 volumes of the Tris buffer just before use. The content of bound ABT was determined by measuring UV absorption spectrum of the gel suspension after dilution of the gel with ethylene glycol to 1:20 and was found to be 4-5 pmol/ml of gel ( Fig. 1).

Membrane Preparation
Five hundred grams of frozen porcine cerebrums were thawed and homogenized in 2 liters of 0.32 M sucrose, 5 mM Hepes buffer (pH 7.5), 50 FM CaC12 with a Waring blender for 30 s and then with a Silverson homogenizer three times for 30 s each. The homogenate was diluted with 2 liters of the same buffer and centrifuged at 11,000 x g for 90 min. The pellet was homogenized with a Silverson ho- mogenizer in 4 liters of 50 mM Tris-HC1 buffer (pH 7.5) and centrifuged a t 11,000 X g for 90 min, and this process was repeated once more. In these procedures, the loosely packed pellet was collected and the darker brown portions of the pellet which stuck tightly to the walls of the bottles were discarded. The final pellet was suspended in 0.5 liter of 0.32 M sucrose, 20 mM Tris-HC1 buffer (pH 7.5) and stored a t -80 "C. Concentrations of proteins and [3H]QNB-binding sites were 16.3 mg/ml and 21.5 pmol/ml, respectively.

Solubilization Procedure
About 70 ml of the membrane preparation were thawed and extracted by the addition of 2% digitonin, 20 mM Tris-HC1 buffer (pH 7.5) supplemented with 0.5 mM phenylmethylsulfonyl fluoride, 0.5 pg/ml of pepstatin, 5 pg/ml each of leupeptin, chymostatin, antipain, and aprotinin in a total volume of 300 ml. These protease inhibitors were added to eliminate a possible action of endogenous proteases on muscarinic receptors. The suspension was stirred for 60 min at 4 "C and then centrifuged a t 98,000 X g for 60 min. The supernatant was stored a t -80 "C. Concentrations of proteins and [3H]QNB-binding sites were 1.5-2.3 mg/ml and 1.1-1.5 pmol/ml, respectively. The recoveries of protein and the [3H]QNB binding activity from the membrane preparation were 37-57 and 21-28%, respectively.

Affinity Chromatography
Typically 100 ml of the solubilized preparations were shaken with 10 ml of the ABT-agarose gel in the presence or absence of 0.1 M NaCl for 2 h at 0 "C. The incubated gel was packed in a column and was washed with 150-300 ml of the buffer containing 0.1 or 0.15 M NaC1, 0.1% digitonin, 20 mM Tris-HC1 buffer (pH 7.5) a t 4 "C. In some experiments, protease inhibitors used in the solubilization were included in the washing medium. Muscarinic receptors were eluted with 50-100 ml of 1 mM BP or AB& at 4 "C and collected in polypropylene tubes. All procedures after the elution with muscarinic ligands were carried out in plastic tubes or plastic columns instead of glass vessels. Muscarinic ligands used for the elution were removed from receptors by using small columns (0.7 x 5 cm, 2 ml) of Sephadex

Binding Assays
Solubilized preparations (usually 0.2 ml) or the effluents from the Sephadex columns following the elution from the ABT-agarose gel (usually 0.4 ml) were incubated with 2.5-3 nM ['HIQNB in the digit0nin:phosphate or digitonin:Tris buffer in a total volume of 0.5 ml a t 30 "C for 60 min. No appreciable difference was found between the two buffers with regard to ['HIQNB binding activity. After the incubation, 0.2 ml of the reaction mixture was applied in duplicate on the same Sephadex column used for the separation of muscarinic ligands and the bound form of ['HIQNB was eluted with 1.1 ml of the buffer. The whole eluate was collected in a vial, and the vial was assayed in a liquid scintillation spectrometer with an efficiency of 38%.

Estimation of Protein Concentrations
Protein concentrations were determined as described by Lowry et al. (13) using bovine serum albumin as standard. Special care was taken for the fraction eluted from the ABT-gel by muscarinic ligands because protein content was very low. The effluent from the Sephadex column equilibrated with the digit0nin:phosphate buffer and bovine serum albumin dissolved in the same buffer (0.2 ml each) was reacted with Folin regents in polypropylene tubes in a total volume of 1.3 ml, and the absorbance a t 700 nm was read with Union-Gigen High Sensitivity Spectrophotometer SM 401 at a full scale setting of 0.1.

Synthesis of Ligands and Their Affinities for the Receptor-
In a series of initial attempts, muscarinic ligands with an amino group and an ether bond were synthesized by the standard procedure starting with nitro compounds followed by halogenation, dehydrohalogenation, and reduction of a nitro group, which is more common than the procedure reported here. However, a large amount of by-products was formed during the dehydrohalogenation, and tedious purification steps were required, which resulted in a low yield. Thereafter, we found that dehydration in concentrated sulfuric acid could be used for the reaction including amino compounds. It was essential that the temperature was kept low when the reagents were mixed with sulfuric acid. The structures of synthesized compounds were characterized and verified by NMR and mass spectroscopy. The procedure reported here was very simple, gave high yields, and appears to be generally applicable for the synthesis of amino compounds which are useful for affinity chromatography and affinity labeling.
The binding of ["HIQNB to digitonin-solubilized preparations was inhibited by newly synthesized ligands, and the displacement curves showed an apparent Hill coefficient of unity (data not shown). Concentrations giving a half-maximal inhibition were 630 nM for BP, 250 nM for ABQ, and 50 nM for ABT in the presence of 2.6 nM ['HIQNB; the Kd values were calculated to be 84, 33, and 7 nM, respectively, by assuming a simple mass action.
Adsorption of the Receptor to the ABT-Agarose Gel-The adsorption of solubilized receptors to the ABT-agarose gel was examined in the presence of different concentrations of NaCl (Fig. 2). More than 90% of [3H]QNB binding activity was adsorbed to the gel in the absence of NaCl whether the solubilized preparations were preincubated with ['HIQNB or not. In the presence of 0.1 M NaC1, however, less than 5% of the ["HIQNB-receptor complex was adsorbed to the gel when the solubilized preparations were pretreated with ['HIQNB, in marked contrast with the adsorption of 75% of the ['HI QNB binding activity to the gel without pretreatment. About 55% of proteins in solubilized preparations were adsorbed to the gel in the absence of NaCl, but less than 10% were adsorbed in the presence of 0.1 M NaC1.
Adsorption of the ["HIQNB-receptor complex as well as proteins to the ABT gel in the absence of NaCl may be caused by nonspecific electrostatic interaction with a positive charge of ABT. On the other hand, the adsorption of the ligand-free receptor in the presence of 0.1 M or higher concentrations of NaCl appears to be due to the biospecific ligand-receptor interaction because the ["HjQNB-receptor complex is hardly adsorbed under the conditions.
The adsorbed receptor was found to be eluted with 30 nM ["HIQNB, 0.1 mM atropine, 10 mM oxotremorine, and other muscarinic ligands in small scale experiments. The recovery of ["HIQNB binding activity was dependent on species and concentrations of ligands and was the highest among conditions examined when 1 mM B P or ABQ was used (data not shown).
Affinity Chromatography- Fig. 3 indicates elution profiles of protein and the ['HIQNB binding activity when 100 ml of digitonin-solubilized preparations were applied to 10 ml of the ABT-agarose gel in the presence of 0.1 M NaCl. About 98% of proteins and 29% of ['HIQNB binding activity in the solubilized preparations were not adsorbed to the gel and recovered in a flow-through fraction. Very small amounts of proteins and [,"H]QNB binding activity were eluted by further washing of the gel with buffers containing 0.1 or 0.15 M NaC1. The gel was then eluted with the buffer containing 1 mM BP, and the eluates were passed through Sephadex columns to remove BP. The sharp elution of the ['HIQNB binding activity was detected without elution of appreciable amounts of proteins. The ['HIQNB binding activity in these fractions amounted to 22% of that applied to the gel unless the correction was made for the recovery through the Sephadex column and to 26% when the correction was made by using the recovery factor obtained with crude solubilized preparations.
The reason for the nonquantitative recovery of the ['HI QNB binding activity is not known at present. In connection with this problem, it should be mentioned that the ["HIQNB binding activity in the fraction eluted from the ABT-agarose column with muscarinic ligands was greatly decreased when glass vessels were used or when digitonin was omitted from the medium of the Sephadex column for the binding assay, although these factors hardly affected the /'H]QNB binding activity of crude solubilized preparations. During the course of this study, Andre et al. (14) reported an application of affinity chromatography with dexetimide as a ligand for purification of the muscarinic receptor. They suggested an extreme instability of the purified receptors because the radioligand binding activity was not detected in the purified preparations eluted by atropine. In our experiments, the ['HjQNB binding activity of the purified preparations was hardly decreased during the storage at 4 "C for 2 days, although the acitivity was very likely to be lost by contact with glass vessels or by omission of digitonin from the medium as described above. QNB binding that is the difference of the binding in the absence and presence of 1 PM atropine. The dissociation constant was calculated to be 408 pM, and the maximum binding capacity to be 65 fmol, which corresponded to 0.6 nmol/mg of protein.
Protein concentrations of the fraction eluted with B P were too low to be estimated in a routine assay using the Lowry (13) procedure. We tried to estimate an approximate amount by statistical analysis of the scattered data that were obtained when minute amounts of proteins were subjected to the Lowry procedure. When 0 , 0. for 0 and 0.2 kg of albumin were significantly different with p < 0.05 and the values for 0 and 0.4 pg of albumin with p < 0.001. When 0.2 mi of the fraction with the highest ["HIQNB binding activity was assayed, an average absorbance and standard deviation was 0.048 & 0.005 ( n = 7 ) and the amount of protein was estimated to be 0.54 pg from the above correlation equation.
In preliminary experiments, sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the purified fraction followed by silver staining revealed several bands, suggesting the necessity for further purification steps.
Binding of the Purified Receptor with Muscarinic Ligands- Fig. 4 shows the binding of ['HIQNB with the purified receptor a t different concentrations of ['HIQNB. A straight line obtained by the Scatchard analysis was consistent with a single class of binding sites with a Kd for ['HIQNB of 408 pM. The Kd value was comparable to those reported by others (15) and obtained for unpurified soluble receptors. The specific ['HIQNB binding activity was estimated to be 0.6 nmol/mg of protein by using the protein concentration determined as described above. This value means a 920-fold purification of the muscarinic receptor as the specific [3H]QNB binding activity of unpurified soluble preparations is 0.65 pmol/mg of protein. Fig. 5 shows the displacement curves by atropine, pirenzepine, and acetylcholine of the ['HIQNB binding with the purified receptor.
Apparent Kd values were calculated by assuming a simple mass action and found to be 3 nM for atropine, 144 nM for pirenzepine, and 67 p M for acetylcholine. These values were comparable with those determined for unpurified soluble preparations.
In conclusion, affinity chromatography using the ABTagarose gel newly developed and described here should be useful for large scale purification of the muscarinic receptor.