Characterization of a Novel Iodocyanopindolol and SM-11044 Binding Protein, Which May Mediate Relaxation of Depolarized Rat Colon Tonus*

Studies under blockade of α-, β1-, and β2-adrenoreceptors revealed a good correlation between the responses of rat colon relaxation of depolarized tonus and of rat adipocyte lipolysis elicited by catecholamines or BRL-37344, a selective β3-adrenoreceptor agonist, suggesting β3-adrenoreceptor stimulation. In contrast, SM-11044, a nonselective β-adrenoreceptor agonist, stimulated colon relaxation more efficiently than lipolysis; its effects were differently antagonized by cyanopindolol withpA2 values of 8.31 in colon and of 7.32 in adipocytes. Binding studies in rat colon smooth muscle membranes using [125I]iodocyanopindolol under blockade of adrenaline and serotonin receptors revealed the existence of a single class of sites (K d = 11.0 nm,B max = 716.7 fmol/mg protein). The specific binding was saturable and reversible and was displaced by SM-11044 but not by BRL-37344, isoproterenol, noradrenaline, adrenaline, serotonin, nor dopamine. This binding site was photoaffinity labeled using [125I]iodocyanopindolol-diazirine. The labeling was prevented by SM-11044 but not by BRL-37344. The amino-terminal amino acid sequences of the high performance liquid chromatography-purified peptides generated by enzymatic and chemical cleavages of the affinity labeled 34-kDa protein confirmed that the novel iodocyanopindolol or SM-11044 binding protein of rat colon smooth muscle membranes is different from known adrenaline, serotonin, or dopamine receptors. Its functional role might include the relaxation of depolarized colon.

authors generally suggested that these responses were mediated by an "atypical" ␤-AR, which was identified as the ␤3-AR after the cloning and sequencing of its gene or cDNA in man (7), in mouse (8), or in rat (9,10). Manara and Bianchetti (3) actually reported that the phenylethanol aminotetraline-stimulated rat colon relaxation paralleled rat adipocyte lipolysis, confirming that this response predominantly involved the ␤3-AR.
In contrast, the nonselective ␤-AR agonist SM-11044 has been shown to stimulate guinea pig ileum relaxation of KClinduced tonus more efficiently than rat white adipocyte lipolysis (11), implying the existence in guinea pig ileum of yet another atypical receptor different from the adipocyte ␤3-AR. SM-11044 and BRL-35135A, a potent ␤3-AR agonist, display the additional property of inhibiting leukotriene B4-induced guinea pig eosinophil chemotaxis, whereas isoproterenol and BRL-37344 had no such effect (12,13). This inhibition was unaffected by the non-selective ␤-AR antagonist, propranolol, but was antagonized by alprenolol, a ␤1-,␤2-AR antagonist/ ␤3-AR partial agonist (12,13). These observations confirmed the existence of a functional site in guinea pig ileum and eosinophils that is different from ␤1-, ␤2-, and ␤3-AR.
In the present study, we examined the heterogeneity of the effects of the ␤-AR ligands on rat colon. While this tissue indeed contains ␤3-AR (14) in addition to ␤2-AR and a small population of ␤1-AR (15), our results also strongly suggest the existence of a novel functional SM-11044 or iodocyanopindolol binding site in rat colon. This site was characterized by ligand binding, photoaffinity labeling, and amino acid sequencing, revealing a binding protein designated here SM-11044, or iodocyanopindolol binding protein (SMBP), which is different from known monoamine receptors. Relaxation of Rat Colon-Rat colon segment (2 cm) was suspended in an organ bath containing 10 ml of modified Tyrode's solution (11). The Tyrode solution contained 0.5 M atropine, 0.5 M demethylimipramine, 30 M hydrocortisone, 30 M ascorbic acid, 10 M phentolamine, and 1 M propranolol throughout the study to inhibit spontaneous contraction, neuronal and extra-neuronal uptake of norepinephrine, oxidation of catecholamines, and possible ␣-, ␤1-, and ␤2-AR effects, respectively.

Chemicals
The relaxant action of agonists was determined by measuring relaxation of KCl (100 mM)-induced tonus evoked by cumulative addition of the agonists as described previously (11). In the case of testing the effect of cyanopindolol, it was added 5-10 min before the addition of agonist.
Lipolysis in Rat White Adipocytes-White adipocytes were isolated from epididymal fat pads of male Wistar rats (190 -230 g), and lipolysis was determined according to the previous report (11). The cells were preincubated for 5 min at 37°C in the presence of 30 M ascorbic acid, 10 M phentolamine, and 1 M propranolol. Agonists were then applied and incubated for 90 min. In the case of testing the effect of cyanopindolol, it was added 5 min before the addition of agonist.
Schild Plot-Agonist concentration ratios were determined from the EC 50 values of the concentration-response curves of agonists with or without cyanopindolol, according to the method of Arunlakshana and Schild (16). Linear regression analysis was used to estimate the pA 2 value and slope of the line, after confirming that the regression was linear and the slope was not significantly different from unity (Cochran-Cox test, p Ͼ 0.05). The EC 50 values were calculated using the computer program, InPlot.
Statistical Analysis-Results are expressed as mean Ϯ S.E. Statistical significance between two data sets was examined by Student's t test or Cochran-Cox test, depending on the homogeneity of the variances. Duncan's multiple range test was used for multiple data sets. A probability level of p Ͻ 0.05 was considered to be significant.
Rat Colon Membrane Preparation-Membranes from the colon smooth muscle were prepared from male Wistar rats (300 -360 g) as essentially described by Ek and Nahorski (15). The colon segment was washed in ice-cold Tris/saline (10 mM Tris/HCl, 154 mM NaCl (pH 7.4)) and cut open longitudinally, and the mucosa was removed by scrubbing with a glass slide on an ice-cold plastic plate. The smooth muscle preparations were homogenized with a Polytron homogenizer for 1 min. The homogenate was filtered through a gauze and centrifuged (1,500 ϫ g for 20 min at 4°C), and the supernatant was collected and centrifuged (50,000 ϫ g for 20 min at 4°C). The pellet was resuspended in Tris/ saline and kept at Ϫ80°C until used.
Binding Assays in Rat Colon Membranes-Saturation binding studies were performed in a final volume of 200 l of Tris/saline containing 50 g of membrane proteins and different concentrations (0.05-25 nM) of [ 125 I]ICYP, supplemented with 10 M serotonin (5-HT), 10 M phentolamine, 20 M propranolol, and 1.1 mM ascorbic acid (pH 7.4), to block possible 5-HT receptors, ARs, and oxidation of catecholamines, respectively. The [ 125 I]ICYP was used after removing methyl alcohol by compressed air to avoid the influence of the solvent. Incubations were carried out at 37°C for 30 min in a shaking water bath incubator and terminated by addition of 4 ml of ice-cold Tris/saline followed by rapid filtration under vacuum on a Whatman GF/B filter presoaked in Tris/ saline containing 0.1% polyethyleneimine (pH 7.4). The filters were washed three times with 4 ml of ice-cold Tris/saline, transferred to plastic tubes, and counted in a ␥-counter.
Competition assays were performed against 1 nM [ 125 I]ICYP. Nonspecific binding was determined in the presence of 100 M SM-11044.
The inhibition constant, K i , of a ligand was calculated using the equation described by Cheng and Prusoff (17). The Hill coefficient was calculated by linear regression using saturation experiment data. The pseudo-Hill coefficient and IC 50 were determined by the computer program, InPlot (GraphPad Software).
Photoaffinity Labeling Studies-Photoaffinity labeling was performed in a final volume of 1 ml of Tris/saline containing rat colon membranes (0.5 mg of membrane protein) or adipocytes (1 ϫ 10 6 cells), 1.5 nM [ 125 I]ICYP-diazirine, supplemented with 10 M 5-HT, 10 M phentolamine, 20 M propranolol, and 1.1 mM ascorbic acid (pH 7.4). Photoaffinity labeling of Chinese hamster ovary cells stably transfected with mouse ␤3-AR (8) was performed using intact cells (1 ϫ 10 6 cells) in the absence of 5-HT, phentolamine, propranolol, and ascorbic acid. Incubations were carried out in the presence or absence of competitor at 37°C for 45 min in the dark in a shaking water bath incubator and terminated by addition of 10 ml of ice-cold Tris/saline followed by a rapid centrifugation (150,000 ϫ g for 10 min at 4°C). The membranes were irradiated with an UV lamp for 5 min with cooling by circulating water (18). The labeled membranes were diluted with 10 ml of ice-cold Tris/saline, centrifuged (150,000 ϫ g for 30 min at 4°C), and the pellet was resuspended in Tris/saline and kept at Ϫ80°C.
Preparative SDS-PAGE and Extraction of the Photoaffinity Labeled Proteins-Preparative SDS-PAGE was performed with a large size (160 mm width ϫ 200 mm height ϫ 3 mm thickness) of 12% separating and 4% stacking polyacrylamide gels (40% T, 2.6% C) under reducing conditions essentially according to the methods of Laemmli (19). After electrophoresis, the gels were packed in a plastic bag and autoradiographed for 3 days at 4°C on X-OMAT AR film (Kodak). The photoaffinity labeled proteins were extracted by passive extraction, as follows. The radioactive 34-kDa band was cut out and crushed to small pieces of less than 3 ϫ 3 ϫ 3 mm 3 by squeezing using 10-ml disposable plastic syringe (Terumo, Japan). The gels were immersed in 2 ϫ volume of 100 mM Tris/HCl (pH 8.0) containing 0.1% SDS (extraction buffer) and incubated for 16 h at 37°C with rotating. The extract was recovered using a SPIN-XII (0.45-m pore size, Costar) at 1500 ϫ g for 30 min. The remaining gel pieces were again immersed in 2 ϫ volume of extraction buffer, incubated for 2 h at 37°C with rotating, and the extract was recovered as described above. The two extracts were combined and concentrated to a maximum of 0.5 ml using Centriprep 10 and Centricon 10 (Amicon) and kept at Ϫ20°C.
Cleavage by Endoglycosidase or N-Glycopeptidase F-Photoaffinity labeled membranes in the presence of 10 M 5-HT, 10 M phentolamine, and 20 M propranolol were treated with N-glycopeptidase F or endoglycosidase (Endo Hf), using kits according to the manufacturer's specifications (New England Biolabs, Beverly, MA). Briefly, the membranes were solubilized in 0.5% SDS and 1% 2␤-mercaptoethanol, and 40 g of membrane proteins were incubated with 5000 units of Nglycopeptidase F in the presence of 1% Nonidet P-40 or with 2000 units of Endo Hf for 1 or 3 h at 37°C. The digested samples were subjected to SDS-PAGE of 12% polyacrylamide gels. The gels were then dried and submitted to autoradiography as described above.
Wheat Germ Agglutinin (WGA)-Sepharose Chromatography-Photoaffinity labeled membranes in the presence of 10 M 5-HT, 10 M phentolamine, and 20 M propranolol were solubilized in 1% Triton X-100 in Tris/saline at 4°C for 16 h. The solubilized material was collected after centrifugation (200,000 ϫ g for 1 h at 4°C) and diluted to 0.1% Triton X-100 by Tris/saline. One milliliter gel bed volume of WGA-Sepharose 6MB (Sigma) was washed and equilibrated with 30 ml of 0.1% Triton X-100 in Tris/saline (buffer A), and 1 ml of solubilized material containing 200 g membrane proteins was loaded at room temperature. The unretained fraction was recycled three times. After washing with 10 ml of buffer A, the bound material was eluted with 5 ml of 300 mM N-acetyl-D-glucosamine (Merck) in buffer A. The fractions were subjected to SDS-PAGE of 12% polyacrylamide gels. The gels were then dried and submitted to autoradiography as described above.
Chemical Cleavage of the Extracts from Preparative SDS-PAGE and Purification by HPLC-The photoaffinity labeled proteins extracted from the preparative SDS-PAGE were washed twice by distilled water using Centricon 10 and lyophilized. Chemical cleavage was performed at 5 mg protein/ml of 10% cyanogen bromide, 70% formic acid (CNBr/ formic acid) for 24 h at room temperature or 70% formic acid for 72 h at 37°C in the dark. The cleaved products were diluted with 500 l of distilled water and lyophilized. This washing procedure was repeated three times. The cleaved products were dissolved in SDS-reducing buffer and neutralized by addition of aliquots of 30% NaOH until changing the coloration to blue, and were separated by Tricine/SDS-PAGE. The gels were dried and autoradiographed. The labeled bands were cut out, passively extracted, and blotted on PVDF membranes by centrifugation (ProSpin, Applied Biosystems). The membranes were washed 3 times with 1 ml of 20% methanol to remove SDS and salts. The fragments were extracted by 200 l of 75% hexafluoro-isopropanol. Each elution was dried to 20 l in vacuum concentrator, dissolved in 75 l of Me 2 SO, and 75 l of starting buffer (15% acetonitrile, 15% isopropyl alcohol, 0.5% trifluoroacetic acid; buffer A) and loaded on a C4 reverse phase column (Aquapore Butyl BU-300, 2.1 mm inner diameter, 10 mm length, Applied Biosystems). Separation was carried out by a 120-min gradient elution at 40°C with 50% acetonitrile, 50% isopropyl alcohol containing 0.5% trifluoroacetic acid (buffer B) at a flow rate of 0.35 ml/min using a Waters 625 LC System. The gradient started from 30 to 98% buffer B. The elution of fragments was monitored by the absorbance at 210 and 275 nm, and the elution of radioiodinated products was monitored by ␥-counting of the fractions.
Tryptic Cleavage-The photoaffinity labeled membranes were subjected to SDS-PAGE of 12% polyacrylamide gels. The gels were then dried and submitted to autoradiography as described above. The radioactive band at 34 kDa was excised, immersed in distilled water, and minced to small pieces (2 mm width ϫ 2 mm height). The isolated gel pieces corresponding to 800 g of membrane proteins was digested in 500 l of 100 mM Tris/HCl (pH 8.0) containing 0.1% SDS and 50 g of trypsin (EC 3.4.21.4, type IX from porcine pancreas, Sigma) for 24 h at 37°C according to the method of Kawasaki et al. (20). After digestion, the supernatant was recovered and filtered using a SPIN-X filter (0.45 mm pore size, Costar). The gel pieces were crushed through a nylon mesh (200 mesh) by centrifugation for 10 min at 14,000 ϫ g. A 2-fold volume of 100 mM Tris/HCl containing 0.1% SDS was added to the crushed gels, and a second extraction was performed by incubation for 2 h at 37°C with rotating. After incubation, the supernatants were recovered by SPIN-X filter. The two extracts were combined, vacuum concentrated, and submitted to Tricine/SDS-PAGE.
Tricine/SDS-PAGE-Tryptic and chemically cleaved fragments were separated on a Tricine gel system under reducing conditions (21) using 18% polyacrylamide separating gel containing 10.7% glycerol. After electrophoresis, the gels were stained with 0.25% Coomassie Brilliant Blue R-250 (Sigma) in 40% methanol and 10% acetic acid and destained in 10% acetic acid. The gels were then dried and submitted to autoradiography as described above.
Amino Acid Sequencing-Amino acid sequence determination was performed by Edman degradation (22) with an Applied Biosystems 473A protein sequencer. Samples were applied to precycled filters, coated with Polybrene (Biobrene, Applied Biosystems) to reduce peptide-wash-out and to improve initial yields.

Functional Studies in Rat Colon and White Adipocytes-
Under blockade of ␣-, ␤1-, and ␤2-ARs (in the presence of 10 M phentolamine and 1 M propranolol), a number of ␤-AR agonists relaxed KCl-induced tonus in rat colon smooth muscle segment, with a rank order of potencies of BRL-37344 Ͼ SM-11044 Ͼ Ͼ isoproterenol Ͼ Ͼ norepinephrine ϭ epinephrine (Table I). The intrinsic activity value of SM-11044 was significantly higher than that of isoproterenol (Duncan's multiple range test, p Ͻ 0.05), indicating different modes of action. In rat white adipocytes, the same agonists stimulated lipolysis with a rank order of potencies of BRL-37344 Ͼ Ͼ SM-11044 ϭ isoproterenol Ͼ norepinephrine Ͼ epinephrine (Table I). The linear regression line for isoproterenol, norepinephrine, epinephrine, and BRL-37344 reveals a significant correlation (r ϭ 0.97, p Ͻ 0.05) between agonist-induced rat colon relaxation and adipocyte lipolysis (Fig. 1), suggesting that both effects predominantly involve ␤3-AR stimulation. In contrast to these four ligands, SM-11044 stimulated colon relaxation more efficiently than adipocyte lipolysis (Fig. 1). Indeed, the correlation coefficient ceased to be significant when linear regression was analyzed with all agonists including SM-11044 (r ϭ 0.87, p Ͼ 0.05). These data suggest that SM-11044 acts on ␤3-AR and an additional functional site that mediates relaxation in rat colon.
Antagonism of cyanopindolol for SM-11044 and for isoproterenol was compared in both preparations. Cyanopindolol itself, up to the concentration of 10 M, had no effect on the degree of tonus induced by KCl in rat colon and did not stimulate lipolysis in rat white adipocytes. Cyanopindolol antagonized agonist-induced rat colon relaxation in a concentrationdependent manner, with pA 2 values for SM-11044 of 8.31 (slope ϭ 0.78) and for isoproterenol of 7.65 (slope ϭ 1.03) (Table  II). Cyanopindolol also antagonized agonist-induced rat white adipocyte lipolysis in a concentration-dependent manner, with pA 2 values for SM-11044 of 7.32 (slope ϭ 0.96) and for isoproterenol of 7.44 (slope ϭ 1.08) ( Table II). The similar pA 2 values for isoproterenol in colon (7.65), SM-11044 in adipocytes (7.32), and isoproterenol in adipocytes (7.44), with the slopes close to unity, reveal competitive antagonism of cyanopindolol for both agonists binding to ␤3-AR. Slopes of Schild regression lines were not significantly different from unity. Only the slope for SM-11044 in rat colon (0.78) seemed to be lower than unity with high pA 2 value (8.31), suggesting that SM-11044 and cyanopindolol compete not only for binding to ␤3-AR but also for binding to an additional functional site on rat colon.
Binding Assays in Rat Colon Membranes-To identify this site, binding studies in rat colon smooth muscle membranes were performed using [ 125 I]ICYP as the radioligand and SM-  (Table I).  In competition binding studies, specific binding was not displaced by isoproterenol, norepinephrine, epinephrine, dopamine, nor 5-HT, up to the concentration of 1 mM (Fig. 4A, Table  I). The competition binding by isomers of SM-11044 was stereoselective, SM-14011 (the racemic threo-isomer, K i 2.0 M) being 15 times more effective than SM-14010 (the racemic erythroisomer, K i 29.3 M) (Fig. 4B, Table III). The ␤1-AR antagonist, CGP-20712A, and the ␤3-AR agonist, BRL-37344, did not displace the specific binding up to the concentration of 100 M; the ␤2-AR antagonist, ICI-118551, was effective with a relatively high K i (28.5 M) (Table III). Cyanopindolol was the most effective competitor with a K i of 0.11 M, and pindolol had no effect up to the concentration of 100 M. Carazolol, a ligand structurally related to cyanopindolol, was less effective, despite being more lipophilic (Table III). Interestingly, BRL-35135A (methyl ester of BRL-37344) and ICI-198157 (methyl ester of ICI-201651; ICI-215001, a (S)-enantiomer of ICI-201651) displaced the specific binding, whereas the corresponding acid metabolites were inactive (Table III).
Photoaffinity Labeling Study-Photoaffinity labeling was performed to visualize the specific binding site in rat colon membranes using [ 125 I]ICYP-diazirine. In the presence of 10 M 5-HT and 10 M phentolamine, but in the absence of propranolol, a single dense band of 34 kDa was visualized in addition to two broad bands with apparent molecular masses of 50 and 70 kDa (Fig. 5, lane 1). In contrast, in the presence of 20 M propranolol, 10 M 5-HT, and 10 M phentolamine, that is in the same conditions of the competition binding assay with [ 125 I]ICYP, only the 34-kDa band remained visible (Fig. 5, lane  2). These results suggest that the two broad bands are ␤-ARs. Moreover, the 34-kDa band was not displaced by 100 M BRL-37344 but was displaced by 100 M SM-11044 (Fig. 5, lanes 3  and 4, respectively). These data support the results of the competition binding assay, suggesting the existence of a single specific binding site for [ 125 I]ICYP and SM-11044. On the other hand, in adipocytes, the labeled ␤-ARs (50 -70 kDa) and also 34-kDa bands were detected, but the 34-kDa band was not displaced by SM-11044 (Fig. 6, lanes 1-4). Labeling of mouse ␤3-AR indicated that the apparent molecular mass of the ␤3-AR was 50 kDa (Fig. 6, lane 5).
Two-dimensional PAGE of the photoaffinity labeled membranes confirmed the labeling of a single 34-kDa polypeptide chain corresponding to a pI of 6.0 (Fig. 7).
Extraction of the Photoaffinity Labeled SMBP-Two grams of membrane proteins were prepared from colon smooth muscle isolated from 600 rats.  matography such as reverse-phase HPLC with a C4 column (Aquapore Butyl BU-300, Applied Biosystems), ion exchange chromatography (Aquapore Weak Anion AX-300, Applied Biosystems), chromatofocusing (PBE 94 and Polybuffer 74, Pharmacia), and hydroxyapatite chromatography (Bio-Gel HPHT, Bio-Rad). Therefore, preparative SDS-PAGE was performed to isolate photoaffinity labeled SMBP. Fifty mg of the labeled membranes were loaded on a set of polyacrylamide gels without excessive diffusion of the 34-kDa labeled SMBP (Fig. 8). Extraction of 34-kDa bands yielded 79.3-86.2% of the total radioactive proteins in gels. Enzymatic and Chemical Cleavages, Purification, and Sequencing-The molecular size of the photoaffinity labeled 34-kDa protein was not modified by the enzymatic treatments with endoglycosidase or N-glycopeptidase F, whereas both enzymes reduced the molecular size of ovalbumin from 43 to 40 kDa (data not shown).
Solubilized and photoaffinity labeled 34-kDa proteins (373,298 cpm) were applied to a WGA-Sepharose column. The unretained fraction contained 35.7% of the radioactivity, and washed out fractions contained 53.3% of the radioactivity. The specific sugar, 300 mM N-acetyl-D-glucosamine, eluted only 2.3% of the radiolabeled material. The eluted fraction was subjected to SDS-PAGE after concentration, but the photoaffinity labeled 34-kDa band was not detected (data not shown).
A single 7-kDa labeled peptide was generated upon digestion of the photoaffinity labeled 34-kDa protein with trypsin ( Fig.  9). Recovery yields in final extracts from the gel pieces were 62.7% for the labeled 34-kDa protein and 90.4% for the in situ generated tryptic peptides.
Chemical cleavage has some advantage in contrast to proteolytic digestion; it avoids contamination by protease itself and produces limited numbers of large fragments. Analytically, each 1 mg of the extracted, labeled 34-kDa protein was treated with CNBr in 70% formic acid or in 75% trifluoroacetic acid to compare the effect of the acids. In formic acid, CNBr generated three labeled fragments of 8, 10, and 12 kDa, and formic acid alone generated a single 8-kDa labeled fragment. In the acid condition with trifluoroacetic acid, most of the labeling was dissociated by acid itself, a single 10-kDa labeled fragment was observed by CNBr cleavage (Fig. 10).
The extract of the labeled 34-kDa protein from 400 mg of membranes (411,794 cpm) was preparatively cleaved by CNBr/ formic acid, and an aliquot of the cleaved products was resolved on Tricine/SDS-PAGE gels. Three labeled fragments, a major   one of 12 kDa and two minor ones of 8 and 10 kDa, were observed on autoradiograms of Coomassie Blue-stained gels (Fig. 11). The main radioactive 12-kDa fragment (total 39, 683 cpm) in preparative scale was extracted by passive extraction from Tricine/SDS-PAGE gels without Coomassie Blue staining. The labeled fragment was then blotted on PVDF membranes (35514 cpm). The fragment was extracted (17,040 cpm) after removing SDS and further purified by reverse-phase HPLC. Two radioactive peaks, a minor and a major one, were observed at 62% buffer B (fraction numbers 27 and 28; total 789 cpm) and at 65% buffer B (fraction numbers 30 -32; total 5372 cpm), respectively (Fig. 12). Total recovery yield of the initial radio-activity was 71.3%. The peak fractions were submitted to the protein sequencer, and the resulting amino acid sequences were further analyzed. Sequences of minor and major peaks were almost identical, and Sequence 1 was as follows: 6. SDS-PAGE followed by autoradiography of the photoaffinity labeled adipocytes (lanes 1-4) or Chinese hamster ovary-mouse ␤3-AR cells (lanes 5 and 6)   In parentheses are the putative amino acid residues; X indicates the undetermined amino acid residue; I/Y, I or Y. (Swiss-Prot accession number P80969). A search in the data base (non-redundant GenBank and EMBL sequences) using the BLASTp and tBLASTn program (23) did not reveal any known protein that could match this sequence.
The labeled 34-kDa proteins from 400 mg of membrane (38,1198 cpm) were cleaved by formic acid, and an aliquot of the cleaved products was resolved on Tricine/SDS-PAGE gels. A single labeled fragment of 8 kDa was observed on autoradiograms of Coomassie Blue-stained gels (Fig. 13A). The radioactive 8-kDa fragment (total 21,400 cpm) was extracted by passive extraction from Tricine/SDS-PAGE gels without Coomassie Blue staining and was blotted on PVDF membranes (19,581 cpm). The fragment was extracted from PVDF membranes (10,045 cpm) and further purified by reverse-phase HPLC. One radioactive peak was observed at 62% buffer B (fraction numbers 27 and 28; total 3,239 cpm, Fig. 13B). Total recovery yield of the initial radioactivity was 91.6%. The peak fractions were submitted to protein sequencer, and the resulting amino acid sequence was determined as shown in Sequence 2: In parentheses is the putative aspartic acid residue; X indicates an undetermined residue. (Swiss-Prot accession number P80969). A search in the same data bases as above did not reveal any known protein that could match this sequence. DISCUSSION Functional studies were performed under blockade of ␣-, ␤1-, ␤2-ARs, comparing rat colon smooth muscle segments and white adipose cells. A good correlation was obtained between the efficacy of colon relaxation and adipocyte lipolysis with four reference agonists, isoproterenol, norepinephrine, epinephrine, and BRL-37344, suggesting that both types of responses predominantly involve ␤3-AR. However, SM-11044, another ␤-AR agonist, stimulated colon relaxation more efficiently than adipocyte lipolysis with a higher intrinsic activity value than isoproterenol, a full agonist for the rat ␤3-AR (9). These results suggest that SM-11044 stimulates not only ␤3-AR but also another functional site on rat colon. Cyanopindolol competitively antagonized isoproterenol-induced colon relaxation, lipolysis, and SM-11044-induced lipolysis with similar pA 2 values and slopes close to unity. These pA 2 values (7.32-7.65) are close to the reported values of 7.63 in guinea pig ileum (24) and 7.12 in rat colon (4). Thus, cyanopindolol competitively antagonizes binding of both ligands to ␤3-AR on colon and adipocytes. In contrast, cyanopindolol competes only with SM-11044 for relaxation of colon, with a higher pA 2 value (8.31) along with the low slope of Schild regression line (0.78). These studies suggest that in colon, cyanopindolol and SM-11044 compete not only for binding to the ␤3-AR but also to another functional site that modulates KCl-induced depolarized colon tonus.
[ 125 I]ICYP is known to bind to ␤1-, ␤2-, ␤3-ARs, serotonin 5-HT1A, and 5-HT1B receptors (25,26). The present study demonstrates that this ligand binds to yet another, specific, saturable, and reversible site in rat colon smooth muscle membranes under blockade of the above receptors. The affinity of this novel binding site for [ 125 I]ICYP (K d ; 11 nM) is almost 10 times lower than that of ␤3-AR (K d ; 1.3 nM for [ 125 I]ICYP) (10), and the B max of 716.7 fmol/mg protein is 15 times greater than those of ␤1and ␤2-ARs in rat colon membranes (total 48 fmol/mg protein; ␤1 versus ␤2 ϭ 16 -21% versus 79 -84%) (15). The high pA 2 value of cyanopindolol with low slope of Schild plots observed in SM-11044-induced colon relaxation can be explained by the existence of high (␤3-AR) and low (SMBP) affinity sites under blockade of ␤1and ␤2-ARs. Scatchard   FIG. 12. Reverse-phase HPLC purification of the photoaffinity labeled, cyanogen bromide-cleaved 12-kDa fragment. The fragment isolated from Tricine/SDS-PAGE gels was further purified by reverse-phase HPLC, as described in the text. Fragment was eluted from the C4 column with a linear gradient of 30 -98% buffer B in 120 min (---). Radioactive profile for 12-kDa labeled fragment was shown (q). Based on the amount of recovered radioactivity, HPLC column recovery was 71.3%. analysis of the data indicate that the ligand binds to apparently homogeneous sites as shown by the linearity of the Scatchard plot and the Hill coefficient close to unity. The stereo specificity of the binding is evident from the 15-fold affinity ratio between the threo-and erythro-isomers of SM-11044.
Photoaffinity labeling was used to visualize and confirm the homogeneity of the specific binding site in the presence of monoamine receptor blockers. The displacement by 20 M propranolol of the broad bands migrating at 50 -60 kDa in rat colon and adipocytes suggests that these bands correspond to ␤-ARs. The 34-kDa band in rat colon was clearly displaced by SM-11044 but not by BRL-37344, indicating the existence of the specific binding site observed in binding studies. In adipocytes, a similar band was observed at 34 kDa, but it could not be displaced at all, suggesting that it corresponded to nonspecific binding. Two-dimensional PAGE revealed labeling of a single acidic (pI ϭ 6.0) protein, rather than multiple proteins possessing the same affinity. These results support the Scatchard and Hill plots analyses, suggesting that the specific binding site constitutes a single class of sites. Trypsin digestion generated a labeled peptide of 7 kDa.
Analysis of CNBr fragments indicated that cleavage at the methionine residue in the presence of trifluoroacetic acid, which improves the cleavage at CNBr-resistant bonds such as Met-Thr or Met-Ser (27), generated a single 10-kDa fragment. In formic acid conditions, CNBr generated three labeled fragments of 8, 10, and 12 kDa, and formic acid alone generated a single 8-kDa labeled fragment. These data suggest that the 12-kDa fragment contains a CNBr-resistant methionine residue cleaved in CNBr/trifluoroacetic acid thus creating the 10-kDa fragment and that the 8-kDa fragment is a product by cleavage at an acid-sensitive bond such as Asp-Pro.
The partial amino acid sequences did not appear to display any homology with known proteins by a search in two data bases, the non-redundant GenBank and EMBL sequences. In a data base of G protein-coupled receptors, the most homologous protein was human platelet-activating factor receptor with 50% in 14 amino acid residues from the fifth transmembrane domain, whereas [ 125 I]ICYP is known to have affinity for ␤1-, ␤2-, ␤3-ARs, serotonin 5-HT 1A , and 5-HT 1B receptors (25,26). The uniqueness of the amino acid sequences presented here supports the hypothesis that SMBP is different from the known monoamine receptors.
The classification of the novel SMBP functional binding site appears to be difficult, because of the binding of several synthetic ␤-AR ligands and lack of binding of the typical ␤-AR agonist, isoproterenol, the natural AR agonists, norepinephrine, epinephrine, and 5-HT. Dissociation of the specific binding by GTP suggests that iodocyanopindolol itself is a partial agonist, as was also observed for the ␤3-AR and that this binding site is a member of the G protein-coupled receptor family. The molecular size of 34 kDa seems small in comparison with ␤-ARs or other reported G protein-coupled receptors (28) but may be explained by the absence of N-linked glycosylation. Such absence was also reported for human and rat ␣2B-AR (29 -31). The molecular mass of human angiotensin II type 2 receptor was reported as 33 kDa after deglycosylation of 66 -70-kDa protein (32).
SM-11044 and cyanopindolol are synthetic compounds. Therefore, the question arises concerning the identity of the endogenous ligand. To address this question will require further studies similar to those performed for opioid receptors. For instance, although the benzodiazepine receptor has been characterized by synthetic ligands, the endogenous ligand has not yet been determined.
Cyanopindolol behaves as a "nonconventional" partial agonist for ␤3-AR depending on the number of receptors expressed on cells or on the experimental condition used for estimation of the stimulatory response. McLaughlin and MacDonald (4) actually observed that cyanopindolol behaved as an antagonist for ␤3-AR without any agonist potency up to the concentration of 10 M in KCl-induced depolarized rat colon tonus under the same experimental conditions as those used in the present studies.
SM-11044, an agonist for this binding site, has been shown to stimulate guinea pig ileum relaxation of KCl-induced tonus more efficiently than rat white adipocyte lipolysis (11). SM-11044 and BRL-35135A display the additional property of inhibiting leukotriene B4 induced-guinea pig eosinophil chemotaxis, whereas isoproterenol and BRL-37344 had no such effect (12,13). This inhibition was unaffected by propranolol but was antagonized by alprenolol (12,13). The inhibitory effect of SM-11044 on eosinophil chemotaxis was significantly antagonized by cyanopindolol. 2 These observations suggest the existence in guinea pig ileum and eosinophils of the same functional binding site as the one presented here. While guinea pig eosinophil possess ␤2-AR coupled to adenylate cyclase, isoproterenol FIG. 13. A, preparative acid cleavage of SMBP. The isolated, labeled proteins of 34 kDa were incubated with 70% formic acid for 72 h at 37°C. An aliquot of the cleaved products was resolved on Tricine/SDS-PAGE followed by autoradiography. Arrow shows 8-kDa labeled fragment. B, reverse-phase HPLC purification of the photoaffinity labeled formic acid-cleaved 8-kDa fragment. The fragment isolated from Tricine/SDS-PAGE gels was further purified by reverse-phase HPLC, as described in the text. The fragment was eluted from the C4 column with a linear gradient of 30 -98% buffer B in 120 min (---). The radioactive profile for the 8-kDa labeled fragment was shown (q). Based on the amount of recovered radioactivity, HPLC column recovery was 91.6%. did not inhibit chemotaxis (12,13), indicating that the second messenger of the putative receptor is not cyclic AMP. Smooth muscle contraction induced by KCl and eosinophil chemotaxis involves a common signal transduction pathway leading to an increase in intracellular Ca 2ϩ (33). Binding of the agonist SM-11044 may cause the inhibitory effect, through the specific binding site, by blocking intracellular Ca 2ϩ accumulation. SM-11044 will be a good tool to characterize a further functional role of this binding site. In contrast, to the other predicted agonists such as the esterified compounds, BRL-35135A and ICI-198157, are converted in vivo to acid metabolites that appear to be inactive toward SMBP.
The elucidation of the structure of the ␤3-AR was initially considered as the answer to all the questions raised by the atypical ␤-AR responses described in a variety of tissues including fat, muscle, and colon. In fat, the ␤3-AR was indeed shown convincingly to be coupled to the lipolytic response (34). In colon, only a few attempts were made to demonstrate that the ␤3-AR was mediating relaxation. The present findings raise the question whether in colon, it is actually SMBP that displays binding properties different from cloned ␤-ARs but mediates relaxation.
In conclusion, the present study demonstrates the existence in rat colon smooth muscle of a novel SM-11044 or iodocyanopindolol binding protein, SMBP, different from known biogenic amine receptors and that may mediate relaxation of depolarized colon. Further studies will be needed to characterize the functional importance of this receptor including identification of its endogenous ligand and signal transduction system.