Photochemical identification of transmembrane segment IVS6 as the binding region of semotiadil, a new modulator for the L-type voltage-dependent Ca2+ channel.

To identify the binding domain of a new Ca2+ antagonist semotiadil on L-type Ca2+ channels from skeletal muscle, photolabeling was carried out by using an azidophenyl derivative of [3H]semotiadil. Photoincorporation was observed in several polypeptides of membrane triad preparations; the only specific photoincorporation was in the alpha1 subunit of the Ca2+ channel. After solubilization and purification, the photolabeled alpha1 subunit was subjected to proteolytic and CNBr cleavage followed by antibody mapping. Specific labeling was associated solely with the region of transmembrane segment S6 in repeat IV. Quantitative immunoprecipitation was found in the tryptic and the Lys-C/Glu-C fragments of 6.6 and 6.1 kDa, respectively. Further CNBr cleavage of the Lys-C digests produced two smaller fragments of 3.4 and 1.8 kDa that were included in the tryptic and Lys-C/Glu-C fragments. The smallest labeled fragments were: Tyr1350-Met1366 and Leu1367-Met1381 containing IVS6, a possible pore-forming region. The data suggest that semotiadil binds to a region that is overlapped with but not identical to those for phenylalkylamines, dihydropyridines and benzothiazepines. The present study also provides evidence that region IV represents an important component of a binding pocket for Ca2+ antagonists.

and benzothiazepines (BTZ), which are represented by the parent compounds, nifedipine, verapamil, and diltiazem, respectively. These drugs bind to different sites on the ␣ 1 subunit of Ca 2ϩ channels (1), and logically explain the well known allosteric interactions with one another (2). Using photoaffinity labeling and antibody mapping techniques, all three drugs have been shown to bind to different regions in more than one motif. Several other Ca 2ϩ antagonists have different chemical structures and somewhat different pharmacological actions than DHP, PAA, and BTZ (3)(4)(5). Semotiadil (SD-3211) is a novel Ca 2ϩ antagonist with a unique 1,4-benzothiazine ring structure (3) (Fig. 1).
The benzothiadine ring is homologous to the benzothiazepine ring of diltiazem whereas the ring components of the two drugs might contribute different properties in the action on Ca 2ϩ channels. Studies on structure-function relationships of diltiazem (reviewed in Ref. 6) suggest that the acetoxy and 2-(dimethyamino)ethyl groups play important roles in the calcium antagonistic activity. It is likely that the benzothiazepine ring of diltiazem is a structure on which various side groups can be inserted, which may change the position of the ring in binding and subsequent inhibition of the Ca 2ϩ channel. For example, the hydrophobic 4-methoxyphenyl group as well as the acetoxy and 2-(dimethyamino)ethyl groups, probably confer specific activities of diltiazem and other BTZs. In contrast, the calcium antagonist activity of semotiadil depends, in part, on the long side chain of Ar-O-CH 2 CH 2 CH 2 -N(Me)-CH 2 CH 2 -O-Ar at the C-3 position of the 1,4-benzothiazine. This idea is supported by the comparison of the three-dimensional structures between semotiadil and diltiazem based on their conformational analyses by x-ray crystallography and spectroscopy in solution (7,8). There is no apparent similarity in the orientation of the side chains as well as in the common methoxyphenyl group between two drugs, when the phenyl ring of the benzothiazepine and 1,4-benzothiazine are overlaid by the computer. The hypothesis that the long side chain at the C-3 position of the 1,4-benzothiazine ring is a part of the pharmacophore for calcium antagonist activity (13) is supported by the fact that a similar structural component: Ar-C(R 1 R 2 )-CH 2 CH 2 CH 2 -N(Me)-CH 2 CH 2 -Ar exists in verapamil and other PAAs. It is apparent that the 1,4-benzothiazine ring of semotiadil plays an additional role that contributes to the enhanced potency.
For example, in considering the pharmacological characteristics as a Ca 2ϩ antagonist, semotiadil is longer-lasting than diltiazem and nifedipine and shows a higher selectivity for blood vessels compared with cardiac tissues than diltiazem but lower selectivity than nifedipine (9,10 (11)(12)(13).
Localization of the semotiadil binding site would provide information about a putative new class of Ca 2ϩ antagonists but more importantly might uncover overlapping binding region(s), if any, with conventional Ca 2ϩ antagonists. The binding sites for DHP, PAA, and BTZ have been localized by photoaffinity labeling of Ca 2ϩ channels followed by defined proteolysis and antibody mapping using sequence-directed antibodies (14 -18). By comparing the results of the latter, with those derived from mutagenesis experiments (19 -25), one can demonstrate that sequence stretches photolabeled by DHP, PAA, and BTZ indeed contain amino acid residues that directly participate in binding. However, some recent mutagenesis experiments (26 -28) have revealed sites that are not labeled by photoligands. As an initial work to identify the binding site for semotiadil, we employed techniques of photoaffinity labeling of Ca 2ϩ channels isolated from rabbit skeletal muscles with [ 3 H]D51-4700, an azidophenyl derivative of [ 3 H]semotiadil (29), and the localizing of the site(s) of photolabeling and comparing with those for DHP, PAA, and BTZ.
Membrane Preparation-Triad membranes were isolated from rabbit skeletal muscle as described by Mitchell et al. (31).
Photoaffinity Labeling and Purification of Rabbit Skeletal Ca 2ϩ Channels-Rabbit triad membranes (300 pmol of [ 3 H](ϩ)-PN200 -110 binding sites, 20 mg of proteins) were incubated with 100 nM [ 3 H]D51-4700 in 10 ml of binding buffer (25 mM Tris-HCl (pH 7.2), 0.1 mM phenylmethylsulfonyl fluoride, 1 g/ml pepstatin A, 1 g/ml leupeptin, 10 g/ml soybean trypsin inhibitor) in the presence and absence of 10 M semotiadil at 30°C for 60 min. The incubation mixture was transferred into a glass Petri dish on ice, and irradiated for 20 min with a 100 watt black light/blue lamp (Ultra-Violet Products, Inc., San Gabriel, CA) at distance of 10 cm. After photolysis, the [ 3 H]D51-4700-labeled Ca 2ϩ channels were solubilized in 1% (w/v) digitonin and purified by affinity chromatography on WGA-Sepharose 4B according to the described method (30). The sample was dialyzed against 1 mM Tris-HCl (pH 7.3) and lyophilized.
Reductive Carboxymethylation and Gel Permeation High Pressure Liquid Chromatography-The photolabeled and lyophilized protein was resuspended in 0.1 M Tris-HCl (pH 8.0), 1% (v/v) 2-mercaptoethanol, 1.5% (w/v) SDS (final volume of 0.3 ml). After incubation at room temperature for 30 min, iodoacetic acid was added to a final concentration of 84 mM. After incubation for 1 h, the photolabeled ␣ 1 subunits were further purified by gel permeation liquid chromatography as described (14). Fractions corresponding to the ␣ 1 subunit were pooled, lyophilized, and stored at Ϫ30°C until use.
Proteolytic and CNBr Cleavage of [ 3 H]D51-4700-labeled ␣ 1 Subunits-The photolabeled ␣ 1 subunit was dissolved in deionized water (0.5 ml) and dialyzed against 6 M urea as described (14), followed by dialysis against 0.01% Triton X-100 for 6 h. The sample was digested with Lys-C (50 g/ml) in 50 mM Tris-HCl (pH 9.0) containing 0.05% (w/v) SDS and 0.01% (v/v) Triton X-100 (final volume of 100 l) at 37°C for 6 h. For trypsin digestion, the sample was incubated with TPCKtrypsin (100 g/ml) at 37°C for 12 h in 50 mM Tris-HCl (pH 8.0) containing 0.01% (v/v) Triton X-100 and 2 mM CaCl 2 . The reaction was stopped by heating at 90°C for 3 min. Prior to Lys-C/Glu-C digestion and CNBr cleavage, Lys-C digests were dialyzed against H 2 O for 6 h using a microdialyzer apparatus with a 1 kDa cut-off dialysis tube (Spectra/Por 6, Spectrum). For Lys-C/Glu-C digestion, the dialyzed sample was incubated with Glu-C (0.5 mg/ml) in 50 mM sodium phosphate buffer (pH 7.8) containing 0.05% (w/v) SDS for 12 h at 37°C. For CNBr cleavage, the dialyzed sample was lyophilized and then incubated with CNBr (5 mg/ml) in 70% (v/v) formic acid for 12 h at 37°C. After incubation, the mixture was lyophilized.
Immunoprecipitation-Antibodies were bound to protein A-Sepharose CL-4B gel by incubating 1 volume of antiserum with 1 volume of the swollen gel in the buffer A (10 mM Tris-HCl (pH 7.2), 150 mM NaCl, 0.1% (v/v) Triton X-100 and 1 mg/ml bovine serum albumin) for 2 h at 4°C. The gel was washed with the ice-cold buffer A before addition of digested or nondigested [ 3 H]D51-4700-labeled ␣ 1 subunits. After incubation for 2 h at room temperature, the gel was washed with buffer A. Immunoprecipitated radioactivity was directly determined by liquid scintillation counting of the protein A-Sepharose CL-4B gel containing 100 mM sodium citrate (pH 3.0). Immunoprecipitated labeled fragments were extracted from the gel with a sampling buffer for SDS-PAGE (50 mM Tris-HCl (pH 6.8), 4% (w/v) SDS, 2% (v/v) 2-mercaptoethanol and 12% (v/v) glycerol) for 3 min at 90°C and analyzed by SDS-PAGE. To determine the immunoprecipitated fragments size, the antibody-protein A Sepharose complex was cross-linked with dimethyl pimelidate as described by Schneider et al. (32).
Radioluminography and Gel Slicing-Instead of fluorography, a higher sensitive visualization method ("radioluminography") of the tritiated proteins and peptides was used. In brief, the gel after electrophoresis was electrophoretically transferred onto a polyvinylidene difluoride membrane in a transfer buffer (25 mM Tris, 193 mM glycine, 10% methanol) by using a semidry blotting assembly. The blotted membrane was stained with Coomassie Brilliant Blue R250, followed by drying completely in air. The membrane was then placed in contact with an imaging plate, BAS-TR2040S (Fuji Photo Film Co.) in a cassette at room temperature for 2 days. The imaging plate was scanned and analyzed by a Bio-Imaging Analyzer BAS 1000 model (Fuji Photo Film Co.). Scanning conditions were at a sensitivity 10,000, latitude 4, gradation 1024, and resolution 100. Printouts were performed by a high quality pictorial copy apparatus. Alternatively, individual gel lanes were manually cut into 3-mm slices and radioactivity was determined in ACSII with 3% (v/v) H 2 O 2 . [ 3 H]D51-4700 Labeling Occurs Only within Repeat IV-To determine the localization of photolabeled site within the ␣ 1 subunit, we first subjected the photolabeled ␣ 1 subunit to protease digestion with an endoprotease Lys-C and probed the Lys-C fragment by immunoprecipitation with a series of sequence-directed antibodies (see "Experimental Procedures") against different regions of ␣ 1 . The Lys-C digestion of ␣ 1 is shown in Fig. 3, resulting in a labeled fragment of 8.3 Ϯ 0.7 kDa (n ϭ 5) (Fig. 3A, lane 1). The fragment contained 92 Ϯ 4% (n ϭ 5) of the ␣ 1 -associated radioactivity as determined by gel slicing (not shown). Immunoprecipitation with sequence-directed antibodies revealed that only two antibodies directed against epitopes located near segment S6 in repeat IV (anti-(1338 -1351) and anti-(1382-1400), see Fig. 7) immunoprecipitated the photolabeled fragments, whereas anti-(1320 -1332) and anti-(1401-1414) did not immunoprecipitate at all (Fig.  3B). Other antibodies against repeat I, repeat III, and repeat IV efficiently immunoprecipitated the nondigested labeled ␣ 1 but did not immunoprecipitate Lys-C fragments (not shown). About 56 -68 and 57-75% of the ␣ 1 -associated labeling were associated with a fragment recognized by anti-(1338 -1351) and anti-(1382-1400), respectively (Fig. 3B). After the immunoprecipitated radioactivities were normalized with respect to the radioactivities immunoprecipitated in nondigested samples (100%), the values calculated were 125-147 and 124 -163%, respectively (Fig. 3B). The reason why the calculated values were over 100% will be discussed later (see "Discussion"). The radioactivity applied was recognized quantitatively by anti-(1338 -1351) and anti-(1382-1400) suggesting that both of the antibodies were immunoprecipitating the same 8.3-kDa band. This was confirmed by SDS-PAGE analysis of the antibody bound radioactivity (Fig. 3A, lane 2 and 3). Since the extracellular ␣ 1 (1338 -1351) or intracellular ␣ 1 (1382-1400) epitope is located within a single Lys-C fragment that contains IVS6 and intracellular residues, or IVS6 and extracellular residues, respectively, the 8.3-kDa fragment represents the correct digested product at Lys 1336 and Lys 1403 (calculated mass 7.9 kDa, see Fig. 7).

H]D51-4700 Labeling Is Located in Tryptic Fragments
Containing the S6 Segment in Repeat IV-Since the Lys-C fragment contains cleavable sites by trypsin, the photolabeled ␣ 1 subunits were digested with TPCK-trypsin to refine the photolabeled sites. SDS-PAGE revealed two smaller labeled fragments with apparent molecular masses of 8.3 Ϯ 0.8 (n ϭ 3) and 6.6 Ϯ 0.7 kDa (n ϭ 3). A radioluminogram of a gel where two peaks are clearly separated is shown in Fig. 4A. 84 Ϯ 8% of the ␣ 1 -associated radioactivity was recovered in these peaks and no other smaller fragments were observed as determined by gel slicing (not shown). Location of the photolabeled tryptic fragments was assessed by immunoprecipitation using anti-(1338 -1351) and anti-(1382-1400). About 53-61% of the ␣ 1 -associated labeling were associated with fragments recognized by anti-(1338 -1351) (Fig. 4B). The immunoprecipitated peptides were 8.3 and 6.6 kDa, determined by SDS-PAGE analysis (Fig. 4A, lane 2). The 6.6-kDa peptide was immunoprecipitated to a greater extent than the 8.3-kDa peptide, which is in accordance with the fact that the 6.6-kDa band was the major labeled peptide (Fig. 4A). Therefore, both peptides must contain the full epitope sequence of anti-(1338 -1351) and the 6.6-kDa peptide must be the smallest labeled peptide obtained by trypsin digestion.
In contrast, immunoprecipitation by anti-(1382-1400) decreased markedly to 12-18% (Fig. 4B) compared with the results obtained with the Lys-C fragment. The epitope of anti-(1382-1400) contains an arginine residue at 1389 that was cleavable by trypsin. Therefore, the major labeled fragment (6.6 kDa) must be generated by trypsin cleavage at Arg 1389 and is not recognized by anti-(1382-1400) (see Fig. 7). The antibody only recognizes the minor labeled fragment (8.3 kDa) that is a partially trypsin-digested product containing the epitope region (1382-1400).
Since no smaller fragments than 8.3 and 6.6 kDa were obtained, the 6.6 kDa must be the smallest labeled peptide obtained by trypsin digestion. The peptide contains the epitope of anti-(1338 -1351) but loses the epitope of anti-(1382-1400). The 6.6-kDa labeled peptide, therefore, is derived by trypsin cleavage at Lys 1336 and Arg 1389 (calculated molecular mass as 6.1 kDa) and contains IVS6 together with adjacent extracellular and cytoplasmic amino acid residues.
Isolation and Characterization of Smaller Photolabeled Fragments by Glu-C Digestion-Since the Lys-C fragment also contains potential cleavable sites by Glu-C, the photolabeled Lys-C fragment was subsequently digested with endoprotease Glu-C to further restrict the photolabeled sites. As shown in Fig. 5A, a radioluminogram of a gel revealed two smaller labeled fragments with apparent molecular masses of 7.8 Ϯ 0.9 (n ϭ 3) and 6.1 Ϯ 0.7 kDa (n ϭ 3). The ␣ 1 -associated radioactivity was recovered in 86 Ϯ 7% in these peaks and no other smaller fragments were observed as determined by gel slicing (not shown).
In immunoprecipitation experiments, anti-(1382-1400) retained its binding activity (62%) to the Glu-C digested fragments whereas anti-(1338 -1351) showed only 20% immunoprecipitation of the digests (Fig. 5B). Since the epitope of anti-(1338 -1351) contains three glutamic acid residues (at positions 1341, 1348 and 1349) that are susceptible to Glu-C cleavage, the low efficiency in immunoprecipitation by anti-(1338 -1351) must result from the cleavage. The photolabeled and immunoprecipitated peptides by anti-(1382-1400) were analyzed by SDS-PAGE (Fig. 5A, lane 2). The 6.1-kDa photolabeled peptide was observed in addition to a small portion of 7.8-kDa fragments, indicating that the 6.1-kDa peptide was the smallest labeled fragment after successive digestion with Lys-C and Glu-C. According to our estimation of molecular mass, the cleavage site by Glu-C most likely corresponds to Glu-1349 (calculated molecular mass 6.2 kDa). Fig. 7 shows the position of the smallest photolabeled fragment by Lys-C/Glu-C digestions within the linear alignment near IVS6 segment.
Isolation and Characterization of Smaller Photolabeled Fragments by CNBr Cleavage-Since the Lys-C fragment contains two methionine residues, the photolabeled Lys-C fragment was subsequently treated with CNBr to further restrict the photolabeled sites. As shown in Fig. 6A, a radioluminogram of a gel revealed three smaller labeled fragments with apparent molecular masses of 5.7 Ϯ 0.6 (n ϭ 3), 3.4 Ϯ 0.4 (n ϭ 3), and 1.8 Ϯ 0.3 kDa (n ϭ 3). During the incubation with CNBr in 70% formic acid, almost 70% of the photolabeled radioactivity was liberated and migrated to the dye front position on SDS-PAGE (Fig. 6B). However, the liberated radioactivity was not blotted on the polyvinylidene difluoride membrane sheet, and therefore it did not interfere with the analysis of newly generated labeled fragments in the radioluminogram (Fig. 6A). In the immunoprecipitation experiments, anti-(1338 -1351) showed apparent binding activity (10 Ϯ 3%, n ϭ 3) to the total radioactivity applied after CNBr cleavage, whereas anti-(1382-1400) did not immunoprecipitate at all. As the radioactivity associated with peptide fragments was only 30% of the radioactivity in the applied sample, the immunoprecipitated value of 10% can be corrected to 33%. This value is further corrected to 73% after normalization with respect to immunoprecipitation avidity of anti-(1338 -1351) in uncleaved ␣ 1 subunits (45%).
The photolabeled and immunoprecipitated peptides by anti-(1338 -1351) were analyzed by SDS-PAGE (Fig. 6A, lane 2). The 3.4 and 5.7-kDa photolabeled peptides were observed, but the 1.8-kDa fragment was not immunoprecipitated. The results indicate that the 3.4-and 5.7-kDa fragments contain the epitope of anti-(1338 -1351). Therefore, we assign the labeled 3.4-kDa fragment to Leu 1337 -Met 1366 (calculated molecular mass 3.6 kDa) and the 5.7-kDa fragment to Leu 1337 -Met 1381 (calculated molecular mass 5.3 kDa) that is a partially cleaved product at Met 1381 but not cleaved at Met 1366 . On the other hand, the nonimmunoprecipitated labeled fragment of 1.8 kDa must be Leu 1367 -Met 1381 (calculated molecular mass 1.7 kDa) that contains no sequence for the epitope of anti-(1338 -1351). The smallest labeled fragments are 3.6 kDa (Leu 1337 -Met 1366 ) and 1.8 kDa (Leu 1367 -Met 1381 ). Fig. 7 shows the position of these photolabeled fragments by CNBr cleavage within the linear alignment near segment IVS6.

DISCUSSION
Semotiadil Receptor Site of the ␣ 1 Subunit-[ 3 H]D51-4700, a photoaffinity probe of semotiadil, selectively labeled the ␣ 1 subunit of Ca 2ϩ channels in skeletal triad membranes. In the absence of unlabeled semotiadil, the probe labeled several polypeptides including the ␣ 1 subunit. This may explain the observation that reversible binding of [ 3 H]D51-4700 to triad membrane preparations is rather difficult to show due to the high level of nonspecific binding (not shown). However, the photoincorporation to the ␣ 1 subunit occurred in a specific manner since the Ca 2ϩ channels purified by WGA column showed a single photolabeled band of 170 kDa, and the label was totally blocked by excess unlabeled semotiadil. The specifically photolabeled site was localized within the ␣ 1 subunit by an antibody mapping method employed previously for the DHP-, PAA-, and BTZ-binding domains (14 -18). As shown in the results of Lys-C digestion, we observed that the normalized values of the immunoprecipitated percentage of the proteasedigested fragment gave more than 100% with respect to those of the nondigested samples. Similar results were reported in the literature (16) where the labeled site was localized to a single peptide fragment. This is probably due to the fact that higher reactivity of the anti-peptide antibody occurs to the -4700-labeled ␣ 1 was digested with Lys-C followed by CNBr cleavage and analyzed on the Schä gger and von Jagow gel. Labeled peptide fragments were detected by liquid scintillation counting of 3-mm gel slices. About 70% of the radioactivity was migrated at the dye front due to its liberation from the labeled sites during CNBr treatments. Three runs were carried out. peptide fragment rather than to the nondigested polypeptide ␣ 1 .
Only a single labeled fragment of 8  (11,12). This suggests that the binding site for semotiadil is similar but not identical to that for PAA. The present results are consistent with this interpretation.
The two labeled fragments by [ 3 H]D51-4700 are not only overlapped with the [ 3 H]LU49888 labeled site but also are part of the labeled regions by DHP (14,15) and BTZ (17,18). The association of the newly identified semotiadil site with those of the three typical Ca 2ϩ antagonists (DHP, BTZ, and PAA) within the pore-forming regions of the channel allows allosteric interactions among semotiadil and these drugs. Although a few reports are available concerning the pharmacological interac-tion of semotiadil and other Ca 2ϩ antagonists (11)(12)(13), the observed negative allosteric effect of semotiadil on the binding of DHP, PAA, and BTZ to canine skeletal muscle membranes (12) suggests that the binding sites for all these drugs are in close apposition in the Ca 2ϩ channel but not identical. This is clearly consistent with the present photoaffinity labeling results.
In contrast to the photolabeled sites for DHP (14,15) and BTZ (18), the identified fragments for photolabeling with [ 3 H]D51-4700 do not contain any peptides in repeat III. It is tempting to conclude that the semotiadil binding site is different from those for DHP and BTZ. However, there are complexities between the results obtained by photoaffinity labeling and those obtained by molecular biological techniques. In BTZ, for example, IIIS6 as well as IVS6 were identified as the photolabeled fragments (18), whereas only the IVS6 was shown to be sufficient for BTZ sensitivity for L-type Ca 2ϩ channels (23). With regard to PAA, only the IVS6 with the adjacent extracellular and intracellular stretches were identified by the photolabeling technique, whereas not only IVS6 (24,25) but also IIIS6 appear to be determinants of high affinity binding for (Ϫ)D888, a PAA drug, using molecular biological techniques (28). The DHP situation is more clearly understood and shows reasonable correlation between photolabeling and mutation methods in which IVS6 and IIIS6 have been identified as molecular determinants of binding (14, 15, 19 -22). Interestingly, IIIS5 may also be an important region for DHP binding as shown by site-directed mutagenesis (26,27). These controversies may result partly from flexible photoreactive side chains that are not able to photoincorporate into all contact regions of the drug molecules. Therefore, it is necessary to employ mutagenesis to survey the regions that are involved in semotiadil binding.
Taken together, the present results indicate that IVS6 is an important region for semotiadil binding. This agrees closely with the observations that all of the binding domains so far identified as Ca 2ϩ -sensitive antagonists contain IVS6. One can, with caution, suggest that repeat IV is perhaps a common region for pharmacological consequences of Ca 2ϩ channel modulator drugs. This region may be considered as an intrinsic portion of a binding region or "pocket" that contributes to drug interactions. In the second line, particular amino acid residues that are potential cleavage sites by the following protease digestions and CNBr treatment are shown as single letters. In the third to sixth lines, the smallest labeled fragment(s) observed by each protease digestion (Lys-C or trypsin) or its combination with subsequent proteolysis (Lys-C/Glu-C) or CNBr treatment (Lys-C/CNBr) are indicated with its size (kDa) and the N and C terminus amino acid residues. Consequently, the smallest labeled fragments can be deduced as Tyr 1350 -Met 1366 (Y1350-M1366) and Leu 1367 -Met 1381 (L1367-M1381).