Analysis of the Dihydropyridine Receptor Site ofl-type Calcium Channels by Alanine-scanning Mutagenesis*

The dihydropyridine Ca2+antagonist drugs used in the therapy of cardiovacular disorders inhibitl-type Ca2+ channels by binding to a single high affinity site. Photoaffinity labeling and analysis of mutant Ca2+ channels implicate the IIIS6 and IVS6 segments in high affinity binding. The amino acid residues that are required for high affinity binding of dihydropyridine Ca2+ channel antagonists were probed by alanine-scanning mutagenesis of the α1C subunit, transient expression in mammalian cells, and analysis by measurements of ligand binding and block of Ba2+ currents through expressed Ca2+ channels. Eleven amino acid residues in transmembrane segments IIIS6 and IVS6 were identified whose mutation reduced the affinity for the Ca2+ antagonist PN200-110 by 2–25-fold. Both amino acid residues conserved among Ca2+ channels and those specific to l-type Ca2+ channels were found to be required for high affinity dihydropyridine binding. In addition, mutation F1462A increased the affinity for the dihydropyridine Ca2+ antagonist PN200-110 by 416-fold with no effect on the affinity for the Ca2+ agonist Bay K8644. The residues in transmembrane segments IIIS6 and IVS6 that are required for high affinity binding are primarily aligned on single faces of these two α helices, supporting a “domain interface model” of dihydropyridine binding and action in which the IIIS6 and IVS6 interact to form a high affinity dihydropyridine receptor site on l-type Ca2+ channels.


From the Department of Pharmacology, University of Washington, Seattle, Washington 98195-7280
The dihydropyridine Ca 2؉ antagonist drugs used in the therapy of cardiovacular disorders inhibit L-type Ca 2؉ channels by binding to a single high affinity site. Photoaffinity labeling and analysis of mutant Ca 2؉ channels implicate the IIIS6 and IVS6 segments in high affinity binding. The amino acid residues that are required for high affinity binding of dihydropyridine Ca 2؉ channel antagonists were probed by alanine-scanning mutagenesis of the ␣ 1C subunit, transient expression in mammalian cells, and analysis by measurements of ligand binding and block of Ba 2؉ currents through expressed Ca 2؉ channels. Eleven amino acid residues in transmembrane segments IIIS6 and IVS6 were identified whose mutation reduced the affinity for the Ca 2؉ antagonist PN200-110 by 2-25-fold. Both amino acid residues conserved among Ca 2؉ channels and those specific to L-type Ca 2؉ channels were found to be required for high affinity dihydropyridine binding. In addition, mutation F1462A increased the affinity for the dihydropyridine Ca 2؉ antagonist PN200-110 by 416-fold with no effect on the affinity for the Ca 2؉ agonist Bay K8644. The residues in transmembrane segments IIIS6 and IVS6 that are required for high affinity binding are primarily aligned on single faces of these two ␣ helices, supporting a "domain interface model" of dihydropyridine binding and action in which the IIIS6 and IVS6 interact to form a high affinity dihydropyridine receptor site on L-type Ca 2؉ channels.
Voltage-gated Ca 2ϩ channels mediate Ca 2ϩ influx in response to membrane depolarization and thereby initiate cellular activities such as secretion, contraction, and gene expression. Several types of voltage-gated Ca 2ϩ channels have been distinguished by their physiological and pharmacological properties and have been designated L, N, P/Q, R, and T (for review see Refs. [1][2][3]. L-type Ca 2ϩ channels are the molecular targets for the dihydropyridine, phenylalkylamine, and benz(othi)azepine classes of calcium channel blockers that are widely used in the therapy of cardiovascular diseases. These drugs are thought to bind to three separate but allosterically coupled receptor sites on L-type Ca 2ϩ channels (4,5).
The dihydropyridines (DHPs) 1 are allosteric modulators that act on L-type Ca 2ϩ channels as either agonists or antagonists (reviewed in Refs. 17 and 18). Charged DHPs are thought to traverse an extracellular pathway to gain access to the DHP receptor site located within the lipid bilayer 11-14 Å from the extracellular surface of the cell membrane (19 -22). Photoreactive DHPs specifically label the ␣ 1 subunit of the Ca 2ϩ channel (23)(24)(25)(26)(27)(28)(29). The predominant site of labeling corresponds to transmembrane segment IIIS6 and a portion of the extracellular segment connecting IIIS5 and IIIS6, and approximately 15-30% of the total photolabeling is incorporated into a peptide fragment composed primarily of transmembrane segment IVS6 (30 -33). In addition to these membrane-associated regions, a site in the intracellular carboxyl-terminal domain has been photoaffinity-labeled by photoreactive DHPs (34). Analysis of chimeric Ca 2ϩ channels implicated the IIIS5 (35), IIIS6 (35), and IVS6 (35)(36)(37) transmembrane segments in DHP binding, in general agreement with previous photoaffinity labeling results (30,31).
Site-directed mutagenesis of single amino acid residues that differ between L-type and non-L-type Ca 2ϩ channels revealed multiple amino acid residues on one side of the IIIS6 and IVS6 putative transmembrane ␣ helices that are important determinants of high affinity binding of DHP agonists and antagonists to L-type Ca 2ϩ channels (37,38). In addition, mutation of a Tyr residue (Tyr 1048 in ␣ 1S ) in segment IIIS6 that is conserved in all Ca 2ϩ subtypes was found to have the largest effect on DHP affinity (38). Mutation of this residue to Phe caused a 12-fold reduction in binding affinity, and mutation to Ala prevented measurable high affinity DHP binding. In the experiments reported here, we have used alanine-scanning mutagenesis to map both conserved and L-type-specific amino acid residues that contribute to the high affinity of the DHP receptor site.
Construction of Mutant Ca 2ϩ Channels-For the construction of mutant ␣ 1CII subunits (10), a 1.5-kilobase EspI fragment of the ␣ 1CII subunit of the rat brain Ca 2ϩ channel was subcloned into the bacteriophage M13mp19 for recovery of single-stranded DNA template. The resulting construct was used as a template for the generation of single-stranded DNA for introduction of mutations into transmembrane segment IVS6 of ␣ 1CII . These mutations were transferred to the full-length ␣ 1CII subunit cDNA by subcloning the 272-base pair DraIII/DraIII fragment. Mutations in transmembrane segment IIIS6 were made with a singlestranded DNA template containing the 1.7-kilobase AvrII/EcoRI fragment of the ␣ 1CII cDNA. The IIIS6 mutations were transferred to the full-length ␣ 1CII construct using the 1.5-kilobase SpcI/DraIII and the 272-base pair DraIII/DraIII fragments in a three-way ligation. Sitedirected mutagenesis was performed using established procedures (39), mutations were verified by sequence analysis of approximately 200 base pairs surrounding the site of mutagenesis, and the integrity of the full-length mutant Ca 2ϩ channel cDNA was confirmed by extensive restriction digest analysis.
Expression of Ca 2ϩ Channels-Human tsA-201 cells, a simian virus 40 (SV40) T-antigen expressing derivative of the human embryonic kidney cell line HEK293, were maintained in Dulbecco's modified Eagle's medium/Ham's F-12 medium (Life Technologies, Inc.) enriched with 10% fetal bovine serum. Human tsA-201 cells were cotransfected with wild type or mutant ␣ 1CII and cDNA encoding the ␤ 1b (40) and ␣ 2 ␦ Ca 2ϩ channel subunits such that the molar ratio of the plasmids was 1:1:1. Cells were transfected by calcium phosphate precipitation according to the procedures outlined by Margolskee et al. (41), and cells were harvested 30 -40 h following transfection.
Preparation of Membranes-Cells were washed two times, scraped, and homogenized using a glass-teflon homogenizer in Buffer A (50 mM Tris, 100 M phenylmethanesulfonyl fluoride, 100 M benzamidine, 1.0 M pepstatin A, 1.0 g/ml leupeptin, and 2.0 g/ml aprotinin, pH 8.0). The homogenate was centrifuged at 700 ϫ g for 5 min. The resulting pellet was discarded, and the supernatant was centrifuged 30 min at 100,000 ϫ g. The supernatant was discarded, and the membrane pellet was washed and homogenized in Buffer A. The resulting membrane homogenate was divided into aliquots and stored at Ϫ80°C for up to 3 months with no detectable loss of (ϩ)-[ 3 H]PN200-110 binding activity.
Electrophysiology-Cultures of tsA-201 cells at 75% confluence in 35-mm cell culture dishes were transfected with a total of 4 g of DNA consisting of an equimolar ratio of the three channel subunit cDNAs and 0.8 g of CD8 cDNA. After the addition of CaPO 4 -DNA, cells were incubated overnight at 37°C in 5% CO 2 . Twenty h after transfection, the cells were removed from culture dishes using 2 mM EDTA in phosphate-buffered saline and replated at low density for electrophysiological analysis. Transfectants were recognized by labeling with anti-CD8 antibody-coated beads (M450 CD8 Dynabeads, Dynal, Inc.). Barium currents through L-type Ca 2ϩ channels were recorded using the whole cell configuration of the patch clamp technique. Patch electrodes were pulled from VWR micropipettes and fire-polished to produce an inner tip diameter of 4 -6 m. Currents were recorded using a List EPC-7 patch clamp amplifier and filtered at 2 kHz (8-pole Bessel filter, Ϫ3 db). Voltage pulses were applied, and data were acquired using Fastlab software (Indec Systems). Linear leak and capacitance currents have been subtracted using an on-line P/4 subtraction paradigm. (ϩ)-PN200-110 was applied to cells by the addition of 0.2 ml of a 6ϫ stock to a 1-ml bath. Inhibition of Ca 2ϩ channel current (carried by Ba 2ϩ ) during application of (ϩ)-PN200-110 was monitored by 100-ms depolarizations to ϩ10 mV from a holding potential of Ϫ60 mV applied every 10 s until block had reached steady-state (1-3 min). The bath saline contained 150 mM Tris, 2 mM MgCl 2 , and 10 mM BaCl 2 . The intracellular saline contained 130 mM N-methyl-D-glucamine, 10 mM EGTA, 60 mM HEPES, 2 mM MgATP, and 1 mM MgCl 2 . The pH of both solutions was adjusted to 7.3 with methanesulfonic acid. All experiments were performed at room temperature (20 -23°C).

RESULTS
Transmembrane segments IIIS6 and IVS6 contain important molecular determinants for high affinity binding of DHPs as determined by photoaffinity labeling (Fig. 1A; Refs. 30 and 31) and by analysis of chimeric and mutant Ca 2ϩ channels ( Fig.  1B; Refs. [35][36][37][38]. Previous molecular biological studies that support this conclusion have focused almost exclusively on amino acid residues that are present in DHP-sensitive L-type Ca 2ϩ channels but not in DHP-insensitive non-L-type channels. In the present experiments, alanine-scanning mutagenesis of single amino acid residues in transmembrane segments IIIS6 and IVS6 of ␣ 1CII was employed to identify all of the amino acids, both conserved and L-type-specific, that are important determinants for DHP binding. Most of the amino acids in IIIS6 and IVS6 of ␣ 1CII were systematically changed to Ala and coexpressed with ␣ 2 ␦ and ␤ 1b in tsA-201 cells (Fig. 1, C and D, shaded boxes). To assess the relative importance of Ala 1157 and Ala 1467 , these amino acids were changed to Pro and Ser, respectively, as found in DHP-insensitive Ca 2ϩ channels. Ala was chosen for these experiments because its substitution in ␣ helices in the core of proteins has minimal effects on secondary structure (42). Substitution of Ala is therefore expected to reduce the hydrophobicity and size of the amino acid residue in each position in these putative ␣ helices without causing global conformational changes.
Wild type and mutant Ca 2ϩ channels were expressed transiently in tsA-201 cells, membrane preparations were isolated by differential centrifugation, DHP binding was measured in filter binding assays using the radiolabeled DHP antagonist (ϩ)-[ 3 H]PN200-110, and K d values for high affinity DHP binding were determined by Scatchard analysis. A high ratio of specific to nonspecific binding was observed in binding experiments with both wild type Ca 2ϩ channels (not shown) and mutant channels having substantially reduced affinity (for example, mutant N1472A; Fig. 2A). The linearity of the Scatchard plots (Fig. 2B) indicates a single class of high affinity binding sites with a K d of 55 pM for wild type and 470 pM for the mutant N1472A.
Mutation of several amino acid residues in transmembrane segments IIIS6 and IVS6 to Ala had large effects on DHP binding affinity (Fig. 3). Of the mutations in IIIS6, the largest effects on PN200-110 binding were observed when the L-typespecific residues Ile 1153 , Ile 1156 , Met 1160 , and Met 1161 were changed to Ala, resulting in K d values that are larger than wild type by 6.2-, 17-, 3.5-, and 9.6-fold, respectively (Fig. 3A). In addition, mutation of three residues in IIIS6 that are conserved between L-type and non-L-type Ca 2ϩ channels to Ala had large effects on DHP binding. The mutants F1158A and F1159A have K d values that are 4.5-and 4.3-fold larger than wild type, respectively. The mutant Y1152A, which corresponds to the conserved residue Tyr 1048 in ␣ 1S , which we previously showed was essential for DHP binding (38), exhibits no detectable DHP binding at concentrations of (ϩ)-[ 3 H]PN200-110 up to 25 nM.
Of the amino acid residues screened by alanine-scanning mutagenesis in transmembrane segment IVS6, mutations Y1463A, M1464A, and I1471A, which affect L-type-specific res-idues, and N1472A, which affects a conserved residue, exhibited the largest effects, with K d values that were larger than wild type by 2.9-, 1.6-, 2.7-, and 9-fold, respectively (Fig. 3B). The comparatively small effects of mutation of residues in IVS6 compared with IIIS6 suggest that this transmembrane segment is on the periphery of the DHP receptor site and has relatively weak interactions with bound DHPs compared with IIIS6.
We were unable to detect specific DHP binding to the ␣ 1C mutants Y1152A, Y1458A, and F1462A using concentrations of (ϩ)-[ 3 H]PN200-110 up to 25 nM. To determine if these mutants can form functional Ca 2ϩ channels and to measure their affinity for DHPs, these mutants were expressed and analyzed electrophysiologically (Fig. 4). The mutant F1458A did not express Ba 2ϩ currents, so it is likely that the lack of DHP binding observed with this mutant is caused by failure to express functional Ca 2ϩ channels. The mutants Y1152A and F1462A both expressed Ba 2ϩ currents but less effectively than wild type (Fig. 4, B and C). The voltage dependence of activation of these mutants was comparable with that of wild type as assessed from current-voltage relationships. In contrast, the voltage dependence of inactivation was shifted for these mutant Ca 2ϩ channels (Fig. 4D). For mutant Y1152A, the voltage dependence of inactivation was shifted to more negative membrane potentials and became significantly less steep than wild type. For mutant F1462A, the voltage dependence of inactivation was shifted 10 mV to more positive membrane potentials. As DHP antagonists bind with higher affinity to inactivated Ca 2ϩ channels, changes in the extent of inactivation at the membrane potential of our experiments could influence DHP binding and block. However, as shown below, the changes in DHP affinity caused by these two mutations are in the opposite directions from those expected on the basis of the changes in the voltage dependence of inactivation so the changes in affinity are likely to result from molecular changes in the DHP binding site rather than from indirect allosteric effects due to the change in inactivation gating.
The DHP antagonist PN200-110 inhibits wild type and mutant Ca 2ϩ channels in a concentration-dependent manner (Fig.  4, A-C). Analysis of concentration-effect relationships indicates that the IC 50 for inhibition of Y1152A by PN200-110 at Ϫ60 mV is approximately 25-fold higher than wild type (Fig. 4E). Thus, Tyr 1152 is indeed an important determinant for DHP binding. This large decrease in binding affinity coupled with low levels of expression are likely to prevent detection of specific (ϩ)-[ 3 H]PN200-110 binding to this mutant.
In contrast, the IC 50 for block of mutant F1462A by PN200-110 was decreased from 6.8 Ϯ 1.2 nM to 16 Ϯ 5.8 pM, indicating that this mutation results in an ␣ 1 subunit with an affinity for PN200-110 at Ϫ60 mV that is at least 400-fold higher than wild type (Fig. 4, B and E). The Ba 2ϩ currents for this mutant are very small and fewer cells express detectable Ba 2ϩ current. Therefore, the failure to detect (ϩ)-[ 3 H]PN200-110 binding to F1462A is likely due to a combination of low expression levels and expression in fewer cells rather than a reduction in DHP affinity. Surprisingly, the effect of the DHP agonist Bay K8644 to enhance Ca 2ϩ currents was not altered in this mutant (EC 50 ϭ 151 Ϯ 84 pM for wild type; EC 50 ϭ 111 Ϯ 45 pM for F1462A, n ϭ 3). Thus, mutation at this position has a large and specific effect on high affinity binding of a DHP antagonist without FIG. 1. Location of the DHP receptor site. A, a transmembrane folding model for domains III and IV of the L-type Ca 2ϩ channel ␣ 1 subunit is used to illustrate the peptide segments critical for DHP binding. Bold and shaded regions refer to segments of the receptor site found to be important for the actions of DHPs by photoaffinity labeling (30 -32). Numbers refer to the amino acid residues in ␣ 1C analogous to the NH 2 and COOH termini of the photoaffinity-labeled peptides. B, a transmembrane folding model for domains III and IV of the L-type Ca 2ϩ channel ␣ 1 subunit is used to illustrate the peptide segments critical for DHP binding. Bold and shaded regions refer to segments of the receptor site that were identified using chimeric Ca 2ϩ channels (35)(36)(37). Numbers refer to amino acid positions in the sequence of ␣ 1C that are analogous to those identified by site-directed mutagenesis and radioligand binding experiments as being critical for the binding of DHP agonists and antagonists (38). C and D, sequence alignment of transmembrane segments IIIS6 (C) and IVS6 (D) from DHP-sensitive (␣ 1S , ␣ 1C , and ␣ 1D ) and DHP-insensitive (␣ 1A , ␣ 1B , and ␣ 1E ) Ca 2ϩ channel ␣ 1 subunits. Asterisks refer to L-type-specific amino acids in IIIS6 and IVS6 of ␣ 1S that were analyzed by mutation, expression, and radioligand binding previously (38). Also indicated by an asterisk is the conserved Tyr 1048 , which was the most essential residue identified by Peterson et al. (38). Shaded boxes refer to amino acids that were analyzed by alanine scanning mutagenesis and radioligand binding in this study. Circles refer to amino acids found to be critical for the binding and/or actions of DHP agonists and antagonists. F1462A is identified by a white circle. effect on agonist binding and action. Analysis of additional mutations at this position will be of interest to further define the structural basis for the specific increase in affinity for DHP antagonists.

Mapping of the DHP Receptor Site by Alanine-scanning Mu-
tagenesis-A variety of techniques have been used to localize the DHP receptor site. Results from experiments using charged DHPs, photoreactive DHPs, chimeric Ca 2ϩ channels, and mutations of L-type-specific amino acid residues all indicate that the core of the DHP receptor site is formed primarily by amino acid residues in IIIS6 and IVS6 (Fig. 1, A and B). In this study, most of the amino acids in IIIS6 and IVS6 of ␣ 1C were systematically changed to Ala, and intrinsic Ala residues were changed to the amino acids present in corresponding positions in non-L-type Ca 2ϩ channels (see Fig. 1, C and D, shaded  boxes). The amino acid residues found by alanine-scanning mutagenesis to be important for DHP binding are indicated with circles in the second lines of panels C and D of Fig. 1, and they are compared with those identified by mutations of L-typespecific residues in our previous studies in the top lines of panels C and D in Fig. 1 (38). Our results using alaninescanning mutagenesis have identified four new conserved amino acid residues and two new L-type-specific amino acid residues within transmembrane segments IIIS6 and IVS6 whose mutation has significant effects on DHP binding affinity. The positions of these critical amino acid residues within the IIIS6 and IVS6 ␣ helices are illustrated in Fig. 5 where black residues indicate a reduction of DHP binding affinity by 5-fold or more and shaded residues indicate a significant reduction in binding affinity that is less than 5-fold by mutation to Ala.
In segment IIIS6, the largest effects were observed when amino acid residues Tyr 1152 , Ile 1153 , Ile 1156 , and Met 1161 were changed to Ala, resulting in K d values that are larger than wild type by 25-, 6.2-, 17-, and 9.6-fold, respectively (Fig. 5, black  residues). The mutants Phe 1158 , Phe 1159 , and Met 1160 had K d values for (ϩ)-[ 3 H]PN200-110 binding that were 3.5-4.5-fold higher than wild type (Fig. 5, shaded residues). The COOHterminal portion of IIIS6 including residues Phe 1158 , Phe 1159 , Met 1160 , and Met 1161 was not analyzed in our previous studies; therefore, all of these amino acids are newly identified residues that are important determinants of high affinity DHP binding in transmembrane segment IIIS6.
In segment IVS6, mutations of residues Tyr 1463 , Met 1464 , Ile 1471 , and Asn 1472 had the largest effects, with increases in K d values for (ϩ)-[ 3 H]PN200-110 of 2.9-, 1.6-, 2.7-, and 9-fold, respectively. In addition, mutation F1462A caused a large increase in affinity for a DHP antagonist with no change in affinity for the DHP agonist Bay K8644. In our previous study, all mutations of L-type-specific amino acid residues had similar effects on the binding affinity for DHP agonists and antagonists (38). Thus, Phe 1462 may have unique interactions with DHP antagonists that both determine their affinity and differentiate between agonists and antagonist drugs.  3. Amino acid residues in IIIS6 and IVS6 required for high affinity DHP binding. A, the indicated individual amino acids in segment IIIS6 were systematically changed to alanine by site-directed mutagenesis. DHP affinity was determined by radioligand binding using (ϩ)-[ 3 H]PN200-110 as described under "Experimental Procedures" and in Fig. 2. No detectable DHP binding was detected for the mutant Y1152A. B, the indicated individual amino acids in segment IVS6 were systematically changed to alanine by site-directed mutagenesis. DHP affinity was determined by radioligand binding using (ϩ)-[ 3 H]PN200-110 as described under "Experimental Procedures" and in Fig. 2. No specific binding was detected for mutants Y1458A and F1462A.
Asn 1472 in transmembrane segment IVS6 is conserved between Ca 2ϩ and Na ϩ channels. When this residue is changed to Ala in Na ϩ channels (N1769A), the resulting mutant has a 15-fold increased affinity for local anesthetic binding to the resting state of the channel (43). It was hypothesized that this mutation creates a local conformation of the drug receptor site in the resting state of the channel which resembles the high affinity local conformation of that receptor site in the inactivated state (43). Because the affinity for drug binding is highest in the inactivated states for both Na ϩ and Ca 2ϩ channels, the increase in affinity for block of resting calcium channels by DHPs observed with the F1462A mutation in this work may have a similar molecular basis as the state-dependent increase in affinity for local anesthetics caused by mutation Asn 1769 in segment IVS6 of Na ϩ channels.
The DHP Receptor Site Involves Both Conserved and L-typespecific Amino Acid Residues-The largest effects on DHP binding affinity were caused by alteration of conserved amino acid residues. The largest decrease in affinity (25-fold) was seen for mutation of Tyr 1152 in IIIS6. Mutation of the conserved residue Asn 1472 in IVS6 also caused a major (9-fold) decrease in DHP binding affinity. Alteration of a third conserved residue, Phe 1462 , resulted in a 416-fold decrease in IC 50 for block by PN200-110, consistent with greatly increased DHP affinity of resting Ca 2ϩ channels. These results suggest that DHP binding involves conserved amino acid residues that are important for common aspects of the structure and function of all voltagegated Ca 2ϩ channels. These conserved residues are likely to be largely responsible for the low affinity binding of DHPs found in non-L-type Ca 2ϩ channels (44).
We have also identified several L-type-specific amino acid residues in IIIS6 and IVS6 that are critical molecular determinants for DHP binding. Alteration of four L-type-specific residues in IIIS6 (Ile 1153 , Ile 1156 , Met 1160 , and Met 1161 ) and three L-type-specific residues in IVS6 (Tyr 1463 , Met 1464 , and Ile 1471 ) resulted in channels with significantly increased K d values for (ϩ)-[ 3 H]PN200-110 binding. It is likely that these residues are largely responsible for the high affinity of L-type Ca 2ϩ channels for DHPs and for the requirement of the IIIS6 and IVS6 transmembrane segments for the transfer of high affinity DHP binding to non-L-type Ca 2ϩ channels as demonstrated in experiments with channel chimeras (35)(36)(37)(38). Recent studies also implicate two additional L-type-specific amino acid residues in transmembrane segment IIIS5 in high affinity DHP binding and in the transfer of high affinity DHP binding to non-L-type Ca 2ϩ channels (45).
The Domain Interface Model of Ca 2ϩ Agonist and Antagonist Binding-All three classes of Ca 2ϩ antagonist drugs and the DHP Ca 2ϩ agonist drugs are allosteric ligands that exhibit state-dependent binding and are influenced in a reciprocal manner by binding of the other classes of drugs to distinct but interacting sites. Their effects on Ca 2ϩ channel function and their interactions with each other are similar to the actions of allosteric effectors on enzymes. In most well studied cases, binding of allosteric effectors to enzymes is thought to take place at subunit interfaces (46 -48). Similarly, strong evidence now indicates that the agonists of pentameric nicotinic acetylcholine receptors bind at sites that are formed at the interfaces of the ␣ subunits with the ␤ and ␦ subunits (49). Thus, there is increasingly strong precedent for binding of allosteric ligands at structural interfaces, possibly because the structure of multimeric proteins is most flexible in those regions. Our results indicate that the core of the receptor site for DHP antagonists is shared between transmembrane segments IIIS6 and IVS6, in agreement with the predictions from photoaffinity labeling studies (30 -32). Amino acid residues in analogous positions in segments IIIS6 and IVS6 appear to contribute to high affinity binding of DHPs (Fig. 5). Our results, therefore, are consistent with a "domain interface model" for DHP action proposed previously (32). We hypothesize that DHP binding at the interface between domains III and IV influences domain-domain interactions that are important for channel gating. Interdomain interactions have been observed for voltage-dependent activation of K ϩ channels formed from four identical subunits (50). Such interactions may also take place among the four different domains of Na ϩ and Ca 2ϩ channels during the voltage-dependent activation process.
Site-directed mutagenesis experiments demonstrate that Ca 2ϩ binding in the pore of the Ca 2ϩ channel is allosterically coupled to DHP binding (51,52). In the context of the domain interface model, DHP binding between transmembrane segments IIIS6 and IVS6 may cause conformational changes that are transmitted to the pore region of the channel resulting in increased affinity for Ca 2ϩ . We have postulated that this increased affinity for Ca 2ϩ stabilizes a blocked state of the channel having Ca 2ϩ tightly bound in the pore (51).
The domain interface model of drug action can now be generalized to the phenylalkylamine and benz(othi)azepine receptor sites as well as the DHP receptor site based on recent studies (37, 38, 53-55). All three classes of Ca 2ϩ antagonist drugs interact with receptor sites formed by the IIIS6 and IVS6 transmembrane segments and are therefore well positioned to alter protein-protein interactions at the interface between domains III and IV. Thus, the modulation of Ca 2ϩ channel function by all three classes Ca 2ϩ antagonist drugs appears to be structurally analogous to the regulation of allosteric enzymes by their effectors.
Allosteric Interactions among Dihydropyridines, Phenylalkylamines, and Benz(othi)azepines-Ligand binding studies indicate that the phenylalkylamines, benz(othi)azepines, and DHPs bind to three separate receptor sites that interact allosterically (4,5). This requires that all three classes of drugs be able to occupy their receptor sites simultaneously. Consistent with the existence of separate receptor sites for phenylalkylamines versus DHPs and benz(othi)azepines, permanently charged DHPs and benz(othi)azepines approach their receptor sites from the extracellular side, whereas charged phenylalkylamines approach their receptor site from the intracellular side (19 -22, 56 -58). The close proximity of the receptor sites for DHPs, phenylalkylamines, and benz(othi)azepines (37,38,(53)(54)(55) raises the possibility that individual amino acid residues may be required for high affinity binding of more than one of these ligands. In agreement with that expectation, amino acid residues required for high affinity binding of two or more classes of these drugs have been identified (e.g. Tyr 1463 , Ala 1467 , and Ile 1471 ; Refs. 37, 38, and 53-55 and this study). In fact, Tyr 1463 is essential for high affinity binding of all three classes of Ca 2ϩ channel antagonists. How can these findings be accommodated with earlier results indicating separate binding sites for phenylalkylamines, benz(othi)azepines, and DHPs? Phenylalkylamines are considered to block Ca 2ϩ channels directly by occluding the transmembrane pore through which Ca 2ϩ ions move. In contrast, DHPs are thought to bind to identical or overlapping sites at which agonists increase Ca 2ϩ channel activity and antagonists reduce Ca 2ϩ channel activity, so they cannot bind in a manner that blocks the pore. Therefore, it is our working hypothesis that these different classes of drugs bind to closely spaced sites that have both shared and distinct molecular components and have different spatial relationships to the pore. Binding of DHPs affects pore function allosterically, whereas binding of phenylalkylamines blocks the pore directly. The relationships among the amino acid residues that comprise these receptor sites are defined in more detail in the following paper (59). The amino acid sequences of transmembrane domains IIIS6 and IVS6 are illustrated as ␣ helices. White letters inside black circles represent amino acids that when mutated have a reduction in DHP binding that is greater than 5-fold. Black letters inside shaded circles represent amino acids that when mutated have a significant reduction in DHP binding that is less than 5-fold. The dotted circle represents Phe 1462 , which when changed to alanine has 416-fold increased affinity for block by DHPs. Black letters inside white circles are amino acids that had no effect on DHP binding when mutated. Gray letters inside white circles were not analyzed in this study. Arrows point to residues that have been shown to be critical for block by phenylalkylamines and benz(othi)azepines (53)(54)(55). A schematized DHP ligand is depicted contacting the key binding determinants in the two transmembrane segments.