The Role of Region IVS5 of the Human Cardiac Calcium Channel in Establishing Inactivated Channel Conformation

The role of inactivated channel conformation and use dependence for diltiazem, a specific benzothiazepine calcium channel inhibitor, was studied in chimeric constructs and point mutants created in the IVS5 transmembrane segment of the L-type cardiac calcium channel. All mutations, chimeric or point mutations, were restricted to IVS5, while the YAI-containing segment in IVS6, i.e. the primary interaction site with benzothiazepines, remained intact. Slowed inactivation rate and incomplete steady state inactivation, a behavior of some mutants, were accompanied by a reduced or by a complete loss of use-dependent block by diltiazem. Single channel properties of mutants that lost use dependence toward diltiazem were characterized by drastically elongated mean open times and distinctly slower time constants of open time distribution. Mutation of individual residues of the IVMLF segment in IVS5 did not mimic the complete loss of use dependence as observed for the replacement of the whole stretch. These results establish evidence that amino acids that govern inactivation and the drug-binding site and other amino acids that are located distal from the putative drug-binding site contribute significantly to the function of the benzothiazepine receptor region. The data are consistent with a complex “pocket” conformation that is responsive to a specific class of L-type calcium channel inhibitors. The data allow for a concept that multiple sites within regions of the α1 subunit contribute to auto-regulation of the L-type Ca2+ channel.

L-type Ca 2ϩ channels of cardiac, skeletal, and smooth muscle play a central role in excitation-contraction coupling. It is also believed that these channels may participate in the pathophysiology of some cardiac arrhythmias, hypertension, and angina pectoris (1)(2)(3)(4). Increases in [Ca 2ϩ ] i have been linked to a variety of changes in membrane properties that contribute to the changes in excitability, conduction, refractoriness, automaticity, and vascular resistance (1)(2)(3)(4). In terms of excitation-contraction coupling, Ca 2ϩ enters the cell during depolarization through voltage-activated calcium channels and triggers contraction but also is responsible for activation of other cellular functions. In the course of sustained depolarization, Ca 2ϩ currents progressively undergo voltage-dependent inactivation with specific kinetics, primarily regulated by the membrane potential.
Ca 2ϩ channels are the target for three main classes of drugs that include dihydropyridines (DHP), 1 phenylalkylamines (PAA), and benzothiazepines (BTZ). These molecules bind specifically to different sites of the Ca 2ϩ channel ␣ 1 subunit, and their binding domains are allosterically linked to each other (5)(6)(7). Ca 2ϩ channel antagonists, particularly diltiazem and verapamil, block calcium channels in a voltage-and use-dependent manner; thus the binding of the drug facilitates protein conformational changes that alter channel function by interfering with channel gating. These drugs serve as useful tools in dissecting molecular mechanisms of channel inactivation. Several amino acid residues that are involved in Ca 2ϩ antagonist binding, when mutated, change the electrophysiological properties of the channel, such as voltage-dependent inactivation. Single amino acids in segments IIIS6 and IVS6 (5, 8 -11) have been identified as determinants of inactivation. Mutation of key amino acids in these segments alters not only the high affinity drug-binding sites but also the complexity of the kinetic behavior of the channels, consequently changing use-dependent block particularly for the PAAs and BTZs (for review, see Ref. 12).
Our previous finding showed that a segment in IVS5 of the human ␣ 1C (Ca v 1.2) subunit is critically involved in inactivation of the channel (13). Consequently, mutants constructed in this region lost the characteristic use-dependent block by PAA and BTZ and recovered from inactivation significantly faster after drug block compared with the wild type channel. However, [ 3 H]PN200 -110 (isradipine) binding and allosteric interaction assays revealed that the DHP and BTZ receptor sites maintained normal coupling in the chimeric mutant channels. In the present study, we analyze the impact of individual amino acids in IVS5 (Ile, Val, Met, Leu, Phe) on the inactivation of the human ␣ 1C subunit. Our observations with point mutants in IVS5 strongly suggest that amino acid substitutions outside of or distal to the IIIS6 and IVS6 segments, where the key positions for drug interactions are located, play a critical ancillary role in determining inactivation kinetics and use dependence of the channel.

EXPERIMENTAL PROCEDURES
Site-directed Mutagenesis of Human Heart ␣ 1C -Amino acids that reside in the IVS5 region of human heart ␣ 1C (Ile, Val, Met, Leu, Phe) were sequentially mutagenized. A two-step PCR method, termed "megaprimer" PCR was utilized. The first set of oligonucleotides (forward primers) carrying the desired bases had at least 15 nucleotides on  either side of the mutation and had the sequence GTG GCC CTC CTG  ATC TTC ATG CTG TTC TTC ATC; CTC CTG ATC GTG ATG GTG  TTC TTC ATC TAC GCG; CTG ATC GTG ATG CTG ATG TTC ATC  TAC GCG GTG; CCC TAT GTG GCC CTC CTG CTC TTC CTG GTG  TTC TTC ATC TAC GCG for V1339F, L1341V, F1342M, and HHT-5421, respectively. The reverse primer GGT TGA TGA TCA GGA AGG C was designed around the BclI site (4311) of hHT-1 (14). PCR was performed on the hHT-1 template using in each reaction one mutant primer and a BclI primer. The amplification products (526 bp) were isolated from high range agarose gel.
After purification, these products were utilized as megaprimers, and PCR amplifications were done between a primer designed around the AatII site (3807) and the individual megaprimers. The products were gel-isolated again and ligated into the EcoRV site of pBlueScript KS (Ϫ), and DNA from a number of resultant clones was sequenced to identify correct mutant cDNAs. Finally, the mutant cassettes were liberated by AatII and BclI restriction cleavage, isolated, and ligated into AatII/BclI-cleaved hHT-1 to replace the corresponding wild type segment of ␣ 1C .
Expression of Calcium Channels in Xenopus Oocytes-Expression of the wild type and mutant calcium channels was done as previously described (15). In vitro synthesized cRNA was made using the mMessage mMachine synthesis kit (Ambion). Xenopus oocyte isolation and cRNA injection were performed as published elsewhere (16). Briefly, female Xenopus laevis (purchased from Xenopus I, Ann Arbor, MI) frogs were anesthetized by exposing them for 15-20 min to 0.15% methanesulfonate salt of 3-aminobenzoic acid ethyl ester (MS-222; Sigma) solution before pieces of the ovary were removed. The follicular layers of the oocytes were digested with 2.0 mg/ml collagenase (Type IA; Sigma) dissolved in OR-2 medium (in mM): 82.5 NaCl, 1 KCl, 1 MgCl 2 , and 5 HEPES, pH 7.5. Stage V-VI oocytes were incubated at 19°C in P/S medium (in mM): 96 NaCl, 2.0 KCl, 1.0 MgCl 2 , 1.8 CaCl 2 , 5 HEPES, 2.5 sodium pyruvate, and 0.5 theophylline at pH 7.5. The P/S medium was supplemented with 100 units/ml penicillin and 100 g/ml streptomycin.
Whole-cell Ba 2ϩ Current Recordings-The recording medium was a Ca 2ϩ -and Cl Ϫ -free solution composed of (in mM): 40 Ba(OH) 2 , 50 Nmethyl-D-glucamine, 2 KOH, 5 HEPES, pH adjusted to 7.4, with methanesulfonic acid. Voltage and current electrodes were filled with 3 M KCl and had a resistance of 0.5-1.5 megohms. Currents were recorded using an Axoclamp-2A (Axon Instruments Inc., Foster City, CA) amplifier. Whole cell leakage and capacitive currents were subtracted on line using the P/4 procedure (19). Currents were digitized at 1 kHz after being filtered at 1 kHz. The pClamp software (version 5.6 Axon Instruments) was used for data acquisition, and version 6.0.3 was used for analysis. Ba 2ϩ currents were elicited by a 350-ms-long depolarizing pulse from a holding potential of Ϫ80 mV to test potentials between Ϫ30 mV and ϩ40 mV in 10-mV increments to determine the peak potentials of the current voltage relationship of the Ca 2ϩ channel construct.
Use-dependent block was determined as the inhibition of peak I Ba during trains of 15 test pulses of 80-ms duration applied at 0.5 Hz from a holding potential of Ϫ60 mV to test potentials ϩ20 mV positive to the peak potential of the I-V curves. Identical pulse protocols were used in the presence of drugs. Diltiazem (racemic) was perfused in the bath (for a 2.5-min period) at concentrations of 200 M. This concentration was selected because it provided sufficient block for the use-dependent measurements in Xenopus oocytes (13).
A double pulse protocol was used in those experiments measuring the voltage dependence of steady-state inactivation. The protocol consisted of a 5-s prepulse that ranged from Ϫ80 to ϩ50 mV at a holding potential of Ϫ80 mV followed by a 30-ms return pulse to Ϫ80 mV. Finally a 400-ms-long test pulse was applied at ϩ20 mV. The curves were fit to the Boltzmann equation Recovery from Inactivation-Recovery of I Ba from inactivation was studied after depolarizing the calcium channels during a 3-s prepulse to ϩ20 mV. The time course of I Ba recovery from inactivation was estimated at a holding protential of Ϫ80 mV by applying a 400-ms test pulse to ϩ20 mV at various time intervals after the conditioning prepulse. Peak I Ba values were normalized to the peak current amplitude measured during the prepulse. After the double-pulse protocol, the membrane was hyperpolarized to Ϫ100 mV for 3 min to permit complete recovery from inactivation and block.
Single Channel Recordings-Single channel recordings were obtained from cell-attached patches using an AXON 200 amplifier (20). After mechanically removing the vitelline membrane of Xenopus oocytes (after about 5 min of incubation in hypertonic solution composed of in mM: 200 potassium-aspartate, 20 KCl, 1.0 MgCl 2 , 10 EGTA, 10 HEPES), the membrane potential of the oocytes was "zeroed" with a high potassium medium composed of (in mM): 140 KCl, 2 MgCl 2 , 5 EGTA, and 5 HEPES, pH 7.4. The patch electrodes were coated with Sylgard, then fire-polished, and had resistances between 5 and 20 megohms. The pipette solution contained (in mM): 110 BaCl 2 and 10 HEPES, pH 7.4. Single channel currents were low pass-filtered at 2 kHz, digitized at 10 kHz, and stored for off-line analysis. Only well resolved openings were measured; therefore, all single channel experiments were conducted in the presence of 1 M Bay K8644 at room temperature (22-23°C). To cancel capacitive transients and leak currents, blank records with no openings were subtracted from those with openings. Patch potentials were maintained at Ϫ80 mV, and 500 depolarizing pulses (180 ms long) to ϩ20, ϩ30, and ϩ40 mV, respectively, were delivered. Open and closed transitions were detected by the halfamplitude threshold criterion. Open time histograms were fitted to multiple exponential functions by using the maximum likelihood method. Analysis was performed on 3-10 patches in which we succeeded in recording from 500 up to 1000 sweeps.

RESULTS
To examine the role of individual amino acid residues in governing the participation of transmembrane segment IVS5 in inactivation and use-dependent block, we substituted each amino acid of the wild type ␣ 1 in the region where previously we identified a segment with a chimeric construct (HHT-5411) that is a determinant of inactivation ( Fig. 1). By analyzing the macroscopic Ba 2ϩ currents, we noticed in mutants I1338L and M1340L that the half-maximal voltage for activation (V 0.5 act ) was slightly but significantly shifted to negative potentials ( Table I). The reversal potential (data not shown), and peak potential (data not shown), were not significantly affected by any of the mutations. There was, however, a slight but significant change in the slope factor (k act ) for I1338L, M1340L, and HHT-5421 (Table I). On the other hand, we observed more widespread alterations in the inactivation properties of mutants. The V 0.5 act of the steady-state inactivation drastically shifted to hyperpolarized potentials for L1341V and the HHT-5421 chimera. Also, the inactivation rate was significantly slower for L1341V, F1342M, and the HHT-5421 compared with the wild type.
Ca 2ϩ Channel Block by Diltiazem Determined by Amino Acids in IVS5-In the next set of experiments, we tested whether the alterations in inactivation properties of certain mutant channels correlated with the use dependence properties. Indeed, the reduced rate in current inactivation of HHT-5421 and L1341V was accompanied by a substantial reduction in use-dependent block of I Ba by diltiazem during 80-ms steps at 0.5 Hz. On the contrary, mutation V1339F enhanced use-dependent block by diltiazem (and also by verapamil, data not shown) during low frequency pulse trains compared with wild type (WT) channel. The single phenylalanine to methionine substitution (F1342M) and the other two mutants, M1340L and I1338L in segment IVS5 of ␣ 1 , induced similar use-dependent I Ba inhibition by diltiazem compared with the wild type (Fig. 2, B and C).
Interrelationship between Channel Inactivation and Block Development by Diltiazem-The voltage dependence of steadystate inactivation, induced by a 5-s prepulse, showed that WT, I1338L, V1339F, M1340L, and F1342M mutant channels are nearly completely inactivated at positive voltages (Table I). In contrast, the steady-state inactivation curves for HHT-5421 and L1341V exhibited a slowed rate of inactivation and also a residual current that appears associated with a loss of voltage dependence. The half-maximal voltage for inactivation (V 0.5 inact ) was shifted toward more positive potentials for mutant HHT-5421 (Ϫ0.4 Ϯ 1.7 mV) and for L1341V (2.4 Ϯ 4.9 mV) compared with that of the wild type channel (Ϫ7.6 Ϯ 1.8 mV). The slope factor of the curve was not significantly affected for L1341V (11.0 Ϯ 0.7 mV), but a change occurred for HHT-5421 (7.1 Ϯ 0.3 mV) compared with the WT (11.9 Ϯ 0.5 mV) (Fig. 3A). These latter two mutant channels inactivated completely after diltiazem addition, and no sustained current remained at the end of the depolarizing pulse (Fig. 3B). This observation indicates that the transition between the open state and inactivated state and/or the stability of the inactivated state are impaired by these mutations.
The accumulation of channels in inactivation during a pulse train depends on how fast inactivation is removed between pulses. To provide a more detailed analysis of the role of channel inactivation in block development, we investigated the effects of the mutations on the time course of recovery from inactivation by employing a double pulse protocol at a holding potential Ϫ80 mV (Fig. 3C). In general, I Ba recovered to 80 -90% of control within 28 s, whereas the remaining current did not recover within the 28-s period analyzed. The effect of diltiazem appeared in an overall slowing of the recovery time course in the WT and mutant channels. The recovery time courses in the absence of drug were similar for the WT ( 1 ϭ 1655.3 Ϯ 172.9 ms, n ϭ 8) and mutant channels, I1338L, M1340L, and F1342M (see Table II), while HHT-5421 ( 1 ϭ 541.6 Ϯ 180.2 ms, n ϭ 3) and L1341V ( 1 ϭ 782.5 Ϯ 231.9 ms, n ϭ 8) recovered with a faster time course. The recovery of V1339F followed a slower time course than the WT when approximated with a single exponential. The tendency remained similar in the presence of diltiazem (Table II). Approximating the recovery by two exponentials gave more insight into the biphasic nature of the procedure. None of the mutations showed significant differences compared with the WT in the early faster pulse ( 1 ) of the recovery from inactivation when tested in the absence of drug. However, in the absence of drug, mutation V1339F recovered according to a slower 2 , while L1341V showed a 2 faster than WT in the slow phase of recovery. This is in good agreement with that observed for the single exponential approximation. Addition of diltiazem drastically slowed the fast phase of recovery from inactivation for WT ( fast ϭ 1531.  Table II); however, this behavior was observed only for L1341V and HHT-5421 in the slow phase of recovery (Fig. 3D).
Single Channel Behavior of IVS5 Mutants-If mutant channels enter the inactivated state from the open state more slowly than the wild type one can assume that these channels will display an increased mean open time compared with the wild type. To obtain further insight into the mechanism as to how the impaired inactivation is established in certain mutants, six TABLE I Activation and inactivation parameters of wild type and mutant calcium channels The voltage dependence of activation was determined from I-V curves obtained by step depolarizations from a holding potential Ϫ80 mV to various test potentials. To obtain the voltage of half-maximal I Ba activation (V 0.5 act ) current-voltage relation curves were fitted to the Boltzmann function. k act is the slope factor of the activation curve. The voltage dependence of inactivation (steady-state inactivation) was investigated for WT and mutant channels in control and in the presence of 200 M diltiazem. The half-maximal voltage for steady state inactivation (V 0.5 inact ) and the slope factor (k inact ) were calculated by fitting the data to the Boltzmann equation. The inactivation rates for V m ϭ ϩ20 mV were estimated from a single trace at time t ϭ 3 s. The inactivation time course of whole cell Ba 2ϩ traces was determined by a double exponential fitting. The p values (Ͻ0.05) for comparison to the wild type channels are marked as * next to each electrophysiological value.   0.49 ms). Surprisingly, the mean open time constants for chimeric mutant HHT-5371 at ϩ30 mV were faster ( fast ϭ 1.05 Ϯ 0.39 ms, slow ϭ 3.56 Ϯ 0.74 ms) than WT ( fast ϭ 1.33 Ϯ 0.09 ms, slow ϭ 5.19 Ϯ 0.49 ms); however, this channel had a high frequency of short reopenings throughout the depolarizing pulse. Despite the decreased use-dependent block (Fig. 2), the mean open time for HHT-5371 was also shorter at both ϩ20and ϩ30-mV depolarizing pulses compared with the WT (Fig. 5A, Table III). Interestingly, when valine was replaced with the more hydrophobic phenylalanine in mutant V1339F, the open time constants were shortened compared with the WT. This finding suggests crucial roles for size and hydrophobicity of the residue at position 1339 for inactivation. The single channel data for V1339F are supported by an enhanced use dependence and slowed FIG. 3. Voltage-dependent inactivation and the recovery of I Ba from inactivation for single amino acid mutants and chimeric constructs within transmembrane segment IVS5. A, steady-state inactivation curves are shown for wild type, I1338L, V1339F, M1340L, L1341V, F1342M, and HHT-5421 in the absence of drug. Currents were elicited by test pulses to ϩ20 mV following 5-s conditioning pulses to various potentials from a holding potential of Ϫ80 mV. Voltage dependence of inactivation of ␣ 1C wild type and mutants coexpressed with ␤ 2a and ␣ 2 /␦ subunits is shown in A, in which normalized peak current during test pulse is plotted as a function of conditioning prepulse. Experimental points were fitted by a Boltzmann distribution. I Ba ϭ 1/{1ϩexp [(V Ϫ V 0.5 )/k]}, where V 0.5 is the voltage at half-maximum of inactivation, and k is a slope factor. B, effects of wild type and mutants, L1341V and HHT-5421, on the voltage dependence of inactivation in the presence of 200 M diltiazem. C, recovery from inactivation of wild type, I1338L, V1339F, M1340, L1341V, F1342M, and HHT-5421 were measured by a two-pulse protocol in the absence of diltiazem. The currents were inactivated by a 3-s prepulse to ϩ20 mV. I Ba recovery at Ϫ80 mV was measured by applying a sequence of test pulses at various times after the prepulse. Peak currents of the test pulses were normalized to peak currents of the prepulse and against time. Plotted lines represent fits of the mean data by single exponentials. D, recovery of I Ba from inactivation in the presence of 200 M diltiazem of wild type, L1341V, and HHT-5421 mutant channels. recovery from inactivation. In contrast, substituting leucine for valine at position 1341 had diverse effects on single channel recordings that were anticipated from the reduced use-dependent block for diltiazem. These results are summarized in Table  III. The longest open time was observed for the mutants HHT-5411 (6.45 Ϯ 0.6 ms, n ϭ 5) and HHT-5421 (11.17 Ϯ 1.33 ms, n ϭ 7), 123 and 261% of the wild type, respectively (Table III, Fig. 5B). Thus, these mutants exhibited slowed inactivation from the open state (Fig. 5B). These observations are also consistent with the data from the inactivation of macroscopic currents (Table I). The average latencies to the first opening for all mutants was similar to that of the wild type, suggesting that these mutations have little effect on the activation process (Fig. 5C). DISCUSSION We have carried out a systematic analysis of the role of a stretch of amino acid residues in IVS5 of the human ␣ 1C channel in establishing slow, incomplete inactivation and also establishing a correlation with use-dependent block by diltiazem.
In previous experiments, after expression in Xenopus oocytes, chimeric mutants HHT-5371 and HHT-5411 (13) decreased use-dependent block and shifted voltage dependence of steadystate inactivation to positive membrane potentials, and the inactivation appeared incomplete. For these mutants, the recovery from inactivation was accelerated at negative membrane potentials. Here we have shown that the destabilized inactivation is due both to the decreased rate of inactivation from the open state and an increased rate of exit from the inactivated state. The mean open time was increased for HHT-5421 and HHT-5411 compared with wild type suggesting a defect in the process of inactivation. Despite the slow inactivation kinetics of HHT-5371 at the macroscopic level, the mean open time was not changed compared with that of the wild type channel. Single channel analysis revealed frequent reopenings for HHT-5371 indicating a substantial impairment of the stability of the inactivated state.
We attempted to introduce a wild type-like inactivation into HHT-5411 through single amino acid exchanges. We constructed the HHT-5421 mutant channel by introducing a phenylalanine at position 1340 in HHT-5411, which is recognized as being important for the inactivation. This phenomenon is similar to that found in Na ϩ channels, where at the corresponding position, a phenylalanine was shown to play an essential role in fast inactivation (23). Interestingly, the single amino acid difference from HHT-5411 to HHT-5421 did not exert the anticipated wild type-like inactivation but showed a more extensive use-dependent block by diltiazem compared with HHT-5411. Recent evidence provided further proof (24) that Ile-1485 and Met-1487, also called an IFM cluster in the Na ϩ channel, contribute to stabilizing the hydrophobic inactivation particle for fast inactivation. In our study, however, because none of the mutations were introduced in the binding site region, we speculate that mutations in HHT-5421 and HHT-5411 may cause a conformational modification of the binding site that slows the formation of the inactivation gate or decreases the drug access to the binding site of the channel by steric hindrance. We assume that the cluster of LFLVM and LFLVF amino acids enters into a hydrophobic interaction with other amino acids in the intracellular mouth of the pore during the inactivation process. On the other hand, replacement of valine with a bulkier phenylalanine (V1339F) that has a higher molecular weight than valine enhanced the use-dependent block, while the inactivation rate remained similar to that of the wild type. In mutant L1341V, however, we observed a decreased use-dependent block by diltiazem, slowed Ca 2ϩ current decay, and facilitated I Ba recovery from inactivation. Leucine and valine are both aliphatic, hydrophobic amino acids, but valine has a smaller molecular weight and is less hydrophobic than leucine. However, contrary to previous studies concentrating on segment IIIS6 in the ␣ 1 subunit of L-type Ca 2ϩ channels (11,25,26), our results suggest that the valine in this position actually destabilizes the inactivated state. L1341V and HHT-5421 (HHT-5411 and HHT-5371) (13) shifted the voltage dependence of steady-state inactivation to more positive membrane potentials and inactivated incompletely during the pulse, and a substantial sustained current remained at the end of the depolarization. Therefore, it is possible that these mutant channels might increase the energy requirement from the closed states to the inactivated state (27). After replacement of phenylalanine by methionine to create mutant F1342M, we observed a very small effect on inactivation, which is in accordance with results from previous studies on the Na ϩ channel (28). At this point, however, we cannot confirm that just one critical amino acid is linked to the impaired inactivation in the IVS5 region and is responsible for channel gating. In fact, we feel that the importance of the hydrophobic amino acids, phenylalanine, valine, isoleucine, and leucine, and their apparent involvement as critical elements for gating, are evident. It is also possible that four or five amino acids in the IVS5 of the human ␣ 1C subunit segment represent a domain that is critical for the faster inactivation. Among the six new mutants investigated, mutants HHT-5421 and L1341V caused the most pronounced effect on channel inactivation kinetics as well as diltiazem sensitivity. The effect of the HHT-5421, HHT-5411, and  L1341V mutations, i.e. slow entry into the inactivated state, implies that these amino acid residues participate in a conformational change that may be required to form an effective inactivation gate receptor.
Taken together, our results provide convincing evidence that IVMLF mutations in the IVS5 segment affect overall ␣ 1C Ca 2ϩ channel inactivation kinetics compared with wild type. Previous findings described changes in the inactivation properties by site-directed mutations in the IVS6 and IIIS6 segments in the pore region, in the binding motifs, or near the binding pocket of voltage-gated Ca 2ϩ channels. Numerous previous studies focused on inactivation determinants in segment IVS6 and IIIS6 that participate in the formation of the binding pocket for PAA and BTZ (12). In 1995, Hockermann et al. (29) identified three pore-oriented amino acid residues in IVS6 (Tyr-1463, Ala-1467, and Ile-1470) as critical determinants for high affinity block by PAA. Later, Johnson et al. (30) analyzed the pharmacological effect of mutations at the PAA receptor site in IVS6 of the ␣ 1 subunit in tsA201 cells. The combined mutation of residues Tyr-1463, Ala-1467, and Ile-1470 reduced use-dependent block and accelerated the rate of recovery from inactivation. The result of their studies suggested that the mutant YAI residues interact directly with verapamil and introduce steric hindrance to drug access and binding.
Degtiar et al. (5,31) showed that mutation I1811M (AL25/-I) in transmembrane segment IVS6 had the highest impact, reducing use dependence and displaying the slowest inactivation time course. They suggested that this pore-lining isoleucine in IVS6 played a key role in the formation of the PAA receptor site. Berjukov et al. (11) also reported that two amino acids from segment IVS6 in rabbit heart ␣ 1C-a (V1504A, L1381I) are strong inactivation determinants and showed that the reduced use-dependent block of V1504A caused by a reduced (ϩ)-cisdiltiazem sensitivity and could be explained by an allosteric modulation of the drug binding process.
Kraus et al. (26) investigated the contribution of individual IIIS6 amino acid residues for diltiazem sensitivity by employing alanine-scanning mutagenesis. Mutations of IIIS6 residues Phe-1164 and Val-1165 slowed inactivation kinetics and accelerated the recovery from drug block and were found to be major determinants for use-dependent diltiazem block. They proposed that the time course of recovery from channel block was critically determined by steric orientation of the receptor determinants in the pore region (9). This study was supported by recent observations by Kraus et al. (32) in familial hemiplegic migraine mutation, which demonstrated that pore-forming residues valine and isoleucine play an important role for P/Q-type Ca 2ϩ channel inactivation and can be responsible for neuronal instability in patients with hemiplegic migraine mutation.  outside the putative drug-binding region in the intracellular loop between domains I and II, the use-dependent block, was also modulated block by (Ϫ)-gallopamil. Later, they presented experimental evidence that different ␤-subunit compositions of Ca v 1.2 affect the channel inactivation kinetics and PAA sensitivity as well (33,34). Higher sensitivity to PAA was also observed with Ca 2ϩ as a charge carrier (34). In the latter case, however, the slower recovery from PAA-induced channel conformation and not the inactivation played a significant role in the enhanced use dependence. We also identified amino acid residues outside of the drug-binding domain, at the cytoplasmic end of the putative IVS5 ␣-helix, as important determinants for the inactivation mechanism. Our results support the notion that residues in IVS5 indirectly control diltiazem sensitivity in a frequency-dependent manner by slowing channel inactivation and by facilitating recovery from drug block and channel unblock between individual test pulses of train.
In summary, a number of results provide evidence that more than one region of the ␣ 1 -subunit of the L-type Ca 2ϩ channel contributes to use-dependent block by PAA and BTZ drugs. These regions include the IVS6, IVS5, IIS6, and I-II intracellular connecting loop and also the nature of the ␤-subunit and the charge carrier involved. The common underlying mechanism of the phenomenon is probably in the voltage-dependent inactivation and inactivated conformation of the channel molecule. Thus far, there is no information about the three-dimensional structure of calcium channels. An increasing body of information, however, points to the movement of the voltage sensor and associated membrane segments (35,36). Thus, it is conceivable that regions of the inactivated channel, including the PAA and BTZ "receptor pocket" that are otherwise buried in membrane, move out into the extracellular space and provide better access to the drug. This process occurs regardless of which mechanism has generated the accumulation of the inactive channel.