Inactivation determinant in the I-II loop of the Ca²⁺ channel α₁-subunit and β-subunit interaction affect sensitivity for the phenylalkylamine (−)gallopamil

Abstract

1. The role of calcium (Ca²⁺) channel inactivation in the molecular mechanism of channel block by phenylalkylamines (PAAs) was analysed in a PAA-sensitive rabbit brain class A Ca²⁺ channel mutant (α_1A-PAA). Use-dependent barium current (I_Ba) inhibition of α_1A-PAA by (−)gallopamil and Ca²⁺ channel recovery from inactivation and block were studied with two-microlectrode voltage clamp after expression of α_1A-PAA and auxiliary α₂-δ- and β_1a- or β_2a-subunits in Xenopus oocytes.

2. Mutation Arg387Glu (α_1A numbering) in the intracellular loop connecting domains I and II of α_1A-PAA slowed the inactivation kinetics and reduced use-dependent inhibition (100 ms test pulses at 0.2 Hz from -80 to 20 mV) of the resulting mutant α_1A-PAA/R-E/β_1a channels by 100 μm (−)gallopamil (53 ± 2 %, α_1A-PAA/β_1avs. 31 ± 2 %, α_1A-PAA/R-E/β_1a, n ≥ 4). This amino acid substitution simultaneously accelerated the recovery of channels from inactivation and from block by (−)gallopamil.

3. Coexpression of α_1A-PAA with the β_2a-subunit reduced fast I_Ba inactivation and induced a substantial reduction in use-dependent I_Ba inhibition by (−)gallopamil (25 ± 4 %, α_1A-PAA/β_2a; 13 ± 1 %, α_1A-PAA/R-E/β_2a). The time constant of recovery from block at rest was not significantly affected.

4. These results demonstrate that changes in channel inactivation induced by Arg387Glu or β_2a-α₁-subunit interaction affect the drug-channel interaction.

Calcium (Ca²⁺) channel inhibition by drugs such as phenylalkylamines (PAAs), benzothiazepines (BTZs) and mibefradil increases during repetitive depolarisation of the membrane (Lee & Tsien 1983; McDonald et al. 1984; Bezprozvanny & Tsien 1995; Aczél et al. 1998). Such a ‘use-dependent’ channel inhibition reflects distinct drug interactions with the resting, open and inactivated channel states. It is believed that state-dependent Ca²⁺ channel block plays an important role in the therapeutic action of PAAs and BTZs as antiarrhythmics (Hondeghem & Katzung, 1984).

Functional studies on mutant Ca²⁺ channels enabled the first insights into the molecular architecture of the Ca²⁺ channel drug-binding domains (see Hockerman et al. 1997b and Striessnig et al. 1998 for review). The available data suggest that three different classes of Ca²⁺ channel antagonists (PAAs, BTZs and 1,4-dihydropyridines) bind in close proximity within the pore region of L-type Ca²⁺ channel α₁-subunits (Striessnig et al. 1998). Three amino acids in segment IVS6 (Tyr1463, Ala1467, Ile1470) and four residues in transmembrane segment IIIS6 (Tyr1152, Ile1153, Phe1164 and Val1165) have been identified as crucial L-type determinants of the PAA sensitivity (Hockerman et al. 1995, 1997a). Insertion of three ‘L-type-specific’ residues (Tyr1463, Ala1467 and Ile1470) into segment IVS6 of the only weakly PAA sensitive class A (α_1A) Ca²⁺ channel transferred PAA sensitivity to the corresponding α_1A mutant (here called α_1A-PAA; Hering et al. 1996).

An unequivocal identification of the PAA binding determinants by mutational analysis of α₁ Ca²⁺ channel subunits is, however, complicated by an apparent interdependence between Ca²⁺ channel block and inactivation gating (see Hering et al. 1998 for review). In particular, transfer of the IVS6 L-type determinants of PAA sensitivity (Tyr1463, Ala1467 and Ile1470) to class A Ca²⁺ channels accelerated inactivation (Hering et al. 1996; Degtiar et al. 1997). Moreover, introduction of an additional L-type amino acid (Met1464 into IVS6 of α_1A-PAA) facilitated channel inactivation and enhanced use-dependent channel block by (−)gallopamil (Hering et al. 1996). Accordingly, substitution of a PAA determinant in α_1A-PAA (Ile1470 by the corresponding class A channel Met, α_1C-a numbering) which substantially reduced channel inactivation induced an about 30-fold decrease of the apparent association rate for (−)devapamil (Degtiar et al. 1997) and alanine substitutions of three L-type amino acids localised close to the inner channel mouth on segment IIIS6 and IVS6 reduced Ca²⁺ channel inactivation and simultaneously BTZ and PAA sensitivity (Hering et al. 1997; Berjukow et al. 1999).

In our previous studies on the role of Ca²⁺ channel inactivation in channel block by PAAs and BTZs we have focused on residues that are located on segments IIIS6 and IVS6 (Degtiar et al. 1997; Hering et al. 1997; Berjukow et al. 1999). Here we analyse in a PAA-sensitive class A Ca²⁺ channel mutant (α_1A-PAA) expressed in Xenopus oocytes if inactivation determinants localised outside the channel pore of an α₁-subunit influence Ca²⁺ channel block by (−)gallopamil.

We demonstrate that a single amino acid substitution (Arg387Glu, α_1A numbering) in the intracellular loop between domains I and II slows channel inactivation and reduces sensitivity for (−)gallopamil. Furthermore, a reduced inactivation caused by coexpression of α_1A-PAA with β_2a- instead of the β_1a-subunit reduced Ca²⁺ channel block by (−)gallopamil even more dramatically. Our study clearly demonstrates that inactivation determinants that are localised outside the putative drug binding regions in the channel pore affect the molecular mechanism of use-dependent Ca²⁺ channel block by (−)gallopamil.

METHODS

Generation of α_1A-constructs

The construction of the PAA-sensitive triple rabbit brain class A Ca²⁺ channel mutant AL25 (named herein α_1A-PAA) was previously described (Hering et al. 1996). The derived mutant α_1A-PAA/R-E was constructed by introducing a single point mutation (R387E, α_1A numbering) into α_1A-PAA cDNA by the ‘gene SOEing’ technique (Horton et al. 1989). The point mutation was verified by sequence analysis. All constructs were inserted into the polyadenylating transcription plasmids pSPCBI 2 (a kind gift of Dr O. Pongs, University of Hamburg).

Electrophysiology

Female Xenopus laevis (NASCO, Fort Atkinson, WI, USA) were anaesthetised by exposing them for 15 min to a 0.2 % MS-222 (methane sulfonate salt of 3-aminobenzoic acid ethyl ester; Sandoz) solution before surgically removing parts of the ovaries. The frogs were then allowed to recover and returned to their tank. Each frog was reused up to two times and subsequently killed by decapitation under anaesthesia. The interval between the operations was longer than 4 months. Follicle membranes from isolated oocytes were enzymatically digested with 2 mg ml⁻¹ collagenase (Type 1A, Sigma). Calcium channel currents (I_Ba) were studied 2 to 7 days after microinjection of approximately equimolar cRNA mixtures of α₁ (0.3 ng per 50 nl)-β_1a(β_2a) (0.1 ng per 50 nl)-α₂δ (0.2 ng per 50 nl) with two-microelectrode voltage clamp of Xenopus oocytes with 40 mM Ba²⁺ as charge carrier in a bath solution containing (mM): 40 Ba(OH)₂, 40 N-methyl-D-glucamine, 10 Hepes, 10 glucose, adjusted to pH 7.4 with methanesulfonic acid as previously described (Grabner et al. 1996). Endogenous chloride currents of the oocytes were suppressed by injecting 20-40 nl of a 0.1 M BAPTA solution 30-240 min before the voltage clamp measurements. Voltage-recording and current-injecting microelectrodes were filled with 2.8 M CsCl, 0.2 M CsOH, 10 mM EGTA, 10 mM Hepes (pH 7.4) and had resistances of 0.3-2 MΩ.

Drug sensitivity was estimated as use-dependent Ca²⁺ channel block during 20 test pulses (100 ms) applied at 0.2 Hz from -80 mV to 20 mV corresponding to the peak current voltage of the current-voltage relationships of all studied Ca²⁺ channel mutants. Use-dependent block was measured after a 3 min equilibration of the oocytes in drug-containing solution. To estimate the accumulation of Ca²⁺ channels in inactivation under control conditions similar pulse trains were applied in the absence of drug. Resting channel block was measured in an individual set of experiments as peak I_Ba inhibition during 100 ms test pulses from -80 to 20 mV after a 5 min equilibration in drug-containing solution.

Recovery from inactivation was studied at a holding potential of -80 mV after depolarising Ca²⁺channels during a 3 s prepulse to 20 mV by applying 30 ms test pulses to 20 mV at various time intervals after the conditioning prepulse. Peak I_Ba values were normalised to the peak current measured during the prepulse. The time course of I_Ba recovery from inactivation was fitted to a biexponential function:

Initial rates of I_Ba decay (see Fig. 1B) were estimated by calculating the maximum derivative of mono- (α_1A-PAA/β_2a, α_1A-PAA/R-E/β_2a) or biexponential (α_1A-PAA/β_1a, α_1A-PAA/R-E/β_1a) fits to current inactivation during a 3 s depolarisation from -80 to 20 mV.

Figure 1. Mutation Arg387Glu and α₁-β_2a -subunit interaction reduce channel block by (−)gallopamil.

A, normalised I_Ba of mutants α_1A-PAA/β_1a, α_1A-PAA/R-E/β_1a, α_1A-PAA/β_2a and α_1A-PAA/R-E/β_2a during 3 s depolarising test pulses applied from a holding potential of -80 mV to 20 mV. B, comparison of the time course of I_Ba inactivation of mutants α_1A-PAA/β_1a, α_1A-PAA/R-E/β_1a, α_1A-PAA/β_2a and α_1A-PAA/R-E/β_2a (in control) measured as initial rate for I_Ba decay during a 3 s pulse from -80 mV to 20 mV (n ≥ 7, see Methods). Statistically significant reduction in the rate of current inactivation compared with α_1A-PAA/β_1a is indicated by the asterisk (* P < 0.05). C, use-dependent I_Ba inhibition of α_1A-PAA/β_1a, α_1A-PAA/R-E/β_1a, α_1A-PAA/β_2a and α_1A-PAA/R-E/β_2a by 100 μm (−)gallopamil during trains of 20 pulses (100 ms) applied at 0.2 Hz from a holding potential of -80 mV to 20 mV. Vertical bars, 0.5 μA. D, comparison of the use-dependent I_Ba block of α_1A-PAA and α_1A-PAA/R-E expressed with either the β_1a- or β_2a-subunit by 10 and 100 μm (−)gallopamil. The block of I_Ba was measured as cumulative peak current inhibition (as a percentage) during 20 depolarising pulses (100 ms, 0.2 Hz) in control (□) or in the presence of 10 μm () and 100 μm (−)gallopamil (▪). Bars represent the means ±s.e.m. (n = 5-8). * Significantly different from I_Ba block of α_1A-PAA/β_1a; † Significantly different from α_1A-PAA/β_2a. E, initial rate of peak I_Ba inhibition by 100 μm (−)gallopamil (see C for details of the pulse protocol). Smooth lines are single-exponential fits to averaged peak current decays (n≥ 5) with time constants of 13.0 ± 2 s (▴, α_1A-PAA/R-E/β_2a), 12.5 ± 1 s (•, α_1A-PAA/β_2a), 6.5 ± 0.5 s (▾, α_1A-PAA/R-E/β_1a) and 6.0 ± 0.5 s (▪, α_1A-PAA/β_1a). F, inactivation curves of α_1A-PAA/β_1a (○) and α_1A-PAA/R-E/β_1a channels (□) in control. Curves are drawn according to the equation I/I_max=I_ss+ (1 - I_ss)/(1 + exp[(V - V_0.5)/k]) (see Methods) with V_0.5= -11.3 ± 0.5 mV, k = 7.4 ± 0.4 mV, I_ss= 0.03 ± 0.01, n = 5 for α_1A-PAA/β_1a and V_0.5= -3.1 ± 0.1 mV, k = 4.20 ± 0.08 mV, I_ss= 0.07 ± 0.01, n = 4 for α_1A-PAA/R-E/β_1a.

Voltage dependence of I_Ba inactivation (‘inactivation curve’) was measured as normalised peak current during a 30 ms test pulse that was applied 3 ms after a 3 s conditioning prepulse to a given voltage. Conditioning and test pulses were applied every 60 s from a holding potential of -100 mV. Inactivation curves were fitted to the equation:

Data are given as means ±s.e.m. Statistical significance was calculated according to Student's unpaired t test (P < 0.05).

RESULTS

Mutation Arg387Glu in the domain I-II loop of the PAA-sensitive α_1A-PAA-subunit and β_2a-subunit interaction affect channel inactivation and sensitivity for (−)gallopamil

We have previously reported that amino acid substitutions in transmembrane segments IIIS6 and IVS6 of class A and class C α₁-subunits affect Ca²⁺ channel inactivation kinetics and simultaneously sensitivity for PAA and BTZ (Hering et al. 1998). In order to further characterise the role of channel inactivation in Ca²⁺ channel block by PAAs we substituted arginine by glutamine in position 387 of the intracellular loop between domains I and II of the PAA-sensitive class A channel mutant α_1A-PAA (see Herlitze et al. 1997).

As expected from previous studies on wild-type class A channels (Herlitze et al. 1997), the I-II loop mutation Arg387Glu reduced the rate of current decay of the resulting quadruple mutant α_1A-PAA/R-E and shifted the mid-point of the inactivation curve to more positive potentials (Fig. 1A, B and F). I_Ba of α_1A-PAA/R-E/β_1a displayed less use-dependent I_Ba inhibition by (−)gallopamil (10 and 100 μm) compared with α_1A-PAA/β_1a (Fig. 1C andD).

Next we determined if a modulation of the current decay by different β-subunits would affect Ca²⁺ channel block by (−)gallopamil. It is well established that coexpression of the β_2a-isoform induces a slow rate of voltage-dependent inactivation of Ca²⁺ channels (Stea et al. 1994). Accordingly, coexpression of the mutants α_1A-PAA and α_1A-PAA/R-E with the β_2a-subunit dramatically decreased I_Ba inactivation (Fig. 1A and B). Again, slower inactivation kinetics induced significantly less use-dependent block of α_1A-PAA/β_2a and α_1A-PAA/R-E/β_2a compared with α_1A-PAA/β_1a and α_1A-PAA/R-E/β_1a channels (Fig. 1C and D). Interestingly, peak I_Ba inhibition by (−)gallopamil occurred at a faster rate if Ca²⁺ channels were coexpressed with the β_1a-subunit (Fig. 1E). There was little additional effect of the point mutation on inactivation rate once the β_2a-subunit was coexpressed (Fig. 1A); however, there was still a substantial effect on drug sensitivity (Fig. 1D).

We did not observe significant resting channel block (<5 %) by 10 μm (−)gallopamil. Resting channel inhibition induced by 100 μm (−)gallopamil was 15.0 ± 1.7 % (α_1A-PAA/β_1a), 6.7 ± 1.1 % (α_1A-PAA/R-E/β_1a, P < 0.01 compared with α_1A-PAA/β_1a), 20.1 ± 1 % (α_1A-PAA/β_2a) and 18.2 ± 1 % (α_1A-PAA/R-E/β_2a) (n≥ 4).

Mutation Arg387Glu and α₁-β_2a-subunit interaction differently affect recovery of Ca²⁺ channels from block

To elucidate the molecular basis of the different (−)gallopamil sensitivities of α_1A-PAA and α_1A-PAA/R-E coexpressed either with the β_1a- or β_2a-subunit, we analysed the drug-induced changes in I_Ba recovery from inactivation and block. The time courses of I_Ba recovery were fitted to a double-exponential function (see Methods). Under control conditions the slow component in Ca²⁺ channel repriming reflects recovery of Ca²⁺ channels from a slow or ultra-slow inactivated state (Boyett et al. 1994). (−)Gallopamil dose-dependently slowed recovery in all mutants (Fig. 2) whereas the time constant of recovery from fast inactivation was not significantly affected by the drug. The latter finding suggests that slow recovery reflects the fraction of drug-modified Ca²⁺ channels.

Figure 2. Recovery of Ca²⁺ channels from inactivation and block by (−)gallopamil.

I_Ba recovery of α_1A-PAA/β_1a (A), α_1A-PAA/R-E/β_1a (B), α_1A-PAA/β_2a (C) and α_1A-PAA/R-E/β_2a (D) was measured by a two-pulse protocol in the absence (▪) and presence of 10 (•) and 100 μm (−)gallopamil (▴). Test pulses were applied at various times (between 50 ms and 15 s) after a conditioning prepulse (see Methods). I_Ba values were normalised to peak I_Ba of the conditioning prepulse and plotted as a function of time. Data points are fitted by a double-exponential function (see Table 1 for mean values).

Mutation Arg387Glu did more than just slow the rate of current inactivation (see Fig. 1A). As shown in Fig. 2A and B, α_1A-PAA/R-E/β_1a channels recovered significantly faster from inactivation (τ_fast= 0.20 ± 0.04 s, τ_slow= 1.9 ± 0.3 s, n = 7) compared with α_1A-PAA/β_1a (τ_fast= 0.34 ± 0.04 s, τ_slow= 2.8 ± 0.4 s, n = 7, P < 0.05, Table 1). Faster I_Ba recovery from inactivation was accompanied by significantly faster recovery of α_1A-PAA/R-E/β_1a channels from block by (−)gallopamil (α_1A-PAA/R-E/β_1a: τ_slow, 100 μm= 4.4 ± 0.2 s, n = 6; α_1A-PAA/β_1a: τ_slow, 100 μm= 7.3 ± 0.6 s, n = 6, P < 0.05, Table 1).

Table 1.

Time constants (τ, s) and corresponding amplitude coefficients (A) of double-exponential I_Ba recovery from inactivation (see Fig. 2)

Construct	Conditions	τfast	Afast	τslow	Aslow	n
α1A-PAA/β1A	Control	0.34 ± 0.04	0.54 ± 0.03	2.8 ± 0.4	0.33 ± 0.03	7
	10 μm	0.34 ± 0.06	0.21 ± 0.02	5.0 ± 0.3	0.57 ± 0.01	5
	100 ± M	0.34*	0.11 ± 0.02	7.3 ± 0.6	0.59 ± 0.01	6
α1A-PAA/R-E/β1A	Control	0.20 ± 0.04	0.43 ± 0.04	1.9 ± 0.3	0.39 ± 0.04	7
	10 μm	0.23 ± 0.04	0.24 ± 0.02	3.3 ± 0.2	0.61 ± 0.02	4
	100 μm	0.2*	0.10 ± 0.01	4.4 ± 0.2	0.75 ± 0.01	6
α1A-PAA/β2A	Control	0.23 ± 0.05	0.073 ± 0.007	3.0 ± 0.2	0.221 ± 0.007	4
	10 μm	0.4 ± 0.2	0.09 ± 0.02	5.1 ± 0.4	0.43 ± 0.02	4
	100 μm	0.23*	0.08 ± 0.01	6.3 ± 0.4	0.54 ± 0.01	4
α1A-PAA/R-E/β2A	Control	0.24 ± 0.07	0.09 ± 0.01	2.8 ± 0.4	0.15 ± 0.01	5
	10 μm	0.28 ± 0.09	0.08 ± 0.01	3.5 ± 0.2	0.32 ± 0.01	5
	100 μm	0.24*	0.071 ± 0.008	4.9 ± 0.2	0.53 ± 0.01	5

Data were obtained by fitting the kinetics of I_Ba recovery to a biexponential function (Fig. 2, see Methods).

^*In the presence of 100 μm (−)gallopamil τ_fast was fixed to τ_fast,control.

As shown in Fig. 1A, coexpressing α_1A-PAA and α_1A-PAA/R-E with the β_2a-subunit almost completely diminished fast I_Ba inactivation. This finding was confirmed by the corresponding recovery experiments. During the 3 s conditioning pulse less than 10 % of α_1A-PAA/β_2a or α_1A-PAA/R-E/β_2a channels accumulated in fast inactivation compared with more than 50 % of α_1A-PAA/β_1a and α_1A-PAA/R-E/β_1a (Fig. 2). Subsequent application of 10 or 100 μm (−)gallopamil substantially accelerated the I_Ba decay of all channel constructs (Fig. 3) and attenuated the slow component in I_Ba recovery (Fig. 2). Furthermore, drug-induced acceleration of the current decay was more prominent in α_1A-PAA/β_2a and α_1A-PAA/R-E/β_2a channels. As previously observed for α_1A-PAA/R-E/β_1a, recovery of α_1A-PAA/R-E/β_2a channels from block was more rapid than recovery of α_1A-PAA/β_2a (Fig. 2 and Table 1).

Figure 3. (−)Gallopamil-induced acceleration of I_Ba decay.

Normalised I_Ba of α_1A-PAA/β_1a, α_1A-PAA/R-E/β_1a, α_1A-PAA/β_2a and α_1A-PAA/R-E/β_2a in control and in the presence of 10 and 100 μm (−)gallopamil are superimposed to illustrate channel block during 1 s (α_1A-PAA/β_1a, α_1A-PAA/R-E/β_1a, A and B) or 3 s (α_1A-PAA/β_2a, α_1A-PAA/R-E/β_2a, C and D) test pulses from -80 mV to 20 mV.

DISCUSSION

Photoaffinity labelling experiments, radioligand binding studies and alanine scanning mutagenesis of L-type transmembrane segments IIIS6 and IVS6 suggest that the drug binding pockets for PAA, BTZ and 1,4-dihydropyridines are located near the Ca²⁺ channel pore (see Striessnig et al. 1998 for review). However, substitutions of inactivation determinants that have been identified in close proximity to the putative drug binding determinants on pore forming S6 segments significantly modulate sensitivity for PAA (see Hering et al. 1997) and BTZ (Berjukow et al. 1999). The latter findings suggest that those amino acids form either part of the drug-binding site or, alternatively, affect PAA and BTZ sensitivity in an indirect manner (i.e. via conformational changes modulating drug access, drug trapping or the steric orientation of the receptor determinants in the pore region, see Hering et al. 1998 for review). We have, therefore, investigated if inactivation determinants that are localised outside the putative drug-binding region have similar modulatory effects on sensitivity for the phenylalkylamine (−)gallopamil.

Here we demonstrate that an amino acid substitution (Arg387Glu) in the intracellular loop between domains I and II of the PAA-sensitive class A Ca²⁺ channel mutant α_1A-PAA and coexpression of α_1A-PAA with the β_2a-subunit slow the rate of I_Ba inactivation and, simultaneously, reduce use-dependent Ca²⁺ channel block by (−)gallopamil (Fig. 1).

It is highly unlikely that Arg387 forms part of the PAA-binding site. Hence, arginine in position 387 of the α_1A-PAA-subunit was mutated to the corresponding glutamate of the L-type α_1C-a-subunit. Since L-type channels carry the high-affinity PAA-binding site, transfer of a L-type amino acid to α_1A-PAA would be expected to increase and not, as shown here, to decrease PAA sensitivity (Fig. 1C, D and E).

It appears more likely that mutation Arg387Glu and the association of the β_2a-subunit with a motif on the I-II loop known as the ‘alpha interaction domain’ (AID; Pragnell et al. 1994; see also Walker et al. 1998 for a second β-interaction motif located on the carboxyl terminus of α_1A) affect drug sensitivity indirectly by modulating channel inactivation. In other words, our data are consistent with the hypothesis that the observed effects on gallopamil sensitivity are secondary to the change in inactivation and recovery rates. It appears, therefore, likely that the different inactivation rates of naturally occurring Ca²⁺ channel splice variants (Zuhlke et al. 1998; Bourinet et al. 1999) affect their pharmacological properties.

Such a mechanism clearly differs from observations of Zamponi et al. (1996) indicating that the I-II loop of class A channels forms part of the piperidine receptor.

As shown in Fig. 2B, α_1A-PAA/R-E/β_1a channels display a faster recovery from inactivation than α_1A-PAA/β_1a (see also Table 1). α_1A-PAA/R-E/β_1a also recovered more rapidly from block by (−)gallopamil suggesting that mutation Arg387Glu affects sensitivity for (−)gallopamil by accelerating channel unblock at rest (Table 1).

The β_2a-subunit-induced changes in α_1A-PAA inactivation (Fig. 1) and the consequences for use-dependent channel block by (−)gallopamil were even more dramatic (Fig. 2). Coexpression of the β_2a-subunit almost completely diminished fast inactivation during a depolarising test pulse (Fig. 1A). However, the strong time-dependent block of α_1A-PAA/β_2a channels indicates that nearly complete lack of fast inactivation does not prevent Ca²⁺ channel block (Fig. 3C and D). This finding is in line with the comparable resting channel inhibition by 100 μm (−)gallopamil observed for α_1A-PAA/β_2a, α_1A-PAA/R-E/β_2a and α_1A-PAA/β_1a.

The time constants of α_1A-PAA/β_2a recovery from fast and slow inactivation did not significantly differ from α_1A-PAA/β_1a. Accordingly, drug bound α_1A-PAA/β_2a channels recovered at nearly the same rate as the ‘higher sensitive’α_1A-PAA/β_1a channels (Fig. 2 and Table 1), and the reduced use-dependent inhibition of α_1A-PAA/β_2a channels is, in line with the slower rate of I_Ba inhibition (Fig. 1E), caused by a reduced block development during membrane depolarisation.

Arg387Glu also diminished use-dependent channel block by (−)gallopamil (Fig. 1D). However, contrary to the effect of the β_2a-subunit, this amino acid substitution affected not only the rate of block development but also the rate of recovery from block at rest. ‘Lower PAA sensitivity’ of α_1A-PAA/R-E/β_1a (see Fig. 1) is, therefore, mainly caused by a faster channel unblock between individual pulses (Fig. 2A and B). This hypothesis fits nicely with Arg387Glu-induced changes in channel inactivation (Fig. 1B and F) that would reduce drug trapping in inactivation and facilitate recovery from block at rest (Fig. 2A and B).

Under control conditions, the effect of the Arg387Glu substitution on I_Ba recovery of α_1A-PAA/R-E/β_2a was less evident than in α_1A-PAA/R-E/β_1a channels. However, a crucial role of Arg387Glu for channel unblock was confirmed by significantly faster recovery of α_1A-PAA/R-E/β_2a (compared with α_1A-PAA/β_2a) in the presence of 10 and 100 μm (−)gallopamil (Fig. 2 and Table 1). In other words, reduced use-dependent block of α_1A-PAA/R-E/β_2a (compared with α_1A-PAA/β_2a, Fig. 1D) is also due to facilitated channel unblock between individual test pulses of a train (Table 1).

Taken together, use-dependent block was reduced either by slowing channel inactivation (caused by β_2a-interaction) or by speeding recovery (mutation Arg387Glu). The additive and kinetically different effects of the β_2a-subunit interaction and mutation Arg387Glu on channel block (Fig. 1D) and recovery (Fig. 2) indicate that the corresponding conformational changes in the α₁-subunit are distinct and independent. Interestingly, only α_1A-PAA/R-E/β_1a, but not α_1A-PAA/β_2a or α_1A-PAA/R-E/β_2a, displayed significantly different resting channel block compared with α_1A-PAA/β_1a channels.

It is tempting to speculate that reduced use-dependent block of α_1A-PAA/β_2a channels during a train (Fig. 1D) reflects β_2a-induced changes in the steric orientation of the putative PAA binding determinants on pore forming S6-segments during a membrane depolarisation whereas Arg387Glu-induced changes in inactivation promote accelerated channel unblock. A detailed characterisation of the functional consequences of an amino acid substitution (i.e. possible modulation of fast and/or slow inactivation gating) is, therefore, a key requirement for future mutational studies directed towards the identification of drug-binding domains.