Effects of 2,3-butanedione monoxime (BDM) on calcium channels expressed in Xenopus oocytes

Abstract

1. We examine the actions of a chemical phosphatase, 2,3-butanedione monoxime (BDM), on endogenous and expressed Ca²⁺ channel currents in Xenopus oocytes. In previous studies on L-type Ca²⁺ channel currents in cardiomyocytes and dorsal root ganglia, the inhibitory effects of BDM were attenuated by activation of protein kinase A.

2. Ba²⁺ currents (I_Ba) through a human wild-type L-type Ca²⁺ channel complex (i.e. hα_1C, α₂-δ_a and hβ_1b) are inhibited by BDM with an IC₅₀ of 16 mm, with 10 mm producing a 36.1 ± 2.2 % inhibition. I_Ba through endogenous oocyte N-type Ca²⁺ channels, upregulated by exogenous α₂-δ_a and hβ_1b subunits, are inhibited to a similar degree by BDM.

3. To examine whether the action of BDM is dependent on PKA-dependent phosphorylation, a clone of hα_1C deficient in all five serine PKA consensus sites (hα_1C-SA5) was co-expressed with α₂-δ_a and the human cardiac hβ₃ subunit, which naturally lacks PKA consensus sites. This complex exhibited a sensitivity to BDM that was similar to the wild-type complex, with 10 mm BDM producing 31.6 ± 1.5 % inhibition.

4. As limited proteolysis upregulates Ca²⁺ channels in cardiomyocytes and renders them less sensitive to BDM, experiments were performed with a carboxyl terminus deletion mutant, hα_1C-Δ1633. I_Ba through this subunit showed a sensitivity to BDM that was similar to the wild-type complex, with 10 mm BDM producing 31.3 ± 1.4 % inhibition. However, co-expression with α₂-δ_a and hβ₃ subunits reduced potency, and is reflected by an increased IC₅₀ of 22.7 mm.

5. The actions of BDM were examined on a rat brain rbA-1 Ca²⁺ channel clone, α_1A, co-expressed with α₂-δ_b and β_1b subunit homologues from rat brain. BDM inhibited the current through this channel complex to a similar degree to that seen for cardiac wild-type channels, with 10 mm BDM causing a 33.1 ± 3.5 % inhibition.

6. The effects of BDM were compared at two holding potentials, -80 and −30 mV, using the hα_1C-Δ1633, α₂-δ_a and hβ₃ subunit combination. At −30 mV BDM is more potent with 10 mm BDM reducing I_Ba by 39.8 ± 2.7 %, compared with 20.8 ± 2.2 % at −80 mV.

7. The data suggest that BDM may not exert its inhibitory action by means of a chemical phosphatase effect, but by channel block. The similar potency observed between α_1C, α_1A and endogenous (N-type) channels may help point towards a possible site of action; differences with the carboxyl deletion mutant may help further to define a locus of interaction.

It is well established that stimulation of the cAMP-dependent protein kinase (PKA) cascade increases the activity of L-type calcium channels in cardiac myocytes (Trautwein & Hescheler, 1990). This is the central observation that underlies both inotropic and chronotropic actions of β-adrenoceptor stimulation; the increase in activity of PKA by cAMP leads to phosphorylation of specific proteins with the cascade cycling via the opposing action of endogenous phosphatases. However, it still remains unclear on which of many sites targeted by PKA, functional changes in Ca²⁺ channels depend (Hofmann, Biel & Flockerzi, 1994). The molecular structure of the cardiac L-type Ca²⁺ channel is minimally composed of four subunits, α_1C, α₂-δ and β: the α_1C subunit forms the ion conducting pore and has the voltage sensor and sites for drug binding, whilst the auxillary subunits exert synergistic effects on kinetics, functional expression and assembly. In addition, the multimeric complex contains many putative phosphorylation sites (Hofmann et al. 1994; Mikala, Mershon & Schwartz, 1996). That a multiplicity of sites is implicated in the functional regulation of native cardiac Ca²⁺ channels by phosphorylation is suggested by several observations, e.g. changes in modal behaviour following β-adrenoceptor stimulation (Ochi & Kawashima, 1990), and changes in channel gating following inhibition of endogenous phosphatases (Wiechen, Yue & Herzig, 1995).

Observations supporting this notion that several phosphorylation sites may be associated with different functional changes in Ca²⁺ channel gating arise from the use of oximes; these agents can dephosphorylate proteins (Aldridge & Reiner, 1972). Oximes exert potent and reversible negative inotropic actions on the heart (Mulieri, Hasenfuss, Ittleman, Blanchard & Alpert, 1989), but their actions are not yet clearly defined; they can also, for example, act directly on the myofibrillar apparatus and may prove useful as cardioprotective agents (Ebus & Stienen, 1996). In cardiomyocytes, oximes, in particular 2,3-butanedione monoxime (BDM), inhibit whole-cell I_Ca and increase the rate of inactivation (Coulombe, Lefevre, Deroubaix, Thuringer & Coraboeuf, 1990). The sensitivity of the current to BDM is reduced in the presence of β-adrenergic stimulation and more so when ATPγS is included in the pipette to promote thiophosphorylating conditions (Chapman, 1993). BDM-induced inhibition of neuronal murine dorsal root ganglia L-type Ca²⁺ channels is also antagonized by PKA activation (Huang & McArdle, 1992), as is the transient outward current in ventricular myocytes (Xiao & McArdle, 1995). A specific ‘phosphatase’ action of BDM in cardiomyocytes is implied, since charged oximes only exert their effect when applied via the patch pipette (Chapman, 1993); furthermore, when trypsin is applied intracellularly, a protocol which leads to an increase in I_Ca and a loss of β-adrenergic stimulation (Hescheler & Trautwein, 1988), the inhibition induced by BDM is largely lost (Chapman, 1995). In a single channel study on cardiomyocytes, Allen & Chapman (1995) proposed a phosphatase action by BDM which reveals that phosphorylation is involved in the basal level of Ca²⁺ channel activity, in β-adrenergic activation and also in components of the current inactivation, and that several sites may be implicated. This notion is consistent with the observations mentioned earlier using phosphatase inhibitors (i.e. Wiechen et al. 1995).

Whilst inhibition of Ca²⁺ channel and transient K⁺ current by BDM is clearly affected by the PKA-regulated state of the cell, there is also clear evidence that oximes can affect other ionic channels, including nicotinic ACh channels at the motor endplate (Tattersall, 1993). Furthermore, in smooth muscle cells (Lang & Paul, 1991) and sympathetic ganglia neurons (Zhu & Ikeda, 1993), cells in which PKA fails to regulate I_Ca, BDM still exerts an inhibitory effect. These observations thus support the contention that BDM may exert its effect on Ca²⁺ channels via channel block, rather than as a ‘chemical phosphatase’.

As the molecular structure for L-type Ca²⁺ channels has largely been resolved, we have an opportunity to explore these possibilities further; a number of questions can be raised and are pursued in the present study. If functional phosphorylation events occur on sites associated within the Ca²⁺ channel complex (i.e. α_1C, α₂-δ and β), what are the effects of BDM on (i) a wild-type channel, which possesses consensus PKA sites on both α_1C and β, (ii) a mutant channel complex lacking PKA sites, and (iii) a mutant channel with the carboxyl terminus partially deleted, which mimics the native channel following limited partial proteolysis? We also examined the effects of BDM on non-cardiac Ca²⁺ channels in oocytes formed by endogenous N-type (Lacerda, Perez-Reyes, Wei, Castellano & Brown, 1994) or exogenous neuronal α_1A (Stea et al. 1994) channels.

A preliminary report of some of these results has been communicated to the Oxford meeting of The Physiological Society (Allen, Mikala, Wu, Schwartz & Dolphin, 1995).

METHODS

Preparation of cRNA and cDNA

The preparation of capped cRNA transcripts was performed as described in Contreras, Cheysen, Dayane & Fiers (1982), with slight modifications. Briefly, cRNAs were synthesized from pBluescript-derived plasmids bearing the subunit cDNAs in T7 orientation. The plasmid containing hα_1C (hHtα₁; Schultz et al. 1993) was linearized by XbaI, hβ_1b (previously designated as β_a, Collin, Wang, Nargeot & Schwartz, 1993) by HindIII, hβ₃ (Klöckner, Mikala, Varadi, Varadi & Schwartz, 1995) by EcoRV and α₂-δ_a (Ellis et al. 1988) by XhoI. The five-point mutant for the PKA consensus sites on hα_1C has been described (Eisfeld, Mikala, Schwartz, Varadi & Klöckner, 1996). The hα_1C carboxyl terminus truncation mutant (hα_1C-Δ1633) was constructed by using a ‘long primer’ mutagenesis protocol, introducing a stop codon after Gly1633, followed by an XbaI restriction site. The forward primer harboured the Sse 8387 I (6143) site. The sequence-verified PCR product was then used to replace the Sse 8387 I (6143)/XbaI (polylinker) fragment of hHtα₁. The linearized cDNAs (5 μg each) were transcribed at 37°C in a volume of 50 ml containing (mm): 40 Tris (pH 7.2), 6 MgCl₂, 10 dithiothreitol, 4 spermidine, 0.5 each of ATP, CTP, UTP and 7-methylguanosine-triphospho-guanosine, 0.05 GTP, supplemented with 200 U T7 RNA polymerase (Epicentre Technologies, Madison, USA) and 100 U rRNasin (Promega, Madison, USA). After 90 min of incubation, the GTP concentration was raised to 0.5 mm and the incubation continued for an additional 60 min. The cRNA products were purified by phenol-chloroform extractions, recovered by ethanol precipitation, redissolved and reprecipitated. For microinjection, cRNA concentrations were 0.2–0.6 μg μl⁻¹ in sterile water.

The vertebrate expression plasmids pAGS-3 and pMT2 were also used for expression of Ca²⁺ channel clones in Xenopus oocytes. The deletion mutant cDNA for hα_1C-Δ1633 was transferred into pAGS-3, using HindIII/NotI sites. Construction of pAGS-SKα₂ and pAGS-hβ₃ has been described elsewhere (Klöckner et al. 1995). Rat brain rbA-1 Ca²⁺ channel, α_1A, was examined with co-expression of rat brain clones for α₂-δ_b and β_1b (Stea et al. 1994) using the pMT2 expression vector.

Experimental procedures

Adult Xenopus laevis were either first anaesthetized with 0.2 % tricaine or cervically dislocated, then decapitated and the ovarian lobes surgically removed. Lobes were treated for 40–60 min with 1 mg ml⁻¹ collagenase (Type 1A, Sigma) in Ca²⁺-free solution (82 mm NaCl, 2 mm KCl, 1 mm MgCl₂, 5 mm Hepes, pH 7.5). Stage V-VI oocytes were mechanically defolliculated and transferred to 1.8 mm Ca²⁺ incubation solution (containing (mm): 100 NaCl, 2 KCl, 1 MgCl₂, 1.8 CaCl₂, 2.5 sodium pyruvate, 5 Hepes (pH adjusted to 7.5 with NaOH), supplemented with 100 U ml⁻¹ penicillin and 100 μg ml⁻¹ streptomycin). Oocytes were normally injected with 20–40 nl of a 1 : 1 : 1 cRNA mixture; however, for cDNA injection, oocytes were briefly centrifuged (∼200 g for 10–20 min) to expose the nucleus and then injected intra-nuclearly with 10 nl of 1 : 1 : 1 cDNA mixture (1 μg μl⁻¹). Oocytes were maintained at 18°C for at least 2 days before current recordings, at which time they were transferred to a recording chamber and continuously superfused with solution at a flow rate of 1–2 ml min⁻¹.

Dihydropyridine-sensitive inward currents were measured using conventional two-electrode voltage clamp with either an Axoclamp-2A or GeneClamp 500 amplifier (Axon Instruments), data acquisition and analyses were performed with pCLAMP software. Electrodes were fabricated from borosilicate glass, filled with 1 m KCl and had resistances of 0.5–1.0 MΩ.

The bath solution was a Ca²⁺- and Cl⁻-free solution composed of (mm): 40 Ba(OH)₂, 50 N-methyl glucamine (NMG), 2 KOH, 5 Hepes (pH adjusted to 7.4 with methane sulphonic acid), and 0.4 mm niflumic acid (diluted from a 20 mm stock; pH adjusted to 7 with tetraethylammonium hydroxide) was also included to inhibit Ca²⁺-dependent Cl⁻ channels (osmolarity, 226 mosmol l⁻¹). No compensation was made for the direct addition of up to 20 mm BDM to the bathing solution. However, as 100 mm BDM caused a small increase in solution osmolarity (of ∼20 mosmol l⁻¹; see also Zhu & Ikeda, 1993), NMG was lowered to 47 and 45 mm for 30 and 50 mm BDM, respectively. In those experiments in which 4 mm Ba²⁺ was used as the charge carrier NMG was raised to 95 mm. All chemicals were purchased from either Sigma or BDH.

Ba²⁺ currents were routinely elicited by voltage pulses from a holding potential of −80 mV at 0.1 Hz, unless otherwise indicated. Data were filtered at 1 kHz before digitization at 10 kHz and storage on computer for later analysis. A P/4 protocol was used for leak subtraction. All experiments were performed at room temperature (∼22°C).

Curve-fitting procedures

The following curve fitting procedures were performed using Fig.P (BioSoft, Cambridge, UK):

Current-voltage (I-V) relationships

The I-V relationships, and normalized I-V relationships (to the peak of the I-V curve) were fitted using the following modified Boltzmann relationship (see Allen, 1996):

(1)

where I is the peak (or normalized) inward current at a given potential V (mV), V_a is the potential for maximum activation slope (mV), V_r is the reversal potential (mV), k_a is the activation slope factor (mV), and G_max is the macroscopic slope conductance (μS); all curves are drawn to these derived parameters.

Steady-state inactivation

Barium currents at 20 mV, following a 2.5 s prepulse to various potentials, were normalized to those elicited with a prepulse to −50 mV (I_{−50 mV}):

(2)

where V is the preconditioning pulse potential (mV), V_i is the half-inactivation potential (mV) and k_i is the slope factor.

Dose-response curves

These were fitted with the Hill equation:

(3)

where I_Ba is the percentage inhibition of I_Ba, [BDM] is the concentration of BDM (mm), IC₅₀ is the half-maximal blocking BDM concentration (mm) and P is the Hill coefficient.

Data analysis

Statistical analyses have been routinely performed using a parametric (ANOVA) test or Student's paired t test (Instat, GraphPad Software, San Diego, CA, USA). All data are presented as means ±s.e.m., symbols in the figures indicate means, and the vertical bars, where larger than the symbol, indicate 1 s.e.m.

RESULTS

BDM inhibits a human wild-type cardiac Ca²⁺ channel expressed in oocytes

The oxime 2,3-butanedione monoxime (BDM) reversibly inhibited I_Ba in Xenopus oocytes co-expressing hα_1C, α₂-δ_a and hβ_1b subunits. Figure 1 illustrates the general experimental approach used to determine the action of BDM on inward Ba²⁺ currents through expressed channels; sequential inward currents in response to depolarizing pulses to 20 mV were recorded and peak currents plotted as a function of time. In all oocytes, Ba²⁺ currents showed some degree of ‘run-down’; determinations for the action of BDM take this into account, by time-averaging control currents determined in the absence of BDM (see legend to Fig. 1). Furthermore, due to differences in absolute expression levels between oocytes, the results are presented as mean percentage inhibition. Figure 1A shows peak current measured at 20 mV as a function of time, and the effects of bath perfusion of 2, 10 and 20 mm BDM. This causes a clear and reversible decrease in inward currents. The gaps in the time course are periods when I-V curves were determined; these are blanked for clarity in all figures. In twelve to twenty-one oocytes, BDM at 2, 10 and 20 mm inhibited I_Ba by 9.5 ± 2.0, 36.1 ± 2.2 and 55.4 ± 1.9 %, respectively (P < 0.001 compared with control).

Figure 1. BDM inhibits a human wild-type cardiac Ca²⁺ channel expressed in oocytes.

Oocyte injected with cRNA encoding hα_1C, α₂-δ_a and hβ_1b subunits. A shows time course of peak inward I_Ba recordings from an oocyte elicited by depolarizations to 20 mV. Gaps in time course are periods of I-V determination. Dotted line indicates zero current level and dashed line time course of current run-down; used for time-averaging control currents. Boxes at top of figure denote periods during which oocyte was exposed to 2, 10 or 20 mm BDM. B, original currents obtained from voltage pulses to 20 mV for control (a), and 2 mm (b), 10 mm (c) and 20 mm (d) BDM from periods denoted by corresponding letters in A. C, sections i and ii compare the action of BDM on currents from two different oocytes: section i, nifedipine-sensitive currents; section ii, endogenous currents in an oocyte only injected with α₂-δ_a and hβ_1b subunits. Currents shown are for control and 2, 10 and 20 mm BDM. Dotted lines indicate zero current level.

The inhibition produced by BDM is associated with an increase in the rate of inactivation of I_Ba (Fig. 1B); the inactivation time constant at 20 mV, τ₁, decreases from 198 ms in control to 110 ms in the presence of 10 mm BDM, and 68 ms in the presence of 20 mm BDM. In twenty oocytes, 10 mm BDM decreased τ₁ from 218.1 ± 28.3 ms in control to 140.2 ± 12.2 ms (P < 0.001). This change is more evident following subtraction of the endogenous Ca²⁺ channel current, obtained after inhibition of the expressed current with 20 μm nifedipine (Fig. 1C); the inactivation time constant decreased from 302 ms in control to 178 ms in the presence of 10 mm BDM and 75 ms in the presence of 20 mm BDM. The inhibition by 10 mm BDM is not significantly changed if the inactivation time course is extrapolated back to the onset of the voltage pulse; the amplitude of τ₁ is reduced by 40.5 ± 2.1 % (n= 20) compared with 36.1 ± 2.2 % observed with the peak current. Steady-state inactivation curves (fitted using eqn (2)) were determined with a test pulse to 20 mV in five oocytes, normalized to the test pulse from a holding potential of −50 mV (Fig. 2A); 10 mm BDM caused a significant 4.6 mV shift to the left of V_i, from -14.2 ± 0.8 to -18.8 ± 0.5 mV (P < 0.01; k_i= 9.4 ± 0.7 and 8.2 ± 0.4, respectively).

Figure 2. Summary of BDM-induced inhibition of a human wild-type cardiac Ca²⁺ channel expressed in oocytes.

A, steady-state inactivation curves from 5 oocytes expressing the hα_1C, α₂-δ_a and hβ_1b channel complex: control (^) and in presence of 10 mm (▪) BDM. B, normalized I-V relationships from 11 oocytes, for control (^), and 2 mm (•), 10 mm (□) and 20 mm (▪) BDM. Mean currents at 20 mV were -0.98 ± 0.18 μA. C, BDM dose-response curve obtained at 20 mV pulse potential from n= 5–21 oocytes; line is fitted to eqn (3).

BDM also inhibited endogenous Ba²⁺ currents in a dose-dependent manner in oocytes injected only with α₂-δ_a and hβ_1b subunits; in four oocytes, BDM at 2, 10 and 20 mm inhibited I_Ba at 20 mV by 7.8 ± 0.8, 29.6 ± 1.8 and 53.9 ± 2.9 %, respectively (P < 0.001 compared with absence of BDM). Mean inward currents at 20 mV were -0.11 ± 0.01 μA, and 10 mm BDM decreased τ₁ from 164.8 ± 17.3 ms in control to 86.7 ± 2.3 ms (P < 0.05) (see Fig. 1C). Therefore, the sensitivity to BDM of endogenous currents through N-type channels, which are upregulated by exogenous α₂-δ and β subunits (Lacerda et al. 1994; Tareilus et al. 1997), is not statistically different from currents through the expressed human hα_1C, α₂-δ_a and hβ_1b subunit complex.

The inhibition of I_Ba produced by BDM also affected I-V relationships of expressed wild-type Ca²⁺ channels. Whilst normalized data from eleven oocytes, each exposed to 2, 10 and 20 mm BDM, are shown in Fig. 2B, if one compares control with 10 mm BDM from the parameters obtained from fitting individual I-V relationships using eqn (1), the following is revealed (Table 1); V_a and k_a significantly increase whilst G_max decreases (P < 0.005). Mean inward currents at 20 mV were -0.98 ± 0.18 μA (n= 11, from four batches of oocytes for which BDM was examined at these three concentrations). The inhibition by BDM of I_Ba was curve fitted by using the Hill equation and yielded a half-maximal inhibitory concentration (IC₅₀) of 16 ± 1.1 mm with Hill coefficient of 1.2 ± 0.1 (see Fig. 2C).

Table 1.

Comparison of BDM effects on parameters obtained from I-V relationships

	Va	ka	Gmax
hα1C+α2-δ+ hβ1b (n= 11)
Control	3.9 ± 1.4	6.7 ± 0.2	23.0 ± 4.0
10 mm BDM	7.1 ± 1.3 **	7.7 ± 0.2 **	15.8 ± 2.4 *
hα1C-SA5+α2-δ+ hβ3 (n= 7)
Control	3.5 ± 1.5	7.3 ± 0.4	16.3 ± 4.1
10 mm BDM	3.2 ± 1.2	7.7 ± 0.5	10.9 ± 3.0 *
α1A+α2-δ+β1b (n= 5)
Control	−4.4 ± 2.5	6.2 ± 0.3	9.2 ± 3.1
10 mm BDM	−2.8 ± 2.3 *	7.2 ± 0.4 **	6.3 ± 2.3 *
hα1C-Δ1633 (n= 4)
Control	18.6 ± 1.7	7.4 ± 0.7	19.2 ± 5.6
10 mm BDM	19.3 ± 1.5	8.5 ± 1.0	14.9 ± 4.6 *
hα1C-Δ1633+α2-δ+ hβ3 (n= 4)
Control	8.5 ± 2.3	5.3 ± 0.3	22.6 ± 6.5
10 mm BDM, −80 mV (4 mm Ba2+) †	8.8 ± 2.6	5.6 ± 0.3	19.0 ± 5.6 *
hα1C-Δ1633+α2-δ+ hβ3 (n= 4)
Control	−20.2 ± 1.1	4.5 ± 0.6	31.7 ± 3.2
10 mm BDM, −30 mV (4 mm Ba2+) †	−18.9 ± 0.8	4.8 ± 0.6 *	26.1 ± 2.8
hα1C-Δ1633+α2-δ+ hβ3 (n= 4)
Control	−13.7 ± 0.9	4.6 ± 0.1	19.3 ± 3.9
10 mm BDM	−12.7 ± 0.6	4.7 ± 0.3	13.0 ± 3.3 *

Parameters are obtained from fitting of I-V relationships using eqn (1). V_a= potential for maximum activation slope (mV); k_a= activation slope factor (mV); and G_max= macroscopic slope conductance (μS). Subunit combinations are as given in text

^†Data obtained from varying holding potentials in 4 mm Ba²⁺.

^*P < 0.05 or

^**P < 0.005 significant difference from control value in the absence of 10 mm BDM.

Does an L-type Ca²⁺ channel complex lacking consensus PKA sites alter inhibition by BDM?

The data presented in the previous section illustrate that expressed wild-type Ca²⁺ channel currents (through the hα_1C, α₂-δ_a and hβ_1b subunit complex) are inhibited by BDM, but what is the mechanism underlying this inhibition? A potential phosphatase action may be directed at α_1C and β_1b subunits. The functional state of expressed channels remains unclear; it has been proposed that expressed Ca²⁺ channels in Xenopus oocytes are already functionally upregulated due to a high level of basal phosphorylation, such that PKA-mediated changes are only observed following prior inhibition of PKA (Singer-Lahat, Lotan, Biel, Flockerzi, Hofmann & Dascal, 1994; Perets, Blumenstein, Shistik, Lotan & Dascal, 1996). This possibility is given support by our observation that the diterpene activator of adenylyl cyclase, forskolin, at 10 μm does not affect expressed Ca²⁺ channel currents (n > 10 oocytes, data not shown). However, in four oocytes, exposure to 10 mm BDM in the presence of forskolin results in 24.5 ± 1.0 % inhibition of I_Ba. This is significantly different (P > 0.05) from inhibition of I_Ba effected by BDM in the absence of forskolin (i.e. 36.1 %), which suggests either that raising cAMP may attenuate the effects of BDM, or that forskolin itself may do so.

We can examine more directly whether BDM-induced inhibition is consistent with a phosphatase action on the channel complex, by expressing a channel complex lacking consensus PKA sites; the hα_1C subunit possesses five consensus PKA sites (Schultz et al. 1993) and the hβ_1b subunit has two more (Collin et al. 1993). Whilst Perets et al. (1996) and Gao et al. (1997) have suggested that functional modulation by PKA may occur on a site on the carboxyl terminus of rabbit α_1C (Ser1928), it has also been suggested that functional modulation by PKA may reside on the β subunit (Klöckner, Itagaki, Bodi & Schwartz, 1992; Haase, Bartel, Karczewski, Morano & Krause, 1996). Our experimental approach to circumvent these possibilities was to co-express an α_1C mutant that lacks all consensus PKA sites, i.e. hα_1-SA5 (Eisfeld et al. 1996), with α₂-δ_a and a β subunit also lacking consensus PKA sites, i.e. hβ₃ (Klöckner et al. 1995).

Figure 3 shows examples of I_Ba recordings elicited by depolarizations to 20 mV from an oocyte co-expressing the hα_1-SA5, α₂-δ_a and hβ₃ subunit combination, and the inhibitory effects of 2, 10 and 20 mm BDM. In six to thirteen oocytes, BDM at 2, 10 and 20 mm inhibited I_Ba measured at 20 mV by 9.4 ± 0.9 % (n= 6), 31.6 ± 1.5 % (n= 13) and 58.7 ± 1.9 % (n= 11), respectively (P < 0.001 compared with control). As observed with oocytes co-expressing hα_1C, α₂-δ_a and hβ_1b subunits, inhibition is associated with an increase in the inactivation of I_Ba (Fig. 3A); τ₁ decreased from 182 ms in control to 137 ms in the presence of 10 mm BDM, and 94 ms in the presence of 20 mm BDM. In twelve oocytes, 10 mm BDM decreased τ₁ from 251.5 ± 26.1 ms in control to 151.8 ± 11.5 ms (P < 0.001). Normalized data from I-V curves collected from seven oocytes are summarized in Fig. 3B; mean control currents at 20 mV were -0.67 ± 0.14 μA. From the parameters obtained from fitting individual I-V relationships using eqn (1), comparing control data with those for 10 mm BDM, only G_max changes significantly (Table 1).

Figure 3. Removal of consensus PKA sites on hα_1C has little effect on inhibition by BDM.

A illustrates original currents for control and 2, 10 and 20 mm BDM from an oocyte injected with cRNA encoding hα_1C-SA5, α₂-δ_a and hβ₃ subunits. This combination therefore lacks consensus PKA sites on both α₁ and β subunits. B shows normalized data from I-V curves collected from 7 oocytes for control (^), and 2 mm (•), 10 mm (□) and 20 mm (▪) BDM; mean currents at 20 mV were -0.67 ± 0.14 μA. The means of parameters obtained from the fit of individual I-V curves are given in Table 1.

The inhibition produced by 10 mm BDM is not statistically different between the three channel complexes thus far presented (36.1 ± 2.2 % with hα_1C, α₂-δ_a and hβ_1b compared with 31.6 ± 1.5 % with hα_1-SA5, α₂-δ_a and hβ₃, and 29.6 ± 1.8 % for the endogenous channel). The same follows when BDM-induced inhibition was examined in three oocytes co-expressing hα_1-SA5 with the hβ_1b subunit rather than hβ₃; I_Ba was inhibited by 38.4 ± 1.3 % with 10 mm BDM. Thus, cloned Ca²⁺ channels lacking consensus PKA sites in Xenopus oocytes not only retain sensitivity to BDM-induced inhibition, but its potency is largely unchanged compared with a wild-type channel complex.

Effects of hα_1C carboxyl terminus deletion and hβ subunits on BDM-induced inhibition

As mentioned in the previous section, functional up-regulating effects of PKA-dependent phosphorylation may be directed at site(s) located at the carboxyl terminus of rabbit heart α_1C, as indicated in recent studies (for example, Perets et al. 1996; Gao et al. 1997). This has also been suggested previously in studies on neuronal L-type α_1C subunit from rat brain, where a carboxyl terminal truncated isoform proved not to be a substrate for PKA (Yoshida, Takahashi, Nishimure, Takeshima & Kokubun, 1992). These latter observations prove attractive as they support the earlier findings that Ca²⁺ current in cardiac myocytes is increased following limited proteolysis; this upregulation is not additive with subsequent β-adrenergic stimulation (Hescheler & Trautwein, 1988), and currents through channels formed by expressed hα_1C are increased by intracellular trypsin or carboxypeptidase exposure (Klöckner et al. 1995). As currents through carboxyl terminus mutants of α_1C bear certain similarities (e.g. an increase in open probability) to those of upregulated, i.e. PKA-phosphorylated, channels in cardiac myocytes (see Klöckner et al. 1995), it was of interest to examine the action of BDM in Xenopus oocytes using a mutant Ca²⁺ channel, hα_1C-Δ1633, which lacks two distal consensus PKA sites at the truncated carboxyl terminus (amino acids S1940 and S1641). To determine whether β subunits can influence BDM-induced inhibition, we examined hα_1C-Δ1633 alone and following co-expression with hβ₃ subunit.

Oocytes co-expressing the hα_1C-Δ1633 subunit with or without the auxiliary α₂-δ_a and hβ₃ subunits exhibited robust Ba²⁺ currents (see also Klöckner et al. 1995). Firstly, oocytes were injected only with cRNA encoding hα_1C-Δ1633; Fig. 4A illustrates currents from one of these oocytes. Due to the absence of exogenous β subunit the test pulse potential was changed to 30 mV due to the associated shift in peak I-V. In three to four oocytes, BDM at 2, 10 and 20 mm inhibited I_Ba at 30 mV by 2.0 ± 1.5, 31.3 ± 1.4 and 48.9 ± 1.8 %, respectively. The inhibition produced by 10 mm BDM is not statistically different from that seen in wild-type channels co-expressing auxiliary subunits. Figure 4B illustrates normalized I-V curves collected from three oocytes; the mean current at 30 mV was -0.60 ± 0.27 μA. From the parameters obtained from fitting individual I-V relationships using eqn (1), and comparing control data with those for 10 mm BDM, V_a does not significantly change, but k_a increases and G_max decreases (Table 1).

Figure 4. Effects of hα_1C carboxyl terminus deletion and hβ₃ subunit co-expression on BDM-induced inhibition.

A, inhibition of I_Ba in an oocyte injected with cRNA to express hα_1C-Δ1633 subunit, by 2, 10 and 20 mm BDM; pulses were elicited each 15 s to +30 mV. B, normalized I-V curves from 3 oocytes expressing hα_1C-Δ1633 subunit: control (^), and 2 mm (•), 10 mm (□) and 20 mm (▪) BDM. Note shift in peak of I-V to 30 mV. C, inhibition of I_Ba at 20 mV in an oocyte expressing hα_1C-Δ1633, α₂-δ_a and hβ₃ subunits (by 2, 10 and 20 mm BDM). D, normalized I-V curves from 4 oocytes expressing hα_1C-Δ1633, α₂-δ_a and hβ₃ subunits: control (^), and 2 mm (•), 10 mm (□) and 20 mm (▪) BDM.

To examine the possibility that the actions of BDM may be influenced by co-expression of auxiliary subunits, oocytes were injected to express hα_1C-Δ1633, α₂-δ_a and hβ₃ subunits. In five to ten oocytes, BDM at 2, 10 and 20 mm inhibited I_Ba by 3.1 ± 2.8 % (n= 5), 20.6 ± 1.7 % (n= 10) and 39.7 ± 4.0 % (n= 5), respectively. These data are pooled from oocytes injected with either cDNA or cRNA encoding these subunits: with 10 mm BDM there was no significant difference in the inhibition of I_Ba between these different means of channel expression (cDNA, 21.2 ± 3.3 %; cRNA, 20.1 ± 1.5 %; n= 5).

BDM also enhanced inactivation of the inward Ba²⁺ current. In Fig. 4C, τ₁, the time constant of inactivation, decreases from 139 ms in control to 69.7 and 42.3 ms with 10 and 20 mm BDM, respectively. In nine oocytes, 10 mm BDM decreased τ₁ from 249.6 ± 39.78 to 161.2 ± 29.5 ms (P < 0.001). Figure 4D compares the normalized I-V curves from four oocytes exposed to 2, 10 and 20 mm BDM. At 20 mV the peak I_Ba was -1.03 ± 0.27 μA. From the parameters obtained from fitting individual I-V relationships using eqn (1), and comparing control data with those for 10 mm BDM, V_a and k_a do not change significantly, but G_max does (Table 1).

With 10 mm BDM there is a significant difference between inhibition of the hα_1C-Δ1633, α₂-δ_a and hβ₃ and wild-type (i.e. hα_1C, α₂-δ_a and hβ_1b) channel complexes (P < 0.01), with the deletion mutant exhibiting less sensitivity. The IC₅₀ for BDM was increased to 22.7 ± 1.12 mm and the Hill coefficient to 1.83 ± 0.19 (Fig. 5). The difference in the IC₅₀ and Hill coefficient values obtained with BDM for these two channel complexes is illustrated by including in Fig. 5 the dose-response curve previously shown in Fig. 2 for the wild-type channel. This difference persists if the curves are constructed from peak I_Ba obtained at 50 mV from I-V curves; the IC₅₀ and Hill coefficient values become, respectively, 14.3 mm and 1.1 for the wild-type and 21.0 mm and 1.7 for the hα_1C-Δ1633,α₂-δ_a and hβ₃ complex.

Figure 5. Comparison of BDM-induced inhibition of carboxyl terminus deletion mutant and wild-type Ca²⁺ channel complexes.

▪, mean BDM dose-response curve for the hα_1C-Δ1633, α₂-δ_a and hβ₃ Ca²⁺ channel complex (n= 4–10 oocytes). •, curve taken from Fig. 2 for the hα_1C, α_2a-δ_a and hβ_1b channel complex. □, data at 10 mm BDM from oocytes expressing hα_1C-SA5, α₂-δ_a and hβ₃ subunits. Lines are fits to eqn (3).

Does BDM affect Ca²⁺ channels formed by non L-type neuronal α_1A?

Whilst there is good evidence that BDM inhibits L-type Ca²⁺ channels in dorsal root ganglia (Huang & McArdle, 1992) and smooth muscle cells (Lang & Paul, 1991), no data to date have been presented on cloned channels other than representatives from the dihydropyridine-sensitive α_1C class. The α_1A Ca²⁺ channel exhibits properties quite distinct from L-type channels and possesses consensus PKA sites (Hofmann et al. 1994; Stea et al. 1994), and thus offers an opportunity to examine the action of BDM on an alternative expressed Ca²⁺ channel. Figure 6 illustrates data from an oocyte injected intra-nuclearly with cDNA encoding α_1A, and also rat brain α₂-δ_b and β_1b subunits. The rat brain α₂-δ_b and β_1b subunits are homologues of those used in earlier sections of this paper. Figure 6 shows the time course of I_Ba recordings elicited by depolarizations to 20 mV; as found with oocytes co-expressing cardiac Ca²⁺ channel subunits, BDM causes a reversible inhibition. In five oocytes, BDM at 2, 10 and 20 mm inhibited I_Ba when measured at 10 mV, at the peak of the I-V relationship, by 9.1 ± 2.3 % (n= 3), 38.4 ± 2.3 % (n= 5) and 59.7 ± 3.3 % (n= 4), respectively. These values are not significantly different from those obtained from wild-type cardiac Ca²⁺ channels.

Figure 6. Effects of BDM on currents through neuronal α_1A Ca²⁺ channels.

A shows time course of I_Ba recordings elicited by depolarizations to 20 mV. Boxes denote periods when the oocyte was exposed to 2, 10 or 20 mm BDM. B illustrates original currents elicited by pulses to 10 mV for control and 2, 10 and 20 mm BDM (decreasing trace order) from times indicated by the letters in A. Oocyte injected with cDNA to express α_1A, α₂-δ and β_1b subunits. C summarizes normalized I-V curves from 3 oocytes for control (^), and 2 mm (•), 10 mm (□) and 20 mm (▪) BDM. The mean current at 10 mV was -0.67 ± 0.18 μA.

Figure 6B illustrates original current traces for control and 2, 10 and 20 mm BDM recordings from the experiment illustrated in Fig. 6A. Inhibition at 10 mV was associated with an increase in the inactivation of I_Ba; τ₁ decreased from 127 ms in control to 81 ms in the presence of 10 mm BDM, and 55 ms in the presence of 20 mm BDM. In five oocytes, 10 mm BDM decreased τ₁ from 125.8 ± 5.2 ms in control to 95.2 ± 6.5 ms (P < 0.05). Figure 6C summarizes normalized I-V curves from three oocytes which were exposed to 2, 10 and 20 mm BDM, the mean current at 10 mV was -0.67 ± 0.18 μA. Consistent with previous reports in Xenopus oocytes (i.e. Stea et al. 1994), when α_1A is co-expressed with β_1b the I-V peak lies close to 10 mV. From the parameters obtained in five oocytes by fitting individual I-V relationships using eqn (1), comparing control data with those for 10 mm BDM, V_a and k_a increase, and G_max decreases (Table 1).

Does holding potential affect BDM-induced inhibition of Ca²⁺ channels?

To determine further the action of BDM on cloned α_1C channels, we investigated the effects of altering holding potential by utilizing the hα_1C-Δ1633, α₂-δ_a and hβ₃ subunit combination, and comparing inhibition at normal and depolarized holding potentials; i.e. -80 and −30 mV (Fig. 7). This combination of subunits produced large robust currents, and thus allowed examination using 4 mm Ba²⁺ as charge carrier; however, as a result of this reduction in concentration a leftward shift in the peak of the I-V relationship occurred (see Fig. 7B and C). Figure 7A shows the time course of I_Ba from an oocyte previously injected with cRNAs encoding hα_1C-Δ1633, α₂-δ_a and hβ₃ subunits; pulses were to 0 mV. The oocyte was first held at −30 mV to examine the action of BDM; the holding potential was then changed to −80 mV and the applications repeated. As shown in preceding sections, BDM evoked a reversible inhibition of inward Ba²⁺ currents. At a holding potential of −30 mV, data from eight oocytes revealed that BDM at 2, 10 and 20 mm inhibited I_Ba by 9.9 ± 2.0 % (n= 6), 39.8 ± 2.7 % (n= 8) and 66.6 ± 1.6 % (n= 4), respectively, when measured at −10 mV, which is close to the peak of the I-V relationship. When examined at a holding potential of −80 mV, BDM at 2, 10 and 20 mm inhibited I_Ba by 2.9 ± 1.5 % (n= 3), 20.8 ± 2.2 % (n= 4) and 45.2 ± 1.6 % (n= 3), respectively. These data reveal that BDM is significantly more effective on channels when a depolarized holding potential of −30 mV is used (P < 0.005 at both 10 and 20 mm BDM). The data also show that lowering the Ba²⁺ concentration from 40 to 4 mm does not affect inhibition by BDM; 10 mm BDM gave rise to 20.8 ± 2.2 % inhibition of I_Ba (n= 4), whilst at 40 mm Ba²⁺ inhibition was 20.6 ± 1.7 % (n= 10; hα_1C-Δ1633, α₂-δ_a and hβ₃ complex).

Figure 7. Does holding potential affect BDM-induced inhibition?

A shows time course of peak I_Ba, carried by 4 mm Ba²⁺, for an oocyte injected with cRNA to express hα_1C-Δ1633, α₂-δ_a and hβ₃ subunits. Initially holding potential was −30 mV, then this was changed to −80 mV. Pulse potential, 0 mV. B and C illustrate normalized I-V curves from 3 oocytes: control (^), and 2 mm (•), 10 mm (□) 20 mm (▪) BDM. Note that due to the charge carrier, 4 mm Ba²⁺, there is a shift to the left in the I-V peak. For these oocytes, at a holding potential of −80 mV, the mean inward current at −10 mV was -1.61 ± 0.26 μA, and at −30 mV this decreased to -0.58 ± 0.18 μA.

Figure 7B and C illustrates normalized I-V curves from three oocytes which were exposed to all three concentrations of BDM at both holding potentials. For these oocytes, at a holding potential of −80 mV (Fig. 7C), the mean inward current at −10 mV was -1.49 ± 0.25 μA, whilst at −30 mV this decreased to -0.58 ± 0.18 μA (Fig. 7B). Comparing control data with those for 10 mm BDM in four oocytes held first at −30 mV then at −80 mV, from the parameters obtained by fitting individual I-V relationships using eqn (1), at −30 mV G_max decreases (Table 1), whilst at −80 mV k_a increases and G_max decreases (Table 1).

DISCUSSION

This paper describes the effects of 2,3-butanedione monoxime (BDM) on Ca²⁺ channel currents expressed in Xenopus oocytes. BDM has previously been suggested to inhibit I_Ca via a chemical phosphatase activity, and we examine its action on a number of different Ca²⁺ channels: (i) endogenous Xenopus oocyte Ca²⁺ channels (i.e. N-type), upregulated by expression of exogenous α₂-δ_a and hβ_1b subunits, (ii) a wild-type human cardiac L-type Ca²⁺ channel complex (i.e. hα_1C, α₂-δ_a and hβ_1b), (iii) a mutant of hα_1C deficient in consensus sites for PKA and co-expressed with the α₂-δ_a subunit and a β subunit which naturally lacks PKA consensus sites (i.e. hα_1C-SA5, α₂-δ_a and hβ₃), (iv) a partial carboxyl terminus deletion mutant of hα_1C (i.e. hα_1C-Δ1633) expressed with or without exogenous α₂-δ_a and hβ₃ subunits, and (v) a rat brain rbA-1 Ca²⁺ channel clone co-expressed with auxiliary subunit homologues from rat brain (i.e. α_1A, α₂-δ_b and β_1b).

Effects of BDM on I_Ca

When data obtained with 10 mm BDM, a concentration which lies close to the IC₅₀ found for native basal L-type Ca²⁺ currents in cardiomyocytes (Chapman, 1993; Allen & Chapman, 1995), murine dorsal root ganglia (Huang & McArdle, 1992) and smooth muscle cells (Lang & Paul, 1991), were compared, the Ca²⁺ channels examined exhibited a marked similarity in sensitivity, with the notable exception of the hα_1C-Δ1633, α₂-δ_a and hβ₃ complex. In general, the inhibitory effects of BDM are largely unaltered whether or not consensus PKA sites are present on both human α_1C and β_1b subunits, the rat α_1A and β_1b subunits, or the endogenous oocyte N-type channel, and are not significantly affected when the PKA site-deficient mutant hα_1C-SA5 is co-expressed with hβ₃. Furthermore, expression of the partial carboxyl terminus deletion mutant hα_1C-Δ1633, which retains three of five consensus PKA sites, little changes BDM sensitivity. However, the co-expression of α₂-δ_a and hβ₃ with hα_1C-Δ1633 causes a significant, albeit small, reduction in sensitivity of the peak current to BDM, with an IC₅₀ shift from 16 to 22.7 mm.

These observations provide data which expand on the results of Eisfeld, Mikala, Varadi, Schwartz & Klöckner, (1997), who examined the action of BDM using a cultured mammalian cell expression system. In this latter study similar hα_1C mutants were used, but in each case these were examined following co-expression with the PKA site-deficient β subunit, hβ₃, and no significant change in the IC₅₀ for BDM was found. Our data add to this observation by showing that a wild-type human cardiac L-type Ca²⁺ channel complex (which comprises consensus PKA sites on both hα_1C and β_1b subunits) expressed in Xenopus oocytes, exhibits similar sensitivity (IC₅₀ of 16 mm) to hα_1C and hβ₃ subunits expressed in HEK 293 cells (IC₅₀ of 15.3 mm; Eisfeld et al. 1997). This observation is significant in two respects: firstly, it has been suggested that functional PKA-dependent regulation may depend or reside on the β subunit (Klöckner et al. 1992; Haase et al. 1996). If this were so, the lack of change in IC₅₀ observed by Eisfeld et al. (1997) might be accounted for by the use of the hβ₃ subunit, which lacks consensus PKA sites (Klöckner et al. 1995). Secondly, it has been suggested that the expressed α_1C channel in Xenopus oocytes is already upregulated by PKA (Singer-Lahat et al. 1994; Perets et al. 1996), such that only manipulations to inhibit PKA activity alter channel current. The expressed α_1C channel in cell lines differs in the important respect of being insensitive to manipulations that either increase or decrease PKA activity (Zong et al. 1996; Eisfeld et al. 1997).

A phosphatase action of BDM may therefore manifest itself quite differently, when data obtained from Xenopus and HEK 293 cells are compared; the IC₅₀ for BDM for the ‘upregulated’ Ca²⁺ channel in Xenopus oocytes could be quite different, since upregulated channels exhibit a reduced sensitivity to BDM (Huang & McArdle, 1992; Chapman, 1993). This is not the case, however, and will be addressed in detail in later sections. Furthermore, with the expressed Ca²⁺ channel in cell lines being insensitive to interventions designed to alter PKA activity (Zong et al. 1996; Eisfeld et al. 1997), one would not expect to see changes in the IC₅₀ for BDM between the wild-type hα_1C and the mutants hα_1C-SA5 and hα_1C-Δ1673 (Eisfeld et al. 1997); i.e. one would not expect a difference in sensitivity between the wild-type channel complex and a ‘phosphorylation deficient’ mutant.

Effects of partial deletion of the carboxyl terminus on the action of BDM

The objective for using a deletion mutant in the present work arises from the observation in cardiomyocytes that when trypsin is applied intracellularly (a protocol which leads to an increase in I_Ca and a loss of β-adrenergic stimulation; Hescheler & Trautwein, 1988), the inhibition by BDM is largely lost (Chapman, 1995). Partial removal of the carboxyl terminus of a cardiac form of α_1C increases expressed currents by increasing the channel open probability (Klöckner et al. 1995), and studies on a neuronal form of α_1C subunit from rat brain show that a carboxyl terminal truncated isoform proved not to be a substrate for PKA (Yoshida et al. 1992). These observations, taken together, lead to the suggestion that functional changes in activity effected by PKA and an inhibitory action of BDM may be linked to sites located on the α_1C carboxyl terminus section, especially as charged oximes only exert their effect on I_Ca in cardiomyocytes when applied via the patch pipette (Chapman, 1993).

The hα_1C-Δ1633, α₂-δ_a and hβ₃ complex when expressed in Xenopus oocytes gave rise to large currents which allowed examination of the action of BDM in media containing 4 mm Ba²⁺. This manoeuvre itself did not affect the potency of BDM, indicating that interaction is not likely to occur around conducting ion binding sites lining the channel pore. The α_1C-Δ1633 deletion mutant lacks approximately 75 % of the carboxyl terminus; whilst the wild-type channel has four consensus sites for PKA located on the carboxyl terminus (Schultz et al. 1993), the deletion mutant retains two. This channel mutant hα_1C-Δ1633 differs only slightly from the mutant hα_1C-Δ1673 used in the study of Eisfeld et al. (1997), which showed that when the hα_1C-Δ1673, α₂-δ_a and hβ₃ complex is directly compared with the hα_1C, α₂-δ_a and hβ₃ complex, the IC₅₀ for BDM was not significantly changed. In the present work, however, a further deletion of the carboxyl terminus clearly results in an increased IC₅₀ for BDM of 22.7 mm. This shift is notable when the α₂-δ_a and hβ₃ subunits are co-expressed: the hα_1C-Δ1633 subunit alone exhibits a sensitivity similar to the wild-type complex with 10 mm BDM. These data indicate a potential ‘protective’ role for the β and/or α₂-δ_a subunit in the interaction with BDM, and may give clues to possible relevant sites for the action of BDM. Whilst β subunits have an established interaction site sequence in the second intracellular loop on the α₁ subunit (Pragnell, De-Waard, Mori, Tanabe, Snutch & Campbell, 1994), there is also the possibility of an interaction site on the carboxyl terminus, since deletions in this region can affect the rate of current inactivation (Klöckner et al. 1995) and different β subunits can also affect this kinetic parameter (Hofmann et al. 1994; Mikala et al. 1996). BDM enhanced the time course of current inactivation and also shifted steady-state inactivation curves to negative potentials.

It should be noted, however, that Tareilus et al. (1997) have shown that exogenous α_1C or α_1E subunits will couple to endogenous oocyte auxiliary subunits; if this were indeed the case in the present study, then our data obtained using the α_1C-Δ1633 subunit may not represent the expression of this α₁ subunit moiety entirely alone. One implication from this possibility is that endogenous auxiliary subunits are not as effective as exogenous subunits in modifying BDM-induced inhibition of the α_1C-Δ1633 subunit.

Can the action of BDM still be regarded as ‘phosphatase-like’ on Ca²⁺ channels?

The data presented indicate that BDM-induced inhibition of expressed channels could be considered simply to reflect channel block, rather than a ‘phosphatase-like’ action; as Eisfeld et al. (1997) found, inhibition by BDM is independent of the presence of consensus PKA sites. A mechanism for channel block by BDM remains unclear, but an interaction at the carboxyl terminus of α_1C is implied. Experiments with further deletions in this region could help clarify this possibility; in cardiomyocytes, the inhibitory effects of BDM are lost following introduction of trypsin into the sarcoplasm (Chapman, 1995). Results from experiments using cloned channels in which the holding potential is altered suggest that block is potentiated at depolarized potentials, which indicates open channel block; this may be consistent with the enhanced time course of inactivation seen in the presence of BDM. However, this notion is not consistent with the observations of Ferreira, Artigas, Pizarro & Brum (1997), who examined gating currents from cardiomyocytes and determined that the action of BDM is not by open channel block but by promoting voltage-dependent inactivation. Indeed, this observation, coupled with the data presented in this paper, and the previously described role of the α_1C carboxyl terminus in voltage-dependent inactivation (see Hofmann et al. 1994; Klöckner et al. 1995; Mikala et al. 1996), does provide strong evidence for the α₁ carboxyl terminus as a potential site of action for BDM.

The effects of BDM on Ca²⁺ channels have been studied using cardiomyocytes (Coulombe et al. 1990; Chapman, 1993; Allen & Chapman, 1995), smooth muscle cells (Lang & Paul, 1991), sympathetic ganglion neurons (Zhu & Ikeda, 1993), sensory dorsal root ganglion neurons (Huang & McArdle, 1992), and HEK 293 cells expressing cardiac channel isoforms (Eisfeld et al. 1997). In each case BDM inhibits the currents examined; however, what must be addressed is why, in those systems where PKA regulation clearly affects Ca²⁺ channel activity, increasing PKA activity antagonizes BDM-induced inhibition, thus implying a phosphatase action for this drug.

PKA regulates Ca²⁺ channel activity in cardiomyocytes (Trautwein & Hescheler, 1990), and a thiophosphorylating environment causes an increase in IC₅₀ of BDM from 6 to 44 mm (Chapman, 1993). The Ca²⁺ channel current in sensory dorsal root ganglion neurons is also regulated by PKA and manoeuvres to increase its activity completely restore the BDM-inhibited current (Huang & McArdle, 1992); the IC₅₀ for BDM was approximately 20 mm, but it was not noted whether this parameter changed following activation of PKA. Whilst these observations may suggest that BDM acts as a phosphatase, results from smooth muscle cells, which are not sensitive to PKA and not affected by thiophosphorylating conditions (Hofmann et al. 1994), illustrate that BDM retains an inhibitory action (IC₅₀, 10 mm; Lang & Paul, 1991). The same follows for I_Ca in sympathetic ganglion neurons, which were not upregulated by PKA (IC₅₀, 18.3 mm; Zhu & Ikeda, 1993).

Taken together, the above findings and the results in the present work can be regarded as supporting the idea postulated by Allen & Chapman (1995) that BDM may exert two effects, a phosphatase action (with ‘high phosphorylated’ channels) and channel block (with ‘low phosphorylated’ channels). Ca²⁺ channels which are not upregulated by PKA exhibit channel block by BDM, while channels which are upregulated become more resistant to channel block by BDM and a phosphatase-like effect of the drug is observed. There is a possibility, therefore, that sites necessary for channel block by BDM become occluded or inaccessible when native channels are upregulated by PKA; in this scenario one does not need to propose an additional ‘dephosphorylating’ action for BDM. Accumulated evidence implicates the α₁ carboxyl terminus in this hypothesis, but other channel components, such as β subunits, may still play a central role.

Further implications

Functional changes in activity effected by PKA are not readily observed for expressed Ca²⁺ channels (Zong et al. 1996; Eisfeld et al. 1997), and if they are observed, primarily by use of PKA inhibitors, these changes are quite small (Yoshida et al. 1992; Singer-Lahat et al. 1994; Perez-Reyes, Yuan, Wei & Bers, 1994; Perets et al. 1996). The present work using BDM demonstrates that the expressed channel, even if affected by PKA, fails to function in the same way as native Ca²⁺ channels which are regulated by PKA. This may reflect inadequacies of the expression system, such as differences in post-translational modification (Yoshida et al. 1992; Perets et al. 1996), or, for example, that a component of the channel is missing, a possibility which is implied by size differences between the purified and the expressed Ca²⁺ channel (Hofmann et al. 1994; Mikala et al. 1996). A further possibility may be a requirement for intimate association of regulatory factors; PKA regulation of cardiac Ca²⁺ channels may require the presence of anchored cAMP-dependent protein kinase (Gao et al. 1997).

As the Ca²⁺ channel complex can be phosphorylated by a number of protein kinases (Trautwein & Hescheler, 1990; Hofmann et al. 1994; Mikala et al. 1996), it has been suggested that sites targeted by protein kinases other than PKA may be involved in the ‘phosphatase’ action of BDM (Allen & Chapman, 1995). However, a role for protein kinase C is unlikely; in sensory dorsal root ganglia neurons the action of BDM was not antagonized in the presence of PKC activators (Huang & McArdle, 1992). It is tempting to speculate that, were cloned Ca²⁺ channels to display behaviour normally associated with functional regulation by PKA, such as modal gating (Ochi & Kawashima, 1990; Wiechen et al. 1995), one might indeed see significant differences in the action of BDM.