Voltage-dependent calcium channels are widely distributed in excitable membranes and are involved in the regulation of many cellular functions. These channels can be modulated by neurotransmitters and drugs. There is one particular type of calcium channel in cardiac cells (L-type) whose gating is affected in different ways by β-adrenoceptor and 1,4-dihydropyridine agonists. We have analysed single calcium channel currents (i) in myocytes from rat hearts in the absence and presence of isoproterenol or 8-bromo-cAMP. We have found that both compounds have similar effects on calcium channel properties. They increase the overall open state probability (po) of individual calcium channels while i remains unaffected. Analysis of the gating kinetics of calcium channels showed: (a) a slight increase in the mean open times of calcium channels, (b) a reduction in time intervals between bursts of channel openings, (c) an increase in burst length and (d) a prominent reduction in failures of calcium channels to open upon depolarization. These kinetic changes caused by isoproterenol and 8-bromo-cAMP can account for the increase in po. Since the macroscopic calcium current, ICa, can be described by Ica = N×po×i, the increase in po accounts for the well-known increase in Ica, by β-adrenergic catecholamines. Cyclic AMP-dependent phosphorylation of calcium channels is a likely metabolic step involved in this modulation. Another class of drug that modulates calcium channel gating is the 1,4-dihydropyridines which can either enhance or reduce Ic,, either by prolonging the open state of the channels or by facilitating the inactivated state. Both effects depend strongly on membrane potential and are independent of cyclic AMP-dependent phosphorylation reactions.
Single-channel recordings by means of the patch-clamp technique provide a method for observing the kinetic properties of individual ion channels in cell membranes (Sakmann & Neher, 1983). Three distinct types of voltage-dependent Ca2+ channels have so far been characterized in muscle cells and neuronal tissues: T-type channels are responsible for a rapidly inactivating, transient Ca2+ current, while L-type channel openings produce a slowly inactivating, long-lasting Ca2+ current (Carbone & Lux, 1984; Nilius, Hess, Lansman & Tsien, 1985; Armstrong & Matteson, 1985; Bean, 1985; Fedulova, Kostyuk & Veselovsky, 1985). N-type channels have the kinetic features of neither T-nor L-type channels (Nowycky, Fox & Tsien, 1985). In addition to differences in their gating and conductance properties, the three types of channels can also be distinguished by their different sensitivities to pharmacological interventions (Bean, 1985; Nilius et al. 1985; Nowycky et al. 1985). The respective distributions of the various types of Ca2+ channels seem to vary from tissue to tissue. For example, more L-type Ca2+ channels may open during excitation in cardiac and smooth muscle cells than in nerve cells, where in turn the other channel types may prevail. In view of their functional significance, it will be of considerable interest to explore in greater detail the respective densities and sensitivities to neurotransmitters and drugs of various Ca2+ channels in different tissues. In this review we summarize briefly the basic properties of cardiac Ca2+ channels and the modulatory effects of catecholamines and 1,4-dihydropyridines on these channels.
PROPERTIES OF T- AND L-TYPE CALCIUM CHANNELS
The existence of transient and maintained Ca2+ currents has been shown in several cell types. However, direct comparisons of single-channel measurements with wholecell recordings have only recently been done in a few cases (Carbone & Lux, 1984; Nilius et al. 1985; Nowycky et al. 1985). Fig. 1A-D shows examples of such recordings, while in Fig. 1E the current/voltage characteristics of T- and L-type Ca2+ channels have been plotted. The data in Fig. 1C–E have been obtained from the same cardiac cell. With Ba ions as charge carriers the single-channel conductance of T-type Ca2+ channels is only approximately one-third that of L-type channels. T-type Ca2+ channels pass Ba2+ and Ca2+ ions about equally well, while L-type channels have a higher permeability for Ba2+ ions (Bean, 1985). Correspondingly, with Ca2+ ions as charge carriers, T-channels may have a higher conductance than L-channels (Carbone & Lux, 1984), while the opposite holds true with Ba2+ ions as charge carriers (Fig. 1E) (Nilius et al. 1985). Moreover, T-type channels can be activated by smaller depolarizations (to about −60mV) than L-type channels (to about 𢈒40mV) (Fig. 1E). This agrees with Ca2+ current measurements in wholecell recordings (e.g. Bean, 1985). T-type Ca2+ channels require a rather negative holding potential, VH, for their activation by depolarizing voltage steps. They are inactivated at VH positive to −40 mV. L-type channels inactivate much more slowly and over a more positive voltage range (Reuter, Stevens, Tsien & Yellen, 1982; Cavalié, Ochi, Pelzer & Trautwein, 1983). Openings of T-type Ca2+ channels occur primarily at the beginning of a voltage step (Fig. 1D) and thus produce the transient appearance of this current component. In contrast to L-type Ca2+ channels, T-type Ca2+ channel activity persists in excised membrane patches (Carbone & Lux, 1984; Nilius et al. 1985). Both channel types have rather different sensitivities to pharmacological agents. β-Adrenoceptor agonists and 1,4-dihydropyridines seem to have an effect primarily on L-type channels (Bean, 1985; Mitra & Morad, 1985; Nilius et al. 1985).
The L-type Ca2+ channel is the most common and the one that has been best explored so far. It seems to be present in many cell types including cardiac cells. With isotonic BaCl2 solution it has a conductance between 15 and 25 pS (see Fig. 1E). This type of voltage-dependent Ca2+ channel has distinct modes of gating (Reuter et al. 1982; Cavalié et al. 1983; Hess, Lansman & Tsien, 1984). Elementary currents flowing through a single Ca2+ channel which have been recorded from a cultured neonatal rat heart cell are shown in Fig. 2. The charge carriers were Ba2+ ions. Depolarizing step pulses of 70 or 80 mV amplitude were applied from a holding potential identical to the resting potential (approx. −70mV). The probability,p, of the channel being in the open state during each trace is indicated on the left-hand side of each record. With p = 1·0 the channel would be open throughout the voltage step, while with p = 0 it would remain shut. Ca2+ channel openings often occur in bursts separated by longer intervals of closures (mode 1; Hess et al. 1984). Quite rarely very long openings can also be observed (mode 2; Hess et al. 1984; see second and third current traces with ΔV = 70 mV in Fig. 2). Sometimes the channel does not open at all during repetitive depolarizations (mode 0; Hess et al. 1984; see current traces with p = 0·00 in Fig. 2). The probability of the channel entering an open state increases with increasing depolarization. Open times of Ca2+ channels during bursts (mode 1) are exponentially distributed with time constants (τo) that are only slightly voltage-dependent (Fig. 3). However, it should be noted that the rare very long openings (mode 2) last longer than would be expected from the exponential fits. Closed time histograms of Ca2+ channels can be fitted best by double exponentials (Fig. 3). The short time constant, τc1, of this double exponential curve corresponds to brief closures during the bursts of openings, while the longer time constant, τC2, describes the distribution of closed time intervals between the bursts. While τc1 does not seem to be very voltage-dependent, τC2 decreases greatly with increasing depolarizing step pulses. Failures of the channel to open (nulls; mode 0) are often grouped (Hess et al. 1984). Similar results have been obtained in other cell types (Fenwick, Marty & Neher, 1982; Hagiwara & Ohmori, 1983; Lux & Brown, 1984).
MODULATION OF L-TYPE CALCIUM CHANNELS
L-type Ca2+ channels in cardiac cell membranes were the first potential-dependent ion channels that were shown to be modulated in their activity by neurotransmitters (see Reuter, 1983). Modulation of Ca2+ channels by β-adrenoceptor agonists causes an increase in Ca2+ current in intact cardiac preparations (Reuter, 1974). It occurs through a cascade of events (Reuter, 1979, 1983) finally leading to a cyclic AMPdependent phosphorylation of one of the subunits of the channel protein (Curtis & Catterall, 1985). Fig. 4 shows membrane currents obtained from averaging many single-channel current traces, such as those illustrated in Fig. 2, in the absence and presence of isoproterenol (0-5μmoll−1), respectively. Isoproterenol approximately doubles the average currents. Assuming that the individual Ca2+ channels involved in this effect have identical properties and open and close independently, the average Ca2+ channel current, Ica, can be described as Ica− N×po×i, where N is the number of Ca2+ channels in the patch, po is the overall opening probability of these channels, and i is the current that flows through individual channels when they are open.
Several groups have analysed which of these factors leading to an increase in Ica are affected by catecholamines (Reuter, Cachelin, de Peyer & Kokubun, 1983; Bean, Nowycky & Tsien, 1984; Brum, Osterrieder & Trautwein, 1984) or 8-bromo-cAMP (Cachelin, de Peyer, Kokubun & Reuter, 1983; Reuter et al. 1983). Fig. 5 shows the principal results. From the current/voltage relationships, in the absence and presence of isoproterenol (Fig. 5A), it is clear that the unitary channel current, i, is not affected by the catecholamine. However, averaging the open state probability of many single-channel traces at each pulse potential shows a clear increase in po by the drug. This accounts for the increase in the average current in Fig. 4. The effect of β-adrenergic catecholamines on po is bigger if the opening probability of a channel is low in the absence of the drug. Very similar effects on single Ca2+ channels have been obtained with 8-bromo-cAMP (Cachelin et al. 1983). Opening and closing kinetics (gating) of Caz+ channels are affected by catecholamines in a characteristic way: (a) there is a slight increase in the mean open times (increase in τo), (b) time intervals between bursts of opening are reduced (decrease in τc2), (c) burst lengths are prolonged, and (d) the number of failures of the channels to open upon depolarization is considerably reduced. All these changes together account for the increase in po (Cachelin et al. 1983; Reuter et al. 1983; Brum et al. 1984). The main argument, that the opening probability rather than the number, N, of available Ca2+ channels is affected, comes from two observations. (1) Our group and Brum et al. (1984) have never been able to observe Ca2+ channel activity in the presence of the drug in patches where no channel openings were seen in the controls. This argues against the recruitment of previously silent channels in the patch. (2) If several channels were present in a patch, the superposition of opening levels of these channels was not larger than expected from an increase in po. However, as Tsien et al. (1986) pointed out, a distinction between channel availability and overall opening probability is to some extent a matter of definition of N and po. Failures of channels to open (mode 0) can be interpreted either as channels not being in a functional state, or as channels having a very low overall opening probability.
What is the mechanism by which po is increased by the catecholamines? In some elegant experiments Osterrieder et al. (1982) and Kameyama, Hofmann & Trautwein (1985) have injected the catalytic subunit of cyclic AMP-dependent protein kinase into single cardiac cells and found effects on ICa identical to those seen with isoproterenol application or cyclic AMP injection. This strongly supports the hypothesis that openings of Ca2+ channels are facilitated if the channels are phosphorylated.
Another group of drugs that modulates the voltage-dependent gating of L-type Ca2+ channels is the 1,4-dihydropyridines. Dihydropyridines are highly lipid-soluble compounds which seem to have access to individual Ca2+ channels through the lipid phase of the membrane (Kokubun & Reuter, 1984). Chemically closely related dihydropyridine derivatives can have opposite effects on Ca2+ channels. One group, which also has therapeutic applicability, reduces Ca2+ currents (antagonists), while the other group increases Ica (agonists). Even two enantiomers of the same compound can have opposite effects on Ca2+ currents (Williams et al. 1985; Reuter, Porzig, Kokubun & Prod’hom, 1985a). The blocking effects of dihydropyridine compounds such as nifedipine, nitrendipine or nimodipine result from a reduction of the opening probability of the channel by favouring mode 0 behaviour of gating (Hess et al. 1984). The agonist compounds (CGP28392, BAYK8644) increase Ca2+ currents by promoting long channel openings (mode 2; Fig. 6) (Kokubun & Reuter, 1984; Hess et al. 1984). Both effects depend strongly on membrane potential. The binding affinity for antagonists and agonists is greatly increased by membrane depolarization (Reuter et al. 1985a,b). Correspondingly, the blocking potency of antagonistic compounds is also increased (Bean, 1984; Reuter et al. 1985a), while the activator effect of the agonists turns into a blocking effect by increasing the failure rate of channels to open (mode 0) (Reuter et al. 1985a).
It is interesting to note that dihydropyridine agonists not only produce changes in Ca2+ channel kinetics which are distinctly different from those induced by catecholamines, but these changes are also independent of the cyclic AMP level in the cell (Kokubun & Reuter, 1984; Böhm et al. 1985). This shows that ‘openness’ of Ca2+ channels can be modulated by very different mechanisms and in different ways.