ABSTRACT
Extracellular ATP appears to have a widespread role as a neurotransmitter or neuromodulator in mammals (Gordon, 1986; Burnstock, 1990), but little is known about any similar functions in invertebrates. During studies of the effects of cyclic nucleotides on electrically excitable salivary cells of the leech, we found that cyclic GMP produced a rapid (less than 1min) reduction of spike duration, suggesting an extracellular effect (Wuttke and Berry, 1991). We now show that micromolar concentrations of ATP (and higher doses of other nucleotides) also reduce spike duration, and that this is caused by depression of a specific voltage-dependent Ca2+ conductance. Selective modulation of Ca2+ current by external ATP has rarely been found, and the effect is also unusual because it changes the kinetics of inactivation rather than those of activation.
Experiments were performed on isolated anterior salivary glands of the Amazon leech Haementeria ghilianii (de Filippi). The glands were pinned to Sylgard in a Perspex experimental bath (volume 0.25ml) and immersed in a continuous flow of physiological saline containing (in mmoll−1): NaCl, 115; KCl, 4; CaCl2, 2; MgCl2, 1; glucose, 11; Tris maleate, 10 (pH7.4) at room temperature (18–22°C). The following substances (Sigma) were applied by perfusion at known concentration: ATP (adenosine 5′-triphosphate), AMP (adenosine 5′-monophosphate), cyclic AMP (cyclic 3′,5′-adenosine monophosphate), adenosine, GMP (guanosine 5′-monophosphate), cyclic GMP (cyclic 3′,5′-guanosine monophosphate) and guanosine. Ba2+ (substituting for Ca2+) or tetraethylammonium and 4-aminopyridine (replacing equimolar Na+) were used to block K+ channels, and Co2+ was used to block Ca2+ channels (Hille, 1984; see Marshall and Lent, 1984, and Wuttke and Berry, 1991, for effects on the salivary cells of Haementeria). Solutions flowed through the bath at a rate of about 10bathvolumesmin−1 and could be changed rapidly without altering recording conditions.
Individual gland cells were impaled with two bevelled KCl-filled microelectrodes (resistance 10–20 MD) mounted on high-speed steppers (Digitimer SCAT-02).
Membrane voltage and current were measured with a voltage-clamp amplifier (Axoclamp 2A, Axon Instruments). Signals were monitored on a storage oscilloscope (Tektronix 5111) and pen recorder (Brush 2200S) and stored on tape (Thorn EMI 3000 FM tape recorder). For further details, see Wuttke and Berry (1992). No corrections were made for the small, linear leakage currents. The gland cells are electrically isolated (Wuttke and Berry, 1988) so these experiments are not complicated by electrical coupling between cells.
The salivary gland cells have resting membrane potentials ranging from −40 to −70mV and produce overshooting Ca2+-dependent action potentials with a duration of 100–400ms. Sodium ions make no contribution to the action potential (Wuttke and Berry, 1988). Application of 10−4 moll−1 ATP produced little or no change in resting membrane properties: in five cells the membrane depolarized by 1.25±0.6mV (S.D.) and there was no measurable change in input resistance as measured by injection of constant-current hyperpolarizing pulses (Fig. 1). The duration of action potentials, however, was reversibly reduced by 20–30% (Fig. 2). Much larger effects were produced on artificially prolonged action potentials, such as those recorded in Ca2+-free saline (Fig. 3A; the bathing solution was nominally Ca2+-and Mg2+-free and contained 1mmol l−1 EGTA). In the absence of Ca2+, sodium ions pass through the Ca2+ channels (Wuttke and Berry, 1988) and the action potentials become Na+-dependent, reaching 0mV and sometimes lasting for tens of seconds.
The effects of ATP on the membrane currents that underlie action potential shortening were studied by voltage-clamping the cells (Fig. 3B). Stepping from −70 mV to −50mV elicited a two-phase inward current consisting of an initial fast, transient component (<500 ms) followed by a longer-lasting component that decayed to zero in less than 30 s. ATP (10−4 mol l−1) had little effect on the amplitude of the inward current but increased its rate of inactivation within 1min (Fig. 3B). The lowest dose of ATP tested (10−6 mol l−1) reduced action potential duration in Ca2+-free saline to 71±7% of control values ( N=5) within 1min. An example of its effect on membrane current is shown in Fig. 4.
Depolarising steps applied from more positive levels (approximately −40mV) elicited the longer-lasting component on its own, indicating the presence of at least two types of Ca2+ channels: low-voltage-activated (LVA) and high-voltage-activated (HVA). ATP (10−4 mol l−1) in saline containing Ca2+ increased the rate of inactivation of the HVA current (reducing the half-decay time to 32±12% of control, N=6) and reduced its amplitude (to 51±10%, N=6), but had little effect on its rate of activation (Fig. 5A). Fig. 5B shows that 10−4 mol l−1 ATP reduced the duration, but not the amplitude, of the dual (LVA+HVA) inward current in the presence of external Ca2+ (2mmol l−1, cf. Fig. 3B). The preparation was bathed in saline containing 50mmol l−1 tetraethylammonium and 10mmol l−1 4-aminopyridine in order to separate the Ca2+ currents from K+ currents; this also ensured that the effects of ATP on inward current were not caused by an increase in outward K+ current. Responses to ATP were unaffected by replacement of Ca2+ with 10mmol l−1 Ba2+, which also blocks K+ currents (N=4; data not shown). Furthermore, when inward current was blocked with 10mmol l−1 Co2+, ATP did not potentiate the voltage-dependent outward current (N=5) (Fig. 6; in this cell the current was slightly reduced by ATP).
Qualitatively similar results were obtained with AMP, cyclic AMP, GMP, cyclic GMP and guanosine (data not shown; see Wuttke and Berry, 1991, for the effects of cyclic GMP on spike duration in Ca2+-free saline). Relative potencies were assessed by measuring the effects on the half-amplitude duration of prolonged spikes (4–6.5 s) in saline containing 50 mmol l−1 tetraethylammonium and 10 mmol l−1 4-aminopyridine. Control action potentials of constant duration were elicited by a 0.4s depolarizing pulse applied at 2-min intervals. Nucleotides were passed into the bath at increasing concentrations and spike duration was measured after 10min of exposure to each dose. There was a 20min washout between applications, which allowed full recovery from any response. Each nucleotide was tested at all doses on at least four cells. The following order of potency was observed: ATP>AMP>GMP>cyclic GMP>cyclic AMP????????guanosine>adenosine. At 10−6 mol l−1, only ATP was effective, reducing spike duration by 18±3.5%. At 10−5 mol l−1, this reduction increased to 47±6 %; AMP and GMP caused reductions of 26±5.5% and 9±3%, respectively, and the other substances had no effect. Cyclic GMP produced a reduction of 31±3.5% at 10−4 mol l−1 but cyclic AMP, guanosine and adenosine remained ineffective. At 10−3 mol l−1, cyclic AMP and guanosine reduced spike duration by 35±4% and 51±11 % respectively; adenosine had no effect on four cells but reduced spike duration by 10–13.5% in three others.
Our results show that extracellular ATP reduces action potential duration in leech salivary gland cells by selectively decreasing the amplitude and increasing the rate of inactivation of a high-voltage-activated Ca2+ current. The low dose at which the effect occurs (<10−6 mol l−1) suggests the presence of a receptor that is similar to the vertebrate P2 purinoceptor in its relative insensitivity to adenosine. P2 receptors are not usually coupled to adenylate cyclase, and this enzyme does not appear to be activated by ATP in the salivary cells; for example, 8-bromo-cyclic AMP (presumably acting intracellularly) has the opposing effect of lengthening the normal action potential by blocking outward K+ currents (Wuttke and Berry, 1991), and we have recently found that 3-isobutyl−1-methylxanthine (a phosphodiesterase inhibitor) also prolongs the action potential (W. A. Wuttke and M. S. Berry, unpublished results). Extracellularly applied cyclic AMP shortens action potentials.
ATP appears to be equipotent in normal saline and in solutions lacking free divalent cations, indicating that both free ATP4-and divalent-cation-complexed ATP can cause the effects outlined above. This also indicates that ectokinase or ecto-ATPase activity is not involved in the mechanism of receptor activation because phosphate transfer reactions require divalent cations.
ATP is unlikely to act indirectly by a Ca2+-induced inactivation of Ca2+ current (e.g. Xiong et al. 1991) because its effects persist in the absence of external Ca2+. Release of Ca2+ from internal stores cannot be excluded, but an effect on membrane conductance might then be expected (e.g. activation of a Ca2+-dependent K+ conductance; Wuttke and Berry, 1991).
ATP may be a neurotransmitter in the rat salivary gland; it activates Ca2+-permeable channels, and the resultant increase in intracellular Ca2+ ultimately stimulates fluid secretion (Soltoff et al. 1990). However, the mechanism of secretion by the salivary gland of Haementeria is quite different (Wuttke et al. 1989), and we have no evidence of any physiological role for ATP. We cannot yet explain why these exocrine cells have the unusual feature of a variety of voltage-gated channels: K+ (delayed rectifier) (Wuttke and Berry, 1991), Na+/K+ (inward rectifier) (Wuttke and Berry, 1992) and Ca2+ (HVA), all of which are modulated by putative neurotransmitters.
ACKNOWLEDGEMENTS
This work was supported by SERC (grant no. GR/F/17087) and the SmithKline (1982) Foundation.