A novel role for membrane potential in the modulation of intracellular Ca²⁺ oscillations in rat megakaryocytes

Abstract

1. The effect of membrane potential (V_m) on ADP-evoked [Ca²⁺]_i oscillations was investigated in rat megakaryocytes, a non-excitable cell type recently shown to exhibit depolarisation-evoked Ca²⁺ release from intracellular stores during metabotropic purinoceptor stimulation.

2. Hyperpolarising voltage steps caused a transient fall in [Ca²⁺]_i and either abolished Ca²⁺ oscillations or reduced the oscillation amplitude. These effects were observed in both the presence and absence of extracellular Ca²⁺ and also in Na⁺-free saline solutions, suggesting that hyperpolarisation leads to a reduction in the level of ADP-dependent Ca²⁺ release without a requirement for altered transmembrane Ca²⁺ fluxes.

3. In the presence of Ca²⁺ oscillations, depolarising voltage steps transiently enhanced the amplitude of Ca²⁺ oscillations. Following run-down of Ca²⁺ oscillations, depolarisation briefly restimulated oscillations.

4. Simultaneous [Ca²⁺]_i and current-clamp recordings showed that Ca²⁺ and V_m oscillate in synchrony, with an average fluctuation of approximately 30–40 mV, due to activation and inactivation of Ca²⁺-dependent K⁺ channels. Application of a physiological oscillating V_m waveform to non-oscillating cells under voltage clamp stimulated [Ca²⁺]_i oscillations.

5. Analysis of the relationship between [Ca²⁺]_i and V_m showed a threshold for activation of hyperpolarisation at about 250–300 nM. The implications of this threshold in the interaction between V_m and Ca²⁺ release during oscillations are discussed.

6. We conclude that the ability of voltage to control release of endosomal Ca²⁺ in ADP-stimulated megakaryocytes is bipolar in nature. Our data suggest that V_m changes are active components of the feedback/feedforward mechanisms contributing to the generation of Ca²⁺ oscillations.

Many non-excitable cells undergo oscillations in cytosolic Ca²⁺ concentration ([Ca²⁺]_i) or repetitive transient increases in [Ca²⁺]_i following agonist stimulation of the inositol phosphate signalling cascade. Ca²⁺ oscillations and transitions have been postulated as a mechanism to overcome the long-term desensitisation that may occur with constant elevations of Ca²⁺, and also for frequency encoding of Ca²⁺ signalling. The latter proposal is supported by the observation that mitochondrial [Ca²⁺], and consequently, the activity of Ca²⁺-dependent mitochondrial dehydrogenases, are tuned to oscillating rather than sustained increases in cytosolic [Ca²⁺] (Hajnoczky et al. 1995). Additionally, gene expression mediated via Ca²⁺-dependent transcriptional activating factors has been shown to be encoded by the frequency of Ca²⁺ oscillation (Dolmetsch et al. 1998; Li et al. 1998). Transcriptional activators can also display different sensitivities to Ca²⁺ oscillation frequency, thus providing a mechanism for specificity within activation of gene expression by this ubiquitous second messenger (Dolmetsch et al. 1998).

Numerous models have been proposed to explain the mechanisms underlying Ca²⁺ oscillations in non-excitable cells (for reviews see Jacob, 1990; Berridge, 1992; Fewtrell, 1993; Putney & Bird, 1993; Thomas et al. 1996). Currently, the role of changes in membrane potential (V_m) in the initiation and/or modulation of Ca²⁺ oscillations is unclear. In a number of cell types Ca²⁺ oscillations can be observed under conditions devoid of changes in potential during simultaneous measurement of [Ca²⁺]_i under patch-clamp conditions (Osipchuk et al. 1990; Uneyama et al. 1993b; Thorn, 1995). Such experiments demonstrate that changes in potential are not required for the induction of sustained Ca²⁺ oscillations. On the other hand, oscillating changes in V_m induced by changes in command voltage under patch-clamp conditions have been shown to induce Ca²⁺ oscillations in Jurkat T-lymphocytes and mast cells during sustained activation of Ca²⁺ influx across the plasma membrane (Penner et al. 1988; Lewis & Cahalan, 1989). Such an effect can be ascribed to the influence of electrical driving force on Ca²⁺ influx, with depolarisation resulting in a marked fall in [Ca²⁺]_i and hyperpolarisation a dramatic increase. Thus, physiological changes in potential occurring during activation of Ca²⁺ influx may have important modulatory influences upon steady-state Ca²⁺ and oscillatory Ca²⁺ events. One example is the inhibition of Ca²⁺ oscillations in mitogenically stimulated T cells by K⁺ channel blockers, presumably due to depolarisation or lack of hyperpolarisation and thus reduced Ca²⁺ influx (Grissmer et al. 1992). A second example is the large transitional changes in potential induced by depletion of intracellular Ca²⁺ stores in rat basophilic leukaemia cells that lead to concomitant changes in [Ca²⁺]_i as a result of effects on the Ca²⁺ driving force (Mason et al. 1999).

In contrast to the influence of potential on driving force for Ca²⁺ entry in non-excitable cells, a novel effect of depolarisation has recently been reported in the rat megakaryocyte. In the presence of metabotropic purinoreceptor stimulation, depolarisation evokes a marked transient increase in [Ca²⁺]_i as a result of the release of Ca²⁺ from an intracellular compartment via a mechanism that requires functional IP₃ receptors (Mahaut-Smith et al. 1999). This conclusion is based upon four pieces of evidence. First, rat megakaryocytes lack voltage-gated Ca²⁺ channels, thus ruling out voltage modulation of endogenous Ca²⁺ channel activity (Uneyama et al. 1993a; Somasundaram & Mahaut-Smith, 1994, 1995; Hussain & Mahaut-Smith, 1998; Mahaut-Smith et al. 1999). Second, depolarisation in the presence of ADP stimulates an increase in [Ca²⁺]_i in the absence of extracellular Ca²⁺ (Mahaut-Smith et al. 1999). Third, depolarisation stimulates an increase in [Ca²⁺]_i in the absence of both extracellular and intracellular Na⁺, thereby discounting voltage-dependent modulation of Na⁺-Ca²⁺ exchange activity as the source of this Ca²⁺ increase (Mahaut-Smith et al. 1999). Fourth, voltage-dependent elevations in [Ca²⁺]_i are blocked by heparin (Mahaut-Smith et al. 1999), an IP₃ receptor inhibitor (Bezprozvanny & Ehrlich, 1995). Therefore, changes in V_m may have marked effects upon the temporal pattern of Ca²⁺ signalling in megakaryocytes. While V_m changes in the megakaryocyte are not required for the induction of Ca²⁺ oscillations by purinoceptor stimulation, in as much as oscillations are observed under voltage-clamp control (Uneyama et al. 1993b; Hussain & Mahaut-Smith, 1998), it is important to note that in the absence of voltage clamp, marked oscillations in V_m accompany purinoceptor stimulation (Somasundaram & Mahaut-Smith, 1995; Mahaut-Smith et al. 1999). Therefore, physiological voltage changes may have a significant impact upon the normal temporal pattern of Ca²⁺ oscillations via their ability to modulate release of Ca²⁺ from intracellular stores. In the present study, we have undertaken experiments to investigate the effects of V_m on purinoceptor-induced Ca²⁺ oscillations in rat megakaryocytes.

METHODS

Cell isolation

Male Wistar rats (150-300 g) were killed by exposure to a rising concentration of CO₂ gas followed by cervical dislocation and megakaryocytes were isolated from femoral and tibial marrow as previously described (Mahaut-Smith et al. 1999).

Solutions

The standard external saline contained (mM): 145 NaCl, 5 KCl, 1 CaCl₂, 1 MgCl₂, 10 Hepes, 10 D-glucose, adjusted to pH 7.35 with NaOH. For Ca²⁺-free saline, CaCl₂ was replaced by an equal concentration of MgCl₂ and 0.5 mM EGTA was included. The standard pipette saline contained (mM): 150 KCl, 2 MgCl₂, 0.1 EGTA, 0.05 Na₂GTP, 0.05 K₅fura-2, 10 Hepes, adjusted to pH 7.2 with KOH. In some experiments 0.05 mM K₅fura-2 (Molecular Probes, Eugene, OR, USA) was replaced with 0.05 mM (NH₄)₅fluo-3 (Calbiochem-Novabiochem (UK) Ltd, Nottingham or Molecular Probes). ADP was obtained from Sigma-Aldrich (Poole, Dorset, UK).

Electrical recording

Conventional whole-cell patch-clamp recordings were carried out using an Axopatch 200A amplifier with CV202 headstage (Axon Instruments, CA, USA) in either voltage-clamp or slow current-clamp mode. pCLAMP 6.0 (Axon Instruments) was used to generate voltage steps and oscillating waveform protocols. Patch pipettes were pulled from borosilicate glass tubing (Clark Electromedical Instruments, UK). Series resistance compensation of 70–75 % was achieved. Membrane potentials have not been corrected for the small offset (approximately -3 mV) that results from the liquid:liquid junction potential between internal and external saline solutions.

Fluorescence measurements

A Cairn spectrophotometer system (Cairn Research Ltd, Kent, UK) coupled to a Nikon Diaphot TMD inverted microscope (Nikon, Japan) was used to measure fura-2 or fluo-3 fluorescence during simultaneous whole-cell patch clamp. Light from a 75 W xenon arc lamp passed through a spinning filter wheel containing 10 nm bandpass interference filters to provide alternating 340 and 380 nm excitation for fura-2 experiments or 490 nm for single-wavelength fluo-3 experiments. A ×40 (1.3 NA) Nikon Fluor lens focused excitation light from the epifluorescence port of the microscope onto the cells within a saline-filled chamber. Fluorescence emission from a rectangular area slightly larger than the patch-clamped cell was collected by a photomultiplier tube and sampled at 60 Hz, together with the electrophysiological signals, using Cairn fluorescence software. The emission bandwidth was either 480–600 nm for fura-2 or 528–600 nm for fluo-3 and was achieved by first passing the emission light through a 400 nm dichroic mirror in the case of fura-2, or a 503 nm dichroic mirror in the case of fluo-3, to separate the emission from the excitation light. The light was then passed through a 600 nm dichroic mirror placed immediately in front of a 480 or 528 nm long pass emission filter. Light above 600 nm, achieved by transillumination of the cell with light > 600 nm, passed through the dichroic mirror to a CCD camera for continual visualisation of the cell during fluorescence recording (Mahaut-Smith, 1998). Data were further averaged to give a final acquisition rate of 15 Hz and exported for analysis using IGOR software (Wavemetrics, Lake Oswego, OR, USA). For presentation, smoothing was applied to some experiments using a binomial filter within IGOR. For calibration, fura-2 constants, R_min and R_max (the fluoresence ratios in the absence and presence of saturating Ca²⁺, respectively), were obtained extracellularly since it was difficult to clamp [Ca²⁺]_i at high levels in the megakaryocyte. A calibration kit (Molecular Probes) was used to derive a K_d for fura-2 (258 nM). After application of a viscosity correction factor (0.85) to R_min and R_max (Poenie, 1990), background corrected 340 nm/380 nm values were converted to [Ca²⁺]_i as described by Grynkiewicz et al. (1985). Single wavelength fluo-3 data are presented as raw, background corrected fluorescence, thus avoiding the problems of calibration under conditions of uncertain fluo-3 concentration. All recordings were made at an ambient temperature of 20–25°C.

Solution application

ADP was applied by gravity-driven bath superfusion. Solution changes, as noted in the figures, have been corrected for delays resulting from dead space within the superfusion system.

RESULTS

We have previously reported that rat megakaryocytes held at -75 mV under conventional whole-cell patch clamp using pseudophysiological salines initially respond to ADP with a large transient increase in [Ca²⁺]_i. In the sustained presence of ADP, the response that follows varies in different cells, from an immediate elevated plateau of [Ca²⁺]_i to a variable number of oscillations prior to settling at a plateau level. In Fig. 1 we show the effect of different holding potentials on [Ca²⁺]_i in a cell that displayed prolonged Ca²⁺ oscillations in response to 1 μM ADP. Depolarisation from the initial holding potential of -75 mV to 0 mV was accompanied by a large transient increase in [Ca²⁺]_i as previously reported (Mahaut-Smith et al. 1999). Repolarisation to -75 mV was accompanied by a transient decline in [Ca²⁺]_i and a second transient decline in [Ca²⁺]_i was induced by a further hyperpolarisation to -115 mV. The lower panel of Fig. 1 shows [Ca²⁺]_i changes accompanying these hyperpolarisations on expanded scales. The final potential change, a return to -75 mV from -115 mV, induced a transient increase in [Ca²⁺]_i demonstrating the existence of depolarisation-evoked increases in [Ca²⁺]_i over this more negative voltage range. Hyperpolarisation from 0 to -75 mV during exposure to 1 μM ADP stimulated a transient decline in [Ca²⁺]_i in 33 of 35 cells. The influence of hyperpolarisation was also observable over smaller hyperpolarising ranges. Hyperpolarisation from -40 or -45 mV to -75 or -80 mV induced a transient decline in [Ca²⁺]_i in 18 of 18 cells, while transient declines in [Ca²⁺]_i were detected in 6 of 6 cells following a voltage step from -75 to -115 mV. In the absence of ADP, membrane potential changes over these voltage ranges had no significant effect on [Ca²⁺]_i (data not shown). It is important to note that the effect of changes in V_m on [Ca²⁺]_i cannot be explained by modulation of voltage-gated Ca²⁺ channel activity as rat megakaryocytes are devoid of such channels (Uneyama et al. 1993a; Somasundaram & Mahaut-Smith, 1994, 1995; Hussain & Mahaut-Smith, 1998; Mahaut-Smith et al. 1999). These data extend our original observation of depolarisation-evoked increases in [Ca²⁺]_i (Mahaut-Smith et al. 1999) to include a transient inhibitory effect of hyperpolarisation on ADP-evoked [Ca²⁺]_i elevations. Thus, the modulation of [Ca²⁺]_i by V_m during metabotropic purinoceptor stimulation is bipolar, with depolarisation inducing an elevation in [Ca²⁺]_i and hyperpolarisation inducing a transient decline in [Ca²⁺]_i.

Figure 1. Effect of V_m changes on [Ca²⁺]_i in a rat megakaryocyte.

Upper panel, a megakaryocyte was whole-cell voltage clamped and fura-2 fluorescence simultaneously monitored following dialysis of fura-2 from the patch pipette. Where indicated 1 μM ADP was applied. Lower panel, for clarity a portion of the upper trace showing voltage steps from 0 to -75 mV and -75 to -115 mV has been expanded. In this and all subsequent figures major intracellular and extracellular ionic compositions are denoted in the patch-clamp icon and all scales are linear.

Two additional observations can be made from the data presented in Fig. 1. First, the depolarising step to 0 mV was accompanied by an apparent transient increase in the amplitude of the Ca²⁺ oscillations. Second, the transient decline in [Ca²⁺]_i induced by hyperpolarisation from 0 to -75 mV was accompanied by termination of Ca²⁺ oscillations. Figure 2 shows an experiment designed to further explore the role of V_m changes in the modulation of Ca²⁺ oscillations. In Fig. 2A, a megakaryocyte was whole-cell voltage clamped with a pipette solution containing fluo-3 as the Ca²⁺ indicator. The cell was exposed to 1 μM ADP at a holding potential of -75 mV with the activation of long lasting oscillations of [Ca²⁺]_i. A brief depolarising step to 0 mV from -75 mV was applied approximately 130 s after ADP application. The depolarising step induced two oscillations of larger amplitude, an effect that was clearly reversible, although initially accompanied by a dramatic fall in [Ca²⁺]_i and a temporary loss of clear oscillations as noted above. Depolarisation from -75 or -80 mV to 0 mV increased the amplitude of Ca²⁺ oscillations in 12 of 12 cells in which the depolarising step was of sufficient duration for adequate detection of multiple Ca²⁺ oscillations. The effect of repolarisation from 0 mV to -75 or -80 mV was complex, with hyperpolarisation either terminating Ca²⁺ oscillations or inhibiting the amplitude of the oscillations. Of 14 cells in which the hyperpolarising step was of sufficient duration to allow determination of its effect, 5 cells displayed a permanent (Fig. 1) or temporary loss of clear oscillations (Fig. 2A), while 8 showed a clear reduction in oscillation amplitude. Only 1 cell showed no modulation of Ca²⁺ oscillations by a voltage step of this magnitude.

Figure 2. Effect of V_m changes on [Ca²⁺]_i and Ca²⁺_i oscillations in a rat megakaryocyte.

A, a megakaryocyte was whole-cell voltage clamped and fluo-3 fluorescence simultaneously monitored following dialysis of fluo-3 from the patch pipette. The record starts approximately 130 s after commencing application of 1 μM ADP. B, measurement of fluo-3 fluorescence under whole-cell voltage clamp following cessation of Ca²⁺ oscillations in the same cell as shown in A (approximately 150 s after commencing ADP application).

An effect of depolarisation could also be observed in non-oscillating cells. Figure 2B shows the fluo-3 fluorescence following cessation of Ca²⁺ oscillations in the same cell as shown in Fig. 2A, 150 s after ADP application. In this cell a voltage step to 0 mV triggered oscillations in [Ca²⁺]_i above the plateau phase level. These oscillations declined in amplitude over a period of approximately 15 s (6-7 oscillations) to a sustained value virtually identical to that recorded at -75 mV. Repolarisation to -75 mV was accompanied by a marked transient decline in [Ca²⁺]_i, similar to the effect of hyperpolarisation reported in Fig. 1. Two further long lasting depolarisations to 0 mV induced transient periods of Ca²⁺ oscillations and transient declines in [Ca²⁺]_i during membrane hyperpolarisation, demonstrating the repetitive nature of these responses. The ability of depolarisation to 0 mV to induce more than one Ca²⁺ oscillation in non-oscillating cells was detected in 8 of 14 cells in which the depolarising step was sustained for sufficient duration to allow for the detection of multiple oscillations. Smaller depolarisations were also effective in inducing Ca²⁺ oscillations in non-oscillating cells. A depolarising step from -75 or -80 mV to -40 or -45 mV was effective in inducing multiple oscillations in 6 of 8 cells in which the depolarising step was of sufficient duration for the detection of multiple oscillations. The ability of depolarisation to modulate Ca²⁺ oscillations was observed in both fluo-3- and fura-2-loaded cells.

The data presented in Figs 1 and 2 serve to highlight the bipolar influence of V_m on the temporal pattern of Ca²⁺ changes occurring during ADP application. First, depolarisation is a potent stimulus for the induction of transient Ca²⁺ oscillations in non-oscillating cells and for increasing the amplitude of Ca²⁺ oscillations in oscillating cells. Second, hyperpolarisation is a potent stimulus for terminating oscillations or reducing oscillation amplitude. Taken together, these data are consistent with a strong voltage sensitivity of the Ca²⁺ oscillation mechanism.

We have previously reported that the rise in [Ca²⁺]_i induced by depolarisation during ADP application is a result of release of Ca²⁺ from an intracellular compartment (Mahaut-Smith et al. 1999). This conclusion is based upon the observation that depolarisation stimulates the rise in [Ca²⁺]_i in the absence of extracellular Ca²⁺. Thus, changes in [Ca²⁺]_i stimulated by depolarisation cannot be accounted for by potential-sensitive modulation of Ca²⁺ influx. Experiments were undertaken in Ca²⁺-free saline to rule out modulation of endogenous Ca²⁺ influx pathways in the hyperpolarisation-induced fall in Ca²⁺ during ADP application. An experiment performed in Na⁺-containing, Ca²⁺-free saline supplemented with 0.5 mM EGTA is shown in Fig. 3. Application of 1 μM ADP induced Ca²⁺ oscillations often superimposed on a plateau level similar to that detected in the presence of extracellular Ca²⁺. However, in contrast to experiments in Ca²⁺-containing saline, the plateau level of [Ca²⁺]_i and the oscillations showed a slow return to basal levels. This probably reflects a requirement for Ca²⁺ influx in the maintenance of the plateau phase of elevated [Ca²⁺]_i and the maintenance of longer lasting oscillations. The rate of decline in the plateau level of [Ca²⁺]_i was rapid in many cells making detection of a further hyperpolarisation-induced reduction in [Ca²⁺]_i difficult. In the experiment in Fig. 3, the ADP-evoked [Ca²⁺]_i declined at a slower rate, allowing for investigations of the effect of hyperpolarisation. An initial depolarisation from a holding potential of -75 mV to 0 mV modestly increased the amplitude of the Ca²⁺ oscillations consistent with data presented from experiments performed in the presence of extracellular Ca²⁺. Return to -75 mV was accompanied by a transient decline in [Ca²⁺]_i and a reduction in the amplitude of the Ca²⁺ oscillations. A second depolarising step to 0 mV demonstrates the reproducibility of this response, as Ca²⁺ oscillation amplitude was markedly increased, while a second repolarisation to -75 mV was again accompanied by a concomitant transient decline in [Ca²⁺]_i and the cessation of detectable oscillations. A further change in potential to -115 mV was accompanied by a third fall in [Ca²⁺]_i demonstrating that this phenomenon occurs over different hyperpolarising ranges. In the absence of extracellular Ca²⁺, hyperpolarisation induced a detectable decline in [Ca²⁺]_i in 6 of 6 cells tested. Given that these experiments were performed in Ca²⁺-free saline, the effect of hyperpolarisation on [Ca²⁺]_i, like that of depolarisation (Mahaut-Smith et al. 1999), cannot be explained by V_m modulation of an endogenous Ca²⁺ influx pathway.

Figure 3. Effect of V_m changes on [Ca²⁺]_i in Ca²⁺-free saline.

A rat megakaryocyte was whole-cell voltage clamped and fura-2 fluorescence simultaneously monitored following dialysis from the patch pipette. The cell was superfused with Ca²⁺-free saline supplemented with 0.5 mM EGTA prior to application of 1 μM ADP.

We have previously reported that increases in [Ca²⁺]_i induced by depolarisation are not a result of altered Na⁺-Ca²⁺ exchange activity. Such a conclusion is based upon the finding that depolarisation-evoked Ca²⁺ increases occur in the absence of intracellular and extracellular Na⁺ (Mahaut-Smith et al. 1999). The inhibitory effect of hyperpolarisation on steady-state [Ca²⁺]_i and Ca²⁺ oscillations was also detected in the absence of both intracellular and extracellular Na⁺. A representative experiment performed in the absence of extracellular Na⁺ (Na⁺ replaced with NMDG) is shown in Fig. 4. At a holding potential of -40 mV, application of 1 μM ADP resulted in a large transient increase in [Ca²⁺]_i followed by a plateau level, with small Ca²⁺ transients superimposed. Hyperpolarisation to -80 mV resulted in a transient decline in [Ca²⁺]_i and the termination of the small Ca²⁺ oscillations consistent with the effect of hyperpolarisation in Na⁺-containing saline (Figs 1 and 2). Subsequent depolarisations to -40 mV were accompanied by the onset of oscillations which were again abruptly terminated by a voltage step to -80 mV. It is important to note that before and after exposure to ADP, voltage steps between -40 and -80 mV were without effect on the [Ca²⁺]_i, consistent with our previous observations in Na⁺ saline (Mahaut-Smith et al. 1999). A transient decline in [Ca²⁺]_i induced by hyperpolarisation was observed in 10 of 10 cells studied, while depolarisation resulted in activation of Ca²⁺ oscillations or an increase in oscillation amplitude in 7 of the 10 cells. Hyperpolarisation was effective in inhibiting Ca²⁺ oscillations or oscillation amplitude in 5 of 5 cells in which it was investigated. Given that these experiments were performed in the absence of both extracellular and intracellular Na⁺ (K⁺-containing pipette solution), modulation of Na⁺-Ca²⁺ exchange activity cannot account for the effect of V_m on [Ca²⁺]_i oscillations and steady-state Ca²⁺ levels.

Figure 4. Effect of V_m changes on [Ca²⁺]_i in Na⁺-free saline.

A megakaryocyte was whole-cell voltage clamped in Na⁺-free, NMDG-containing extracellular saline and fura-2 fluorescence simultaneously monitored following dialysis of fura-2 from the patch pipette. Where indicated 1 μM ADP was applied.

The significance of a bipolar influence of V_m on ADP-evoked changes in [Ca²⁺]_i is highlighted by the finding that when measured under current clamp, ADP stimulated marked oscillations in V_m (Mahaut-Smith et al. 1999). To directly define the temporal relationship between V_m and [Ca²⁺]_i, simultaneous measurements of these two variables were undertaken during ADP application. V_m was recorded under whole-cell current clamp, during measurements of [Ca²⁺]_i and a representative experiment is shown in Fig. 5. During application of 1 μM ADP, each oscillatory increase in [Ca²⁺]_i is accompanied by a hyperpolarisation due to the activation of Ca²⁺-gated K⁺ channels (Uneyama et al. 1993a,b). Therefore, each V_m oscillation consists of a hyperpolarising and a depolarising phase which may contribute to the Ca²⁺ oscillation by promoting Ca²⁺ release during the depolarising phase and a decline in Ca²⁺ during the hyperpolarising phase. To investigate if physiological changes in V_m are sufficient to induce changes in [Ca²⁺]_i during ADP application, we have used a typical current-clamp recording of V_m recorded during ADP application as the command potential within a voltage-clamp experiment (Fig. 6). A single megakaryocyte was whole-cell voltage clamped at -45 mV in Na⁺-free NMDG saline and exposed to 1 μM ADP. The cell initially responded with Ca²⁺ oscillations that then subsided into a sustained, plateau phase of elevated Ca²⁺. The cell was then stepped to -75 mV followed by application of an oscillating voltage protocol. The initial rapid hyperpolarisation and each subsequent hyperpolarisation within the protocol were accompanied by a fall in [Ca²⁺]_i below the plateau level. In contrast, each depolarisation to approximately -42 mV, driven by the voltage protocol, was accompanied by a transient elevation in [Ca²⁺]_i above the steady-state plateau value. Termination of the oscillating voltage protocol at a value of -42 mV resulted in cessation of Ca²⁺ oscillations and the maintenance of a sustained plateau phase elevation in [Ca²⁺]_i. Steady-state voltage steps of similar magnitude to those applied during the oscillating protocol were then applied to observe their comparable effects. A hyperpolarisation from -42 to -80 mV was accompanied by a rapid decline in [Ca²⁺]_i of similar magnitude to that induced by the oscillating protocol. The fall in [Ca²⁺]_i was transient, displaying partial recovery during the 15 s duration hyperpolarisation. Subsequent depolarisation to -40 mV was accompanied by a transient increase in [Ca²⁺]_i of similar magnitude to that induced by the oscillating protocol. Application of a waveform voltage protocol of the same amplitude as that shown in Fig. 6 induced Ca²⁺ oscillations in 11 of 12 cells. The magnitude of the [Ca²⁺]_i response to the oscillating V_m protocol was variable, consistent with the marked heterogeneity of the depolarisation-induced Ca²⁺ transient (Mahaut-Smith et al. 1999).

Figure 5. Simultaneous recording of [Ca²⁺]_i and V_m under whole-cell current clamp.

A megakaryocyte was placed under whole-cell current clamp for the simultaneous recording of V_m and fura-2 fluorescence during application of 1 μM ADP. Fura-2 was loaded into the cell via dialysis from the patch pipette. Where indicated 1 μM ADP was applied.

Figure 6. Effect of physiological oscillations in V_m on [Ca²⁺]_i during ADP application.

A megakaryocyte was whole-cell voltage clamped and fura-2 fluorescence simultaneously monitored following fura-2 dialysis from the patch pipette. Following cessation of Ca²⁺ oscillations an oscillating voltage protocol was used as the command potential under voltage-clamp control. Where indicated 1 μM ADP was applied.

It is clear from these results that the effects of depolarisation and hyperpolarisation on [Ca²⁺]_i occur with sufficient speed to respond to physiological changes in potential. Therefore, under physiological conditions, the dynamic swings in potential observed in response to ADP exposure would be expected to contribute directly to both the rising and the falling phase of each Ca²⁺ oscillation. It should be noted that during a voltage-clamp experiment such as that presented in Fig. 6, feedback between [Ca²⁺]_i and V_m, via activation and inactivation of Ca²⁺-gated K⁺ channels, is effectively removed. In the absence of voltage clamp, elevations in [Ca²⁺]_i brought about by depolarisation-induced release of Ca²⁺₁ would activate Ca²⁺-gated K⁺ channels, thus producing a hyperpolarisation and a feedback inhibition of the voltage-dependent Ca²⁺ release. The net result is a damping of any depolarisation-dependent Ca²⁺ release. For depolarisation to significantly contribute to Ca²⁺ increases during oscillations, a lag must exist between the initial rise in [Ca²⁺]_i stimulated by depolarisation and the onset of either Ca²⁺-gated K⁺ channel activity or the resultant hyperpolarisation. To investigate this, the relationship between [Ca²⁺]_i and V_m was analysed from current-clamp experiments such as those presented in Fig. 5. The records of V_m and [Ca²⁺]_i from a single cell during exposure to 1 μM ADP are shown in the upper left panel of Fig. 7. The relationship between [Ca²⁺]_i and V_m has been plotted during the rising phase of [Ca²⁺]_i for oscillations 2, 4 and 6 (Fig. 7, upper right, lower right and lower left panels, respectively). During the onset of each oscillation, [Ca²⁺]_i rises from a starting value of approximately 150 nM to a value of between 200–250 nM with the detection of only a modest hyperpolarisation. Further increases in [Ca²⁺]_i were accompanied by a non-linear hyperpolarisation, consistent with the activation of a Ca²⁺-gated K⁺ conductance in these cells (Uneyama et al. 1993a,b). An increase in [Ca²⁺]_i of 100–200 nM without a marked membrane hyperpolarisation was observed in 7 of 10 cells for each V_m-Ca²⁺ oscillation and for the first V_m-Ca²⁺ response in the 3 remaining cells. In the latter group, the second and subsequent oscillations were superimposed upon an elevated level of [Ca²⁺]_i which may be closer to, or beyond, that required for activation of the Ca²⁺-gated K⁺ conductance.

Figure 7. Relationship between V_m and [Ca²⁺]_i during the rising phase of ADP-evoked Ca²⁺ oscillations.

Upper left panel, a megakaryocyte was placed under whole-cell current clamp and fura-2 fluorescence simultaneously monitored following fura-2 dialysis from the patch pipette. The cell was patched with a pipette internal solution containing 150 mM KCl and an extracellular solution containing 145 NaCl and 1 mM Ca²⁺. The relationship between V_m recorded under current clamp and [Ca²⁺]_i during the rising phase of the 2nd (upper right panel), 4th (lower right panel) and 6th (lower left panel) oscillations are shown.

The initial increase in [Ca²⁺]_i that occurs independently of a significant hyperpolarisation can be explained by the existence of a threshold [Ca²⁺] for the activation of Ca²⁺-gated K⁺ channels. Alternatively, these data may be indicative of the fact that [Ca²⁺]_i is not initially elevated in the immediate vicinity of the Ca²⁺ binding site responsible for opening of the K⁺ channel. Under this latter condition the lag between the initiation of a Ca²⁺ increase and the initiation of the hyperpolarisation may be a result of diffusional constraints. Regardless of the mechanism, these data indicate that depolarisation-induced release of Ca²⁺ from intracellular stores may contribute to the initial rising phase of each Ca²⁺ oscillation before the onset of the membrane hyperpolarisation and inhibition of voltage-dependent Ca²⁺ release. Thus, depolarisation-dependent Ca²⁺ release and the hyperpolarisation-dependent decline in [Ca²⁺]_i may contribute to the temporal pattern of Ca²⁺ signalling induced by ADP.

DISCUSSION

Previous experiments in rat megakaryocytes, a non-excitable cell type, have documented a depolarisation-induced transient increase in [Ca²⁺]_i during exposure to ADP (Mahaut-Smith et al. 1999). This effect of depolarisation is novel, as it results primarily from release of Ca²⁺ from intracellular stores and involves functional IP₃ receptors. The current study extends this observation by demonstrating that during ADP application, hyperpolarisation results in a transient decline in [Ca²⁺]_i. Thus, the influence of V_m on ADP-evoked changes in [Ca²⁺]_i is bipolar.

The question arises as to whether the depolarisation-induced increase in [Ca²⁺]_i and the hyperpolarisation-induced decrease in [Ca²⁺]_i represent a continuum of a single mechanism. Clearly, this cannot be conclusively ascertained until the underlying mechanism accounting for the potential sensitivity is determined. However, it is worth noting that in the presence of ADP, the opposing influences of depolarisation and hyperpolarisation are both detectable in the complete absence of extracellular and intracellular Na⁺ (see Figs 4 and 6 in this study and Mahaut-Smith et al. 1999). Therefore, neither potential effect can be adequately explained by alterations in Na⁺-Ca²⁺ exchange activity. We have also previously reported that depolarisation-induced increases in [Ca²⁺]_i were detected in the absence of extracellular Ca²⁺, thus ruling out the modulation of plasma membrane Ca²⁺ influx pathways by potential (Mahaut-Smith et al. 1999). Hyperpolarisation-induced falls in [Ca²⁺]_i during ADP application were also detected in the absence of extracellular Ca²⁺ (Fig. 3). Therefore, conditions supporting depolarisation-induced elevations in [Ca²⁺]_i also support hyperpolarisation-induced declines in [Ca²⁺]_i. Whether this is indicative of the existence of a single bipolar mechanism is presently unclear. Alternatively, hyperpolarisation during ADP exposure may be stimulating a Ca²⁺ sequestering or extrusion mechanism not normally active in the absence of ADP. However, under no conditions tested did hyperpolarisation result in [Ca²⁺]_i falling below its unstimulated level. Such a finding is inconsistent with hyperpolarisation stimulating a Ca²⁺ sequestering or extrusion mechanism.

In addition to the ability of membrane voltage to modulate steady-state [Ca²⁺]_i, the present results also document the ability of whole-cell potential changes to modulate the pattern of Ca²⁺ oscillations. First, hyperpolarising voltage steps were capable of either reducing the amplitude of the oscillations or completely arresting the oscillations in addition to development of a transient decline of [Ca²⁺]_i. Second, subsequent depolarisations were frequently sufficient to overcome this inhibition, thus regenerating Ca²⁺ oscillations. The physiological importance of V_m changes in the modulation of [Ca²⁺]_i and Ca²⁺ oscillations in megakaryocytes is reinforced by the finding that purinoceptor-induced Ca²⁺ oscillations are accompanied by marked oscillations in V_m over a voltage range documented to influence ADP-evoked Ca²⁺ signalling (Fig. 5 this study and Mahaut-Smith et al. 1999). Furthermore, this physiological voltage waveform induced Ca²⁺ oscillations when applied as a command potential to cells displaying steady plateau Ca²⁺ levels during ADP exposure (Fig. 6). The existence of a threshold level of Ca²⁺ required for the activation of Ca²⁺-gated K⁺ channels complements the ability of oscillatory changes in V_m to contribute directly to Ca²⁺ oscillation. Elevations in [Ca²⁺]_i of approximately 100 nM occur without a marked onset of hyperpolarisation, thereby providing an appropriate temporal lag between the induction of depolarisation-induced release of Ca²⁺ from intracellular stores and the onset of Ca²⁺-dependent K⁺ channel activation. It is not possible using the present experimental protocol to determine which properties of the [Ca²⁺]_i elevation and/or the Ca²⁺-gated K⁺ channel activation are responsible for the increase in [Ca²⁺]_i independent of changes in V_m. The fluorescence measurements in this study employed a photomultiplier tube and thus provide no spatial information of the Ca²⁺ changes. Therefore, the lag between the onset of increases in [Ca²⁺]_i and the onset of membrane hyperpolarisation may result from the time required for Ca²⁺ to diffuse to the site for activation of Ca²⁺-gated K⁺ channels. Alternatively, this lag may be accounted for by the existence of a threshold Ca²⁺ concentration for the activation of the Ca²⁺-gated K⁺ conductance. The Ca²⁺-gated K⁺ channels in the rat megakaryocyte remain to be characterised; however, these are likely to be of the medium conductance class since this is the main type expressed in both platelets (Mahaut-Smith, 1995) and the megakaryocytic cell lines DAMI (Sullivan et al. 1998) and HEL (Lu et al. 1999; M. P. Mahaut-Smith & M. J. Mason, unpublished observations). Studies of this class of Ca²⁺-gated K⁺ channel in lymphocytes show a threshold for activation by [Ca²⁺]_i of approximately 200–300 nM (Mahaut-Smith & Schlichter, 1989; Grissmer et al. 1992). Thus, the depolarising phase of each membrane potential oscillation would stimulate Ca²⁺ release from intracellular stores with a small increase in [Ca²⁺]_i before the required threshold [Ca²⁺] is reached, Ca²⁺-gated K⁺ channels are opened, and the cell hyperpolarises. The hyperpolarisation then stimulates a voltage-dependent decline in [Ca²⁺]_i, contributing to the falling phase of each Ca²⁺ oscillation. In this scheme, depolarisation and hyperpolarisation represent active rather than passive components of the oscillatory cycle. In summary, while the mechanism underlying the bipolar response of V_m on [Ca²⁺]_i is unclear, the physiological significance of the phenomenon has been demonstrated: (a) by the demonstration of the dynamic changes in potential experienced by the megakaryocyte during ADP application and (b) by the sensitivity of [Ca²⁺]_i and Ca²⁺ oscillations to these physiological changes in potential.

The ADP receptor that stimulates Ca²⁺ mobilisation in megakaryocytes is not well characterised; however, several pieces of evidence are consistent with it being a G-protein-linked receptor coupled to phospholipase Cβ. First, the increase in [Ca²⁺]_i induced by ADP is prevented by heparin (Somasundaram & Mahaut-Smith, 1995; Mahaut-Smith et al. 1999), an IP₃ receptor inhibitor (Bezprozvanny & Ehrlich, 1995). Second, intracellular dialysis with IP₃ or GTPγS evokes Ca²⁺ oscillations similar to those induced by ADP (Uneyama et al. 1993b; Hussain & Mahaut-Smith, 1998). Third, megakaryocytes lack Ca²⁺-induced Ca²⁺ release via ryanodine receptors (Uneyama et al. 1993b; M. J. Mason & M. P. Mahaut-Smith, unpublished observations). Changes in V_m could influence any stage of the signalling cascade from receptor/agonist binding to activation of phospholipase C including the modulation of availablity of highly polar substrates, such as ADP, GTP and phosphotidylinositol 4,5-bisphosphate. A voltage dependence to one or more of these events would allow transmembrane potential to modulate IP₃ production. A depolarisation-induced increase in IP₃ levels could account for the present results, resulting in release of Ca²⁺ from endosomal stores and/or the activation of Ca²⁺ oscillations. In addition, hyperpolarisation may cause a reduction in IP₃ levels and thus a decline in Ca²⁺ release which would tend to reduce the [Ca²⁺]_i as a result of the endosomal or plasma membrane Ca²⁺ pumps acting in the presence of less opposition from IP₃-mediated Ca²⁺ release. Such a fall in [Ca²⁺]_i could account for the observed inhibition of Ca²⁺ oscillations. Evidence for depolarisation-stimulated increases in IP₃ formation has been reported in skeletal muscle (Vergara et al. 1985). Furthermore, hyperpolarisation has been proposed to reduce IP₃ production during agonist stimulation of rabbit coronary mesenteric artery smooth muscle cells (Itoh et al. 1992).

Voltage-dependent modulation of Ca²⁺ release from IP₃-sensitive stores has been reported in one class of excitable cell, the coronary artery smooth muscle cell (Ganitkevitch & Isenberg, 1993). Depolarisation-evoked Ca²⁺ release similar to that which we report (Mahaut-Smith et al. 1999; and this paper) has been reported during stimulation of metabotropic cholinergic receptors. Furthermore, a small decline in [Ca²⁺]_i in reponse to hyperpolarisation was also evident, consistent with our findings in megakaryocytes. Importantly, depolarisation-induced release of Ca²⁺ from intracellular stores was observable during whole-cell dialysis with GTPγS. Cell dialysis of GTPγS activates the G-proteins coupling metabotropic receptors to PLC without the requirement for agonist. Such a finding is consistent with the hypothesis that in smooth muscle cells the site of voltage dependence is downstream of receptor/agonist interactions. This is further supported by the finding that voltage-dependent Ca²⁺ release was also observed in the absence of agonist following dialysis with Li⁺, an inhibitor of inositol phosphate metabolism that brings about elevated cytosolic levels of inositol phosphates (Brami et al. 1991). Under these conditions, voltage modulation of basal IP₃ production may have a more dramatic effect upon IP₃ levels due to inhibited inositol metabolism. However, given the lack of direct evidence for potential-dependent modulation of IP₃ levels, further experiments are required to explore this hypothesis in both smooth muscle cells and megakaryocytes. It is interesting to note that in coronary artery smooth muscle cells, oscillations in [Ca²⁺]_i were not observed in response to depolarisation during metabotropic cholinergic receptor stimulation (Ganitkevitch & Isenberg, 1993). However, depolarisation was effective in inducing Ca²⁺ oscillations in response to Li⁺ application (Ganitkevitch & Isenberg, 1993), an effect of depolarisation frequently observed in megakaryocytes during ADP application.

The present data serve to highlight the importance of electrogenic influences during metabotropic purinoceptor-evoked Ca²⁺ signalling. The experiments described here and in Mahaut-Smith et al. (1999) provide a framework whereby physiological changes in V_m feedback to alter Ca²⁺ signalling in a manner distinct and opposite from the conventional effect of potential upon modulation of Ca²⁺ influx from the extracellular environment. The changes in potential observed in megakaryocytes following ADP application appear to be active components of the Ca²⁺ signalling cascade, contributing to both the release of Ca²⁺ from intracellular stores and the decline of Ca²⁺₁ possibly as a result of a bipolar action of V_m on Ca²⁺ release. In conclusion, this study reinforces the importance of further investigations into the interactions between ionotropic and metabotropic agonists in concerted activation of megakaryocytes and other cells.