ABA activates multiple Ca²⁺ fluxes in stomatal guard cells, triggering vacuolar K⁺(Rb⁺) release

Abstract

The mechanisms by which abscisic acid (ABA) activates the release of K⁺(Rb⁺) from the vacuole of stomatal guard cells, a process essential for ABA-induced stomatal closure, have been investigated by tracer flux measurements. The form and timing of the ABA-induced efflux transient could be manipulated by treatments that alter three potential Ca²⁺ fluxes into the cytoplasm, the influx from the outside and two pathways of internal release, those dependent on phospholipase C (inhibited by U73122) and cyclic ADP-ribose (inhibited by nicotinamide). Ba²⁺, acting as a competitive inhibitor of Ca²⁺ influx but also as an inhibitor of internal release, was an effective inhibitor of the transient. The results suggest that a threshold level of cytoplasmic Ca²⁺ is required for the initiation of the minimal efflux transient after a lag period and with a low rate of rise. As conditions improve for the generation of an efflux transient (higher ABA or reduced Ba²⁺), a second threshold is crossed, generating a transient with zero lag and rapid rate of rise. This may reflect different Ca²⁺ levels required for activation of different tonoplast K⁺ channels. In this state, at high ABA, the transient is inhibited by removal of external Ca²⁺, suggesting Ca²⁺ influx makes a major contribution to increase in cytoplasmic Ca²⁺. By contrast, at low ABA, the transient is not inhibited by removal of external Ca²⁺ but is sensitive to either U73122 or nicotinamide, suggesting internal release makes the major contribution, involving both pathways. ABA appears to activate all three processes, and their relative importance depends on conditions.

The ability to reduce transpirational water loss when water is scarce is essential for plant survival, and such control is achieved by regulation of the size of stomatal pores in the leaf surface. Pore size is determined by the turgor of the pair of stomatal guard cells and is modified by changes in guard cell solute content, largely potassium salts, but also other solutes including sugars. The so-called drought hormone, abscisic acid (ABA), is produced by or imported into stomatal guard cells in water stress conditions, changing the activity of a number of ion channels in guard cell membranes to produce net loss of salt from guard cells and loss of turgor with consequent closure of the stomatal pore. The end result is the loss of both K⁺ and associated anion (Cl⁻ and/or malate) from both the vacuole and the cell by up-regulation of appropriate channels in the two membranes. Commonly there is discussion about signaling pathways involved in the process of stomatal closure, as if this were a single process, with debate about whether these pathways are (for example) Ca²⁺ dependent or Ca²⁺ independent. This view is inadequate and seriously misleading. Instead, we need to distinguish the essential contributory processes and their signaling chains, the four separate stimulations of transport activities by which net efflux of K⁺ and anions at the plasmalemma and net flux of K⁺ and anions from vacuole to cytoplasm are achieved.

Guard cell signaling processes are of considerable interest in their own right, but the guard cell has also become a model system for investigation of plant signaling pathways. Considerable progress has been made toward understanding mechanisms of regulation of plasmalemma ion channels. Measurements of cytoplasmic pH and Ca²⁺ by using fluorescent dyes show ABA-induced increase in cytoplasmic Ca^{2 +} (although variable in timing and extent) and in cytoplasmic pH (see ref. 1 for review and references). Electrophysiological studies of guard cell ion currents, either in patch-clamped guard cell protoplasts or by impalement of intact guard cells, have identified a number of ABA-induced changes in plasmalemma ion channels and their potential signaling intermediates. Thus cytoplasmic Ca²⁺, cytoplasmic pH, and complex networks of protein kinases/phosphatases are identified as having roles in ABA signal transduction and its impact on plasmalemma channels.

By contrast, our knowledge of ion channels in the tonoplast and their response to ABA is much less satisfactory, although the release of vacuolar ions (both anions and cations) is absolutely required for stomatal closure, and one of the striking differences between open and closed guard cells is the extent of vacuolar solute accumulation and the vacuolar volume. Guard cells in closed stomata are peculiar among mature plant cells in having an unusually small fraction of cell volume occupied by the vacuole. It is therefore important to shift attention to the other half of the overall process of aperture change, the events at the tonoplast membrane. We need now to investigate whether the same signaling intermediates are involved in the response of tonoplast channels to ABA. This paper is concerned with the mechanisms by which the best-studied “closing” signal, ABA, up-regulates flux from vacuole to cytoplasm.

The study of tonoplast ion channels by patch clamping isolated vacuoles suffers from uncertainty in the appropriate choice of bathing solution to mimic cytoplasmic conditions, with probable omission of cytoplasmic regulatory factors. Several ion channels potentially capable of supporting K⁺ flux from vacuole to cytoplasm have been identified, including the K⁺-specific VK channel and the less-specific (cation) SV channel, both of which are Ca²⁺ activated (2, 3). However there is still uncertainty about the channel for anion release at the tonoplast. A Ca²⁺-activated anion channel has been described (4) but was postulated to be responsible for anion uptake to the vacuole rather than release from vacuole to cytoplasm; although this was true in the conditions of these experiments, it does not seem clear that the channel is incapable of anion transfer from vacuole to cytoplasm when the gradient is appropriate. The identity of tonoplast channels that respond to ABA in the intact cell and the mechanisms of their regulation remain unclear. The Ca²⁺ sensitivity of the tonoplast channels that have so far been identified highlights the need to understand mechanisms for increase in cytoplasmic Ca²⁺, whether by influx from outside or by release from internal stores, and to determine whether Ca²⁺-mediated regulation of tonoplast ion channels is involved in the physiological response to ABA.

Two mechanisms of Ca²⁺ influx have been identified, both activated by ABA. A nonspecific cation channel was observed (5), responsible for ABA-induced spikes of associated inward current and increase in cytoplasmic Ca²⁺, and more recently a hyperpolarization-activated Ca²⁺ channel has been identified that is up-regulated by ABA (6–8), shifting the activation voltage to much more positive values. Activation of this channel produces large increases in cytoplasmic Ca²⁺, which appear to derive from both primary influx of Ca²⁺ from outside and induced release from internal stores.

Two ligand-gated Ca²⁺ release channels have been identified in plant endomembranes, gated by inositol 1,4,5-trisphosphate (IP₃) and cyclic ADP-ribose (cADPR), respectively. IP₃-sensitive channels were first identified in the tonoplast, but it is now clear that IP₃-triggered Ca²⁺ release is associated with a number of membrane fractions in the cell and that only a small fraction of the total activity is associated with the tonoplast fractions (9). Thus IP₃-triggered Ca²⁺ release from either vacuole or endoplasmic reticulum (ER) could be involved in the ABA response but perhaps more likely the ER. The second candidate channel, gated by cADPR, has been shown to be present in the tonoplast (10), but it remains possible that it is also present in the ER, as in animal cells. Whether either of these triggers, IP₃ or cADPR, is involved in the signaling chains leading to specific ABA-induced changes in guard cells, including the vacuolar ion release, needs to be established. A third ligand-gated Ca²⁺ channel in plant endomembranes has recently been characterized; this is in the ER and not in the tonoplast, is activated by nicotinic acid adenine dinucleotide phosphate, and is of unknown function (11).

Guard cells can respond to release of IP₃ from the caged compound by increase of cytoplasmic Ca²⁺ (12) and by Ca²⁺-dependent changes in plasmalemma ion channels (13), and all of the components of the phosphoinositide signaling pathway are present in guard cells (14). Changes in the pattern of ³²P labeling of both inositol lipids and inositol phosphates argued for ABA-induced phosphoinositide turnover in guard cells (15), but the changes in label in putative Ins(1,4,5)P₃ were surprisingly small if this is the major activator of ABA-triggered Ca²⁺ release.

It has recently been shown that the aminosteroid U73122, an inhibitor of phospholipase C (including the enzyme in guard cells), modifies the form of ABA-induced oscillations in the level of cytoplasmic Ca²⁺ but not those induced by external Ca²⁺ (16). The inhibitor did not abolish ABA-induced stomatal closure but reduced it by about 20% and had no effect on closure induced by external Ca²⁺. Although the effect of the inhibitor on the ABA-induced rise in cytoplasmic Ca²⁺ was not determined, the results suggest that IP₃-triggered Ca²⁺ release may contribute to ABA-induced stomatal closure. A second study also observed a partial inhibition of ABA-induced stomatal closure by U73122 (17).

There is evidence for the involvement of cADPR in another ABA-initiated signaling chain, that responsible for ABA-regulated gene expression in tomato (18). By coinjection of GUS-reporter gene constructs from two ABA-inducible genes in Arabidopsis and putative messengers, it was shown that the activation requires increase in cytoplasmic Ca²⁺, mediated through activation of a cADPR-sensitive Ca²⁺ release channel and not through an IP₃-sensitive channel. ABA was also shown to stimulate the formation of cADPR.

In guard cells also there is evidence for a role for cADPR in ABA-induced stomatal closure (19). Injection of 8-NH₂-cADPR (a competitive antagonist of cADPR) or blocking cADPR synthesis by the addition of 20 or 50 mM nicotinamide externally, inhibited ABA-induced stomatal closure but again only partially. Patch clamping of isolated vacuoles also showed effects of cADPR. However, the properties of the Ca²⁺-permeable channel in the tonoplast that was activated by cADPR, in particular its inhibition by Ca²⁺ > 600 nM, suggest involvement of the FV channel and make it unlikely that its activation can be the only source of increased cytoplasmic Ca²⁺. It is, however, possible that cADPR-triggered release of Ca²⁺ from the ER, as in animal cells, is the major contributor, rather than release from the vacuole.

Yet another Ca²⁺-related signaling pathway has recently been discovered (20), with a further potential second messenger, inositol hexakisphosphate (IP₆). It was found that ABA produces rapid changes in the level of IP₆ in guard cells, that submicromolar levels of IP₆ in the patch pipette inhibit the inward K⁺ channel, and that this inhibition is abolished by the inclusion of Ca²⁺ chelator in the patch pipette. These effects are reminiscent of the effects of IP₃ (13), but IP₆ is about 100 times more effective than IP₃ in inactivation of the inward K⁺ current.

Flux studies have the advantage of using intact cells with full cytoplasmic content but have poorer time resolution and are semiquantitative only. However, they do provide information on ABA-induced events in the intact cell that is not available from other sources and in particular allow such events at the tonoplast to be studied. Previous flux studies (21–23) have identified ABA-induced changes in fluxes of both anions and cations at both plasmalemma and tonoplast. In these studies, isolated guard cells of Commelina communis L. were preloaded with ⁸⁶Rb⁺ (used as an analogue for K⁺ and not as a tracer), and the rate of loss of tracer to unlabeled solution followed with time, before and after the addition of ABA; ABA is added after about 40 min of washout, when the tracer remaining is almost entirely vacuolar. ABA induces a transient stimulation of tracer efflux, but the rate of tracer loss in the presence of ABA then returns to a value very similar to that before ABA was added. This behavior suggests that ABA may act by changing the “set point” of a regulatory system rather than by producing a permanent change in ion permeability. In suboptimal conditions, at low concentrations of ABA, there is a lag before initiation of the efflux transient, and the transient rises more slowly to a reduced peak rate (22). Nevertheless, the peak is reached after the loss of the same amount of tracer in the two conditions but after different times; similarly the end of the transient is reached at a given tracer content but at different times. This would be consistent with the cell regulating to a different content (or volume) after ABA application, with ABA reducing the “set point.” The delay introduced in suboptimal conditions suggested that the triggering of the vacuolar ion release requires a threshold of some kind, for example of internal Ca²⁺, to be reached. Thus detailed examination of the form of the ABA-induced efflux transient provides information on the factors affecting the initiation of release of ions from the vacuole after addition of ABA. Experimental manipulation of such transients can therefore be used to identify signaling intermediates in intact cells with intact signaling pathways.

This approach has been extended in the present work, which shows that the form of the ABA-induced efflux transient can also be manipulated by the addition of a number of ion channel blockers. Some preliminary experiments by using Ba²⁺ have already been described (24); these showed that Ba²⁺ had the same effects on the efflux transient as are produced by reducing the concentration of ABA, namely the introduction of a lag period before initiation of the transient, a reduced rate of rise, and a lower peak. In contrast, Cd²⁺, which reduced ⁸⁶Rb⁺ efflux in the absence of ABA by an amount similar to that produced by Ba²⁺, had no effect on the timing of the transient and very little effect on the peak height. The effect of Ba²⁺ has now been more thoroughly investigated in a range of conditions. The results suggest that release of Ca²⁺ from internal stores is critical for initiation of vacuolar K⁺(Rb⁺) release, and that a threshold level of cytoplasmic Ca²⁺ must be reached for initiation of the vacuolar efflux transient.

The effects on the ABA-induced vacuolar efflux transient of putative inhibitors of the potential Ca²⁺ release mechanisms were also determined by using U73122 as an inhibitor of phospholipase C and nicotinamide as an inhibitor of cADPR synthesis. The role of influx of Ca²⁺ was also assessed by measuring efflux transients in the presence and absence of external Ca²⁺. The results suggest that all three mechanisms can contribute to ABA-induced increase in cytoplasmic Ca²⁺, the influx of Ca²⁺ from outside, and both pathways for internal release.

Materials and Methods

Methods followed those used previously (21, 22). Isolated guard cells of Commelina communis L. were prepared by treatment of epidermal strips at low pH (pH 3.9), to kill all cells other than guard cells. All incubations took place in a thermostated cabinet at 25°C, lit by a bank of fluorescent tubes (PAR 140 μmol⋅m⁻²⋅s⁻¹). Cells were loaded overnight (about 16 h) in ⁸⁶Rb⁺-labeled solution (0.7 − 1.3 MBq ml⁻¹, ⁸⁶Rb⁺ isotope from Amersham, containing 2 mM RbCl, 0.1 mM CaCl₂, and 10 mM Pipes buffer, pH 6 (1,4 piperazinediethane sulfonic acid). Effluxes were measured by transferring individual strips to successive 0.75-ml portions of nonradioactive solutions of the same chemical composition, in the wells of plastic culture chambers (on a vibrating shaker) and counting both these washout solutions and the residue left in the tissue at the end of the efflux by using standard scintillation counting methods.

Tracer was expressed as pmol⋅mm⁻² on the basis of the area of each epidermal strip, and the rate of loss was calculated for each time interval. To reduce the variability between different strips arising from differences in number and size of guard cells, the rate constant for exchange (h⁻¹) was calculated as (rate of loss of tracer)/(tracer content) and plotted against time. In constant conditions, the rate constant falls with time as the cytoplasmic exchange proceeds, to reach a steady value equal to the rate constant for vacuolar exchange, when the slow phase of exchange is reached.

Results

Effect of Ba²⁺ on ABA-Induced Efflux Transient.

ABA, added after about 40–45 min of efflux in the absence of ABA when the efflux has reached the slow vacuolar phase, produces a transient stimulation of efflux, best shown in plots of the rate constant for efflux (rate of loss of tracer/total tracer content) against time.

Fig. 1 shows results from an experiment in which efflux transients were measured in 10 and 0.1 μM ABA and in 10, 1, and 0.1 μM ABA in the presence of 1 mM Ba²⁺. In the control, in 10 μM ABA in the absence of Ba²⁺, the efflux transient rose rapidly to a peak at 6.7 ± 0 (n = 4) min with no lag. Each of the other peaks showed a lag, a reduced rate of rise, and a delayed peak time, the lag most marked in 0.1 μM ABA in the presence of Ba²⁺ but the rate of rise very similar in each of the four suboptimal conditions. For clarity, only three of the five curves are shown; the curve for 0.1 μM ABA in the absence of Ba²⁺ (not shown) is very close to that for 10 μM ABA in the presence of Ba²⁺, and that for 1 μM ABA in the presence of Ba²⁺ had a lag intermediate between the two Ba²⁺ curves shown, those at 10 and 0.1 μM ABA.

Figure 1.

Effects of 1 mM Ba²⁺ on ABA-generated efflux transients. Rate constant for efflux plotted against time. Each point shows the mean of four replicate strips. (For clarity, errors are not shown; SEM of 4 strips was 5–10% away from the peaks and 15–25% at the peaks). ABA (10 or 0.1 μM) was added at zero time after 40–45 min of washout in the absence of ABA; Ba²⁺ was added 9–10 min before ABA. ●, 10 μM ABA control; □, 10 μM ABA, Ba²⁺; ▴, 0.1 μM ABA, Ba²⁺.

The effect of Ba²⁺ on the form of the ABA-induced efflux transient was measured in a number of such experiments at different concentrations of ABA and in different batches of tissue. Because the effect of the addition of Ba⁺ on the form of the transient is similar to that produced by reducing the concentration of ABA, a detailed quantitative comparison of their respective effects is presented below. Two features of the transient were used to assess the changes produced, namely the lag period before the initiation of the transient and the rate of rise (the slope of the rising phase, in h⁻²).

The effect of 1 mM Ba²⁺ on the lag period, the time before initiation of the vacuolar release, and on the rate of rise was measured in six different experiments, allowing ten such comparisons at different concentrations of ABA, and the results are shown in Fig. 2. In six of these comparisons with control rates of rise of 2.2 h⁻² or above, the effect of adding 1 mM Ba²⁺ is the introduction or a significant lengthening of the lag and a significant reduction in the rate of rise (to 0.08–0.42 of the value in the control). In two comparisons with control rates of rise 1.23 and 1.19 h⁻², the lag was significantly increased by the addition of 1 mM Ba²⁺ but without further reduction in the rate of rise. However, in the remaining two comparisons, with control rates of rise of 0.75 and 0.62 h⁻² and control lags of 14 and 18.5 min, respectively, the addition of 1 mM Ba²⁺ neither increased the lag period nor further reduced the rate of rise. Thus, if the control transient is already low, the addition of Ba²⁺ produces no further inhibition.

Figure 2.

Effect of Ba²⁺ on lag period and rate of rise of ABA-induced efflux transient. Efflux transients were measured in the same tissue in the presence of 1 mM Ba²⁺ (labeled Ba) and in its absence (labeled C). ABA was added after 40–45 min washout in the absence of ABA, and Ba²⁺ was added 9–22 min before ABA. Each point is the mean of 4 replicate strips; errors are 0–20% in the lags and 10–35% in the rates of rise. ○, 10 or 100 μM ABA; ▵, 0.1 or 0.2 μM ABA.

This distinction between effects on the rate of rise and on the lag is also seen when the effects of different concentrations of Ba²⁺ are compared in the same experiment. In each of five experiments at 10 or 100 μM ABA, the addition of 1 mM Ba²⁺ both reduced the rate of rise and introduced a lag period. However, when the transients induced by 100 μM ABA were compared in 0, 0.5, and 1 mM Ba²⁺ in the same experiment, it was found that 0.5 mM Ba²⁺ reduced the rate of rise to 22% of the control but did not produce a lag period; increase from 0.5 mM Ba²⁺ to 1 mM Ba²⁺ had only a small further effect on the rate of rise but introduced a lag period of 5–7 min. In two further experiments with 0.5 and 0.7 mM Ba²⁺ at 10 μM ABA, the rate of rise was reduced to 0.36–0.37 of that in the control, but only a very small lag of 1–2 min was introduced. Thus the first effect of adding Ba²⁺ is to reduce the rate of rise, but as the degree of impairment of the ability to produce an efflux transient grows more severe, this is also accompanied by a lengthening lag period.

The same distinction is seen in two experiments in which the concentration of ABA was varied in the presence and absence of Ba²⁺. In the absence of Ba²⁺, reduction of ABA from 10 to 0.1 μM both reduced the rate of rise and introduced a lag period [as previously shown in many such experiments (22)]. However, in the presence of 1 mM Ba²⁺, where the rate of rise in 10 μM ABA had already been markedly reduced [to values of 0.99 ± 0.17 (n = 4) h⁻² and 0.88 ± 0.22 (n = 4) h⁻², respectively, in the two experiments, one of which is shown in Fig. 1], the lag was progressively increased as ABA was reduced from 10 μM to 1 μM and to 0.1 μM, but with no further reduction in the rate of rise.

This dual effect can be seen most clearly in a plot of rate of rise against lag period for the individual points for the combined six experiments at different concentrations of ABA, in the presence and absence of Ba²⁺. Fig. 3 Left shows the values for the full transient at 10 μM ABA (or in one experiment 100 μM) in the absence of Ba²⁺; Fig. 3 Right shows the values for the three suboptimal conditions, 10 or 100 μM ABA in the presence of 1 mM Ba²⁺, low concentration of ABA (0.1 or 0.2 μM) with or without Ba²⁺. It appears that, as the transient is progressively impaired, the rate of rise is markedly reduced before a significant lag is introduced, but the lag then increases over the range 4 to 25 min, with no further reduction in the rate of rise. The full transient showed no lag period but had a wide range of values for the rate of rise with a mean of 4.5 ± 0.7 (n = 12) h⁻² and a range of 1.7 to 9.8 h⁻². For inhibited transients with lags of up to 3 min, the mean rate of rise was 1.7 ± 0.3 (n = 14) h⁻². For the inhibited transients with lags in the range 4–25 min, the rate of rise was independent of lag, with a mean value of 1.06 ± 0.09 (n = 42) h⁻²; the regression of rate of rise on lag period had a correlation coefficient r = 0.04.

Figure 3.

Progressive inhibition of ABA-induced efflux transient by reduction in the ABA concentration from 10 or 100 μM to 0.1 or 0.2 μM, or addition of 1 mM Ba²⁺. Plot of rate of rise (h⁻²) against lag period (min) for individual strips. Data are from six experiments. ▾, high ABA; ▴, high ABA, Ba²⁺; ○, low ABA; ●, low ABA, Ba²⁺.

The results suggest there are two thresholds involved in determination of the transient. In the least favorable conditions (low ABA and/or the presence of Ba²⁺), the transient is initiated after a lag period when the first threshold is reached, but in this condition the rate of rise is around 1 h⁻². As conditions for initiation of the transient are improved (increased ABA concentrations or reduction in Ba²⁺ concentration), the lag first decreases without change in the rate of rise, which remains around 1 h⁻². A second threshold is then reached at about the stage of inhibition at which the lag disappears, after which the rate of rise increases above 2 h⁻². Additionally, it should be noted that in the two experiments in which the transient in the absence of Ba²⁺ was most inhibited, the addition of Ba²⁺ had no effect on either lag or rate of rise.

Assessment of Contributions of the Phospholipase C and cADPR Pathways to ABA-Induced Release of Ca²⁺ from Internal Stores.

With the aim of assessing the contribution (if any) of the two potential pathways of Ca²⁺ release to the increase in cytoplasmic Ca²⁺, the effect on the vacuolar efflux transient of two inhibitors was determined: U73122, an inhibitor of phospholipase C, and nicotinamide, an inhibitor of cADPR synthesis, used either separately or together.

At 5 μM ABA, at which a full efflux transient is generated (rate of rise 8.6 h⁻² and zero lag period), the addition of 20 mM nicotinamide had no effect on the rate of rise of the transient but did generate a small delay of 2 min in its initiation. However, when nicotinamide was tested in suboptimal conditions at low concentrations of ABA, there was significant inhibition of the ability to generate an efflux transient with a reduction in the rate of rise and an increase in the lag period, as shown in Fig. 4. [Control peak had a rate of rise of 1.7 ± 0.3 (n = 4) h⁻² and lag 4.2 ± 0 min, compared with 1.15 ± 0.12 (n = 4) h⁻² and 10 ± 2 min in the nicotinamide-treated tissue.] This figure also shows that in this experiment, in the presence of both 20 mM nicotinamide and 1 μM U73122 together, the transient is all but abolished; there is a small increase in efflux rate for a short time above the curve for the two inhibitors in the absence of ABA, but no real transient develops. Thus it appears that in this experiment, both pathways can contribute to the ABA-induced increase in cytoplasmic Ca²⁺ generating the efflux transient.

Figure 4.

Inhibition of efflux transient produced by 0.1 μM ABA by U73122 and nicotinamide. Rate constant for efflux was plotted against time. Each point shows the mean of four replicate strips. (Errors, not shown, 5–20%.) ABA was added at zero time after 35–37 min of washout in the absence of ABA; 20 mM nicotinamide was added 22–25 min before ABA; and 1 μM U73122 was added 7 min before ABA. ●, control; □, nicotinamide; ▵, nicotinamide + U73122; ▿, nicotinamide + U73122, no ABA.

The effects of nicotinamide were consistent in different experiments: in four experiments at 0.1 μM ABA the rate of rise was reduced in the presence of 20 mM nicotinamide to 48 ± 8% of that in the control, and the lag was increased to 220 ± 30% of the control. Twenty millimolar nicotinamide produced a small increase in the rate of efflux before ABA was added, to 125 ± 19% (n = 4) of the control, but in the absence of ABA, nicotinamide gave no further increase over the time course of the experiment, allowing its effect on the response to ABA to be distinguished.

In contrast, the effects of U73122 varied in different batches of tissue. In the first two experiments done, the effects were similar to those seen with nicotinamide. At high ABA (1 μM, with zero lag in the control), the addition of 1 μM U73122 introduced a small lag of 2 min, but the rate of rise was then higher than the control; addition of the inactive analogue U73343 had no effect on the transient. As with nicotinamide, inhibition was observed when the inhibitor was tested at suboptimal concentrations of ABA (0.1 μM), when the control peak was impaired. In the first such test, in which the control peak was similar to that in Fig. 4, the addition of 1 μM U73122 inhibited the transient [control peak, slope 2.0 ± 0.2 (n = 4) h⁻² and lag 8 ± 1 min, peak in the presence of 1 μM U73122, slope 1.36 ± 0.05 (n = 4) h⁻² and lag 11.8 ± 1.3 min]. However, in other experiments the addition of U73122 increased the effectiveness with which ABA generated the efflux transient. In one such experiment, the transient produced by 0.1 μM ABA was reduced by 20 mM nicotinamide, but the addition of U73122 in the presence of nicotinamide restored the transient to the control values; the peak in the presence of U73122 was larger than that in the control. In this experiment, the factor by which U73122 increased the rate of rise at 0.1 μM ABA was 1.7 when used alone and 1.8 when added to nicotinamide. In the same experiment, the addition of both nicotinamide and U73122 at 1 μM ABA increased the rate of rise by a factor of 2.3, and in a further experiment (at 8 μM ABA) the addition of both U73122 and nicotinamide increased the rate of rise by a factor of 3.2. If nicotinamide does not inhibit in these conditions of high ABA, these increases can be attributed to the effect of U73122. The effect of U73122 in the absence of ABA was also variable: the efflux tended to increase but generally after the time in which the ABA-induced peak was generated, allowing the effect of U73122 on the ABA response to be distinguished from the effect of the inhibitor alone.

It is likely that the inhibition of synthesis of cADPR by 20 mM nicotinamide is only partial (19), and the effect of 40 mM nicotinamide was also tested. The effect on efflux in the absence of ABA was larger at this concentration, with stimulation by a factor of 2.7 ± 0.8, an effect that may be osmotic as well as specific to nicotinamide. Forty millimolar nicotinamide abolished the efflux transient at 0.1 μM ABA, but the addition of U73122 together with the nicotinamide restored a small peak (with rate of rise 41% of that in the control and comparable lag) not seen with the two inhibitors in the absence of ABA. In this tissue, the ABA-induced peak was enhanced by the addition of U73122 alone.

Thus U73122 can either inhibit or enhance the ABA-induced efflux transient. The different pattern of behavior is correlated with the rate of rise in the control tissue. In three experiments in which U73122 inhibited the generation of the transient, the control peaks had rates of rise in the range 0.80–2.0 h⁻²; in five comparisons in which U73122 enhanced the transient, the rates of rise in the controls were in the range 3.4 to 11 h⁻². There was one exception to this pattern, an experiment in which U73122 increased the efflux transient despite a low rate of rise in the control (1.1 ± 0.1 h⁻²), but the general pattern is that U73122 increases the ABA-induced efflux transient in tissue in which the control peaks are large, with a high rate of rise, but inhibits in tissue with low rates of rise in the control.

The probable explanation for the general pattern emerged from an experiment in which the efflux transients were compared in the presence and absence of U73122 in the normal solution containing 0.1 mM Ca²⁺ and in Ca²⁺-free bathing solution, with the results shown in Fig. 5. (Ca²⁺ was removed from the walls by treatment with Ca²⁺-free solution containing 2 mM BAPTA, but the tissue was then transferred to Ca²⁺-free solution without BAPTA.) In the presence of Ca²⁺, the efflux transient appears to be enhanced by the addition of U731222, despite the low rate of rise in the control tissue (0.8 ± 0.2 h⁻²). However, there is also in this experiment significant stimulation of efflux in U73122 in the absence of ABA, and the increase attributable to ABA in the presence of U73122 is small; it is therefore argued that U73122 inhibits the ABA effect, as in two other experiments with low rates of rise in the control. However, in the absence of external Ca²⁺, both the stimulation by U73122 in the absence of ABA and the ABA-induced efflux transient in the presence of U73122 disappear. This suggests that U73122 can induce Ca²⁺ influx in the absence of ABA and that U73122 has multiple effects and does not simply inhibit phospholipase C. Such multiple effects have also been observed in animal systems (25). The inability to generate a transient in the presence of U73122 in the absence of Ca²⁺ may simply reflect a quantitative shortfall when the cADPR pathway is the only one available or could indicate a need for Ca²⁺ influx associated with this pathway. The results also suggest that high rates of rise in the control are associated with greater capacity for ABA-induced Ca²⁺ influx, which can be potentiated by U73122, and that in tissue in which the control peaks have low values for the rate of rise, Ca²⁺ influx is not a major contributor.

Figure 5.

Effect of 1 μM U73122 on the efflux transient induced by 0.1 μM ABA in the presence and absence of Ca²⁺. Rate constant for efflux plotted against time. Each point shows the mean of four replicate strips. (Errors, not shown, 5–25%.) ABA was added at zero time after 31–33 min of washout in the absence of ABA; 1 μM U73122 was added 6–7 min before ABA; 0 Ca²⁺ for 21–25 min before ABA (7–9 min in 2 mM BAPTA followed by 0 Ca²⁺ but without BAPTA). ●, control; ■, U73122; □, U73122, no ABA.

The effect of nicotinamide was also tested in the absence of external Ca²⁺, and the results contrast with those observed with U73122 in the absence of Ca²⁺. Nicotinamide inhibits the efflux transient in the presence and absence of external Ca²⁺, by a somewhat larger factor in the absence of Ca²⁺ but the peak is not abolished in nicotinamide in the absence of Ca²⁺ (Fig. 6). Thus when the cADPR pathway is inhibited and internal release must be attributed to the phospholipase C pathway, Ca²⁺ influx is not required to generate an efflux transient.

Figure 6.

Effect of 20 mM nicotinamide on efflux transient induced by 0.1 μM ABA in the presence and absence of Ca²⁺. Rate constant for efflux was plotted against time. Each point shows the mean of four replicate strips. (Errors, not shown, 5–30%.) ABA was added at zero time after 40–43 min of washout in the absence of ABA; 20 mM nicotinamide was added 17–21 min before ABA; 0 Ca²⁺ for 30 min before ABA (10 min in 2 mM BAPTA followed by 0 Ca²⁺ but without BAPTA). ○, Ca²⁺, control; ●, Ca²⁺, nicotinamide; □, 0 Ca²⁺, control; ■, 0 Ca²⁺, nicotinamide.

Contribution of Ca²⁺ Influx to the Generation of ABA-Induced Efflux Transients.

There are two questions to consider: first, whether external Ca²⁺ is required for the ABA-induced efflux transient, and second, whether Ca²⁺ influx makes a significant contribution to ABA-induced increase in cytoplasmic Ca²⁺. In all, 10 comparisons were made of the ABA-induced efflux transients, with and without external Ca²⁺, and in 8 of these, ABA stimulated vacuolar efflux in the absence of external Ca²⁺. Of these, 4 were at 1 or 10 μM ABA, including one in the presence of nicotinamide, and 4 at 0.1 μM ABA, including one in the presence of nicotinamide (already shown in Fig. 6). Thus in general there is no requirement for external Ca²⁺ for the generation of an efflux transient. There were two exceptions to this where removal of external Ca²⁺ abolished the efflux transient. The first is the experiment already shown in Fig. 5, where external Ca²⁺ was required in the presence of U73122 but not in its absence. The second was an experiment in which the tissue was relatively insensitive to ABA, and only a small transient was generated at 0.2 μM ABA (in only 2 of 4 strips); in this instance, the transient was abolished by removal of external Ca²⁺. Thus it would appear that there can be conditions and tissue states where Ca²⁺ influx does make an essential contribution to the increase in cytoplasmic Ca²⁺, presumably when the capacity for internal release is, for whatever reason, restricted.

The effect of removal of external Ca²⁺ on the form and timing of the efflux transient was variable and once again appeared to be correlated with the rate of rise in the control tissue, but in the reverse sense from the correlation seen for the effect of U73122 already discussed. In two experiments with high rates of rise in the control, removal of external Ca²⁺ inhibited the efflux transient, suggesting Ca²⁺ influx did make a significant contribution to increase in cytoplasmic Ca²⁺. Thus in the first experiment, in which the control rate of rise in the presence of Ca²⁺ was 9.4 ± 3.4 (n = 4) h⁻², the removal of external Ca²⁺ inhibited the generation of an efflux transient, lengthening the lag from 2 ± 0 (n = 4) min to 9.5 ± 1.4 (n = 4) min and reducing the rate of rise to 4.2 ± 3.0 (n = 4) h^−2.; also in this tissue, the removal of external Ca²⁺ abolished the minimal efflux transient generated at 0.2 μM ABA, as already discussed. The second experiment, in which the control rate of rise was 22.5 ± 6.3 (n = 4) h⁻², gave similar inhibition on removing external Ca²⁺, reducing the rate of rise by 49% but in this case without introducing a lag; in the same experiment, addition of 1 mM Ba²⁺ had a much more dramatic effect, reducing the rate of rise to 1.9 ± 0.5 (n = 4) h⁻² and introducing a lag of 11 ± 2 min. The stronger inhibition by adding Ba²⁺ indicates there is an effect of Ba²⁺ on the internal release of Ca²⁺ and not simply a competitive inhibition of Ca²⁺ influx. This was confirmed by a further experiment in which the efflux transient was shown still to be sensitive to Ba²⁺ in the absence of external Ca²⁺.

By contrast, in tissue with lower rates of rise in the control in the presence of Ca²⁺, removal of external Ca²⁺ could either enhance the transient or have very little effect. In three comparisons, in tissue with control rates of rise in the range 0.8–1.9 h⁻², the removal of Ca²⁺ had relatively small effects on the transients (as in Fig. 5 in the absence of U73122 or in Fig. 6, with or without nicotinamide), but in one other experiment in which the control rate of rise was 2.7 ± 0.6 (n = 4) h⁻² in 1 μM ABA, removal of external Ca²⁺ significantly enhanced the efflux transient, increasing the rate of rise at both 1 and 0.1 μM ABA. Thus, only in tissue with high rates of rise in the control peak does removal of Ca²⁺ give substantial inhibition of the efflux transient, suggesting that in these conditions, Ca²⁺ influx does make a major contribution to the increase in cytoplasmic Ca²⁺, whereas in conditions with low rates of rise in the control, internal release rather than influx makes the major contribution. The same conclusion was reached by considering the conditions in which U73122 enhances rather than inhibits the transient.

Discussion

Removal of external Ca²⁺ may inhibit, but does not abolish, the ABA-induced efflux transient, and therefore although Ca²⁺ influx may contribute to ABA-induced rise in cytoplasmic Ca²⁺, it is not a prerequisite. The results also show that the two inhibitors, U73122 and nicotinamide, which interfere with the generation of the two recognized second messengers for release of Ca²⁺ from internal stores, IP₃ and cADPR, respectively, can affect the time course of the vacuolar efflux transient by introducing a delay before its initiation and reducing the rate of rise. Taken together, the results suggest that the threshold to be reached before Rb⁺ release is triggered is of the level of cytoplasmic Ca²⁺, and that the form and timing of the efflux transient is an indicator of the level achieved. The results also suggest that all three processes, influx of Ca²⁺ from outside and two processes of internal Ca²⁺ release (downstream of phospholipase C and cADPR respectively), can contribute to ABA-induced increase in cytoplasmic Ca²⁺. This is consistent with the fact that both candidate channels to carry K⁺ currents at the tonoplast, the K⁺-specific VK channel and the less-specific (cation) SV channel, are Ca²⁺ activated (2, 3).

Ba²⁺ is the fourth agent affecting the time course of the vacuolar efflux transient, and its effect seems to be 2-fold, as a competitive inhibitor of Ca²⁺ influx in tissue where Ca²⁺ influx does make a major contribution but Ba²⁺ inhibits also in the absence of external Ca²⁺, implying a second effect on internal Ca²⁺ release. Such inhibition can be expected, because the release of Ca²⁺ from stores is conditional on charge balance by some other ion, and there is direct evidence by using either tetraethyl ammonium or Ba²⁺ as K⁺ channel blockers of dependence of IP₃-triggered Ca²⁺ release on a balancing K⁺ flux in both animals and plants (26–28). It is therefore plausible to expect that Ba²⁺, acting in its capacity as a K⁺ channel blocker, could block Ca²⁺ release from internal stores, thereby interfering with the initiation of the vacuolar efflux transient. On this view, the K⁺ channels are required to carry a current to balance the potential Ca²⁺ flux and, depending on their relative degree of activation and current-carrying capacity, either K⁺ or Ca²⁺ channels could limit the coupled process. In conditions with the potential for a high rate of Ca²⁺ release, inhibition of the charge-balancing K⁺ flux may render it the rate-limiting flux, and the efflux transient will be impaired by the addition of Ba²⁺. However, in conditions where the ability to generate an efflux transient is already strongly impaired, implying a much reduced Ca²⁺ release, partial inhibition of the K⁺ channels may not transfer the rate limitation of the coupled process to the K⁺ channels, and the addition of Ba²⁺ may not further impair the transient. Thus where the transients in the control tissue in the absence of Ba²⁺ are already strongly reduced, with the lowest rates of rise and longest lags, the addition of Ba²⁺ will not produce further inhibition, as is observed.

The form of transients under different conditions indicates that development of the ability to generate an efflux transient is a two-stage process, in which there are two critical thresholds of cytoplasmic Ca²⁺; crossing the first threshold allows the generation of a minimal transient after a long lag period, but if conditions are better (higher ABA or reduced Ba²⁺), the lag is first reduced, and if the second threshold is crossed, a transient with a rapid rate of rise is observed with no lag. The two thresholds may reflect activation of two different channels for vacuolar K⁺(Rb⁺) release, perhaps as a consequence of the different thresholds of Ca²⁺ for activation of the VK and SV channels, with VK activating at much lower concentrations of Ca²⁺ than SV (2, 3). The relative importance of the three processes contributing to increase in cytoplasmic Ca²⁺ appears to differ in the two forms of transient. The full response is observed at high ABA (>1 μM) in the presence of Ca²⁺ and is characterized by zero lag and high rate of rise. In this condition, the transient was inhibited by two treatments: by removal of external Ca²⁺ and, more effectively, by adding Ba²⁺, likely to compete with Ca²⁺ in influx but also likely to inhibit both processes of internal release. In these conditions, inhibition of only one process of internal release by addition of nicotinamide did not inhibit the transient. Thus both influx and internal release contribute to increase in cytoplasmic Ca²⁺, but the full transient can be initiated even when one of the internal processes is inhibited, suggesting there is spare capacity. Further, in tissue with a high rate of rise in the control, the peak is enhanced by U73122, but such enhancement requires external Ca²⁺. The results therefore suggest that in this state the tissue has a high capacity for ABA-induced Ca²⁺ influx, and that U73122 potentiates ABA-induced Ca²⁺ influx. Side effects by which U73122 can potentiate increase in cytoplasmic Ca²⁺ have also been observed in animal tissue (25).

As conditions for initiation of the transient are impaired (by reducing the ABA concentration or adding Ba²⁺), the rate of rise first decreases to around 1–2 h⁻² without introduction of a lag; further inhibition results in an increasing lag but without further change in the rate of rise. In the second stage, the peak is not inhibited by removal of external Ca²⁺, but inhibition of either of the processes of internal Ca²⁺ release by addition of nicotinamide or U73122 can interfere with the generation of the transient, suggesting that internal release rather than Ca²⁺ influx makes the major contribution to increase in cytoplasmic Ca²⁺. In such tissue, the transient was effectively abolished in three conditions: in the presence of both nicotinamide and U73122 (Fig. 4), in U73122 in the absence of external Ca²⁺ (Fig. 5), and in the presence of 40 mM nicotinamide, although in this instance a small transient was restored by the addition also of U73122 (assumed to produce a Ca²⁺ influx). Thus the threshold of Ca²⁺ required to initiate the transient can be achieved by a combination of fluxes from the three sources. Both pathways of internal release appear to make significant contributions, and in this second stage of inhibition the transient is abolished by restriction to one pathway only (in 40 mM nicotinamide or U73122 in the absence of external Ca²⁺ and U73122-induced Ca²⁺ influx). The fact that, in conditions where internal release is important, the inhibition of these two pathways can abolish the transient suggests there are no other major contributors to the Ca²⁺ increase. The question then arises of the relation to the new signaling pathway involving IP₆, shown to inhibit the inward K⁺ channel in a Ca²⁺-dependent manner much more effectively than IP₃ (20). One possibility is that IP₆ is in fact a product of IP₃ and is therefore subsumed into the phospholipase C pathway.

The interrelations of the three pathways need to be established by further work. Ca²⁺ influx through the hyperpolarization-activated Ca²⁺ channel, also activated by ABA, triggers release of Ca²⁺ from internal stores (6). The transient increase in cytoplasmic Ca²⁺ produced by a hyperpolarizing pulse was inhibited by ryanodine, suggesting involvement of cADPR-triggered channels whereas, somewhat surprisingly, treatment with either neomycin sulfate or heparin, inhibitors of phospholipase C, enhanced the Ca²⁺ peaks after a hyperpolarizing pulse (7). These effects have parallels in the effects observed here, that nicotinamide inhibits the efflux transient, but that U73122 can, in some conditions, enhance the efflux transient. Such enhancement is associated with tissue in which the control rate of rise is high, and Ca²⁺ influx appears to make a significant contribution.

Maximal transients, with high rates of rise and zero lags, are inhibited by removal of external Ca²⁺, suggesting influx plays a major role. The range of values for the rate of rise in transients with zero lags may therefore reflect differences in the capacity for ABA-induced Ca²⁺ influx. This implies that the first stage of inhibition, to a reduced rate of rise around 1–2 h⁻² but zero or minimal lag, involves reduction in ABA-induced Ca²⁺ influx. This can be achieved by lowering the ABA concentration from 1 or 10 μM to 0.1 or 0.2 μM or by adding 0.5 mM Ba²⁺. The results therefore suggest that activation of Ca²⁺ influx requires a higher concentration of ABA than does activation of release of Ca²⁺ from internal stores, implying parallel signaling chains from ABA, one involving Ca²⁺ influx (which may also lead to downstream internal Ca²⁺ release) and one involving internal release independent of Ca²⁺ influx.

In summary, the results provide evidence that ABA-induced release of vacuolar ions is Ca²⁺ mediated, a consequence of increase in cytoplasmic Ca²⁺ arising from both influx of Ca²⁺ and its release from internal stores, in pathways dependent on both phospholipase C and cADPR. ABA has multiple effects. Low concentrations of ABA (0.1 or 0.2 μM) activate both processes of internal release and generate a minimal efflux transient after a lag period, with a low rate of rise, but in these conditions Ca²⁺ influx plays a minor role. Concentrations of 1 μM ABA and above also activate Ca²⁺ influx, generating a full efflux transient, with zero lag and a rapid rate of rise. Because discharge of vacuolar ions is one of the essential changes required for ABA-induced stomatal closure, the results suggest that increase in cytoplasmic Ca²⁺ should be required for such closure, even if ABA-induced changes to other ion channels may be Ca²⁺ independent.

Abbreviations

ABA

abscisic acid

ER

endoplasmic reticulum

IP₃

inositol 1,4,5-trisphosphate

IP₆

inositol hexakisphosphate

cADPR

cyclic ADP-ribose