Inositol hexakisphosphate mobilizes an endomembrane store of calcium in guard cells

Abstract

myo-Inositol hexakisphosphate (InsP₆) is the most abundant inositol phosphate in cells, yet it remains the most enigmatic of this class of signaling molecule. InsP₆ plays a role in the processes by which the drought stress hormone abscisic acid (ABA) induces stomatal closure, conserving water and ensuring plant survival. Previous work has shown that InsP₆ levels in guard cells are elevated in response to ABA, and InsP₆ inactivates the plasma membrane inward K⁺ conductance (I_K,in) in a cytosolic calcium-dependent manner. The use of laser-scanning confocal microscopy in dye-loaded patch-clamped guard cell protoplasts shows that release of InsP₆ from a caged precursor mobilizes calcium. Measurement of calcium (barium) currents I_Ca in patch-clamped protoplasts in whole cell mode shows that InsP₆ has no effect on the calcium-permeable channels in the plasma membrane activated by ABA. The InsP₆-mediated inhibition of I_K,in can also be observed in the absence of external calcium. Thus the InsP₆-induced increase in cytoplasmic calcium does not result from calcium influx but must arise from InsP₆-triggered release of calcium from endomembrane stores. Measurements of vacuolar currents in patch-clamped isolated vacuoles in whole-vacuole mode showed that InsP₆ activates both the fast and slow conductances of the guard cell vacuole. These data define InsP₆ as an endomembrane-acting calcium-release signal in guard cells; the vacuole may contribute to InsP₆-triggered Ca²⁺ release, but other endomembranes may also be involved.

InsP₆ (myo-inositol hexakisphosphate), a physiological signal generated in guard cells in response to the drought-stress hormone abscisic acid (ABA), is a potent inhibitor of the K⁺-inward rectifier (I_K,in) (1), one of the fundamental elements of the control of guard cell turgor and hence of stomatal aperture. Whereas much is known of the physiological circumstances and molecular mechanisms by which the second messenger, inositol 1,4,5-trisphosphate [Ins(1,4,5)P₃], acts almost ubiquitously as a calcium-mobilizing signal (2, 3), little is known of the circumstances or mode of action of InsP₆. In animal cells, Ins(1,4,5)P₃ functions to mobilize calcium by acting as a ligand at the cytosolic face of ligand-gated endomembrane calcium-release channels. To date, no candidate Ins(1,4,5)P₃ receptor has yet been identified in the Arabidopsis or yeast (Saccharomyces cerevisiae) genomes. Nevertheless, Ins(1,4,5)P₃ elevates [Ca²⁺]_cyt when released from caged precursor in guard cells, with consequential inhibition of I_K,in and decrease in aperture (4, 5). The calcium reservoir, claimed initially to be the vacuole (6), could also include nonvacuolar stores, because most of the InsP₃-binding capacity in the cell is in nonvacuolar membranes (7). Our ignorance of the role on InsP₆ in signaling chains stems in part from a general lack of responsiveness of InsP₆ to extracellular stimuli in systems other than guard cells (1), fission yeast (8), and hippocampal neurons (9). It is striking that the levels of InsP₆ quickly increase in hyperosmotically stressed Schizosaccharomyces pombe (8), and that in plants, treatment with ABA also quickly increases InsP₆ levels in guard cells (1). It is clearly important to establish the role of InsP₆ in ABA-induced signaling chains. The stomatal guard cell has emerged as a uniquely tractable experimental system in which to study the function of InsP₆ in osmotic-stress biology. By applying patch–clamp electrophysiology in whole-cell and -vacuole mode, we have investigated the cellular effectors of InsP₆ signaling. In guard cell protoplasts (GCPs), inhibition of I_K,in by InsP₆ does not occur in the presence of internal calcium chelators such as EGTA or BAPTA (1), suggesting that InsP₆ is a calcium-mobilizing signal. In the present work, by use of laser-scanning confocal microscopy in dye-loaded patch-clamped GCPs, we show first that laser uncaging of InsP₆ does indeed mobilize calcium and then investigate the source of this increase.

InsP₆-induced calcium transients in guard cells could arise from calcium entry through the plasma membrane or by calcium release from internal stores. In pancreatic β cells, InsP₆ activates an l-type calcium current (10). Recent studies (11–14) have shown that plants harbor calcium-permeable channels [hyperpolarization-activated calcium current (I_Ca)] that activate, in contrast to those in animal cells on membrane hyperpolarization, and that in guard cells, these channels increase their opening probability in response to ABA (13). We have tested the effect of InsP₆ on I_Ca and show that the InsP₆-induced increase in cytoplasmic Ca²⁺ is not a consequence of Ca²⁺ influx from outside but of release of Ca²⁺ from internal stores. Study of tonoplast ion channels in patch-clamped isolated vacuoles shows activation of both fast-activating [fast vacuolar (FV)] and slowly activating [slow vacuolar (SV)] channels by InsP₆.

Methods

Protoplast Isolation. Vicia faba L. cv. (Bunyan) Bunyan Exhibition was grown on vermiculite and GCPs isolated from abaxial epidermal strips of 3- to 4-week-old leaves as described (15). Epidermal strips were floated on medium containing 1.8–2.5% of Cellulase Onozuka RS (Yacult Honsha, Osaka), 1.7–2% Cellulysin (Calbiochem, Behring Diagnostics), 0.026% Pectolyase Y-23, 0.26% BSA, and 1 mM CaCl₂ (pH 5.6 and osmolality 360 mOsm·kg^–1 adjusted with mannitol), and incubated at 28°C, with gentle shaking. After 120–150 min, released protoplasts were passed through a 25-μm mesh and kept on ice for 2–3 min before being centrifuged at 100 × g for 4 min (at room temperature). The pellet of GCPs was resuspended and kept on ice in 1 or 2 ml of fresh medium containing 0.42 M mannitol, 10 mM (Mes), 200 μM CaCl₂, and 2.5 mM KOH (pH 5.55 and osmolality 466 mOsm·kg^–1).

Vacuole Isolation. Freshly made protoplasts were allowed to settle to the bottom of the recording chamber and stick well to the glass, then given a hypoosmotic shock (200 mOsm·kg^–1) by a continuous perfusion of a solution containing (in mM): 100 KCl, 10 Hepes, 2 EGTA, pH adjusted to 8 (KOH). Perfusion was stopped when >50% of the protoplasts had released their vacuoles (usually in ≈20–30 min). Vacuoles were then perfused with a suitable external solution.

Solutions. Protoplasts (or vacuoles) were placed in a 0.5-ml chamber, allowed to settle, then perfused continuously at flow rates of 1–2 ml per min. The bath and pipette solutions were chosen appropriately for the particular membrane conductance measured. Four different conductances were studied: the inward K⁺-rectifier (I_K,in) and the hyperpolarization-activated calcium current (or I_Ca) located at the plasma membrane, and the FV vacuolar current and the SV vacuolar current at the vacuolar membrane. For recording of I_K,in, the solutions contained (in mM): 5 KCl, 2 MgCl₂, 2 K₄EGTA, 10 Mes (KOH) at pH 5.5, osmolality 480 mOsm·kg^–1 (sorbitol) in the bath; and 92 K-glutamate, 2 MgATP, 3.4 CaCl₂, 5 K₄EGTA, 10 Hepes (KOH) at pH 7.5, osmolality 520 mOsm·kg^–1 (sorbitol) in the pipette. For I_Ca, the solutions contained 100 BaCl₂, 10 Mes (Tris) at pH 5.5, osmolality 230 mOsm·kg^–1 in the bath, and 4 K₄EGTA, 10 BaCl₂, 10 Hepes (Tris) at pH 7.5, osmolality 230 mOsm·kg^–1 in the pipette. For recording of both vacuolar channels, the bath contained 200 KCl, 2 MgCl₂, 0.1 CaCl₂, 10 Hepes (Tris) at pH 7.5, osmolality 440 mOsm·kg^–1, with the further addition of 4 mM K₄EGTA for the FV channel; for both SV and FV, the pipette contained 200 KCl, 10 CaCl₂, 10 Mes (KOH) at pH 5.5, osmolality 460 mOsm·kg^–1.

Current-Voltage (I–V) Recording and Analysis. Patch pipettes (5–10 MΩ) were pulled from Kimax-51 glass capillaries (Kimble/Kontes, Vineland, NJ) by using a two-stage puller (Narishige PP-83, Tokyo). Experiments were done at room temperature (20–22°C) by using the standard whole-cell patch–clamp techniques, with an Axopatch 200B Integrating Patch Clamp amplifier (Axon, Union City, CA). Voltage commands and simultaneous signal recordings and analysis were assessed by a microcomputer connected to the amplifier via a multipurpose input/output device (Digidata 1200A, Axon) using pclamp 8.0 software (Axon). After gigaohm seals were formed, the whole-cell configuration was achieved by gentle suction, and the membrane was immediately clamped to a holding voltage (h_v) dependent on the membrane conductance to be recorded. For I_K,in, FV, and SV, h_v was set close to the Nernst potential for K⁺ (E_K). For I_Ca,h_v was set close to E_Cl. GCPs or their vacuoles were left for 3–5 min before starting any current measurements. All current traces shown were low-pass filtered at 2 kHz before analog-to-digital conversion and were uncorrected for leakage current or capacitative transients. Membrane potentials were corrected for liquid junction potential as described in ref. 16. Nernst potentials were calculated after correction for ionic activities (estimated by geochem-pc software, public domain freeware). I–V relationships for I_Ca, FV, and SV were plotted as steady-state currents vs. test potential, but I_K,in was plotted as the time-dependent current (i.e., the steady-state current minus the instantaneous current) vs. test potential. Results are expressed as mean ± SEM.

Confocal Microscopy. Laser-scanning confocal microscopy of dye-loaded patch-clamped GCPs was used to determine the consequences of laser uncaging of InsP₆ on cytosolic calcium. GCPs, isolated from either V. faba or Solanum tuberosum as described above (see also ref. 17), loaded through the patch pipette with P (4, 5)-(o-Nitrobenzyl) inositol hexakisphosphate [P(4,5)-NBZ InsP₆] (100–200 μM) (18) and a calcium-sensitive single wave-length emission dye, Calcium Green-1 (100 μM; Molecular Probes Europe, Leiden, The Netherlands). A Leica TCS-NT-UV CLSM was used to uncage InsP₆ and to follow changes in [Ca²⁺]_i. InsP₆ was uncaged by simultaneous exposure to the 354- and 361-nm lines of an UV laser (10–30 randomly chosen spots of 10-ms duration each). Calcium Green-1 was excited by using the 488-nm line of an ArKr laser with the other lines attenuated to zero. A reflectance short-pass filter of 510 nm was used, and emitted fluorescence was monitored between 515 and 545 nm by using a water immersion 1.2 N.A. objective lens. Ten images of the Calcium Green-1 fluorescence were collected at a resolution of 512 × 512 pixels before uncaging and a further 200 after uncaging at 200-ms intervals. Images were analyzed by using metamorph software (Universal Imaging, West Chester, NY). Each image was pseudocolor-coded, and a scale bar was generated to indicate low (blue) to high [Ca²⁺] (red). GCPs were bathed in solutions (in mM): 10 KCl/0.5 CaCl₂/2 MgCl₂/10 Mes, pH 5.5 (KOH)/sorbitol to give 480 mOsm·kg^–1. The patch pipette contained (in mM): 92 K⁺ glutamate, 2 MgCl₂,2K₂ATP, 10 Hepes at pH 7.5 (KOH), and sorbitol to give 520 mOsm·kg^–1 (some experiments contained 3.4 CaCl₂/5 EGTA to give resting [Ca²⁺]_i 100 nM).

Results

InsP₆, Released from Caged Precursor, Mobilizes Ca²⁺ in Patch-Clamped GCPs. GCPs were loaded with Calcium Green-1 and caged InsP₆ to determine the effects on cytoplasmic Ca²⁺ of uncaging InsP₆. In six of nine dye- and P(4,5)-NBZ InsP₆-loaded GCPs, we observed transient increases in Calcium Green-1 fluorescence on uncaging. One such experiment is shown in Fig. 1 A and B, in which [Ca²⁺]_cyt levels increased to a maximum within1sof uncaging, persisting at maximal level for ≈5 s before decaying to resting levels within 25–30 s. The time course varied in different experiments, and Fig. 1C shows another time course, with a faster return to preflash levels within 6 s, together with a nonflashed control. UV irradiation was without effect in dye-only loaded protoplasts and also in protoplasts coloaded with caged ATP (not shown). The nature of transients induced by InsP₆ release in this study is similar to the transients induced by ABA in GCPs from V. faba (19) and are among the most transitory [Ca²⁺]_cyt signals yet recorded in single protoplasts of a higher plant. We have also investigated the consequence of repetitive uncaging events. Successive release events at intervals of 2–5 min elicited immediate and repetitive excursions in [Ca²⁺]_cyt levels but of diminishing amplitude (not shown). It is therefore clear that uncaging of InsP₆ did not exhaust the reservoir of calcium from which the excursions in [Ca²⁺]_cyt level originate. Similarly, ionomycin treatment of a dye- and P(4,5)-NBZ InsP₆ coloaded and UV laser-responsive protoplast resulted in a massive increase in dye fluorescence, assumed to result from entry of calcium from the bath solution (not shown). This indicates clearly that the GCPs used were able to regulate their internal free calcium both before and after uncaging of InsP₆. These data identify InsP₆ as a calcium-mobilizing agent and provide an unambiguous explanation of the [Ca²⁺]_cyt dependence of inhibition of I_K,in by InsP₆.

Fig. 1.

Transient excursion in cytosolic free Ca²⁺ observed on uncaging of InsP₆ from its caged form. (A) Confocal imaging of a Vicia GCP loaded from a patch pipette containing both [P(4,5)-NBZ InsP₆] (100 μM) and Calcium Green-1 (100 μM). Successive images recorded before and after laser irradiation are shown. Each image was pseudocolor-coded, shown in the scale bar (bottom) indicating low (blue) to high (red) [Ca²⁺]. (B) Integrated fluorescent intensity against time for the images shown. (C) Integrated fluorescent intensity in a second cell flashed at zero time (▪) and in a nonflashed control (□). GCPs were bathed in solutions (mM): 10 KCl/0.5 CaCl₂/2 MgCl₂/10 Mes at pH 5.5 (KOH)/sorbitol to give 480 mOsm·kg^–1. The patch pipette contained (mM) 92 K⁺ glutamate, 2 MgCl₂,2K₂ATP, 10 Hepes at pH 7.5 (KOH), and sorbitol to give 520 mOsm·kg^–1 (some experiments contained 3.4 CaCl₂/5 EGTA to give resting [Ca²⁺]_i ≈100 nM).

InsP₆ Has No Effect on Ca²⁺ Currents at the Plasmalemma. To measure I_Ca, GCPs isolated from V. faba were bathed in 100 mM BaCl₂, pH 5.5 (Mes/Tris), with 10 mM BaCl₂, pH 7.5 (Hepes/Tris) in the pipette. The use of barium instead of calcium has two advantages: that Ba permeates calcium channels more freely than Ca, and that Ba blocks potassium channels more efficiently than Ca. Moreover, in these solutions E_Ba (+29 mV) and E_Cl (–58 mV) are far removed from each other, allowing easy identification of the conducting species. GCPs exhibited large whole-cell hyperpolarization-activated barium-permeable channels (I_Ba) (Fig. 2). Large I_Ba were observed even in the absence of stimuli such as H₂O₂ or ABA, which were essential for channel activity in Arabidopsis GCPs (14). We also confirmed substantial activation by ABA (up to 5-fold) and H₂O₂ (up to 16-fold) of whole-cell currents in Vicia protoplasts (not shown). In Fig. 2A (Lower Inset), Ba-permeable current traces activated instantly without a time-dependent kinetic component, irrespective of voltage (–198 to +42 mV and up to 4-s square wave test pulses). The I–V relationship of whole-cell I_Ba showed a weakly voltage-dependent rectification with current reversing near the predicted Nernstian equilibrium for barium, E_Ba (Fig. 2). External addition of GdCl₃ (20–100 μM) suppressed these currents (Fig. 2B), whereas in other experiments designed to measure I_K,in, similar concentrations of GdCl₃ were without effect (not shown). LaCl₃ also inhibited I_Ba but less potently than GdCl₃. Verapamil, a calcium-channel blocker, was, at 50 μM concentration, much less effective than GdCl₃ or LaCl₃. Moreover addition of 1 mM CsCl, which blocks I_K,in completely, has no or little effect on I_Ba (Fig. 2B Inset). More importantly, we show (Fig. 2C) that InsP₆ (in the range 1–20 μM in the patch pipette) was without major effect on I_Ba (n = 16, control; n = 14, InsP₆). This contrasts markedly with the effects of ABA on I_Ba (13) and also with the strong inhibition of I_K,in manifest in Vicia at submicromolar InsP₆ concentrations (1). The present results therefore indicate that InsP₆-mediated [Ca²⁺]_cyt-dependent inhibition of I_K,in is not a consequence of InsP₆-induced calcium influx via hyperpolarization-activated calcium current.

Fig. 2.

I_Ba is activated by membrane hyperpolarization, blocked by gadolinium (not cesium), and only slightly affected by InsP₆. (A) Whole-cell I–V relationship of I_Ba showing a weak voltage dependence and current reversal at V ≈+30 mV, which coincides with the calculated theoretical E_Ba (arrow). See text for detailed bath and pipette solutions. (Upper Inset) Current response to a 4-s voltage ramp from +42 to –198 mV. (Lower Inset) A family of I_Ba currents in response to square voltage test pulses from +40 to –200 mV (in 20-mV steps); the broken line indicates zero current. The I–V curve is a superimposition of current measurement obtained in response to the voltage ramp protocol (continuous line), and the current measurements in response to the square test pulse protocol (•); each data point is a mean value of the current measured at steady state). (B) I–V curves of I_Ba (4-s voltage ramp) before and after perfusion with 0.1 mM GdCl₃.(Inset) I–V curves of I_Ba (4-s voltage ramp) before and after perfusion with 1 mM CsCl. (C) I–V curves of I_Ba (square voltage pulse; 10 mV steps from +50 to –200 mV) in the presence (▪; n = 14) and absence (•; n = 16) of internal InsP₆.

Further confirmation of the absence of an apoplast-derived calcium influx component in InsP₆-dependent inhibition of I_K,in was afforded by measuring I_K,in in the absence of external calcium. Fig. 3 shows one such experiment performed with 2 mM EGTA in the bath solution. Loading the cell with InsP₆ (0.1–1 μM in the patch pipette), we observed an up to 80% inhibition of I_K,in in 4–6 min (n = 4). We conclude from the foregoing data (Figs. 1, 2, 3) that the InsP₆-induced calcium mobilization and the [Ca²⁺]_cyt dependency of inhibition of I_K,in resides with an InsP₆-sensitive endomembrane store of calcium.

Fig. 3.

InsP₆ inactivates I_K,in in the absence of external calcium. Current recordings from a GCP 1 min (•) and 5 min (▪) after breaking the patch seal to go whole cell. The patch pipette contained 0.1 μM InsP₆. Time-dependent current values (indicated by arrows) were plotted on the corresponding I–V curve; note the inhibition of I_K,in by InsP₆ but absence of effect on I_K,out. The bath contained (mM) 5 KCl, 2 K₄EGTA, 2 MgCl₂, 10 Mes at pH 5.5 (KOH), and sorbitol to give 480 mOsm·kg^–1. The patch pipette contained (mM): 92 K⁺ glutamate, 3.4 CaCl₂/5 EGTA (resting [Ca²⁺]_i ≈ 100 nM), 2 MgATP, 10 Hepes at pH 7.5 (KOH), and sorbitol to give 520 mOsm·kg^–1.

InsP₆ Activates Vacuolar Ion Channels. By patch-clamping of isolated vacuoles, we tested whether the endomembrane-store-dependent inhibition of I_K,in could be explained by mobilization of vacuolar calcium. Three vacuolar conductances have previously been characterized. The fast vacuolar (FV) conductance is active only at low resting calcium concentration and is inhibited by high cytoplasmic calcium concentration (20–22). On the other hand, high [Ca²⁺]_cyt activates a SV channel (21, 23), which is permeated by both potassium and calcium (24), and a vacuolar potassium channel (23).

Vacuoles were perfused with the recording solution, containing either low calcium (2–200 nM) to measure FV or high calcium (100 μM) to measure SV. In low external calcium and in the whole-vacuole mode, excursion of the membrane voltage between –100 and +100 mV initiated a weakly voltage-dependent fast-activating conductance in both inward and outward directions. Challenge of vacuoles with cytosolic InsP₆ (1–5 μM in the perfusing solution) enhanced this conductance without effect on the time-dependent kinetic component (Fig. 4A Left); the activation was reversible on washout of InsP₆ (Fig. 4A Right). The effects of InsP₆ were further confirmed in cytosolicside-out patches (not shown), indicating that InsP₆-dependent activation of the FV conductance is a membrane-delimited process. The I–V plots of vacuolar conductances measured at low cytosolic calcium are shown (Fig. 4B).

Fig. 4.

InsP₆ activates vacuolar conductances measured in either low or high cytosolic calcium. (A) Whole vacuolar current traces from two different patch–clamp experiments obtained in low cytosolic calcium (2 nM–200 nM). (Left) Vacuolar current traces before (•) and 10 min after (▪) addition of 1 μM InsP₆. (Right) Recovery of current to prestimulated levels after washout of InsP₆ (stimulation was achieved in this experiment by 20-min treatment with 5 μM InsP₆). (B) I–V plot (+100 to –100 mV in 20-mV steps) measured from whole vacuolar currents obtained in low cytosolic calcium before (•; n = 3) and after (▪; n = 5) addition of 1–5 μM InsP₆. (C) Whole vacuolar current traces and corresponding I–V curves (+100 to –100 mV in 20-mV steps) obtained in high cytosolic calcium (100 μM) before (•) and after (▪) addition of 5 μM InsP₆. The effect of InsP₆ is reversible (▴). (A and B) Bath contained (in mM): 200 KCl, 2 MgCl₂, 0.1 CaCl2, 4 EGTA (≈2 nM Free Ca²⁺), 10 Hepes at pH 7.5 (KOH); patch pipette composition (in mM): 200 mM KCl, 2 mM MgCl₂, 10 mM CaCl₂, 10mM Mes at pH 5.5 (KOH). Solutions in C were as in A and B but with 100 μM external calcium and no added EGTA. All membrane potentials are specified as the potential on the cytosolic side relative to the vacuolar side.

Experiments in high calcium and in the whole-vacuole mode were also carried out to assess the effects of InsP₆ on SV conductance. Addition of InsP₆ (5 μM) activated SV currents. Fig. 4C shows one such experiment where control, InsP₆, and recovery measurements were obtained from the same vacuole. InsP₆ seems to affect mostly the instantaneous component, with no noticeable effect on the time-dependent activation; as with FV, the effect on SV was also fully reversible on washout (n = 3).

Discussion

The measurements of cytoplasmic Ca²⁺ after the flash release of InsP₆ from its cage establish InsP₆ as an effective Ca²⁺-mobilizing agent in guard cells, to be added to the list of such agents already identified. We now have four agents capable of mobilizing Ca²⁺ in guard cells and need to establish their physiological relevance. Both Ins(1,4,5)P₃ and InsP₆ elicit elevations in [Ca²⁺]_cyt in guard cells and both signal to I_K,in, inhibiting I_K,in in a calcium-dependent manner (1, 5). However, we have shown that InsP₆ is ≈100 times more potent as an inhibitor of I_K,in than Ins(1,4,5)P₃ (1). Two other intracellular signaling agents, cyclic adenosine diphosphoribose (cADPR) (25) and sphingosine-1-phosphate (26), have also been shown to modulate [Ca²⁺]_cyt when introduced into guard cells. Evidence has been presented for the involvement of all four agents in the response of guard cells to ABA (1, 25–29), and an important goal for the future is to establish their relative contributions in different conditions.

The [Ca²⁺]_cyt response of GCPs to uncaging of InsP₆ (Fig. 1) is fundamentally different from the response of guard cells to Ins(1,4,5)P₃, cADPR, and sphingosine-1-phosphate. The instantaneous increase of [Ca²⁺]_cyt observed with InsP₆ and the rapid decay to resting level within 30 s or less contrasts markedly with the sustained increases of [Ca²⁺]_cyt (5–10 min) observed after flash photolysis of caged-Ins(1,4,5)P₃ (4) or the 3- to 5-min periodic elevations of [Ca²⁺]_cyt observed with cADPR (25) and sphingosine-1-phosphate (26). However, it is important to note that the method of application of the agent differs in different studies, and the rapid decay in the response to a pulse of InsP₆ may reflect the diffusion of InsP₆ in the patch pipette or its metabolism; the accurate time course of change in InsP₆ levels after stimulation with ABA in intact cells is not established.

InsP₆-induced calcium transients in guard cells could result from calcium entry through plasma membrane-permeable channels or calcium release from internal stores. Our findings that InsP₆ has no effect on I_Ca and that InsP₆ can inhibit I_K,in in the absence of external Ca²⁺ rule out an effect through activation of calcium influx and point to an endomembrane source for calcium release.

In the context of endomembrane calcium, the vacuole constitutes an important pool of mobilizable calcium in higher plants but may not be the dominant signaling pool. Among the different agents reported to increase guard cell [Ca²⁺]_cyt, both Ins(1,4,5)P₃ and cADPR are believed to target endomembranes (25, 30). Thus both agents activate calcium release from the vacuole (30) but also from other endomembranes such as the endoplasmic reticulum (ER) (7, 31). A further Ca²⁺-mobilizing agent in higher plants, nicotinic acid adenine dinucleotide phosphate, is active exclusively at the ER (32). Our results establish that InsP₆ activates two tonoplast ion channels, the FV and SV channels. The activation of the FV channel is reminiscent of the effects of cADPR (25), which activated a fast tonoplast channel inhibited by cytoplasmic Ca²⁺ above ≈600 nM, but which appeared to be permeable to Ca²⁺, because the reversal potential was shifted by increase in vacuolar Ca²⁺. Our results show activation also of the SV channel by InsP₆, and this channel is permeable to Ca²⁺ (24). Thus our results suggest that the vacuole contributes to the InsP₆-triggered Ca²⁺ release, but other endomembranes may also be involved, as is the case for both InsP₃ and cADPR.

Our study further defines the mechanism by which InsP₆ mediates physiological (ABA-dependent) inhibition of I_K,in. The discovery that InsP₆ is a calcium-mobilizing agent and produces inhibition of I_K,in much more effectively than does InsP₃ may yet explain the absence of InsP₃ receptors in the yeast and Arabidopsis genomes; conversion of InsP₃ introduced into the cytoplasm to InsP₆ would produce the effects observed. The work identifies release from endomembrane stores rather than Ca²⁺ influx as the source of InsP₆-triggered increase in cytoplasmic Ca²⁺, and the demonstration that InsP₆ regulates vacuolar release channel(s) provides further evidence that endomembrane channels are cellular targets of InsP₆. It remains to be established whether this reflects a direct interaction with the channels or action via a regulatory agent. That such channels are involved in osmoregulation adds considerably to our understanding of the physiological and cellular signaling function of this, the most enigmatic of inositol phosphates.