Homeostatic control of slow vacuolar channels by luminal cations and evaluation of the channel-mediated tonoplast Ca2+ fluxes in situ.

Ca(2+), Mg(2+), and K(+) activities in red beet (Beta vulgaris L.) vacuoles were evaluated using conventional ion-selective microelectrodes and, in the case of Ca(2+), by non-invasive ion flux measurements (MIFE) as well. The mean vacuolar Ca(2+) activity was approximately 0.2 mM. Modulation of the slow vacuolar (SV) channel voltage dependence by Ca(2+) in the absence and presence of other cations at their physiological concentrations was studied by patch-clamp in excised tonoplast patches. Lowering pH at the vacuolar side from 7.5 to 5.5 (at zero vacuolar Ca(2+)) did not affect the channel voltage dependence, but abolished sensitivity to luminal Ca(2+) within a physiological range of concentrations (0.1-1.0 mM). Aggregation of the physiological vacuolar Na(+) (60 mM) and Mg(2+) (8 mM) concentrations also results in the SV channel becoming almost insensitive to vacuolar Ca(2+) variation in a range from nanomoles to 0.1 mM. At physiological cation concentrations at the vacuolar side, cytosolic Ca(2+) activates the SV channel in a voltage-independent manner with K(d)=0.7-1.5 microM. Comparison of the vacuolar Ca(2+) fluxes measured by both the MIFE technique and from estimating the SV channel activity in attached patches, suggests that, at resting membrane potentials, even at elevated (20 microM) cytosolic Ca(2+), only 0.5% of SV channels are open. This mediates a Ca(2+) release of only a few pA per vacuole (approximately 0.1 pA per single SV channel). Overall, our data suggest that the release of Ca(2+) through SV channels makes little contribution to a global cytosolic Ca(2+) signal.

Introduction

Slow vacuolar (SV) channels are ubiquitous in all tissues of higher plants with several thousand per vacuole (Hedrich ; Schulz-Lessdorf and Hedrich, 1995; Pottosin ). The SV channel is a non-selective cation channel with significant Ca2+ permeability (Ward and Schroeder, 1994; Gradmann ; Allen ; Pottosin ), encoded by the TPC1 (two-pore calcium channel) singleton gene (Peiter ). The SV channel is undoubtedly the best explored and the most abundant vacuolar channel, as shown by electrophysiological and proteomics studies of the tonoplast (Schulz-Lessdorf and Hedrich, 1995; Carter ; Pottosin and Schönknecht, 2007).

Because the vacuole is the largest Ca2+ store in plant cells, there is an ongoing debate as to the level of participation of the SV channel in Ca2+ signalling. For instance, based on the finding that the SV channel is activated by increased cytosolic Ca2+, Ward and Schroeder (1994) proposed that these channels take part in the so-called CICR (Ca2+-induced Ca2+ release), a potentiation of the initial Ca2+ signal. However, experiments with tpc1-knockout Arabidopsis plants did not reveal any clear changes in the phenotype, thus questioning the participation of the SV channel in Ca2+ responses to a variety of stimuli (Peiter ; Ranf ).

Slow vacuolar channels are regulated by a variety of physiological factors (see Pottosin and Schönknecht, 2007, for a recent review). Most pronounced are the activation by cytosolic Ca2+ (Hedrich and Neher, 1987; Ward and Schroeder, 1994; Schulz-Lessdorf and Hedrich, 1995), and negative modulation by vacuolar Ca2+ (Pottosin , 2004). In the first attempt to quantify the joint impact of luminal and cytosolic Ca2+ on SV channel activity, Pottosin and co-workers (1997) demonstrated that the inhibitory effect of vacuolar Ca2+ prevailed the channel stimulation by cytosolic Ca2+. Hence, at realistic electrochemical Ca2+ gradients across the tonoplast that favour Ca2+ release from the vacuole, only a tiny fraction of SV channels is active.

However, while the range of cytosolic Ca2+ variation is well established, all the information regarding the free vacuolar Ca2+ in higher plants comes essentially from a single study (Felle, 1988). In the absence of Ca2+, other vacuolar cations increase the threshold for the SV channel voltage activation, specifically in the case of Mg2+, Na+, and Cs+, or non-specifically via screening of the negative surface charge (K+, choline, NMDG) (Pottosin , 2005; Ivashikina and Hedrich, 2005; Ranf ). Furthermore, the effects of different vacuolar cations were not additive. For instance, at zero vacuolar Ca2+, increases in the vacuolar K+ caused a positive shift of SV channel voltage dependence, whereas at 0.5 mM vacuolar Ca2+, the same increase in vacuolar K+ caused a negative shift (Pottosin ). Also, cytosolic Mg2+, especially at low cytosolic Ca2+, acted as a positive regulator of the SV channel (Pei ; Carpaneto ). Consequently, the free concentrations of physiologically abundant vacuolar cations such as Na+, K+, Ca2+, and Mg2+, remain to be determined and their joint effect on the SV channel voltage dependence, evaluated.

In light of this, the vacuolar levels of these cations were measured and these values were used to quantify the SV channel activity under a physiological range of cytosolic Ca2+ and Mg2+ activities. This was then compared with the SV channel activity measured in situ from attached patches and with Ca2+ fluxes measured on isolated vacuoles using the MIFE technique. Our results suggest that SV channel mediation of Ca2+ release from the vacuole is insufficient to support global Ca2+ responses in higher plant cells.

Materials and methods

Preparation

Vacuoles were isolated mechanically from Beta vulgaris L. taproot slices as described previously (Pottosin ). Briefly, sections of a taproot were soaked for at least 20 min in a solution essentially identical to the bath solution used in patch-clamp experiments, but made slightly (by 30–40 mOsm) hypertonic. Sections were cut with preparation needles under stereoscopic control, and vacuoles released were collected with a micropipette and transported to the experimental chamber.

Intravacuolar cation activities measurements

Ion-selective microelectrodes for the determination of intravacuolar activities of Ca2+, Mg2+, and K+ were fabricated from 1.5 mm non-filamented borosilicate glass capillaries (GC 150–10, CDR Clinical Technology, Middle Cove, Australia) pulled to the smallest diameter (<1 μm) that still allowed backfilling. Blank capillaries were oven-dried at 225 °C overnight. They were then covered with a steel lid for 10–15 min, before 40–50 μl tributylchlorosilane (90796, Fluka Chemicals) was injected under the lid, and the lid removed 10 min later. Electrode blanks were baked for a further 40 min to ensure complete drying, and allowed to cool before storage in a closed container for up to 6 weeks.

Micropipettes were backfilled with the appropriate solution (Table 1) using a syringe with a thin metal needle. Filling with the ionophore (liquid ion exchanger, LIX) was performed immediately after backfilling by dipping the microelectrode into a wider tipped (30–50 μm) pulled capillary, filled by its submersion in the LIX stock solution. The LIX column obtained in the microelectrode tip was 100–200 μm long. Because the ion-selective microelectrode readings are affected by the ionic strength of the solutions, the activities of major osmotics (K+, Na+) were determined first, and calibration solutions were prepared accordingly.

Table 1.

Liquid ionophores and respective backfilling solutions used for ion selective microelectrodes

Ion	Liquid ion exchanger (LIX)	Backfilling solution
Ca2+	Calcium ionophore I — cocktail A (Sigma I-1772)	500 mM CaCl2
Mg2+	Magnesium ionophore I — cocktail A (Fluka 63048)	500 mM MgCl2
K+	Pottasium ionophore I — cocktail A (Fluka 60031)	500 mM KCl

Potassium electrodes were calibrated in a background of 60 mM NaCl and Ca2+-electrode calibration solutions included 115 mM KCl and 60 mM NaCl. The changes in Ca2+ activity between 150 μM and 350 μM, mimicking the variation in measured vacuolar Ca2+, did not affect the electric potential values reported by Mg2+-electrodes, so no additional corrections for Ca2+ were made.

Microelectrode calibration was performed before and after impalement. Electrodes with slopes less than 50 mV (for K+) or 25 mV per decade (for Ca2+, Mg2+), or a correlation coefficient less than 0.999, were discarded. Typical resistance of the microelectrodes was 1–3 GΩ. Measured microelectrodes voltages were corrected for potential values across the tonoplast and the liquid junction potentials (in mV: 1.3 for the Ca2+-microelectrode, 1.4 for the K+-microelectrode, and 6.6 for the Mg2+-microelectrode). Results of trans-tonoplast electric potential difference, measured in a separate set of experiments with a glass microelectrode filled with 3 M KCl to minimize the liquid junction potentials, were used to correct potentials values reported by the ion-selective microelectrodes. The bath solution was 100 mM KCl, 1 mM HEPES (pH 7.4/KOH), adjusted with sorbitol to be slightly more hyperosmotic than the measured osmolality values for squeezed beet juice (500–650 mOsm).

Due to the low selectivity of the Na+-LIX against K+ and difficulties with front-filling H+-LIX electrodes impalements, Na+ and H+ free activities could not be measured with ion-selective microelectrodes. Instead, centrifuged squeezed juice (mainly vacuolar sap) from beet taproots was used to measure the Na+ concentration using a flame photometer, and H+ activity, with conventional pH-electrodes.

Ca2+ flux measurements with the MIFE technique

Ca2+-fluxes from Beta vacuoles were quantified using the Microelectrode Ion Flux Estimation (MIFE) technique (see Newman, 2001, for details and theoretical background). Microelectrodes fabricated for MIFE were similar to those used for vacuole impalements, except the tip diameter was 2–3 μm and drying time after silanization was 30 min.

Calibration of Ca2+-selective MIFE-electrodes was performed in a background of 100 mM K+. MIFE set-up and experimental procedure was essentially as described previously (Wherrett ). The MIFE electrode was positioned at a distance of 10 μm from the vacuole surface and moved back and forth 50 μm horizontally to avoid the influence of ion leakage fluxes from the underlying glass chamber. Calcium fluxes were calculated by subtracting the background flux determined in the immediate proximity to the vacuole from the measured integral flux, taking into account the diameter of each analysed vacuole. Ca2+-fluxes were estimated against a background of 100 mM KCl, 1 mM HEPES (pH 7.4 with KOH), and either 0, 20, 50, 70, 100, 150, or 250 μM Ca2+, without Ca2+-buffering (background Ca2+-contamination was ∼2 μM, as reported by Ca2+-selective electrodes). The statistical significance of Ca2+-flux changes under treatment with Mg2+ and Zn2+ was evaluated by a one way analysis of variance using SPSS, and mean comparison using Duncan's multiple range test.

Patch-clamp protocols and analyses

Patch pipettes fabrication, patch-clamp set-up, and record acquisition and analyses were as described previously (Pottosin ). The osmolality of all solutions was adjusted by sorbitol to isotonic levels with respect to the vacuolar sap (range 520–680 mOsm), verified using a cryoscopic osmometer (OSMOMAT 030, Gonotec, Germany).

The effects of the major vacuolar cations were examined in a set of experiments on excised patches with 100 mM KCl, 2 mM CaCl2, 10 mM HEPES, pH 7.5, (with KOH) at the cytosolic side, and either 100 mM, 125 or 187 mM KCl, 0 or 62 mM NaCl, 0 or 8 mM MgCl2, with a variable free Ca2+ (0, 0.01, 0.05, 0.1, 0.2, 0.5, 1.0, 2.0 or 10.0 mM), 10 mM MES/KOH (pH 5.5 or 6.4), at the vacuolar side. For measurements in vacuole-attached configuration, patch pipettes were filled with the same cytosolic solution as above, while the bath contained 100 mM KCl, and 0, 0.05, 0.1, 0.2, 0.5, 1, 2 or 10 mM Ca2+, 10 mM HEPES-KOH (pH 7.5). Voltages were corrected for the trans-tonoplast potential from values measured with a 3 M KCl-filled microelectrode in bath solutions containing the same free Ca2+ concentration.

SV currents in vacuole-attached patches and small right-oriented (cytosolic side out) vacuolar vesicles were also analysed at ionic conditions mimicking the physiological cation concentrations. The vacuolar solution used was in mM: 125 K+, 62 Na+, 8 Mg2+, 0.25 Ca2+, pH 5.5 (MES/KOH), and the cytosolic solution contained: 100 mM K+, 10 mM HEPES/KOH (pH 7.5) and 0.1, 1.0, 7.0 or 20 μM free Ca2+, supplemented with either 0.5 or 1.5 mM free Mg2+. Solutions with a certain free Ca2+ and Mg2+ concentrations were prepared using a combination of chelating agents (2 mM of EGTA, EDTA, HEDTA, and/or NTA). The free divalent cation concentration for each condition were calculated on the basis of their total concentration and concentrations of added chelating agents including ATP, taking into the account the ambient pH, temperature and ionic strength (WinMAXC32 v.2.50, Chris Patton, Stanford University).

SV channel currents were evoked by a series of positive voltage steps from a holding potential of –100 mV (cytosol minus vacuole) or –20 mV (in experiments with low cytosolic Ca2+), up to +200 mV in 20 mV increments. Unitary current–voltage (I/V) relationships at different ionic conditions and patch configurations were obtained by the application of 15 ms voltage-ramps between +200 and −200 mV as previously described (Pottosin ). The SV activation curve (number or relative number of open channels versus trans-tonoplast electric potential difference) was obtained as described in detail by Pottosin . To describe the shift of the voltage dependence by vacuolar Ca2+, a threshold voltage value (V1%), i.e. the voltage at which 1% of maximal SV channel activity was reached (P/Pmax=0.01), was estimated at each given condition. Non-linear regression fitting of the data was performed using the GraFit v.6 data analysis and graphics program (Erithacus Software, Ltd).

Results

Vacuolar Ca2+ fluxes and cation concentrations

One of the primary aims of this study was to evaluate net Ca2+ release from sugar beet vacuoles. As one possible route of Ca2+ escape from vacuoles are Ca2+-activated SV channels, by applying the MIFE technique Ca2+ flux measurements were performed at different levels of cytosolic (bath) Ca2+ (Fig. 1A). However, the minimal Ca2+ free concentration achieved was ∼2 μM (reflecting contamination of the bath solution by the vacuolar isolation procedure) due to the inability of the MIFE technique to estimate net Ca2+ fluxes in the presence of Ca2+ buffers. This concentration is referred to as the ‘trace bath Ca2+’ when referring to MIFE results. Under such minimal conditions, the net Ca2+ release from vacuoles was close to the detection limit of the method (∼1 nmol m−2 s−1). This flux tripled and reached its maximum at ∼20 μM free Ca2+ in the bath, while further increases in the free cytosolic Ca2+ caused a gradual decrease of Ca2+ flux, until it reversed polarity at ∼200 μM cytosolic Ca2+. This non-monotonic behaviour is consistent with passive Ca2+ transport, mediated by some (presumably SV) Ca2+-permeable channel, activated by cytosolic Ca2+. Moreover, the observed Ca2+ flux was significantly (P < 0.05) enhanced by 1.5 mM Mg2+ and strongly suppressed by 0.1 mM Zn2+ (Fig. 1B), known activator and blocker, respectively, of SV channels (Hedrich and Kurkdjian, 1988; Gambale ; Pei ; Carpaneto ).

Fig. 1.

Effects of cytosolic Ca2+ on trans-tonoplast Ca2+ fluxes and the membrane potential difference in isolated red beet vacuoles. (A) Ca2+ fluxes from individual vacuoles measured with the MIFE technique (n=6–19 vacuoles for each Ca2+ concentration in the bath). Negative flux implies a net Ca2+ transport from the vacuole to the extravacuolar (cytosolic) side. (B) Vacuolar Ca2+ flux stimulation by cytosolic Mg2+ and inhibition by Zn2+ (n=8–10 vacuoles). (C) Electric potential difference across the tonoplast as a function of bath free Ca2+ (n=8–11 vacuoles). All data are presented as mean ±SEM. Bath: 100 mM KCl, pH 7.5.

The SV channel is controlled by different mono and divalent cations present at cytosolic and vacuolar sides of the tonoplast. Therefore, to link the vacuolar Ca2+ flux to the SV channel activity, a further attempt was made to evaluate the activities of physiologically relevant vacuolar cations. Because single-barrelled electrodes were used to determine the activities of Ca2+, Mg2+, and K+ in fresh isolated sugar beet vacuoles (see Materials and methods for details), electrode readings were corrected for trans-tonoplast electric potential difference with measurements from a separate set of experiments using 3 M KCl-filled microelectrodes (Fig. 1C). However, due to the poor selectivity of all available Na+ resins (Chen ), vacuolar Na+ concentrations were measured using conventional flame photometry. The results of these measurements are summarized in Table 2. Notably, the Ca2+ activity was an order of magnitude lower (Table 2) than previously reported for distinct higher plant preparations (Felle, 1988). Moreover, the same Ca2+ activity value was independently estimated from the reversal condition (respective bath Ca2+concentration) of the Ca2+ flux reported by MIFE, assuming equilibrium conditions for Ca2+ exchange across the tonoplast and correcting the data for the resting tonoplast potential.

Table 2.

Activities of physiological cations in red beet (Beta vulgaris L.) vacuoles

Ion	Activity (mM)	Method
K+	116±2 (n=8)	K+-selective microelectrodes
Na+	62±5a (n=5)	Flame photometry
Mg2+	8.3±0.7 (n=7)	Mg2+-selective microelectrodes
Ca2+	0.20±0.02 (n=7) 0.21	Ca2+-selective microelectrodes MIFE estimate (see Fig. 1)

Concentration instead of activity was determined in this case.

Modulation of the SV channel voltage dependence by vacuolar cations

Our previous study has demonstrated a very strong dependence of the SV channel activation threshold on vacuolar Ca2+ (Pottosin ). However, the above study was performed at a non-physiological vacuolar pH (7.5), while the actual pH of the vacuolar sap is around 5.9±0.3 (n=11 separate preparations). Accordingly, in this work the Ca2+ modulation of the SV channel voltage dependence has been analysed at more physiologically relevant vacuolar pH values (6.4 and 5.5, respectively). The summary of the activation curves at different vacuolar Ca2+ levels and pH 5.5 is presented in Fig. 2A. To make a single-parameter comparison of the Ca2+ dependence at different vacuolar pH values, the membrane potential values at which 1% of available SV channels are open (V1%; see Pottosin , for a justification) have been estimated. This technical activation threshold was plotted as a function of the vacuolar free Ca2+ (Fig. 2B). At vacuolar pH 7.5 within a range from 10 μM to 10 mM Ca2+, the SV channel behaves as a ‘super-Ca2+’ electrode. The curve V1% versus the vacuolar free Ca2+ has a slope of ∼40 mV per 10-fold increase in Ca2+ concentration (as compared to 29.5 mV for an ideal Ca2+ electrode). This is because the SV channel is able to bind more than one Ca2+ ion from the vacuolar side (Pottosin ). Lowering the vacuolar pH desensitizes the SV channel to the vacuolar Ca2+ changes within a physiological range of Ca2+ concentrations (0.1–1 mM). At the same time, lowering the vacuolar pH by itself does not affect the SV channel voltage dependence and V1% at zero vacuolar Ca2+ (Fig. 2).

Fig. 2.

Vacuolar pH modifies the dependence of the SV channel voltage activation on the vacuolar Ca2+. (A) Voltage dependence of the SV channel activation at different vacuolar free Ca2+ concentrations (indicated in mM at the right of the corresponding symbols) and pH 5.5. SV channel activity was analysed in large isolated tonoplast patches (n=4–8 patches for each Ca2+ concentration, means ±SEM). Dashed horizontal line indicates the voltage at which 1% of maximum number of SV channels are open. (B) Dependence of the threshold for SV channel voltage activation (a membrane potential at which 1% of maximal number of SV channels are open) on the vacuolar free Ca2+ at different vacuolar pH. Data for pH 7.5 were calculated from Pottosin . Solution composition was: symmetric 100 mM KCl; 2 mM free Ca2+ at cytosolic side; cytosolic and vacuolar pH values were 7.5 and 5.5, respectively.

Sugar beet is a salt-tolerant species that uses Na+ for maintaining cell turgor. Accordingly, the impact of vacuolar sap Na+ on the voltage dependence of the SV channel at zero and 0.5 mM vacuolar Ca2+ was verified (Fig. 3). At zero Ca2+, addition of 62 mM Na+ caused an increase in the voltage threshold of SV channel activation. This exceeded the effect of the addition of equimolar concentration of K+ (Fig. 3A). However, in the presence of 0.5 mM Ca2+ at the luminal side, addition of Na+ resulted in the opposite effect, i.e. a decrease in the activation threshold (Fig. 3B). Alternatively, the results presented in Fig. 3 could be read in a different manner. Indeed, the activation curves are the same at zero and 0.5 mM Ca2+ in the presence of Na+, indicating that in the presence of 62 mM Na+, a vacuolar Ca2+ increase from zero to 0.5 mM does not have a significant impact on the SV channel voltage dependence. In other words, the presence of physiologically relevant Na+ concentrations in cell vacuoles desensitizes the SV channel to vacuolar Ca2+ variation between zero and 0.5 mM.

Fig. 3.

Vacuolar Na+ has opposite effects on the SV channel voltage dependence at zero and submillimolar vacuolar Ca2+. Voltage dependence of the SV channel activity in large isolated tonoplast patches at two vacuolar free Ca2+ concentrations, c. 10 nM (A) and 0.5 mM (B). Vacuolar K+ and Na+ concentrations (in mM) for each condition are indicated, 100 mM KCl, 2 mM free Ca2+ (pH 7.5) at cytosolic side. Data are means ±SEM, n=3 patches for each condition.

Finally, the modulation of the SV channel voltage dependence by vacuolar Ca2+ at pH 5.5 in the presence of physiological concentrations of K+, Na+, and Mg2+ was analysed. Addition of 62 mM Na+ and 8 mM Mg2+ to the vacuolar solution shifted the activation threshold at zero vacuolar Ca2+ by ∼35 mV to more positive potentials (Figs 4A, 2A). Also the threshold dependence on vacuolar Ca2+ flattened: not only the Ca2+change from 0.1 mM to 1 mM produced little change in the threshold but, in addition, the Ca2+ increase from zero to 0.1 mM caused a lesser V1% shift compared to the case without Mg2+ and Na+ at the luminal side (Fig. 4B). For comparison, threshold potentials were evaluated in vacuole-attached patches, with patch-pipettes filled with the standard cytosolic high Ca2+ solution and variable free Ca2+ in the bath. Threshold V1% values (in mV) were as follows: −14.4±0.1, −11.3±8.7, −14.9±7.4, and +5.3±0.3 (n=4–5 vacuoles) using bath free Ca2+ of 0, 0.1, 0.5, and 1 mM, respectively. Increased threshold at 1 mM free Ca2+ in the bath probably reflects an increase of vacuolar Ca2+, due to Ca2+ ions entering the vacuole from the bath via open Ca2+-activated SV channels. Indirect evidence that such a process takes place comes from the membrane potential measurements (Fig. 1C). At [Ca2+] ≥1 mM, the membrane potential became more sensitive to changes in bath Ca2+, implying a larger contribution of the tonoplast Ca2+-permeable channels. Providing at bath Ca2+ < 1 mM the vacuolar sap composition is unperturbed due to the exchange with the bath, the V1% mean values for vacuole-attached patches would then be in the −15 mV to −11 mV range. These values can be compared with those obtained on excised patches at different vacuolar Ca2+ levels, with the concentrations of other cations at the luminal side mimicking the vacuolar content. For excised patches V1% values between −15 mV and −11 mV were obtained at vacuolar Ca2+ levels between 0.1 mM and 0.7 mM (Fig. 4B). This is in a good agreement with a mean vacuolar free Ca2+ activity ∼0.2 mM reported by ion-selective microelectrodes (Table 2). Thus, in intact isolated vacuoles, the voltage-dependent activity of the SV channel is most likely controlled by a joint action of major vacuolar cations (K+, Na+, Mg2+, Ca2+) and pH.

Fig. 4.

The presence of various cations at the luminal side makes the SV channel almost insensitive to changes in vacuolar Ca2+ from zero up to the millimolar range. (A) Voltage dependence of the SV channel activity in large isolated tonoplast patches at different vacuolar free Ca2+ concentrations. The solution at the vacuolar side contained (in mM): 125 KCl, 62 NaCl, 8 MgCl2 (pH 5.5), and at the cytosolic side: 100 mM KCl, 2 mM free Ca2+ (pH 7.5). (B) Dependence of the threshold for SV channel voltage activation (membrane potential at which 1% of maximum number of SV channels are open) on the vacuolar free Ca2+. Other components of the vacuolar and cytosolic solutions are as above. For comparison, a dependence obtained in 100 mM KCl (pH 5.5) was redrawn from Fig. 2B (dashed line). The range of values obtained in the attached patches is indicated (grey zone). Data are means ±SEM, n=3–5 patches for each condition.

The SV channel activity at low cytosolic Ca2+

All the results illustrated in Figs 2–4 were obtained at non-physiologically high cytosolic Ca2+ concentrations (2 mM). Now, the SV channel voltage-dependent activity was analysed at lower cytosolic Ca2+ (0.1–20 μM), in the presence of physiologically relevant concentrations of major cations and pH at the vacuolar side. The effects of increasing cytosolic Mg2+ from 0.5 mM to 1.5 mM were also checked. Because at high positive potentials, huge SV channel-mediated currents from a typical vacuole tend to saturate the amplifier, small vacuolar vesicles with a visible diameter of 5–10 μm were isolated after achieving the whole vacuole configuration. The advantage of this compared to a standard outside-out configuration, is a larger current magnitude and a higher precision of measurement of the membrane capacitance (range 1–4 pF). This enabled the evaluation of the SV current per unit membrane area.

A typical example of the effect of cytosolic Ca2+ and Mg2+ variation on SV currents is illustrated in Fig. 5A. Both cations increase the SV current measured at high positive potentials. Titration curves for Ca2+ at +100 and +180 mV yielded Kd values of 1.4±0.7 μM and 0.7±0.1 μM, respectively (Fig. 5B), implying that cytosolic Ca2+ binding is essentially voltage-independent. In addition, a detailed analysis of the voltage-dependence at 0.1 μM and 20 μM cytosolic Ca2+ revealed only a small shift in the activation curve at 0.5 mM cytosolic Mg2+, with no shift whatsoever at higher (1.5 mM) Mg2+ (Fig. 5C; Table 3). Furthermore, the slope (determined by the gating charge, z) of voltage dependence increased only slightly upon the increase of cytosolic Ca2+ from 0.1 μM to 20 μM (Table 3). Altogether, this suggests that the contribution of Ca2+ binding from the cytosolic side to the SV channel voltage gating is insignificant. Increasing the cytosolic Mg2+ concentration caused about a 2-fold stimulation of the SV channels. This effect was comparable at all membrane voltages and cytosolic Ca2+ concentrations (Fig. 5; Table 3). Therefore, for this range of cytosolic free cation concentrations (Ca2+: 0.1–20 μM, Mg2+: 0.5–1.5 mM), divalent cations increase the mean number of voltage-activated SV channels without a substantial alteration of their voltage dependence. One may presume, however, that this does not remain true for higher cytosolic Ca2+. The potential at which 1% of voltage-activated SV channels are open at 20 μM cytosolic Ca2+ is about +48 mV (Fig. 5C). For a comparison, at very high (2 mM) cytosolic Ca2+ and physiologic (0.2 mM) vacuolar Ca2+ (Fig. 4B), the V1% value was ∼ −15 mV.

Fig. 5.

Sensitivity of the SV channel to cytosolic Ca2+ and Mg2+ at physiological concentrations of vacuolar cations. (A) An example of a SV current record from a small (c ∼1 pF) outside (cytosolic side)-out vesicle at different concentrations of cytosolic Ca2+. External (cytosolic) solution contained 100 mM KCl (pH 7.4). The composition of a pipette (vacuolar) solution is as in Fig. 4. (B) The Ca2+-dependence of the SV current (nA/pF) at +180 and +100 mV. Cytosolic free Mg2+ was 0.5 mM, and other ionic conditions are as in (A). Data are presented as mean ±SEM (n=6–10 vacuoles). Solid lines are the best fits to the Hill equation, with Kd for cytosolic Ca2+ at 1.4±0.7 μM and 0.7±0.1 μM, and Hill coefficients of 1.4±0.2 and 1.3±0.2 for +100 and +180 mV, respectively. (C) Voltage dependence of the SV current at different cytosolic free Ca2+ and Mg2+ levels. A macroscopic SV current was divided by the single channel current at each potential and by membrane capacitance, yielding the mean number of the open channels per pF. Data are means ±SEM (n=4–8 vacuoles). Solid lines are the best fits to the Boltzmann equation with parameter values given in Table 3.

Table 3.

Parameters of the SV channel voltage dependence at different cytosolic Ca2+ and Mg2+

Cytosolic free Ca2+ and Mg2+ concentrations	Maximum open SV channels number	Gating charge (z)	Midpoint potential (mV)
Outside-outs
0.1 μM Ca2++0.5 mM Mg2+	11±2/pF	1.14±0.20	146±8
0.1 μM Ca2++1.5 mM Mg2+	16±2/pF	1.32±0.18	134±7
20 μM Ca2++0.5 mM Mg2+	87±8/pF	1.41±0.06	133±4
20 μM Ca2++1.5 mM Mg2+	163±5/pF	1.43±0.05	133±2
Attached
0.1 μM Ca2++0.5 mM Mg2+	4.2±0.5	1.18±0.06	136±6a
20 μM Ca2++0.5 mM Mg2+	44±1	1.32±0.15	96±3a

Potential values for the attached patches are referred to the command voltage and should be corrected for the resting tonoplast electric potential difference, ∼+10 mV at these conditions.

Finally, the voltage-dependent activity of the SV channel was analysed in vacuole-attached patches at high (20 μM) and low (0.1 μM) Ca2+. As the membrane capacitance in these experiments was very low (<1 pF), its precise value could not be determined accurately. Thus, the channel density could not be determined either. Instead, the mean number of open SV channels per patch was accessed as a function of membrane potential (Fig. 6; Table 3). The actual electric potential difference across a membrane patch in attached configuration is a sum of the resting trans-tonoplast potential and the command voltage. In a separate set of experiments, using small-tip patch-clamp pipettes filled with 3 M KCl the resting trans-tonoplast potential difference was determined. For bath solutions with 0.1 μM and 20 μM Ca2+, resting potentials were +12±3 mV (n=11 vacuoles) and +9±2 mV (n=9), respectively. Such cytosolic-side positive potentials were due to the absence of H+-pump activity. In the presence of 2 mM Mg-ATP, respective resting tonoplast potentials for 0.1 μM and 20 μM Ca2+ baths became –4±2 mV (n=11) and –12±4 mV (n=11), respectively.

Fig. 6.

Slow-vacuolar channel activity in vacuole-attached patches. Bath and pipette solutions are identical, 100 mM KCl (pH 7.4) and free Ca2+ and Mg2+ concentrations are as indicated at the top of the curves. A macroscopic SV current was divided by the single channel current at each potential, yielding the mean number of open channels per patch (NPo). For comparison, assuming the same channel density as in outside-out patches (Fig. 5), the mean open SV channel number per typical vacuole of 50 μm diameter is given at the second y-axis on the right. Data are means ±SEM (n=4). Solid lines are the best fits to the Boltzmann equation with parameter values given in Table 3. Voltages were command potentials for this case; to convert them to the actual tonoplast electric potential difference, the resting potential value (about +10 mV at these conditions, as verified by 3 M KCl-filled microelectrode impalements) should be totalled.

Correcting the membrane voltage in attached patches for a resting potential difference, the parameters of the SV channel voltage dependence in attached and excised (outside-out) patches became identical at low (0.1 μM) cytosolic Ca2+ (Table 3). However, in contrast to outside-out patches, increasing the cytosolic Ca2+ from 0.1 μM to 20 μM caused a relatively large (40 mV; Table 3) shift of the voltage dependence to more negative potentials. Further on, the 1% voltage threshold (corrected for the resting potential) was about +20 mV for attached patches at 20 μM cytosolic Ca2+, compared to the −5 to −25 mV range obtained for vacuolar-attached patches facing very high (2 mM) cytosolic Ca2+ (Fig. 4B).

Discussion

Other vacuolar cations desensitize the SV channel to changes in luminal Ca2+

In this study, the free vacuolar concentrations of all major physiological cations were determined (Table 2). To our knowledge, this is the first report of free vacuolar Mg2+ concentrations in higher plants. In addition, using two independent approaches, it was found that the mean free vacuolar Ca2+ concentration in Beta taproots was ∼0.2 mM, an order of magnitude lower than previously reported values for Zea roots and Riccia rhizoids (Felle, 1988). Therefore, a re-evaluation of the modulation of SV channel activity by vacuolar Ca2+ under more realistic vacuolar cation concentrations and pH is required.

Our main findings can be summarized as follows: in the presence of cations other than Ca2+ at their physiological concentrations and at acidic pH inside the vacuole, the SV channel voltage-dependent activity becomes insensitive to vacuolar Ca2+ variation within the physiological range (Fig 4B). Part of this effect is non-specific and related to the screening of the negative charge at the membrane surface, thus reducing vacuolar Ca2+ binding at high ionic strength (Pottosin ). However, some monovalent cations such as Na+ and Cs+ appear to exert specific effects on the SV channel voltage dependence when compared to K+ (Ivashikina and Hedrich, 2005; Ranf ). The Na+-induced shift in the SV channel voltage dependence found here is comparable to that reported by Ivashikina and Hedrich (2005) at similar (zero vacuolar Ca2+) conditions. However, in a background of more realistic vacuolar Ca2+ concentrations (0.5 mM in our case), Na+ caused an opposite shift (Fig. 3B). Therefore, accumulation of Na+ in the vacuole during salt stress could ameliorate the inhibitory effects of vacuolar Ca2+ on the SV channel.

The effect of vacuolar Mg2+ was qualitatively similar to that of Ca2+, but smaller in magnitude (Pottosin ). However, bearing in mind the much higher vacuolar Mg2+ concentration compared to that of Ca2+ (Table 3), the actual effects of these two ions on the SV channel voltage dependence may be comparable.

In a previous study (Pottosin ), a detailed kinetic model of stabilization of the closed states of the SV channel by vacuolar Ca2+ and Mg2+ was generated. In particular, it was found that up to three Ca2+ or Mg2+ ions could be bound in one of the closed states: C2. For simplicity, it was proposed that divalent cation binding at this state was non-co-operative and occurred with the same Kd values. In the present work it is demonstrated that, at acidic pH, the dependence of the activation threshold on vacuolar Ca2+ (reflecting mainly the stabilization of C2-state by Ca2+) becomes complex, with a distinct plateau in a physiological Ca2+ range (Fig. 2B).

Curiously, protons by themselves do not affect SV channel voltage dependence at zero vacuolar Ca2+. This may be due to the fact that vacuolar protons bind with equal pKs in closed and open states, so not affecting the distribution between states at any voltage. However, H+ binding obviously affects the vacuolar Ca2+ binding. The effect of acidic pH presented in Fig. 2B may be explained for example, if the protonation of the SV channel protein results in a much higher Kd for binding of the second and third Ca2+ ions. Therefore, a plateau between 0.1 mM and 1 mM Ca2+ may reflect the saturation of the first Ca2+ ion binding in all states. [At these Ca2+ concentrations, in accordance with Kd values for a single Ca2+ ion binding from Pottosin , all states (open and two closed: C1 and C2) will contain a bound Ca2+ ion. Hence, further stabilization of closed states and respective voltage threshold shift is only possible with binding of extra Ca2+ ions to the C2-state.] It should be noted though, that the effects of different cations are not additive. Rather, it seems that they compete with Ca2+ binding, thus diminishing the modulation effect of Ca2+ on the SV channel.

Our working hypothesis therefore, is that the existence of diverse modulatory effects exerted by multiple cations in the vacuolar lumen assists in ‘buffering’ the SV channel activity. Consequently, the variation of vacuolar concentration of a single component, even one as powerful as Ca2+, would not result in a large shift in the SV channel voltage dependence.

The SV channel activity at resting trans-tonoplast potentials

In the absence of externally applied voltage and at resting cytosolic Ca2+ (100 nM), on average 0.01 SV channels per patch will be open, as judged by the activation curve shown for vacuole-attached patches (Fig. 6). This is equivalent to 0.023% of the number of channels open at 20 μM cytosolic Ca2+ and high membrane potentials. Assuming the same channel density as in outside-out patches (see Fig. 5C, 1 pF approximately corresponds to a 100 μm2 membrane area), for a typical vacuole of 50 μm diameter or 7855 μm2 membrane surface area, as an average only 1.5 SV channels per vacuole will be open at any time. At 20 μM cytosolic Ca2+, about 40 SV channels per vacuole will be open at resting potentials (Fig. 6), and this can be taken as an upper estimate (see the dose dependence in Fig. 5B).

The properties of Ca2+ fluxes reported by the MIFE technique, their stimulation by Mg2+ and a strong inhibition by Zn2+, suggests that they are mediated by SV channels. Moreover, the Ca2+ release from isolated beet vacuoles, similar to the SV current, was stimulated by cytosolic Ca2+, reaching a maximum of 2.8±0.6 nmol m−2 s−1 (Fig. 1) at 20 μM cytosolic Ca2+. Assuming a 50 μm vacuole diameter, this equates to a current of 4.3±0.9 pA per vacuole. Comparison of the Ca2+ flux magnitude and the number of open SV channels at 20 μM cytosolic Ca2+, suggests that each SV channel mediates a net Ca2+ current of ∼100 fA at the resting trans-tonoplast potential. This value fits remarkably well the theoretical predictions for Beta SV channel: 100–120 fA of the Ca2+ current per channel at zero voltage, weakly sensitive to pCa for a physiological range of Ca2+ gradients (Gradmann ).

In reality, SV-mediated Ca2+ release could be even smaller. Indeed, the resting potential of the isolated vacuoles in our experiments were depolarized (∼+10 mV for cytosolic bath with 0.1–20 μM Ca2+ plus 0.66 mM Mg2+) in the absence of substrates for vacuolar H+ pumps. The addition of 2 mM Mg-ATP into the bath solutions resulted in more negative membrane potential values, –4 to –12 mV. This is comparable to directly measured values for the tonoplast in intact barley roots and leaves, −10 to −15 mV and ∼0 mV, respectively (Walker ; Cuin ). An even lower negative value of −30 mV was reported for Arabidopsis leaves (Miller ). Based on the parameters values for the SV voltage dependence (Table 3), a negative shift of membrane potential by 16–21 mV results in a 2–3 times lower SV channel activity. However, a more negative potential implies an increase in a driving force (resting potential minus ECa) for vacuolar Ca2+ release. For 20/200 μM Ca2+ gradient the relative increase in the Ca2+ driving force is significant, from –20 mV in the absence of ATP to –41 mV in the presence of 2 mM Mg-ATP (negative sign implies the preferential Ca2+ influx to the cytosol from the vacuole). Bearing in mind that Ca2+ flux under asymmetric conditions is not linear and displays an inward rectification (increase of chord conductance when the potential difference between cytosol and vacuole is made more negative) (Gradmann ), one may expect that at resting tonoplast potential Ca2+ flux through an open SV channel will increase more than the increase in driving force, i.e. >2 times. Therefore, at high (e.g. 20 μM) cytosolic Ca2+ the decrease of the SV channel open probability will be almost compensated by the increase of Ca2+ flux through a single channel. At low cytosolic Ca2+ the relative change of Ca2+ driving force due to a negative shift of resting tonoplast potential will be less significant. Therefore, total Ca2+ release from the vacuole will be reduced in this case, due to a decrease in the mean number of open SV channels.

The cytosolic Ca2+ level of 20 μM considered above may lead to another overestimate of the SV channel activity. Under physiological ionic conditions, half-activation of the SV channel occurs at 0.7–1.4 μM cytosolic Ca2+ (Fig. 5B), close to Kd=1.5 μM defined by Hedrich and Neher (1987). These values are near the peak global free Ca2+ levels reached during plant responses to several abiotic stresses (Ranf ). Thus, the overestimation will be approximately 2-fold. For isolated vacuoles, the SV-mediated Ca2+ release will come closer to that measured by the MIFE technique in the absence of added Ca2+: 1 nmol m−2 s−1 or 1.5 pA per vacuole.

Considering therefore both the more negative trans-tonoplast potential and lower cytosolic Ca2+, the whole vacuole population of SV channels would produce a global Ca2+ release into the cytosol of ≤1 pA. Assuming that for a typical cell the vacuole and the cytosol occupy 80% and 20% of the volume, respectively, and that a vacuole has a diameter of 50 μm, cytosolic volume will be ∼16 pl. Therefore, net cytosol-directed Ca2+ current of 1 pA will tend to change the total cytosolic Ca2+ concentration at 35 μM min−1 rate. However, cytosolic Ca2+ is strongly buffered. For plant cells, the estimated buffering capacity Bmax is 0.2–0.5 mM (Trewavas, 1999) whereas the effective dissociation constant Kd for a cytosolic Ca2+ buffer in different cell types ranges between 0.15 μM and 0.6 μM (Martinez-Serrano ; Kuratomi ). Assuming a simple Michaelis–Menten behaviour, the relation between total and free cytosolic Ca2+ may be expressed as 1 + Bmax/(Cafree + Kd). Therefore, when free cytosolic Ca2+ approaches the Kd, about 1 out of 500 Ca2+ ions as an average could be found in the free form. Thus, the SV channel mediated increase of free cytosolic Ca2+ would be ≤70 nM min−1, which is by several orders of magnitude slower than the Ca2+ responses to most abiotic stresses, and still slower than elicitor-induced Ca2+ signals (Ranf ). Yet, as discussed below, in intact cells the SV channel reactivity could be locally increased due to the formation of a high Ca2+ microdomain resulting from channel gathering, as well as tight contact with Ca2+-permeable channels located at other membranes.

Thus, SV channels make little, if any contribution to global cytosolic Ca2+ responses, as convincingly demonstrated in a recent study by Ranf . This result is expected and perfectly reasonable in view of the fact that the vacuole is an inexhaustible Ca2+ store. Hence, the analogy with the CICR in animal cells may be not valid, because the latter phenomenon is assigned to a different Ca2+ store, such as the endoplasmic reticulum (ER) that can be readily and reversibly depleted. In contrast to the small-sized ER store, depleting a large vacuole, which can occupy up to 90% of the plant cell volume, will inevitably cause irreversible alterations in the cytosolic Ca2+. Therefore, although the activation of SV channels by cytosolic Ca2+ provides a positive feedback system, i.e. the more Ca2+ released via the SV channel, the higher the activity of this and neighbouring SV channels, a strict negative SV channel control by physiological factors (see Pottosin and Schönknecht, 2007, for more details) effectively prevents this chain reaction at the whole cell level.

As Ranf used a cytosolic aequorin reporter, their data set aside the potential role of the SV channel in local Ca2+ signalling, because this could not be detected by their method. Such local Ca2+ signals, the so-called Ca2+ sparks and puffs, have been well studied in animal cells. For instance, in tight junctions of striated muscle, a cross-talk between sarcoplasmic ryanodine receptor channels (RyRs) and voltage-dependent Ca2+ channels of t tubules is initiated. This occurs due to a clustering of RyRs and their positive regulation by cytosolic Ca2+, thereby generating a synchronous response of several RyRs: a Ca2+ spark (Franzini-Armstrong and Protasi, 1997). It should be noted that, with respect to high Ca2+ microdomains, a free Ca2+ concentration at a distance of ∼0.1 μm from the pore entrance of a Ca2+ conducting channel may reach the hundred micromole range (Demuro and Parker, 2006; Rizutto and Pozzan, 2006). Due to its vast area, the tonoplast is often in contact with several intracellular stores, as well as the plasma membrane. Furthermore, tonoplast SV channels are very abundant. Based on the surface density (Fig. 5C), the average distance between neighbouring channels is less than 1 μm. Thus, there should almost certainly be several SV channel copies in the contact zones of the tonoplast with other membranes. Provided the SV channels are clustered so as to increase the cross-activation by Ca2+ ions released from the vacuole into the cytosol, and that at least one channel of each cluster is at a short (∼100 nm) distance from a Ca2+-permeable channel of another membrane (e.g. plasma- or ER membrane), the conditions for Ca2+ spark formation will be met. Therefore, a functionally large vacuole would be split in multiple small circuits, and the vacuolar Ca2+ release would be asynchronous and restricted to local contact zones with other organelles and the plasma membrane. This hypothesis should be addressed in further studies employing such techniques as high-resolution confocal or TIRF microscopy on intact cells (Demuro and Parker, 2006).

In conclusion, a combination of MIFE Ca2+ flux measurements and patch-clamp evaluations of the SV channel activity in isolated vacuoles suggests that only a minor contribution to global Ca2+ signalling is made by SV channel-mediated vacuolar Ca2+ release. Whether the vacuolar SV channels are capable of modulating local Ca2+ responses remains to be elucidated.