Stomatal guard cells play a key role in gas exchange for photosynthesis and in minimizing transpirational water loss from plants by opening and closing the stomatal pore. The bulk of the osmotic content driving stomatal movements depends on ionic fluxes across both the plasma membrane and tonoplast, the metabolism of organic acids, primarily Mal (malate), and its accumulation and loss. Anion channels at the plasma membrane are thought to comprise a major pathway for Mal efflux during stomatal closure, implicating their key role in linking solute flux with metabolism. Nonetheless, little is known of the regulation of anion channel current (ICl) by cytosolic Mal or its immediate metabolite OAA (oxaloacetate). In the present study, we have examined the impact of Mal, OAA and of the monocarboxylic acid anion acetate in guard cells of Vicia faba L. and report that all three organic acids affect ICl, but with markedly different characteristics and sidedness to their activities. Most prominent was a suppression of ICl by OAA within the physiological range of concentrations found in vivo. These findings indicate a capacity for OAA to co-ordinate organic acid metabolism with ICl through the direct effect of organic acid pool size. The findings of the present study also add perspective to in vivo recordings using acetate-based electrolytes.
Stomata are pores that provide the major route for gas exchange across the impermeable cuticle of leaves and stems . They open and close in response to exogenous and endogenous signals and thereby control the exchange of gases, most importantly water vapour and CO2, between the interior of the leaf and the atmosphere. The guard cells surrounding the stomatal pore respond to a number of well-defined signals, including hormones, light and atmospheric CO2 concentration, integrating these signals to regulate the stomatal aperture and balance the conflicting needs for water conservation and for inorganic carbon for photosynthesis. The acquisition of stomata and the leaf cuticle are considered to be key elements in the evolution of advanced terrestrial plants  as these adaptations allow the plant to inhabit a range of different, often fluctuating, environments and still control water content.
A very large body of experimental evidence supports the central role of ionic fluxes, across both the plasma membrane and tonoplast, and of metabolism, notably of organic acids including Mal (malate), in collectively shaping the cyclical changes in osmotic load and turgor pressure that drive guard cells to open and close the stomatal pore [1,3–6]. At maturity, stomatal guard cells lack plasmodesmatal connections with the neighbouring epidermal cells . Thus guard cells define a semi-closed cellular system within the surrounding leaf tissue. Transport of all inorganic ions, K+ and Cl−, both major contributors to the osmotic content of guard cells, along with H+ and Ca2+, takes place across the plasma membrane and co-ordinates with the metabolism of osmotically active organic compounds.
At the plasma membrane, anion channels are key elements that contribute to stomatal closing, and their activities are essential to depolarize the membrane and balance charge with K+ during solute efflux for stomatal closure [3,4,8]. The guard cell anion channels divide between two major groupings, based on their physiological characteristics. The slow, or S-type, anion channels were originally identified with a slow-activating and largely voltage-independent current that is strongly activated by [Ca2+]i (cytosolic-free [Ca2+]) in the micromolar range [9,10]; these channels have since been associated with the SLAC1 gene product in the model plant Arabidopsis [11,12]. The rapid, or R-type, anion channels were first characterized by a strongly voltage-dependent current that activated on increasing membrane voltage beyond the range −100 to −80 mV and inactivated within a few tens to hundreds of milliseconds at more negative voltages [9,13,14].
Both anion channels are closely tied to Mal, contributing to its homoeostasis and most likely to its release to the apoplast during stomatal closing [15,16]. The slac1 loss-of-function mutation is associated with over-accumulations of Mal and other organic acids, as well as K+ , and very high cytosolic concentrations of Mal may itself inhibit the SLAC1 current . The molecular identity of the R-type anion channels remains uncertain, but it has been suggested that ALMT12 (aluminium-activated, Mal transporter 12) and other members of this aluminium-activated Mal transporter family may contribute to the anion current in Arabidopsis . Indeed, extracellular Mal alters the voltage-dependence of R-type anion channels  and similarly of current associated with AtALMT12 (Arabidopsis ALMT12) , in effect promoting anion efflux through the channels in response to the organic anion outside. These observations have been interpreted as an indirect ‘feedforward’ mechanism for CO2 control of the channels , but the implied consequences of the related input from the cytosolic metabolites have been largely ignored.
In the present study, we have have explored the ensemble anion channel current, hereafter designated ICl, and its regulation by metabolically related organic acid anions through experimental manipulations both outside and inside guard cells of Vicia faba L., the broad bean plant widely used as a model in physiological studies of guard cells. We report that Mal action shows a pronounced biphasic characteristic, with higher Mal concentrations outside suppressing ICl. We also find that Mal, as well as its precursor OAA (oxaloacetate), suppress ICl in a voltage-independent manner when present on the cytosolic side of the membrane, the effect of OAA evident well within its physiological concentration range. These findings point to control of the anion channels by OAA, indicating the capacity for feedback control from the cellular metabolic activity and a much more subtle ‘fine tuning’ of the anion channels by metabolism than previously anticipated.
MATERIAL AND METHODS
Plant growth and preparation
Vicia faba L. cv. Bunyard Exhibition was grown on potting mixture and perlite (70:30) at 22 °C and 60% relative humidity with 200 μmol·m−2·s−1 photosynthetic photon flux density under a 16 h day/8 h night cycle. Epidermal strips were prepared from newly expanded leaves of 4–6-week-old plants as described previously . Epidermal peels were affixed to the glass bottom of the experimental chamber after coating the chamber surface with an optically clear and pressure-sensitive silicone adhesive and all operations were carried out on an Axiovert S100TV microscope (Zeiss) fitted with Nomarski differential interference contrast optics. Measurements were conducted in continuous flowing solutions controlled by a gravity-fed system at a rate of 20 chamber volumes/min. The standard perfusion medium contained 15 mM TEA-Cl (tetraethylammonium chloride), 15 mM CsCl and 5 mM Mes, titrated with Ca(OH)2 to pH 6.1 ([Ca2+]=1 mM). The Na+ salts of organic acids and other chemicals were from Sigma unless specified otherwise.
Surface area and volume of impaled guard cells were calculated assuming a cylindrical geometry using Henry IV EP software (http://www.psrg.org.uk/henrys-ep-suite.htm). Single- and double-barrelled microelectrodes were prepared and coated with paraffin to reduce capacitance as described previously [19–21]. Current- and voltage-recording barrels were filled with 100 mM CsCl with additions as indicated. For measurements including organic acids in the microelectrode, K+ salts of the acids were used and the electrolyte solution was adjusted to pH 7.5 with CsOH and HCl as necessary to avoid cytosolic acid or alkaline loading. Microelectrodes were connected to amplfier headstages via 1 M KCl|Ag-AgCl halfcells, and a 1 M KCl agar bridge served as the reference electrode. Voltage-clamp recordings were carried out using Henry IV EP software using a two-electrode clamp circuit [20,21] with an additional μP amplifier (Y-Science). Current records were filtered using an eight-pole active Bessel filter (fc, 0.3 kHz) and clamp current and voltage data were digitized at 1 kHz. Free-running voltages between voltage-clamp recordings were digitized at 10 Hz.
All recordings were analysed and leak currents subtracted using standard methods  with Henry IV software. Where appropriate, data are presented as means±S.E.M. of n observations, and differences were validated by Student's t test or ANOVA. Curve fittings were by non-linear least-squares using a Marquardt–Levenberg algorithm .
Previous studies indicated an activation of anion currents by extracellular Mal in isolated protoplasts , but supporting evidence from intact guard cells has been equivocal . We recorded ICl with a two-electrode voltage clamp in intact guard cells, bathing the cells in 15 mM CsCl and 15 mM TEA-Cl, and using microelectrodes filled with 100 mM CsCl. Diffusion of Cs+ from the microelectrode and its presence with TEA+ outside blocks all K+ currents and Cl− diffusion from the microelectrode provides an effective ‘substrate clamp’ loading the cytosol within the first 5–10 min following impalements to enhance current carried by the anion [24–26]. Under these conditions, the membrane was dominated by ICl, similar to that reported previously with instantaneous current comprising both anion channels and characteristics in the steady-state, predominantly of the S-type anion current [9,26,27].
Biphasic activation of ICl by external Mal
To test the effects of Mal outside, after impalements, we superfused guard cells with malate concentrations up to 20 mM. Figure 1(A) summarizes the results from ten independent experiments, for clarity showing data only for the 0, 1 and 10 mM Mal, including recordings from one guard cell at the three Mal concentrations (insets below). Clamp-voltage steps to +40 mV led to a slow rise in outward current over 10 s (results not shown), and subsequent clamp steps to voltages more negative than approximately −40 mV yielded substantial inward-directed and instantaneous ICl that relaxed to steady-state values near −5 to −10 μA·cm−2 within 3–5 s. Both mean instantaneous and, to a lesser extent, steady-state (inset above) ICl were enhanced by 1 mM Mal. Amplitudes of the mean instantaneous current increased by 70–75% at voltages negative from −100 mV (from −52±7 to −91±9 μA·cm−2 at −220 mV) and near the negative maximum at −75 mV the mean steady-state current increased by approximately 60% from −13±1 to −22±3 μA·cm−2. Analogous to previous results , we observed an approximate −30 mV displacement in the voltage giving maximum steady-state ICl in 1 mM Mal and a marginally greater displacement in 10 mM Mal (inset above). However, in 10 mM Mal both instantaneous and steady-state ICl were reduced close to control values in the absence of Mal, and similar results were obtained with 20 mM Mal (results not shown). To assess further Mal action on ICl gating, we fitted ICl relaxations to derive the halftime, t1/2, for current deactivation at voltages more negative than −40 mV for all three sets of data. Figure 1(B) shows the voltage-dependence of t1/2 for ICl  and underscores the biphasic action of Mal on the current: whereas 1 mM Mal resulted in an increase in t1/2, consistent with a slowing of its relaxation, especially at move negative voltages (inset), additions of 10 mM Mal yielded t1/2 values statistically indistinguishable from the control at all voltages.
External Mal shows a biphasic modulation of ICl in
Vicia guard cells
Cytosolic Mal suppresses ICl
Although implicit in the connection to cellular metabolism, to date almost nothing is known about the influence of Mal on ICl from the cytosolic side of the membrane. Guard cells of many species, including Vicia, accumulate high concentrations of Mal to balance K+ uptake during stomatal opening [6,16,28]. Much of this Mal is distributed to the vacuole, but concentrations of 1–8 mM have been estimated in the cytosol [6,29]. To examine the effect of Mal on ICl from the inside, we impaled guard cells with double-barrelled microelectrodes as before, but with concentrations between 1 mM and 20 mM Mal in addition to 100 mM CsCl filling the microelectrode barrels to load the cytosol by diffusion from the microelectrode. Figure 2 summarizes data from ten or more independent experiments with each loading solution, in each case allowing at least 10 min following impalements for loading from the microelectrode [24–26]. Recordings from guard cells impaled with 1 mM Mal in the microelectrodes showed a statistically significant increase in ICl between −40 mV and −100 mV in the steady-state (Figure 2A), but the effect on the instantaneous current amplitude was marginal (inset) and no effect was evident in mean t1/2 values for ICl deactivation at any voltage (Figure 2B). By contrast, recordings from guard cells impaled with 10 mM Mal in the microelectrodes yielded a reduction in instantaneous ICl compared with the control, albeit with a smaller effect on the steady-state current (Figure 2A) and greatly reduced t1/2 (Figure 2B) independent of membrane voltage. Similar results were obtained with 20 mM Mal (results not shown), and fittings to a hyperbolic function for the t1/2 and amplitude of ICl (see Figure 8) yielded apparent Ki values for Mal of 3±1 mM and 10±3 mM respectively. Thus Mal suppressed ICl on transition to negative voltages, accelerating deactivation and reducing the magnitude of the current, but the effect on current amplitude was evident primarily at the upper end of the phyisological range of Mal concentrations.
Internal Mal shows a marginal stimulation of ICl at low millimolar concentrations in
Vicia guard cells
Cytosolic OAA blocks ICl
OAA is the four-carbon precursor in the formation of cytosolic Mal from phosphoenolpyruvate and is the immediate product of Mal oxidation; it is an intermediate in several metabolic pathways, including glycolysis, amino acid synthesis and in the mitochondrial tricarboxylic acid cycle, and it exchanges with Mal across both chloroplast and mitochondrial membranes [6,29]. Direct measurements of steady-state OAA concentration in the cytosol are not available, but estimates based on energy charge and Mal/OAA exchange transport equilibria suggest values of 0.1–0.5 mM and possibly as high as 3 mM in the light . Given its central importance to Mal balance, we examined the effects on ICl of OAA additions both outside and inside the guard cells. We found that OAA concentrations as high as 10 mM added outside had no appreciable effect on instantaneous or steady-state ICl (Figure 3A), neither did it alter channel gating when quantified using current relaxations on deactivation at negative voltages (Figure 3B). By contrast, cytosolic OAA loads introduced from the microelectrodes had a substantial effect on all three parameters. Adding 1 and 10 mM OAA inside suppressed the instantaneous ICl by approximately 58% and 70% respectively, compared with the control and independent of membrane voltage (Figure 4A). A similar block of steady-state ICl was observed in each case (Figure 4A, inset) and ICl kinetics showed a substantial acceleration in current deactivation, notable even with 1 mM OAA and again largely independent of membrane voltage (Figure 4B). Fitting to a hyperbolic function (see Figure 8) yielded apparent Ki values for OAA in accelerating ICl deactivation and reducing the current amplitude of 0.17±0.08 mM and 0.21±0.05 mM respectively, which is well within the physiological range of OAA concentrations. Additions of OAA, both inside and outside the guard cells, had no significant effect on the reversal voltage for ICl, as was evident from lack of significant difference in currents recorded either side of the apparent reversal voltage (see the legend to Figure 4), indicating that the relative permeability of the channels for OAA is very low compared with that for Cl− and Mal.
Vicia guard cells shows no appreciable sensitivity to external OAA
Internal OAA suppresses ICl at low millimolar concentrations in
Vicia guard cells
Given the potentially counteractive effects of the organic acids in the cytosol, we also examined whether OAA suppression of ICl was rescued in the presence of Mal. In this case, guard cells were impaled with microelectrodes containing 1 mM OAA and 1 mM Mal, and ICl was recorded as described above. Figure 5 summarizes results from 11 independent experiments with both organic acids. We found that OAA suppressed both instantaneous and steady-state ICl (Figure 5A) and accelerated current deactivation (Figure 5B). This effect was only partially ameliorated by the presence of low Mal concentrations which enhanced ICl (see Figure 2) and, again, the influence of both organic acids was voltage-independent (Figure 5B, inset). These results suggest that the immediate metabolic partner of Mal may have an important role in control of ICl under normal physiological conditions.
Internal OAA suppresses ICl in the presence of Mal in
Vicia guard cells
Ac (acetate) electrolytes have an impact on ICl
Finally, we explored the impact of Ac on ICl. The two-carbon acid anion is a substrate for the tricarboxylic acid cycle and fatty acid synthesis, but in a thioester complex with co-enzyme A, and normally occurs as the free acid at concentrations of 0.5–1.5 mM . Nonetheless, previous electrophysiological studies have used the free acid anion in microelectrode filling solutions to avoid cytosolic Cl− loading and its consequences for solute balance and background Cl− leakage across the membrane [24,25,32–34]. In the absence of information on the Ac-sensitivity of ICl, we challenged guard cells with 1 and 10 mM Ac by additions outside and to the microelectrode filling solution. Figures 6 and 7 summarize results of 13 or more independent experiments for each treatment. We found in every case that increasing Ac concentrations suppressed the instantaneous and steady-state ICl, in the latter case most noticeably at voltages between −40 and −100 mV, and it accelerated current deactivation. Unlike the effects of Mal and OAA, however, Ac showed a pronounced voltage-dependence in its effect on ICl kinetics, the reduction in t1/2 being most apparent at voltages near and more positive than −100 mV and when Ac was added to the cytosolic side of the membrane (Figures 6B and 7B), and it led to a shift in the apparent reversal potential for ICl (Figures 6A and 7A, insets). These results indicate a more complex interaction of Ac with the anion channels and we return to this point below.
Vicia guard cells show a sensitivity to external Ac
Vicia guard cells is suppressed by internal Ac
Both inorganic ion transport and organic acid metabolism make major contributions to the changes in osmotic content of the guard cells that drives stomatal movements [3,4,35,36]. During the diurnal cycle of stomatal opening and closing, the guard cells of most plant species accumulate and subsequently lose significant quantities of K+, Cl− and Mal, the latter offsetting Cl− especially later in the daylight period [16,37–39]. Much of the Mal synthesized in the light is transported and accumulates in the vacuole , but is later transported out of the vacuole to be metabolized or released across the the plasma membrane when the stomata close [28,40]. Thus transport, notably of Cl− and Mal, must be regulated in concert with metabolism, both as Mal is synthesized for transport to the vacuole and again as it re-enters the cytosol for export and to feed into the metabolic pools of other organic acids. Very little is known of the mechanisms behind this regulation, although it is clearly central to the osmotic homoeostasis of guard cells. We have examined the impact of cytosolic Mal, its immediate metabolite OAA, and of Ac on ICl, the current carried by anion channels at the plasma membrane of Vicia guard cells. Although we cannot discount some degree of metabolic conversion of the acid anions during these experiments, both the extracellular medium and the microelectrode present virtually infinite volumes, the latter exchanging with the cytosol directly. Thus at steady-state the membrane surface will have been exposed to organic acid concentrations effectively ‘clamped’ close to that of the corresponding solutions [24,25]. We found that all three organic acids affected ICl, but with markedly different characteristics and sidedness to their activities. Most prominent was the suppression of ICl with OAA evident even at low millimolar concentrations. These findings indicate a capacity for co-ordinating organic acid metabolism with ICl through the direct effect of organic acid pool size and they suggest that the effects may be exerted as much by Mal metabolites as by Mal itself. They also add perspective to in vivo recordings using Ac-based electrolytes.
Mal, OAA and ICl regulation
One of the most surprising findings of the present study was that ICl is substantially more sensitive to OAA in the cytosol than it is to Mal (Figure 8). Much as previously reported [14,17], we observed that low concentrations of Mal outside affected the voltage-dependence of the current. We also observed an enhancement of ICl with 1 mM Mal in the cytosol, although elevation of cytosolic Mal to 10–20 mM led to measurable declines in ICl relative to the control (Figure 2). Even additions even of 1 mM OAA, however, resulted in a significant block of ICl consistent with an apparent Ki near 0.1 mM and well within the dynamic concentration range for OAA in vivo . Mal synthesis in the cytosol of guard cells is fed largely from the pools of phosphoenolpyruvate and OAA . The flux through these metabolites will be balanced, in part, with Mal transport across the tonoplast and can lead to large accumulations of Mal in the vacuole, especially in the second half of the diurnal cycle [6,37]. By contrast, stomatal closure at the end of the day or following stimulation with abscisic acid must be accompanied by export of much of the Mal in the vacuole and its elimination through metabolism and loss across the plasma membrane [6,35]. These observations are generally consistent with variations in the low millimolar range previously reported for the pool of cytosolic Mal in leaves . However, such ‘channelling’ of Mal, in the first instance into the vacuole and in the second to the apoplast, implies a substantial hysteresis in Mal flux control overall and a fine control over Mal export both at the tonoplast and at the plasma membrane during the course of the day.
Vicia guard cells shows a high relative sensitivity to OAA (○) within the physiological concentration range compared with that for Mal (●)
If the cytosolic Mal pool remains within the bounds of 0.5–10 mM, how might ICl be regulated to prevent undue Mal efflux during stomatal opening, yet provide a major pathway for its loss during stomatal closure? One obvious mechanism rests with increasing [Ca2+]i which enhances the activity of ICl and both R- and S-type anion channels [9,14,27,41,42], in effect promoting Mal (and Cl−) efflux across the plasma membrane. This mechanism is certainly important in the response of guard cells to abscisic acid and drought stress [3,4], but it is more difficult to reconcile the diurnal cycle of [Ca2+]i with that of Mal synthesis and accumulation in the day and release at night. Guard cells undergo diurnal changes in [Ca2+]i, but the range of free concentrations in the steady-state [43,44] is substantially below the K1/2 of 600–700 nM for ICl activation by [Ca2+]i . Furthermore, the diurnal oscillation in [Ca2+]i is 180° out of phase with the changes that might be expected to modulate Mal in guard cells : [Ca2+]i increases in the day and declines at night, whereas the opposite would be required to facilitate Mal accumulation in the day and its efflux through the anion channels at night. In short, the diurnal changes in Mal do not sit well with the characteristics of ICl modulation by [Ca2+]i in guard cells.
However, the sensitivity of ICl to OAA does lend itself to another mechanism. It is generally thought that much of the fixed carbon in guard cells is imported and passes through glycolysis before condensation with bicarbonate to form OAA and Mal [6,45,46]. OAA is formed through the action of phosphoenolpyruvate carboxylase, which is activated by its phosphorylation in the light thus suppressing its sensitivity to inhibition by Mal [47,48]. Thus we might anticipate that sucrose uptake and metabolism to form Mal in the light will be accompanied by an elevation in the cytosolic OAA pool, even if NAD-Mal dehydrogenase is activated by light ; conversely, carbon passage through Mal and OAA to phosphoenolpyruvate and pyruvate, and transfer of both Mal and pyruvate to the mitochondria at the end of the day  is likely to be accompanied by a relative depletion of this same OAA pool. In other words, diurnal variations in cytosolic OAA are potentially in phase with the expected oscillation between Mal retention and stomatal opening during the day and ICl-mediated Mal release and stomatal closure at night. The Km for OAA of cytosolic NAD-Mal dehydrogenase  is very close to the apparent Ki for OAA of ICl in Vicia guard cells (Figures 4 and 8), thereby ensuring a close co-ordination between the two kinetic processes. We note, too, that low millimolar OAA inside suppresses ICl even in the presence of Mal (Figures 4 and 5). So it is conceivable that diurnal variations in the cytosolic OAA could play an important role in regulating ICl and Mal efflux across the plasma membrane. Indeed, in this context it is of interest to note that the diurnal oscillation of the OAA pool in various CAM (crassulacean acid metabolism) species is also in phase with stomatal movements , albeit with the higher OAA concentration and stomatal opening at night.
Electrophysiological studies with Ac electrolytes
Finally it is of interest that ICl showed a sensitivity to Ac. The monocarboxylic acid is normally present at low concentrations  relative to its action on the current (Figures 6 and 7), and is therefore unlikely to have an impact on ICl activity in vivo. Effects of weak-acid loading with Ac can also be ruled out because, in contrast with the the evidence shown in Figures 6 and 7, cytosolic acidification enhances anion channel activity . The action of higher Ac concentrations is relevant, however, because the caboxylic acid anion is commonly used as an electrolyte substitute to avoid the consequences of Cl− loading via the microelectrode [24,25,32–34]. It is clear that this practice does more than simply avoiding the effects of elevating Cl− in the cytosol: Ac loading leads to a block of ICl. The results of the present study, including parallel effects for Ac added outside and inside, the apparent shift in instantaneous current and current reversal voltage, and the strong voltage-dependence to its action on t1/2 (Figures 6 and 7), implies a complex dependence on Ac residence within the channel pore itself; furthermore, the voltage-dependence to Ac action (Figures 6B and 7B) and the more pronounced effect on instantaneous ICl when added outside (Figure 6A) may indicate a greater sensitivity of the R-type anion channels to Ac. Regardless, however, the block of ICl has had the serendipitous advantage of simplifying the kinetic analysis of the guard cell K+ channels and other currents through their isolation from anion flux through ICl.
From a physiological standpoint, the block of ICl by Ac will not influence other membrane currents under voltage clamp. However, it may well affect the balance of ionic currents under free-running (non-clamped) conditions. For example, suppressing ICl might be expected to promote Ca2+ entry and [Ca2+]i elevation initially by favouring membrane hyperpolarization, even with low millimolar concentrations of K+ outside [4,8,52]. Raising [Ca2+]i above 500–600 nM normally promotes a burst in ICl activity , but this activity is likely to be much suppressed by an Ac load with a consequent loss of the ‘damping’ effect of ICl in promoting membrane depolarization. In short, Ac loading may have led to an overestimate of the normal dynamic range for [Ca2+]i in its coupling with oscillations with membrane voltage [53,54]. Indeed, preliminary analysis using quantitative kinetic modelling of guard cell ion transport and homoeostasis has suggested that [Ca2+]i oscillations are likely to be constrained with an upper limit near 600 nM largely as a result of ICl activation .
In conclusion, we find that the anion channel current of Vicia guard cells shows a compound dependence, from the cytosolic side of the membrane, on several organic acids. Most notably, suppression of ICl by OAA is consistent with a role for this dicarboxylic acid in modulating channel activity for a diurnal cycle of Mal accumulation and release that parallels the cycle of stomatal opening and closing. Additionally, block of ICl by the monocarboxylic acid Ac demonstrates that its use in electrophysiological studies has serendipitous consequences for the study of other ion channels at the guard cell plasma membrane.
Yizhou Wang carried out the experimental work, analysed data and drafted the paper. Michael Blatt helped analyse data and wrote the paper.
We thank Amparo Ruiz-Prado for preparing plant materials and Dr Zhonghua Chen for helpful discussion during the preparation of the paper.
This work was supported by the Biotechnology and Biological Sciences Research Council (BBSRC) [grant numbers BB/F001673/1, BB/F001630/1 (to M.R.B.)]. Y.W. holds a China Scholarship Council Studentship with the University of Glasgow.