The Regulation of Anion Loading to the Maize Root Xylem^,

Abstract

The regulation of anion loading to the shoot in maize (Zea mays) was investigated via an electrophysiological characterization of ion conductances in protoplasts isolated from the root stele. Two distinct anion conductances were identified. In protoplasts from well-watered plants, Z. mays xylem-parenchyma quickly-activating anion conductance (Zm-X-QUAC) was the most prevalent conductance and is likely to load the majority of NO₃⁻ and Cl⁻ ions to the xylem in nonstressed conditions. Z. mays xylem-parenchyma inwardly-rectifying anion conductance was found at a lower frequency in protoplasts from well-watered plants than Zm-X-QUAC, was much smaller in magnitude in all observed conditions, and is unlikely to be such a major pathway for anion loading into the xylem. Activity of Z. mays xylem-parenchyma inwardly-rectifying anion conductance increased following a water stress prior to protoplast isolation, but the activity of the putative major anion-loading pathway, Zm-X-QUAC, decreased. Addition of abscisic acid (ABA) to protoplasts from well-watered plants also inhibited Zm-X-QUAC activity within minutes, as did a high free Ca²⁺concentration in the pipette. ABA was also seen to activate a Ca²⁺-permeable conductance (Z. mays xylem-parenchyma hyperpolarization activated cation conductance) in protoplasts from well-watered plants. It is postulated that the inhibition of anion loading into the xylem (an important response to a water stress) due to down-regulation of Zm-X-QUAC activity is mediated by an ABA-mediated rise in free cytosolic Ca²⁺.

The first manifestation of a plant water deficit is a reduction in cell expansion (Sharp et al., 1997), particularly in the shoot (Hsiao and Xu, 2000). In maize (Zea mays), the primary root will maintain substantial elongation even at a water potential of −1.6 MPa, whereas shoot development is completely inhibited at around −0.8 MPa (Sharp et al., 1997). It has been proposed that this difference between roots and shoots is due to both an increased ability of the roots to reduce their solute potential as well as a decreased ability of shoots to maintain wall extensibility (Hsiao and Xu, 2000). These effects of water deficit on cell expansion occur before any effects on stomatal conductance become apparent (Davies and Zhang, 1991; Sadras and Milroy, 1996).

In this article, we focus on mechanisms underlying reductions in root solute potential in maize under conditions of water deficit. Maize was chosen because of the ease with which the cortex and stele can be physically separated, facilitating study of the separate processes of uptake from the soil and loading into the xylem.

Water stress (and abscisic acid [ABA]) is known to increase solute accumulation within the root by having little effect on initial ion uptake but by significantly inhibiting release of ions into the xylem (Cram and Pitman, 1972; Pitman, 1977; Pitman et al., 1974a, 1974b; Pitman and Wellfare, 1978). Consistent with these classic observations, Roberts (1998) found that inwardly-rectifying K⁺ channel activity of the maize root cortex was insensitive to water stress or pretreatment of plants with ABA, whereas outwardly-rectifying K⁺ channel (KORC) activity of the stele was down-regulated. A rapid reduction by ABA in the concentration of mRNA transcript encoding stelar K⁺ outwardly-rectifying channel (SKOR) in Arabidopsis (Arabidopsis thaliana; Gaymard et al., 1998; a protein homologous in activity and sequence to that in maize; H. Sentenac, personal communication) suggests that some of the decrease in stelar KORC activity observed by Roberts (1998) was due to transcriptional control. Posttranslational control has also been postulated to be likely as application of ABA to a stelar protoplast elicited a decrease in KORC activity by 55% within 5 min, as did an increase in cytosolic free Ca²⁺ ([Ca²⁺]_cyt) concentration from 100 nm to 10 μm (Roberts and Snowman, 2000). Similar control of stelar K⁺ channel activity by [Ca²⁺]_cyt was also observed in barley (Hordeum vulgare; Wegner and De Boer, 1997). It is notable that the inhibition by ABA of outward K⁺ currents in stelar cells (Roberts, 1998) is opposite to the effects of ABA in guard cells (Blatt, 1990; Lemtiri-Chlieh, 1996; Schroeder et al., 2001, and refs. therein).

Although control by ABA of root K⁺ influx and loading into the xylem has been well studied (for review, see De Boer, 1999; De Boer and Volkov, 2003), our knowledge of the control of anion loading is less complete. In barley, three distinct anion conductances have been identified in xylem parenchyma cells: xylem-parenchyma quickly-activating anion conductance (X-QUAC), xylem-parenchyma slowly-activating anion conductance (X-SLAC), and xylem-parenchyma inwardly-rectifying anion conductance (X-IRAC; Kohler and Raschke, 2000). X-QUAC was identified as the conductance most likely to catalyze the loading of anions (NO₃⁻ and Cl⁻) to the xylem (Kohler et al., 2002). The activities of these anion conductances were also differentially regulated by an increase in [Ca²⁺]_cyt; X-QUAC and X-SLAC decreased in activity with increasing [Ca²⁺]_cyt, whereas X-IRAC increased in activity. However, the effects of ABA on anion conductances in the roots remain obscure, despite extensive studies in guard cells (e.g. Schroeder et al., 2001), nor has the mechanism of inhibition by ABA of either cation or anion conductances been probed.

The aim of this study was to identify the pathways for the loading of anions into the xylem in maize roots and to investigate the regulation of these pathways by water stress and ABA.

RESULTS

QUAC

A quickly activating current was detected at voltages both positive and negative of E_Cl across the plasma membrane of protoplasts derived from the stele (Fig. 1A). The quick activation of the current was followed rapidly by a partial inactivation at large voltage deviations from E_Cl, with the inactivation being more pronounced at larger voltages. The pseudo-steady-state current, measured after 4 s, reversed at −9 mV (Fig. 1B). This is close to E_Cl (−4 mV) and clearly positive of E_TEA (−∞) and negative of E_Ca (+136 mV). When E_rev was plotted against E_Cl with varying internal and external ionic conditions, it is clear that E_rev closely follows E_Cl (Fig. 1C). This is consistent with the currents being predominately due to the movement of Cl⁻. The conductance was, therefore, named Z. mays xylem-parenchyma quickly-activating anion conductance (Zm-X-QUAC), following the terminology of Kohler and Raschke (1998, 2000).

Figure 1.

A, Current trace from maize stelar protoplast showing Zm-X-QUAC elicited via the voltage protocol illustrated in the inset. Solutions contained 30 mm CaCl₂ in the bath and 50 mm TEA-Cl in the pipette (ESD:ISC); E_Cl = −4 mV. B, Current-voltage relation of current trace represented in A. •, Instantaneous current; ▪, noninactivating component. Reversal potential for the noninactivating component = −9 mV. C, Reversal potential of Zm-X-QUAC shifts with changes in E_Cl. Data are shown as a mean of E_rev values with SEM (n = 4 for each E_Cl). E_Cl was varied by changing both external and internal solutions. In ascending order, solutions contained (in mM): 60 Cl⁻ in the bath and 50 Cl⁻ in the pipette (ESD:ISC), E_Cl = −4 mV; [Cl⁻]_ext:[Cl⁻]_cyt 50:50 (ESC:ISA), E_Cl = 0 mV; 60:100 (ESD:ISB), E_Cl = +10 mV; and 10:50 (ESE:ISA), E_Cl = +35 mV. Line drawn is E_rev = E_Cl. D, Niflumate inhibits noninactivating component of Zm-X-QUAC (n = 7). Sigmoidal (Hill equation) fit used; EC₅₀ = 17.5 μm, Hill slope = −0.93.

The current-voltage relationship of Zm-X-QUAC is not linear (Fig. 1B). Although outward rectification of the current was observed in all ionic conditions tested, it was clear that substantial amounts of Cl⁻ were able to move in both directions. Therefore, depending upon membrane voltage, Zm-X-QUAC may facilitate either Cl⁻ loading into, or Cl⁻ unloading out from, the xylem.

Ion substitution experiments confirmed that, in addition to Cl⁻, Zm-X-QUAC was highly permeable to NO₃⁻ and I⁻ (Table I). From the mean conductance measured in the presence of various anions, a selectivity sequence could be calculated of NO₃⁻ > Cl⁻ > I⁻ > malate²⁻ > SO₄²⁻ > citrate³⁻. This broadly resembled the positive shifts in the reversal potential observed upon substitution of the external anions (Table I).

Table I.

Permeation of various anions (x) through the quickly activating anion conductance in the plasma membrane of protoplasts isolated from maize stelar parenchyma (Zm-X-QUAC), measured by reversal potentials in biionic solutions, and inward currents and conductances

All experiments were conducted with 50 mM Cl⁻ in the pipette (solution ISC) and external solutions ESN, ESA, ESI, ESK, ESM, and ESL (running from top to bottom) with the maintenance of at least 0.1 mm Ca²⁺. I_x/I_Cl⁻ was calculated for currents at +120 mV. G_x/G_Cl⁻ was calculated from the chord conductance from E_rev to 60 mV positive of E_rev. Values are the average of two measurements.

[Anion] in the Bath		Reversal Potential	Ix/ICl−	Gx/GCl−
mM		mV
Nitrate	20 NO3−	3.8	1.7	1.0
Chloride	20 Cl−	32.5	1.0	1.0
Iodide	20 I−	64.0	0.94	0.59
Sulfate	10 SO42−	82.8	0.11	0.12
Malate	10 C4H6O42−	82.6	0.02	0.18
Citrate	6.6 C6H5O73−	100.3	0.01	0.10

The activity of Zm-X-QUAC also appears dependent on the presence of cytosolic MgATP. When MgATP is excluded from the pipette, currents declined to negligible levels within 5 min of forming the whole-cell configuration, even with [Ca²⁺]_cyt highly buffered to 100 nm with 20 mm 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA; Alexandre and Lassalles, 1992; ISF, n = 5; data not shown; see Gilliham, 2002).

Zm-X-QUAC was reversibly inhibited by the presence of niflumate, a known anion channel blocker, on the external face of the protoplast, with an approximate EC₅₀ of 18 μm with a Hill slope of close to −1 (Fig. 1D). GdCl₃, a known cation channel blocker, had no inhibitory effect on the inward or outward currents either pre- or postinactivation when applied externally at up to 10 mm (data not shown; see Gilliham, 2002).

IRAC

In some protoplasts isolated from maize stelar tissue, another conductance was measured, this being detected as distinct single channel-like opening/closing transition events during whole-cell patch-clamp experiments. These observations suggest either a low channel density or a low probability of opening (Fig. 2A).

Figure 2.

A, Zm-X-IRAC seen as single channel events recorded in whole-cell configuration. Traces are from one protoplast, showing gap-free recordings held at a constant voltage, as indicated. C indicates closed state, O_n indicates open configurations. Two active channels can be observed in most recordings. Solutions contained 1 mm CaCl₂ in the bath and 50 mm CsCl and 50 mm TEA-Cl in the pipette (ESG:ISG). B, Reversal potentials of currents through the channel (E_rev) follow E_Cl. Current-voltage relations from two protoplasts, each with one of the following solution combinations (in mm): •, [Cl⁻]_ext:[Cl⁻]_in = 2:100 (ESG:ISG, E_Cl− = +55 mV, E_rev = +53 mV), unitary conductance, G = 90 pS; and ▪, 10:50 (ESE:ISA, E_Cl = +35 mV, E_TEA = −35 mV, E_rev = +23 mV). G = 80 pS.

Single channel current-voltage relations showed that the current was mainly carried by anions, with E_rev shifting from approximately +23 mV to +53 mV with a shift in E_Cl from +35 mV to +55 mV (Fig. 2B). The reversal potential for the accompanying cation was consistently very negative of E_rev. The unitary conductance determined for the inward current component was 90 pS with 100 mm Cl⁻ in the pipette and 80 pS with 50 mm pipette Cl⁻. The gating of these apparent channels was strongly voltage dependent, leading to a distinct inward rectification of current through the channels when MgATP was present in the pipette solution (Fig. 2). These channels were therefore given the moniker Z. mays inwardly-rectifying anion conductance (Zm-X-IRAC). Zm-X-IRAC was not only permeable to Cl⁻ but also to NO₃⁻ (data not shown; Gilliham 2002).

Occurrence of Anion Conductances

Under varying ionic conditions, but with the maintenance of approximately 100 nm [Ca²⁺]_cyt, Zm-X-QUAC was found in 29% of protoplasts (n = 24/82). Zm-X-IRAC was found in 17% of protoplasts (n = 14/82), and no current was present in 59% (n = 48/82). Both currents were seen in 5% of protoplasts (n = 4/82).

Current densities were calculated at −120 mV (Table II), this being the approximate resting membrane potential of maize stelar tissue (Roberts and Snowman, 2000). Data are presented only from experiments using identical solutions in order to emphasize trends (mM; [Cl⁻]_ext:[Cl⁻]_in 40:50 Cl⁻, 20 BAPTA_in, ESB:ISC). It is clear that the current density for Zm-X-QUAC is much greater than that of Zm-X-IRAC at −120 mV (Table II). It is therefore considered fairly intuitive and reasonable to hypothesize that Zm-X-QUAC is normally likely to catalyze significantly more anion efflux from stelar cells and thus contributes more to xylem anion loading than Zm-X-IRAC. To test this hypothesis further, the effects of water stress and ABA on Zm-X-QUAC activity were investigated to determine whether the conductance may be regulated in a fashion consistent with the reported reduction of anion loading of the xylem by these two factors.

Table II.

Frequency of occurrence of anion currents in protoplasts isolated from maize stelar parenchyma cells and the mean current densities, averaged from all protoplasts

Current densities were measured at −120 mV in the plasma membrane of protoplasts suspended in 20 mm CaCl₂ (solution ESB), with the patch clamp pipette containing 50 mm TEA-Cl and a [Ca²⁺]_cyt of 100 nm (solution ISC). Means provided with SEM, n = 15.

	Frequency of Occurrence	Idensity at −120 mV
		nAm−2
Zm-X-QUAC	5/15 (33%)	−107 ± 39
Zm-X-IRAC	2/15 (13%)	−0.71 ± 0.51
No current	9/15 (60%)

Effects of Water Stress, ABA, and High [Ca²⁺]_cyt

Current-voltage relations of stelar protoplasts obtained from water-stressed plants clearly showed that mean current densities through Zm-X-QUAC were lower at all voltages compared with those measured in protoplasts isolated from well-watered plants (Fig. 3A). A 12-h pretreatment with 20 μm ABA prior to protoplast preparation, or an increase of [Ca²⁺]_cyt from 100 nm to 1 μm, also inhibited the mean current density of Zm-X-QUAC in well-watered plants at all voltages (Fig. 3, A and B). These reductions are summarized in Figure 3C, the reduction being attributed to both a smaller proportion of protoplasts expressing the conductance and a decrease in conductance in the protoplasts in which currents appeared.

Figure 3.

Water stress, ABA, and high [Ca²⁺]_cyt decrease quickly, activating anion conductance (Zm-X-QUAC) activity. Current-voltage relation of the steady-state, postinactivated current through Zm-X-QUAC following manipulation of (A) growth conditions of whole plants prior to protoplast extraction or (B) [Ca²⁺]_cyt. All results obtained using ([Cl⁻]_ext:[Cl⁻]_in) 40:50 (ESB:ISC) except for (▾) 50:50 (ESC:ISE). ▪, Protoplasts obtained from well-watered (control) plants (WW, n = 15); •, protoplasts prepared from whole plants from which water was withdrawn 72 h prior to harvest (WS, n = 10); ▴, protoplasts prepared from whole plants watered with 20 μm ABA 12 h prior to harvest (ABA, n = 8); and ▾, protoplasts prepared from well-watered plants but patch clamped using solution containing 1 μm [Ca²⁺]_cyt (ISE) instead of 100 nm as in all other protoplasts, (Ca²⁺, n = 8). Error bars denote SEM. C, Occurrence of Zm-X-QUAC in protoplasts; protoplasts where current is present (white) compared to total protoplast pool. Current densities of inactivated Zm-X-QUAC at −120 mV; average current density was calculated from all protoplasts within a given treatment. Symbols and abbreviations are as in the first two parts of this figure. A two-way ANOVA gives significance at 0.053.

Rapid Effect of ABA on Zm-X-QUAC

Mean current densities at −120 mV were compared in protoplasts that were patch clamped at different times after exposure of the intact protoplasts to ABA (Fig. 4A). Within 40 min of the application of 20 μm ABA, the activity of Zm-X-QUAC decreased to 20% of the activity in control protoplasts, from −107 ± 39 mAm⁻² (n = 15) to −23 ± 8.8 mAm⁻² (n = 3; Fig. 4A).

Figure 4.

A, ABA rapidly decreases anion conductance (Zm-X-QUAC) activity. Time course of the effect of 20 μm ABA on the pseudo-steady-state current densities of Zm-X-QUAC at −120 mV after 4 s pulses, 5 min after whole-cell establishment. ABA is applied to intact protoplasts in holding solution, which are then patch clamped after the specified times using [Cl⁻]_ext:[Cl⁻]_in, 40:50 (ESB:ISC). Data from all protoplasts are averaged whether or not Zm-X-QUAC could be discerned. B, Zm-X-QUAC activity is decreased by ABA through a Ca²⁺-mediated pathway. Four second voltage pulses at −120 mV were applied to patch-clamped protoplasts every 30 s from a holding voltage of +30 mV (near E_Cl−). Instantaneous current at −120 mV at each time point was divided by the peak (most negative) current exhibited by that protoplast at −120 mV, which occurred approximately 5 min after whole-cell establishment. Twenty micromolar ABA was applied to protoplasts 5 min after the establishment of the whole-cell configuration (indicated by arrow). Data plotted are the running averages of the normalized currents obtained from two protoplasts. Every second data point has been excluded for clarity (see Gilliham [2002] for complete figure). All measurements were made with an external solution containing 20 mm CaCl₂ (ESB). The rundown of the instantaneous current was measured in control protoplasts (□), which contained weakly Ca²⁺ buffered internal solution (5 mm BAPTA, ISA), protoplasts treated the same way but with ABA added (▪), and protoplasts with highly buffered internal Ca²⁺ (20 mm BAPTA, ISC) and to which ABA was added (•). Values shown are the average of n = 2 for each treatment. Note the different time scale for the two parts to this figure.

The time course for ABA control of Zm-X-QUAC was investigated further by the application of 20 μm ABA directly to patch-clamped protoplasts (Fig. 4B). All experiments were performed with a high external Ca²⁺ concentration ([Ca²⁺]_ext = 20 mm CaCl₂) to aid seal integrity. In these conditions, Zm-X-QUAC activity at −120 mV ran down to 66% ± 8% (n = 2) of the maximal current over 20 min. However, if 20 μm ABA was applied to the protoplast 5 min after establishment of the whole-cell configuration (approximately the time at which a protoplast shows maximal Zm-X-QUAC activity), the current at −120 mV was reduced to 26% ± 7% (n = 2) of its maximal value within 15 min of the application of ABA. Therefore, it was clear that ABA significantly increased the rate of rundown of Zm-X-QUAC. If [Ca²⁺]_cyt was highly buffered (i.e. contained 20 mm BAPTA), no appreciable rundown was observed in either control protoplasts or protoplasts to which ABA was added (Fig. 4B). Protoplasts patch clamped with a pipette solution containing high concentrations of BAPTA possessed 94% ± 3% (n = 3) of their Zm-X-QUAC activity after 20 min whether or not they were treated with ABA (Fig. 4B).

It should be noted that a mild water stress or application of 50 μm ABA to well-watered plants 12 h prior to protoplast isolation increased the activity and occurrence of Zm-X-IRAC, as did either increasing [Ca²⁺]_cyt, or applying 20 μm ABA to protoplasts obtained from well-watered plants. Occurrence increased from 13.3% (n = 15) in well-watered plants to 26.6% (n = 10) in water-stressed plants; 25% in ABA pretreated plants (all in ESB:ISC; n = 8); and 50% in plants patch clamped with 1 μm [Ca²⁺]_cyt (ESB:ISE; n = 8; data not shown; see Gilliham, 2002). Twenty micromolar ABA applied directly to protoplasts obtained from a well-watered plant resulted in a rapid increase in the probability of opening of Zm-X-IRAC. However, the resulting increase in current density of Zm-X-QUAC was still 10 times less than that of Zm-X-QUAC at −120 mV (data not shown; see Gilliham, 2002). The up-regulation of Zm-X-IRAC was not investigated further, as the increase in activity was not sufficient to compensate for the conceived reduction in anion flux due to down-regulation of Zm-X-QUAC activity by water stress signals.

Activation of a Ca²⁺ Conductance by ABA

In addition to decreasing the activity of Zm-X-QUAC, ABA also stimulated a conductance at hyperpolarized potentials with time-dependent activation in 82% of cells (n = 18/22; Fig. 5, A and B). These currents were particularly apparent when solutions were designed to maximize the appearance of divalent cation currents. With 20 μm ABA, maximal activation of these currents occurred within 25 to 30 min (Fig. 5, A and B), with average currents in stelar protoplasts in 2.5 mm Ba-HEPES:5 mm BaCl₂ (ESO:ISI) at −160 mV reaching 167 ± 30 pA (n = 18). These currents were insensitive to 5 mm H₂O₂, even after 40 min (n = 5). It should be noted that an ABA-independent activation of the current was also observed in control protoplasts, but at a lower frequency than in ABA-treated protoplasts (25% of protoplasts within 30 min; n = 16).

Figure 5.

ABA-activation of a Zm-X-HACC. A, Current-voltage plot showing a time course for the development of a Zm-X-HACC following the application of 20 μm ABA. Solutions contained 2.5 mm Ba-HEPES:5 mm BaCl₂ (ESO:ISI). Data obtained from a single protoplast displaying typical current responses. B, ABA increases Zm-X-HACC activity over time. Currents at −160 mV were compared to those prior to addition of ABA in protoplasts patch clamped with 2.5 mm Ba-HEPES:5 mm BaCl₂ (ESO:ISI, n = 10). Currents have been normalized by dividing the current at each time point with the original time-dependent current at t = 0. Different symbols indicate measurements from different cells. C, Effects of internal anion and external cation substitution on the ABA-stimulated hyperpolarization-activated current. The top two sets of traces show the small effect of internal anion substitution and the bottom two sets of traces demonstrate the cation selectivity of the currents. Concentrations of main ions indicated on the figures, solutions used were ESO:ISK (top left), ESO:ISJ (top right), ESR:ISL (bottom left), and ESP:ISL (bottom right). Currents shown in response to voltage steps from +40 mV to −160 mV, from a resting potential of 0 mV ± junction potential (*), departures separated by 16 s. Junction potentials were calculated as −1.2, −10.1, −8.0, and −2.1 mV, respectively. D, Current-voltage plot showing increase in inward current with increasing external [Ba²⁺] (using solutions ESW, ESO, and ESX:ISJ). The instantaneous current component was subtracted from the mean current between 900 and 1,000 ms and plotted. E, Current-voltage plots showing greater permeability of Zm-X-HACC to divalent cations over monovalent species. Data plotted were obtained from voltage ramps (voltage protocol inset) executed 20 to 30 min after the application of 20 μm ABA to a patch clamped protoplast. External solutions of equal normality (ESO to ESV; Supplemental Table III) were exchanged in sequence, which was then repeated in reverse order to ensure no time effects could account for the observed changes in currents. Typical results are displayed. Similar results were seen in four other protoplasts.

Activation of this conductance by ABA was repressed by inclusion in the pipette of either 400 μm MgATP (ISM; n = 9/10) or the nonhydrolysable analog MgATP-γ-s (ISN; n = 6/6) with 2.5 mm [Ba²⁺]_ext; if protoplasts were exposed to higher concentrations of divalent cations, seal integrity was lost upon current activation at hyperpolarized potentials. The presence of cytosolic Mg²⁺ had no inhibitory effect upon the activation of the conductance by ABA (n = 5/6; ISP), in contrast to that observed in Vicia faba guard cells (F. Lemtiri-Chlieh, personal communication).

Substitution of pipette Cl⁻ with gluconate did not significantly affect the hyperpolarization-activated current (Fig. 5C). Similar results were seen by substitution of pipette Cl⁻ with HEPES (data not shown). (Backfilling the pipette with a solution containing 10 mm KCl around the AgCl half-cell in the above ionic conditions had no effect on the current characteristics.) As the time-dependent current was still active in protoplasts with only channel impermeant anions on the cytosolic side of the membrane, the inward current was most likely to be due to the influx of the external cation, which in this case was Ba²⁺. In addition, the currents showed clear dependence on the concentration of external cations (Fig. 5D). Accordingly, the conductance allowing this current is referred to as Z. mays xylem-parenchyma hyperpolarization-activated cation conductance (Zm-X-HACC).

The magnitude of Zm-X-HACC was dependent on the identity of the external cation. The permeant cation was exchanged while keeping the normality constant (the concentration of charges, i.e. 5 mm K⁺ = 5 mm of positive charges, as does 2.5 mm Ba²⁺). The greater permeability of divalent cations compared with monovalent cations of Zm-X-HACC can be seen clearly in Figure 5E. At −160 mV, the current ratio through Zm-X-HACC for Ba²⁺:Ca²⁺:Na⁺:K⁺ was 23:18:2.2:1.0.

DISCUSSION

Two anion-selective conductances in the cells of the stele of maize are described, these being distinct in both their activation kinetics and voltage dependence. These conductances share many similarities with two anion conductances found in barley xylem parenchyma cells (Kohler and Raschke, 2000), although a third type of anion current found in barley, X-SLAC, was not found in the maize stele. Unlike the aluminum-activated anion channels characterized in the root tips of maize and wheat (Triticum aestivum), they appear not to allow the permeation of significant amounts of organic acid (Piñeros and Kochian, 2001; Zhang et al., 2001). Zm-X-QUAC and X-QUAC from barley share a profound outwardly-rectifying characteristic with an anion channel characterized in wheat root cortical cells (Skerrett and Tyerman, 1994); however, unlike the wheat cortical channel, they also appear to allow significant efflux of anions out of the cell as well as influx at hyperpolarized and depolarized potentials respectively.

As in barley, it is concluded that X-QUAC is most likely to be responsible for the majority of anion efflux into the xylem as it was more frequently detected than Zm-X-IRAC in well-watered conditions and had a greater current density at physiological resting membrane potentials in all conditions. Although Zm-X-QUAC was the most frequently observed conductance, it was observed in only one-third of protoplasts in well-watered conditions. The limited detection of anion conductances, and the fact that these channels were not observed in excised patches, may suggest a heterogeneous or even polar distribution of the channels within xylem associated cells. This observation is consistent with the xylem transfer hypothesis of Kramer (1981), suggested to likely occur by DeBoer and Volkov (2003), where ions are thought to travel through the symplast of stelar cells until they reach the highly invaginated membrane underlying a xylem vessel pit. From here, ion transporters move ions directly into the xylem vessel lumen. The inability to observe Zm-X-QUAC activity in membrane patches could also indicate the requirement of additional cytosolic factors for activity.

Control of Anion Loading to the Xylem

Water stress, ABA pretreatment, and high [Ca²⁺]_cyt all appeared to inhibit Zm-X-QUAC activity to a similar degree (Fig. 3). If the internal patch solution were only lightly buffered with BAPTA (5 mm), at a level insufficient to maintain [Ca²⁺]_cyt at physiological resting levels (i.e. 100 nm) in the face of high [Ca²⁺]_ext, the activity of Zm-X-QUAC could be seen to decrease over time. The rate of this rundown was increased by application of ABA (Fig. 4B). However, when [Ca²⁺]_cyt was fixed at physiologically resting levels, using 20 mm BAPTA (Alexandre and Lassalles, 1992), both normal rundown of Zm-X-QUAC and that exaggerated by ABA was inhibited (Fig. 4B). The fact that the application of ABA to protoplasts was not seen to inhibit Zm-X-QUAC activity when [Ca²⁺]_cyt is highly buffered precludes the direct regulation of the conductance by ABA. It also appears likely that a rise in [Ca²⁺]_cyt is necessary for ABA to manifest its regulation of Zm-X-QUAC.

Activation of a Ca²⁺ Conductance

One of the first responses of guard cells' membranes when challenged with ABA is a stimulation of HACC activity through a pathway mediated by reactive oxygen species (e.g. Pei et al., 2000). However, although a similar conductance appears to be activated by ABA in the maize stele (Zm-X-HACC), this conductance appeared insensitive to H₂O₂ and did not require the presence of NADPH in the cytosol, in contrast to guard cell HACCs (Pei et al., 2000; Kohler and Blatt, 2002). In V. faba guard cells, (de)phosphorylation cascades have also been implicated in the regulation of HACC activity (Kohler et al., 2003), but these cascades also appear to be irrelevant in stelar cells, as Zm-X-HACC activation could seemingly be inhibited either by MgATP or by its nonhydrolysable analog, and inhibition appears not to require Mg²⁺. This raises the possibility that ATP blocks this conductance in a mechanism akin to that of K_ATP channels (e.g. Tucker et al., 1998). The HACC of the guard cells and root stele also appear to differ in their selectivity, with guard cell HACCs being nonselective for divalent cations over monovalent cations (e.g. Pei et al., 2000; Kohler et al., 2003), but stelar HACCs showing distinct selectivity for divalent cations over monovalent cations (like root hair and suspension cultured HACCs: Véry and Davies, 2000; Gelli and Blumwald, 1997). The ABA sensitivity of these channel types has yet to be investigated, but it has been reported that the Arabidopsis root hair HACC is activated by reactive oxygen species, although in the absence of NADPH in the pipette, H₂O₂ was also unable to activate the HACC on its own (Foreman et al., 2003). Zm-X-HACC may therefore have a novel combination of properties for appropriate regulation within its parent tissue.

Zm-X-HACC activation by ABA within the time course for inhibition of Zm-X-QUAC suggests a potential role for Zm-X-HACC in an increase in [Ca²⁺]_cyt that would facilitate the decrease in Zm-X-QUAC activity. ABA has been reported to increase [Ca²⁺]_cyt in maize root tissue (Gehring et al., 1990), and it would be interesting to test whether activation of Zm-X-HACC could facilitate this increase.

The activity of Zm-X-IRAC was increased by ABA, water stress, and high [Ca²⁺]_cyt. It has been hypothesized that the equivalent channel found in barley xylem parenchyma Hv-X-IRAC could help to regulate the membrane potential of cells by providing a further pathway for counterions to short circuit the stelar proton pump (Kohler and Raschke, 2000). This hypothesis seems to gain credit in light of the response of Zm-X-IRAC to ABA. ABA has been shown to hyperpolarize the cells within the stele of maize (Roberts and Snowman, 2000) and is known to increase the activity of an H⁺-ATPase (Clarkson and Hanson, 1986). Hyperpolarization of the stelar cells would also serve to increase the driving force and magnitude of Ca²⁺ entry into the cytosol through Zm-X-HACC.

Transcriptional Versus Posttranslational Control of Zm-X-QUAC by ABA and [Ca²⁺]_cyt

Some transport proteins in the stele appear to turn over rapidly, in the order of tens of minutes, as revealed by the rapid inhibition of ³⁶Cl⁻ loading into the xylem by addition to intact roots of the amino acid analog, fluorophenylalanine (Schaefer et al., 1975). Likewise, addition of ABA to intact roots inhibits ³⁶Cl⁻ and ⁸⁶Rb⁺ (K⁺) efflux into the stele within 2 h (Cram and Pitman, 1972) and reduces levels of SKOR mRNA within 1.5 h (Gaymard et al., 1998). In addition to the potential for the direct, rapid control of X-QUAC by Ca²⁺ (Figs. 3B and 4B), the reduction in Zm-X-QUAC activity after exposure of intact protoplasts to 20 μm ABA for less than 40 min (Fig. 4A) suggests that there is also the scope for slower Ca²⁺-mediated control by ABA at the translational or transcriptional level such as occurs in guard cells (Webb et al., 2001). The notion that there are two levels of control for the loading of ions to the xylem have also been proposed for K⁺ loading through the SKOR homologs (De Boer, 1999; Tester, 1999; Roberts and Snowman, 2000). However, until proteins responsible for the anion conductances observed in the stele are identified, this remains only a hypothesis.

The requirement for intracellular MgATP for observation of Zm-X-QUAC activity indicates that the conductance could be activated in phosphorylating conditions, a feature common to anion channels (Barbier-Brygoo et al., 2000). Significantly, dephosphorylating conditions have been found to down-regulate the activity of the S-type anion channel and have been implicated in ABA signaling (Pei et al., 1997). Although MgATP is required for maintaining Zm-X-QUAC activity, it cannot sustain activity with high [Ca²⁺]_ext and low [BAPTA]_in. Although ABA retarded the activation of Zm-X-HACC in the presence of [MgATP]_in, this was with quite different solutions, with a much lower [Ca²⁺]_ext to those used for observing ABA inhibition of Zm-X-QUAC. Different solutions had to be used, unfortunately, because seal integrity was lost when similar [Ca²⁺]_ext were used as for studying Zm-X-QUAC in conditions that maximized Zm-X-HACC.

CONCLUSION

During water stress, roots of maize maintain elongation far beyond the point at which the shoot ceases to grow (Sharp et al., 1997). This observation correlates with the accumulation of solutes in the roots upon water stress (Sharp and Davies, 1979) and the reduction of xylem loading of solutes by ABA (Cram and Pitman, 1972). It is proposed that these observations constitute a mechanism by which roots have the resources to sustain growth during water stress and so potentially alleviate water stress, while not contributing to the shoot ion load and exacerbating the effects of water stress in the aerial portion of the plant. We propose that a water deficit increases local ABA concentrations, stimulating entry of Ca²⁺ into the cytosol of stelar cells that down-regulates Zm-X-QUAC activity through an unknown mechanism, so reducing anion entry into the xylem. These effects compare favorably with those observed for other ion transporters in the stele and are summarized in Figure 6. The findings within this article provide an insight into a mechanism by which maize could regulate ion transport between the root and shoot during water stress, contributing to the associated accumulation of ions in the root.

Figure 6.

Simplified model of xylem loading as regulated by water stress in maize. Zm-X-inwardly-rectifying K⁺ channel and other, putative transporters have been omitted. Water stress is indicated by the decrease in soil water potential (ψ_soil). This initiates: (1) increased root ABA concentration, possibly through the detection of cell volume changes within root cells (Jia et al., 2001); (2) activation of HACC (Fig. 5), a process inhibited by cytosolic MgATP and nonhydrolysable ATP analogs; (3) a rise in [Ca²⁺]_cyt; (4) possible activation of Ca²⁺/H⁺ antiport mechanism that would down-regulate SKOR activity through; (5) acidification of the cytosol (stelar symplast); or (6) alkalinization of the apoplast (xylem; Roberts, 1998); (7) an indirect [Ca²⁺]_cyt-dependent inhibition of SKOR activity (Roberts, 1998); and (8) an indirect [Ca²⁺]_cyt-dependent inhibition of QUAC activity (Figs. 3 and 4; QUAC activity is also MgATP-dependent); (9) a [Ca²⁺]_cyt-dependent increase in activity of IRAC (Gilliham, 2002); (10) ABA also appears to activate the H⁺-extruding ATPase in a Ca²⁺-independent manner (Clarkson and Hanson, 1986). This summary is restricted to posttranslational effects of ABA on ion transport, but transcriptional effects have also been demonstrated (for SKOR, by Gaymard et al., 1998) and are proposed (for Zm-X-QUAC, in this work).

MATERIALS AND METHODS

Plant growth, protoplast isolation, and patch-clamp electrophysiology were carried out as described by Roberts and Tester (1995) and Roberts (1998), with the following modifications (for more detailed methods, see Gilliham, 2002). Protoplasts were adhered to a coverslip treated with 75,000 molecular weight poly-d-Lys (Sigma, Gillingham, UK) in an irrigation chamber (of 200-μL volume) with a fast volume exchange (1 mL min⁻¹). The generation of voltage test pulses, recording of whole-cell currents, and storage of data were undertaken using an Axon 200A amplifier under the control of Clampex 8.21 (Axon Instruments, Union City, CA) and a Pentium II computer. Whole-cell currents, digitized at 5 kHz by a Digidata 1200A (Axon Instruments) interface, were filtered with a corner frequency of 2 kHz (8-pole Bessel filter) and stored on the computer. Capacitance and series resistance were compensated using the amplifier. Seal resistance was monitored regularly and data was only analyzed if protoplasts had a seal resistance of over 1 GΩ. Current-voltage relations were obtained from current traces, resulting from voltage step protocols, using the analysis package Clampfit 8 (Axon Instruments). Data analyses were conducted using Excel 2000 (Microsoft, Redmond, WA) and all graphs constructed and data fitted using Origin 5 (Microcal, Northampton, MA) or Prism 3.0 (GraphPad, San Diego). Variations in data are always presented as the se of the mean.

Where appropriate, current densities were obtained per protoplast by dividing the current of a protoplast, either pre- or postinactivation, by its membrane area obtained from a measurement of the protoplast membrane capacitance. The plant membrane was assumed to have a specific capacitance of 7.6 mFm⁻² (Homann and Tester, 1997).

Recordings from outside-out patches were analyzed as in Kohler and Raschke (2000), with steady-state open probabilities of patches with multiple channels determined from a ratio for the total open time to the total recording time, derived from amplitude histograms. Single channel current density was calculated by multiplying current amplitude by the probability of opening as in Kohler and Raschke (2000).

It is worth noting that the water stress imposed in this study was small and slow, with the identical treatment imposed by Roberts (1998) causing only a 2% decrease in relative water content over the 72 h of the withdrawal of watering.

Solution Composition

A variety of solutions were used to investigate both the anion and cation currents in an attempt to reduce the artifacts that may be associated with the use of particular impermeant ion species and channel blockers and hence increase the robustness of the data. The main species can be found in appropriate figure legend or text. TEA⁺, Cs⁺, Ca²⁺, and Gd³⁺ have been used to reduce K⁺ or cation currents, external Ca²⁺ was used to maintain a giga-Ω seal resistance (a requirement for data to be included), niflumate to block anion channels, and Mg-ATP on the cytosolic side to observe Zm-X-QUAC. Ion equilibrium potentials were calculated after correcting for ionic activities using GEOCHEM-PC (Parker et al., 1995) and CALCIUM (Chang et al., 1998). All solutions were filtered (0.22 μm; Millipore, Stonehouse, UK) before use. Pipettes were filled with filtered solutions using a Microfil syringe needle (World Precision Instruments, Sarasota, FL). Junction potentials were measured as described by Ward and Schroeder (1994). Detailed solution composition is listed in the supplementary material (Supplemental Tables I–III); solution acronyms (ESX:ISX) are also consistent with Gilliham (2002), to aid cross referencing.

Variation in the data is always presented as SEM, calculated as the standard deviation divided by the square root of the number of replicates.