Plasma membrane anion channels in higher plants and their putative functions in roots

Authors


Author for correspondence:Stephen K. Roberts Tel: +1524 593145 Fax: +1524 843854 Email: s.k.roberts@lancaster.ac.uk

Abstract

Contents

  •  Summary 647

  • I. Introduction 648
  • II. Overview of higher plant plasma membrane  anion channels 648
  • III. Anion channels in higher plant roots 655
  • IV. Conclusions and prospects 661
  •  Acknowledgements 662

  •  References 662

Summary

Recent years have seen considerable progress in identifying anion channel activities in higher plant cells. This review outlines the functional properties of plasma membrane anion channels in plant cells and discusses their likely roles in root function. Plant anion channels can be grouped according to their voltage dependence and kinetics: (1) depolarization-activated anion channels which mediate either anion efflux (R and S types) or anion influx (outwardly rectifying type); (2) hyperpolarization-activated anion channels which mediate anion efflux, and (3) anion channels activated by light or membrane stretch. These types of anion channel are apparent in root cells where they may function in anion homeostasis, membrane stabilization, osmoregulation, boron tolerance and regulation of passive salt loading into the xylem vessels. In addition, roots possess anion channels exhibiting unique properties which are consistent with them having specialized functions in root physiology. Most notable are the organic anion selective channels, which are regulated by extracellular Al3+ or the phosphate status of the plant. Finally, although the molecular identities of plant anion channels remain elusive, the diverse electrophysiological properties of plant anion channels suggest that large and diverse multigene families probably encode these channels.

I. Introduction

Anion channels are a diverse class of ion channel present in eukaryotic and prokaryotic cells. They are integral membrane proteins that form aqueous pores to allow the rapid transport of anions across the membrane. Anion channels are reported in all plant membranes including the plasma membrane, tonoplast, endoplasmic reticulum, mitochondria and chloroplasts (Schroeder, 1995; Allen & Sanders, 1997; Fuks & Homble, 1999; Lurin et al., 2000; White & Broadley, 2001; Clausen et al., 2004) and these channels have been characterized primarily using a variety of electrophysiological techniques. Despite this, cation channel research in plant cells is far more developed; this is most evident from the much greater number of research articles detailing cation transport in plant cells. There is a particular gulf with respect to the identification of ion channel genes. Many genes encoding cation channels in plants have now been identified and the generation of appropiate knockout mutants has resulted in gargantuan steps towards understanding the roles of cation channels in plant biology (e.g. Mäser et al., 2001; Véry & Sentenac, 2003; Peiter et al., 2005). In contrast, the identification of plant anion channel genes is very much in its infancy. However, in spite of our limited knowledge of anion channels at the molecular level, it is increasingly appreciated that anion channels play fundamental roles in a number of areas of plant biology including turgor regulation, membrane voltage stabilization, signal propagation, toxic metal tolerance and nutrient acquisition.

In this review I will exclusively focus on anion channels in the plasma membrane of plant cells. These have been well documented in stomatal guard cells and (although to a lesser extent) in other leaf cell types, hypocotyls and suspension cultured cells. Plasma membrane anion channels in plants can be broadly classified on the basis of their voltage dependence into depolarization- and hyperpolarization-activated channels and a third class activated by light or membrane stretch. Depolarization-activated anion channels can be subdivided further based on their kinetics and gating properties into R-type, S-type and outwardly rectifying anion channels (Schroeder, 1995; White & Broadly, 2001). For discussion of earlier works on plant anion channels, particularly the classical whole-cell electrophysiological and flux measurements using Characean and other algae, the reader is directed to the review by Tyerman (1992) and references therein.

Until recently, there had been a scarcity of knowledge regarding the role of anion channels in higher plant roots, despite many reports suggesting important functions for these channels in plant nutrition and root/rhizosphere interactions. However, this dearth of knowledge is starting to be addressed, most notably with the publication of several recent electrophysiological reports of anion channels in roots. These studies have resulted in the identification of diverse and sometimes novel anion channel types which are consistent with key roles for anion channels in root physiology. In light of these recent advancements, this review has two main goals. The first is to provide an overview of plant anion channels using primarily electrophysiological (voltage clamp) investigations on leaf cells (including guard cells), hypocotyl cells and suspension cultured cells. This will provide a useful background and framework from which to review the electrophysiological properties of anion channels in root cells (Table 1). The second is to outline and discuss the likely roles of root cell plasma membrane anion channels in root function.

Table 1.  Anion channels in the plasma membrane of higher plant root cells
Cell typeChannel type and gatingKineticsSelectivityRegulatorsInhibitorsReferences
  1. ABA, abscisic acid; ARAC, Arabidopsis root anion channel; A-9-C, anthracene-9-carboxylic acid; Cit, citrate; DIDS, 4,4-diisothiocyanostilbene-2,2-disulfonic acid; DPC, diphenylamine-2-carboxylic acid; Mal, malate; NA, niflumic acid; NPPB, 5-nitro-2-(3-phenylpropylamino)-benzoic acid; PR-ARAC, phosphate-regulated Arabidopsis root anion channel; X-IRAC, inwardly rectifying anion channel; X-QUAC, quickly activiating anion conductance; X-SLAC, slowly activating anion conductance; XPCs, xylem parenchyma cells; cyt, cytosolic; ext, extracellular.

Inorganic anion transport in roots
Wheat root (Triticum aestivum) epidermisDepolarization-activated outward rectifierRapid activationNO3 > Cl > ICa2+cytDIDS > NA > NPPB > Zn2+Skerret & Tyerman (1994); Garrill et al. (1996)
Maize (Zea mays)root apexDepolarization-activated outward rectifierRapid activation (slow inactivation)Cl  Pineros & Kochian (2001); Pineros et al. (2002)
White lupin (Lupinus albus) root epidermisDepolarization-activated outward rectifierSlow activation (slow inactivation)Cl >> Cit3– > H2PO4Weakly dependent on P starvationA-9-CZhang et al. (2004a)
Arabidopsis thaliana root epidermisDepolarization activated R-type (ARAC)Rapid activation (slow inactivation)SO42– > NO3 > ClExtracellular anions; intracellular SO42–NA > NPPB > A-9-CDiatloff et al. (2004)
Root hairs of A. thaliana,Depolarization-activated S-typeSlow activation?Strongly dependent on dessicationDIDSDauphin et al. (2001)
Phaseolus vulgaris and Vigna unguiculate
Barley (Hordeum vulgare) root XPCsVoltage-independent inward rectifier (X-IRAC)Slow single channel gatingCl= NO3ABA; Ca2+cyt Kohler & Raschke (2000)
Barley root XPCsVoltage-dependent inward/outward rectifier (X-QUAC)Rapid activation (slow inactivation)NO3 > Cl > Mal2–Ca2+cyt; NO3extDIDS = IAA-94Kohler & Raschke (2000); Kohler et al. (2002)
Barley root XPCsDepolarization-activated S-type (X-SLAC)SlowClCa2+cyt Kohler & Raschke (2000)
Maize root steleHyperpolarization-activated inward rectifier (X-IRAC) Cl= NO3ABA; Ca2+cyt Gilliham & Tester (2005)
Maize root steleInward/outward rectification (X-QUAC)Instantaneous activation (slow inactivation)NO3= Cl > I > Mal2– > SO42– > Cit3–ATPcyt; Ca2+cyt; ABANAGilliham & Tester (2005)
Organic anion efflux from roots
A. thaliana root epidermisDepolarization-activated R-type (PR-ARAC)Rapid activationCit3–= Mal2– >> ClStrongly dependent on P starvation Diatloff et al. (2004)
White lupin root epidermisHyperpolarization-activated inward rectifierRapid activation (slow inactivation)Cit3–= Mal2– >> ClWeakly dependent on P starvationA-9-CZhang et al. (2004a)
Maize root apexAl3+-activated inward rectifierRapid activationClClext Pineros & Kochian (2001); Pineros et al. (2002)
Maize root apexAl3+-activated inward rectifierRapid activationCl >> Mal2– > Cit2– NA > DIDSKollmeier et al. (2001)
Wheat root apexAl3+-activated inward rectifierRapid activation (slow inactivation)Mal2– > ClClextDPC > NARyan et al. (1997); Zhang et al. (2001)

II. Overview of higher plant plasma membrane anion channels

1. Depolarization-activated anion channels

Rapidly activating anion efflux channels  Rapidly activating anion efflux channels or rapid (R)-type channels have been extensively studied in the plasma membrane of Vicia faba guard cells, Arabidopsis thaliana hypocotyl epidermal cells and tobacco (Nicotiana tabacum) suspension cultured cells. The salient distinguishing features of R-type currents are summarized below.

  • 1The current–voltage (IV) relationship is U-shaped, with peak current activation at voltages negative of the equilibrium potential for the permeant anion (Eanion) and thus favouring anion efflux (Fig. 1a). At resting (hyperpolarized) membrane voltages, R-type channels are in a closed state with membrane depolarization eliciting channel opening.
  • 2Whole-cell currents exhibit rapid (milliseconds) time-dependent activation/deactivation kinetics (Fig. 1c).
  • 3Prolonged depolarizations at activating voltages result in a slow (10–70 s) inactivation, prompting suggestions that these channels mediate a transient efflux of anions (Schroeder, 1995).
  • 4The unitary conductance of R-type channels is small. In V. faba guard cells, the R-type channel [referred to as guard cell anion channel 1 (GCAC1)] has a unitary conductance of 30–40 pS in 154 mm Clcyt: 40 mm Clext (Keller et al., 1989; Marten et al., 1991). This is comparable to 15 pS (150 mm Clcyt[cytosolic chloride]: 100 mm Clext[extracellular chloride]) measured for the R-type channel in tobacco cultured cells [referred to as the tobacco suspension cell anion channel (TSAC); Zimmerman et al., 1994] and 21 pS (150 mm Clcyt: 100 mm Clext) measured for the R-type channel in A. thaliana hypocotyls (Thomine et al., 1995). The magnitude of the GCAC1 unitary conductance is dependent on extracellular Cl and other permeant anions (Hedrich et al., 1994) with conductances ranging from 0 pS in the absence of extracellular Cl (Hedrich & Marten, 1993) up to 67 pS in 154 mm Clext (Dietrich & Hedrich, 1998). This probably reflects an anion binding site on the extracellular side of GCAC1 (apparent Michaelis constant [Km] for Cl ranging from 3 to 16 mm; Hedrich et al., 1994; Dietrich & Hedrich, 1998) which may provide a ‘feedback’ mechanism promoting anion efflux in the face of a reduced anion efflux gradient. However, such a mechanism may be specific to GCAC1, as the unitary conductance of R-type channels in A. thaliana hypocotyls is independent of extracellular Cl (see fig. 7 in Colcombet et al., 2001).
Figure 1.

Distinguishing features of plant anion channels. (a, b) Schematic representation of steady-state current–voltage relationships for (a) R-type (solid line) and S-type (dashed line) anion currents and (b) outwardly rectifying depolarization-activated anion channels (dashed line) and inwardly rectifying hyperpolarization-activated anion channels (solid line). (c, d) Schematic representation of whole-cell activation and deactivation currents for (c) R-type currents and (d) S-type currents. The dashed line represents zero current.

Selectivity  Using Cl as the major cytosolic anion and replacing extracellular Cl with a range of test anions, the relative permeability (test anion compared to Cl; PA/PCl) for GCAC1 was derived from changes in the current reversal voltages (Erev) to give a selectivity sequence of SCN (5.9) > NO3 (4.2) > I (3.9) > Br (1.9) > Cl (1) > acetate (0.44) > propionate (0.34) > malate2– (0.1), and glutamate and gluconate are impermeable (Keller et al., 1989; Hedrich et al., 1990; Hedrich & Marten, 1993; Dietrich & Hedrich, 1994, 1998). However, acetate and propionate permeability is likely to be greatly overestimated in these experiments because efflux currents were barely detectable upon replacing Clcyt with these organic acids (Dietrich & Hedrich, 1998). In contrast, large efflux currents are apparent using malate as the major intracellular anion (Hedrich et al., 1990), indicating that malate is significantly more permeant than the relative permeability value suggests. These observations highlight the importance of directly measuring the permeation of anions through ion channels and the fact that deriving relative permeabilities from changes in current Erev values can be misleading, particularly in determining the physiological relevance of an anion selectivity (see Roberts & Tester, 1997).

The selectivity of R-type channels in A. thaliana hypocotyls was investigated by replacing Clcyt with a range of test anions and the relative conductance (i.e. γACl) determined by comparing the magnitudes of the peak whole-cell current amplitude. The selectivity sequence NO3 (2.6) > SO42– (2.0) > Cl (1.0) > HCO3 (0.8) >> malate2– (0.03) was reported (Frachisse et al., 1999). Although similar to that determined for GCAC1, it differs significantly in that malate is virtually impermeant. However, the key finding of this study is the significant permeability for SO42–. R-type channels in A. thaliana guard cells are also permeable to SO42– (Pei et al., 2000), raising the possibility that SO42– permeation may be a conserved feature of R-type channels.

Regulation by intracellular factors  The regulation of R-type channels is complex and appears to be species and cell-type specific. In V. faba guard cells, elevated intracellular Ca2+ and intracellular nucleotides are essential for GCAC1 activation (Hedrich et al., 1990). Binding of nonhydrolysable forms of ATP increases the open probability of GCAC1 in a voltage-independent manner, an effect that can be described by a Hill function with a coefficient of 3.6 and a Km of 0.4 mm; thus GCAC1 activation appears dependent on the cooperative interaction of four ATP binding sites (Schulz-Lessdorf et al., 1996). An additional level of regulation is evident in that GCAC1 activity is modulated by cytosolic pH (pKa 6.9; Schulz-Lessdorf et al., 1996). It has been suggested that the regulation of anion channels by cytosolic pH plays a specific role in cytosolic pH regulation in plant cells by providing an anion shunt conductance. A shunt conductance is necessary to maintain H+ efflux from the cytosol via the electrogenic H+ pump (Johannes et al., 1998).

Intracellular nucleotides are potent modulators of R-type channel gating in A. thaliana hypocotyl and tobacco suspension cultured cells (Zimmerman et al., 1994; Thomine et al., 1995), although it is noteworthy that nucleotides are not essential for channel activity. Decreasing intracellular nucleotide concentration shifts the activation potential to more negative voltages resulting in channel activation over a wider range of voltages (Zimmermann et al., 1994; Thomine et al., 1995). However, whereas modulation of TSAC gating by ATP is dependent on protein (channel?) phosphorylation (i.e. the effect of ATP is enhanced by the protein kinase inhibitors), modulation of R-type channels in A. thaliana hypocotyls is dependent on voltage-dependent binding of nucleotides to a site within the channel pore, resulting in a voltage-dependent block at hyperpolarized voltages (Thomine et al., 1997; Colcombet et al., 2001). This is reminiscent of voltage-dependent gating of cation channels, most notably K+ selective inward rectifiers in animal cells in which intracellular Mg2+ enters and blocks the pore from the cytosolic face, preventing K+ efflux at voltages positive of potassium equilibrium potential [EK] (Hille, 1992). The physiological significance of the apparently diverse mechanisms by which nucleotides modulate R-type channel activity in plant cells is unclear, although they may be cell-type specific and reflect specific functions for these channels in different cell types. It is also unclear if nucleotide-dependent gating is a conserved property of R-type channels, as A. thaliana guard cells exhibit R-type currents with strong voltage dependence in the absence of intracellular nucleotides (see fig. 1D in Pei et al., 2000).

As a result of the dependence of GCAC1 activity on elevated Ca2+cyt (Hedrich et al., 1990), it has become routine to record R-type channel activity in plant cells in the presence of elevated Ca2+cyt. However, the dependence (if any) of R-type channels in other plant cell types on Ca2+cyt has not been investigated and, in light of the differences in nucleotide regulation of channel activity, differential sensitivities to Ca2+cyt amongst the plant R-type channels would not be surprising. It is interesting that, despite using elevated cytosolic Ca2+ and nucleotides levels (which are sufficient for sustained GCAC1 activity), R-type channels in A. thaliana hypocotyls and TSAC exhibit rapid and complete rundown within 10–30 min during patch clamp experiments, indicating that channel activity is dependent on additional unknown cytosolic factors which are lost during the experiments. Frachisse et al. (1999) showed that intracellular SO42– was able to prevent channel rundown in hypocotyl cells, although it is currently unknown if cytosolic SO42– also positively regulates GCAC1 or TSAC, and it remains unclear whether regulation by SO42– is a universal property of R-type channels (however, see below; Section III, Part 1).

Regulation by extracellular anions  Extracellular anions are potent regulators of R-type channels; typically this manifests itself as a shift in the I–V relationship to more negative voltages with increasing extracellular anion concentration (i.e. modulation of the gating such that channels open at more negative membrane voltages which favour enhanced anion efflux). GCAC1 displays differential sensitivity to a range of extracellular organic anions including auxin (Km = 10–20 µm; Marten et al., 1991), malate (Km = 0.4 mm; Hedrich & Marten, 1993), acetate and propionate (Dietrich & Hedrich, 1998) and the inorganic anions Cl (Km = 15.7 mm) and SCN (Dietrich & Hedrich, 1998). Competition experiments show that extracellular anions probably compete for a common site (Dietrich & Hedrich, 1998). GCAC1 is particularly sensitive to auxin and malate, prompting the proposal that the regulation of GCAC1 by these ligands represented an initial step triggering stomatal movements (Marten et al., 1991; Hedrich & Marten, 1993; Hedrich et al., 1994). However, this has not been universally accepted because of the apparent low specificity of the anion binding site and the finding that the maximal response of GCAC1 to auxin and malate is at supra-optimal concentrations for stomatal closure (Schmidt & Schroeder, 1994; Schroeder, 1995; Schmidt et al., 1995). Interestingly, Colcombet et al. (2001) showed that the modulation of voltage gating of R-type channels by extracellular anions in A. thaliana hypocotyls probably resulted from the entry of anions from the extracellular side and the displacement of nucleotides from the channel pore. However, based on the differing mechanisms by which nucleotides modulate R-type channels, this model is likely to be specific for A. thaliana hypocotyls.

Pharmacology  There is no common pharmacological profile for R-type channels in guard cells, hypocotyls and cultured tobacco cells. GCAC1 has been best characterized. GCAC1 is partially blocked by extracellular ethacrynic acid (EA; 65% block at 100 µm), anthracene-9-carboxylic acid (A-9-C; 50% block at 100 µm) and probenicid (34% block at 100 µm), whereas niflumate (inhibition constant; Ki = 20 µm), IAA-94 [(6,7-dichloro-2-cyclopentyl-2,3-dihydro-2-methy-1-oxo-1H-inden-5-yl)oxyacetic acid; Ki = 7 µm] and NPPB [5-nitro-2-(3-phenylpropylamino)-benzoic acid; Ki = 4 µm) effected > 90% inhibition (Marten et al., 1992). GCAC1 is also highly sensitive to stilbene derivative channel blockers, most notably DIDS (4,4-diisothiocyanostilbene-2,2-disulfonic acid; 50% block at 0.2 µm; Marten et al., 1993). In contrast, TSAC1 is insensitive to EA and is partially blocked by DIDS (50% at 50 µm), A-9-C (50% block at 100 µm) and NPPB (90% block at 100 µm) (Zimmermann et al., 1998). In A. thaliana hypocotyls, R-type channels are almost (97%) completely blocked by niflumate (Ki = 87 µm), although 100 µm NPPB and IAA-94 achieve only 52% and 29% block, respectively, and, in contrast to GCAC1 and TSAC, A-9-C and DIDS are ineffective (Thomine et al., 1997).

It is noteworthy that anion channel blockers shift the I–V relationship to more negative voltages, similar to that reported for a range of inorganic and organic anions (see Regulation by extracellular anions, above). Although little is known of the mode of action for these inhibitors, this observation provides some insights: it is consistent with the anionic species of the inhibitor binding to the extracellular face of the channel and, at least for anion channels in A. thaliana hypocotyls, this may be within the pore region of the channel.

Slowly activating anion efflux channels  Slowly activating anion efflux channels or slow (S)-type channels have been reported in the guard cells of V. faba (Schroeder & Keller, 1992), A. thaliana (Forestier et al., 1998), tobacco (Grabov et al., 1997) and Xanthium strumarium (Linder & Raschke, 1992), in the epidermal cells of A. thaliana hypocotyls (Frachisse et al., 2000) and in coffee (Coffea arabica) suspension cultured cells (Dieudonnéet al., 1997). The salient distinguishing features of S-type whole-cell currents are summarized below.

  • 1Typically, the current–voltage relationship is characterized by a peak current amplitude at voltages negative of the equilibrium potential for the permeant anion (Eanion). However, the peak current is substantially more positive and is less pronounced than that for the R-type currents under the same recording conditions (Schroeder & Keller, 1992; Frachisse et al., 2000), and thus S-type channels activate over a much broader range of voltages than the R-type channels (Fig. 1a).
  • 2Whole-cell currents exhibit slow (up to 1 min) time-dependent activation/deactivation kinetics (Fig. 1d).
  • 3In contrast to R-type currents, S-type channels do not inactivate.
  • 4Few recordings of individual S-type channels have been reported (Linder & Raschke, 1992; Schroeder et al., 1993; Schmidt & Schroeder, 1994; Schmidt et al., 1995) but all show a moderate unitary conductance (33–35 pS using 150 mm Clcyt: 30–50 mm Clext) and long (up to 5 s) open and closed durations. It is noteworthy that spontaneous intervals of rapid flickering between the open and closed states is occasionally observed, indicating that the channel may switch between distinct gating modes (see R- and S-type currents: below, for significance). Note also that S-type channel conductance is not dependent on extracellular anions (Schmidt & Schroeder, 1994).

Selectivity  The selectivity of S-type channels in V. faba guard cells has been investigated by replacing Clcyt with equimolar concentrations of a test anion to reveal both relative permeability (derived from current reversal potentials) and relative conductance (derived by comparing the magnitude of the efflux current for the test anion with that for Cl) (Schmidt & Schroeder, 1994); the relative permeability sequence of NO3 (20.9) > Br (2.4) > F (1.26) > Cl (1.0) > I (0.98) > malate2– (0.24) differed significantly from the relative conductance sequence of Br (3.4) > F (3.3) > NO3 (3.2) > Cl (1.0) > I (0.6) > malate2– (0.3). However, both sequences indicate poor selectivity amongst the halide anions and a relatively small malate permeability. A high relative permeability for NO3 (PNO3/PCl > 20) has been reported for S-type channels in A. thaliana hypocotyl epidermal cells (Frachisse et al., 2000) and in coffee suspension cultured cells (Dieudonnéet al., 1997). Interestingly, in both of these studies, S-type channels are impermeable to SO42–; hence, a high relative permeability for NO3 and an impermeability to SO42– appear to be defining characteristics of S-type channels in plant cells.

Regulation by intracellular factors  The activation of S-type channels is (relatively) weakly dependent on voltage, resulting in their activation across a broad (if not the entire) physiological voltage range for plant cells (Fig. 1a). Thus, it is not surprising that S-type channels are under tight posttranslational control.

S-type channel regulation has been extensively investigated in the abscisic acid (ABA) signal transduction pathway in guard cells. ABA induces stomatal pore closure, which results from the net loss of solutes and turgor from the two guard cells that delimit the stomatal pore; this is thought to be initiated, at least in part, by ABA via the activation of S-type channels and subsequent anion efflux. However, the signal transduction mechanism(s) involved in S-type channel activation appears to be species specific.

Ca2+cyt elevation was originally reported to be a strong activator of S-type channel activity in V. faba guard cells (Schroeder & Hagiwara, 1989), although more recently it has been shown that phosphorylation is the primary positive regulator of S-type channels and that elevation of Ca2+cyt is not essential for GCAC1 activation (Schmidt et al., 1995; Schwarz & Schroeder, 1998; Levchenko et al., 2005). Thus, Ca2+cyt activation of GCAC1 is probably upstream of the phosphorylation event or part of a Ca2+-dependent parallel signalling pathway. Interestingly, ABA is able to maintain and activate S-type channel activity in the absence of cytosolic ATP and elevated Ca2+cyt, indicating that ABA may act by maintaining channel phosphorylation via the downregulation of a membrane-delimited protein phosphatase (Schwarz & Schroeder, 1998).

In A. thaliana guard cells, S-type channel activation by ABA is dependent on a cytosolic Ca2+ elevation (Km = 1.2 µm) and hydrolysable cytosolic ATP (Allen et al., 1999), indicating the likely requirement of Ca2+-dependent protein kinase activity. Allen et al. (1999) showed that kinase inhibitors (K-257a and staurosporine) and strong buffering of Ca2+cyt were effective inhibitors of channel activation by ABA. It is noteworthy that an earlier study on A. thaliana guard cells by the same group (Pei et al., 1997) reported that S-type channel activation was in fact inhibited by protein kinases, a result that lies in juxtaposition to the results obtained in both V. faba and A. thaliana guard cells. These contradictory results apparently arise from the relative timings of inhibitor and ABA application (see Allen et al., 1999 for further discussion) and serve to highlight our relatively basic understanding of the complex nature of S-type channel regulation in guard cells.

Cytosolic nucleotides are essential for S-type channel activation in A. thaliana hypocotyls (Frachisse et al., 2000), although, in contrast to guard cells, S-type channel activation appears to involve ATP hydrolysis and nucleotide binding. The effects of Ca2+cyt on channel activity were not investigated, although experiments were performed in the presence of 1 µm Ca2+cyt.

In summary, most investigations support phosphorylation as a key positive regulator of S-type channels which can be Ca2+ dependent or independent. However, in coffee suspension cultured cells, S-type channel activity appears to be independent of cytosolic ATP and Ca2+ (Dieudonnéet al., 1997).

Pharmacology  The pharmacological profile for S-type currents appears to be dependent on both cell type and species. In V. faba guard cells, S-type channels are irreversibly and completely blocked by 50 µm niflumic acid and reversibly blocked by NPPB (Ki = 7 µm), A-9-C (Ki = 60 µm) and IAA-94 (Ki = 10 µm) but are insensitive to DIDS (Schroeder et al., 1993; Schwartz et al., 1995; Forestier et al., 1998). A-9-C and niflumic acid are also effective inhibitors of S-type currents in tobacco guard cells (Grabov et al., 1997). In A. thaliana guard cells niflumate, NPPB and DIDS were ineffective blockers and only partial (50%; Ki = 50 µm) inhibition was achieved with A-9-C. The most potent blockers were diphenylamine-2-carboxylic acid (DPC) (84% inhibition at 100 µm) and glibenclamide (100% block; Ki = 3.3 µm) (Forrestier et al., 1998). DPC and glibenclamide are established blockers of the ATP binding cassette (ABC) transporter cystic fibrosis transmembrane receptor (CFTR), raising the possibility that S-type channels may be members of the ABC transporter family (Leonhardt et al., 1999). In contrast, S-type channels in A. thaliana hypocotyls are effectively blocked by DIDS (100% block; Ki = 26 µm) and niflumate and NPPB are relatively ineffective.

R- and S-type currents: different functioning modes of the same channel or separate entities?  At least in the guard cells of V. faba and X. strumarium and the epidermal cells of A. thaliana hypocotyls, R- and S-type whole-cell anion currents can be simultaneously recorded in the same conditions (Linder & Raschke, 1992; Schroeder & Keller, 1992; Frachisse et al., 2000; Roelfsema et al., 2004), prompting the suggestion that these two currents represent contrasting kinetic properties of the same enzyme or, at least, that they share common protein subunits (Dietrich & Hedrich, 1994). The most compelling evidence for this is that whole-cell R-type currents are able to switch (in an ATP-dependent manner) to a ‘slow mode’ reminiscent of S-type channel behaviour (Zimmerman et al., 1994; Thomine et al., 1995). Furthermore, transitions between slow and rapid kinetic behaviour of single anion channels have been observed in V. faba guard cells, albeit this switch appears to occur randomly (Dietrich & Hedrich, 1994; Schmidt & Schroeder, 1994). However, a closer assessment of the behaviour of R- and S-type channels reveals significant differences in their kinetics, nucleotide dependence and selectivity. Specifically, these differences are as follows.

  • 1The kinetics of the R-type whole-cell currents in ‘slow’ mode are approximately 10-fold faster than that for the S-type currents under similar recording conditions. For example, in A. thaliana hypocotyls, S-type currents deactivate with a time constant of approx. 6 s (Frachisse et al., 2000) whereas the R-type currents in their ‘fast’ and ‘slow’ modes deactivate in approx. 300 ms.
  • 2Nucleotide regulation of channel activity is distinct between the channel types; specifically, S-type currents are dependent on the presence of intracellular nucleotides, whereas the ‘slow’ mode of the R-type currents is dependent on the absence of ATP (Thomine et al., 1995, 1997; Colcombet et al., 2001).
  • 3Although both R- and S-type currents exhibit broad selectivity for anions, it is significant that S-type channels are not permeable to SO42– while SO42– appears to be a good substrate for R-type channels.

Thus, although it is arguable that a single channel type could exist in two conformations that exhibit significantly different gating kinetics, nucleotide affinity and ion selectivity, the available data are probably best reconciled as indicating that R- and S-type currents result from distinct molecular entities. However, only the cloning of genes encoding the anion channel proteins will provide unambiguous proof.

Outwardly rectifying depolarization-activated anion influx channels  Under most conditions, the anion electrochemical gradient favours passive channel-mediated anion efflux. However, despite this, outwardly rectifying depolarization-activated anion channels (OR-DAACs) mediating anion influx (Fig. 1b) have been reported in maize (Zea mays; Fairley et al., 1991) and A. thaliana (Cerana & Colombo, 1992) suspension cultured cells and the leaf cells of the halophytic angiosperm Zostea muelleri (Garrill et al., 1994). These currents have not been well characterized, although insights into their possible physiological function have been gained from A. thaliana cultured cells; namely, that outward anion currents are induced in high Clext whereas in low Clext outwardly rectifying K+ currents dominate the membrane conductance. This ability to switch between cation and anion conductances probably reflects an ability to stabilize the plasma membrane voltage in diverse ionic environments (Cerana & Colombo, 1992). OR-DAACs in leaf cells of Z. muelleri are proposed to mediate Cl influx for cell turgor regulation as an adaptation to its estuarine habitat which contains high levels of Cl (Garrill et al., 1994).

2. Inwardly rectifying hyperpolarization-activated anion efflux currents

Hyperpolarization-activated anion channel (HAAC) activity (Fig. 1b) has been recorded in the plasma membrane of Amaranthus tricolor cotyledons (Terry et al., 1991), mesophyll and epidermal cells of the pea (Pisum sativum) leaf (Elzenga & Volkenburgh, 1994, 1997), Chara (Chara inflata) (see Tyerman, 1992 for review), the seed coats of Phaseolus vulgaris (Zhang et al., 2004b), suspension cultured cells from carrot (Daucus carota) (Barbara et al., 1994), A. thaliana (Lew, 1991), barley (Hordeum vulgare) (Amtmann et al., 1997) and Asclepias tuberosa (Schauf & Wilson, 1987), the marine phytoplankton Coccolithus pelagicus (Taylor & Brownlee, 2003) and the marine alga Valonia utricularis (Heidecker et al., 1999; Binder et al., 2003). There are few unifying features apparent from these studies, although most currents exhibit voltage-dependent inward rectification which are mediated by large conductance channels (i.e. > 100 pS).

Whole-cell, anion-selective inwardly rectifying currents in the mesophyll cells of pea leaves (Elzenga & Volkenburgh, 1997) exhibit strong voltage-dependent (half maximal activation at 27 mV) exponential activation (time constants ranging from 100 to 400 ms). Channel activity is dependent on Ca2+cyt (Km = 1 µm) and, in the presence of cytosolic ATP, currents exhibit a partial (60%) slow (1–10 s) inactivation which may reflect voltage-dependent blockade by nucleotides entering the pore (reminiscent of the voltage-dependent gating of R-type channels by nucleotides; Frachisse et al., 2000). Large conductance (300 pS in symmetrical 100 mm Cl) channels, which are likely to underlie the whole-cell currents (Elzenga & Volkenburgh, 1994), exhibit a relative conductance (γACl) sequence (determined from the single channel current amplitude) of F (2.6) > I (1.21) > Cl (1.0) > Br (0.96) > malate2– (0.9), indicating a significant malate permeability. This channel was partially (50%) blocked by cytosolic 4-acetamido-4′-isothiocyanostilbene 2′-disulfonate (SITS) but was insensitive to DIDS.

Anion-selective inward currents in carrot culture cells (Barbara et al., 1994) exhibit voltage-dependent rapid (within 20 ms) activation followed by rapid (within 300 ms) and complete inactivation, consistent with this channel mediating transient anion efflux. The currents are mediated by large conductance channels (100 pS; 160 mm Clcyt: 25 mm Clext) which are insensitive to cytosolic Ca2+ and nucleotides. Curiously, channel activity is modulated by slight negative pressure, indicating a potential role in osmoregulation.

In C. pelagicus, V. utricularis and A. tuberosa, the HAACs do not inactivate (Schauf & Wilson, 1987; Binder et al., 2003; Taylor & Brownlee, 2003). In C. pelagicus and V. utriclaris, currents activate rapidly (with time constants of less than 200 ms) and exhibit strong voltage-dependent inward rectification. Based on the voltage dependence of gating, these channels are proposed to play a role in membrane voltage control and action potential propagation. Unique amongst the plant plasma membrane anion channels, the HAAC from C. pelagicus is highly selective for Cl over other anions, and NO3 and SO42– are impermeant. Slower activation kinetics are associated with anion currents in A. tuberosa (time constants of up to 1500 ms). They were completely blocked by 10 µm Zn2+ext and 100 µm EA. Application of 100 µm ABA also inhibited these currents; however, the significance of this result is unclear as 100 µm ABA is not physiological. Furthermore, currents were recorded in other nonphysiological conditions; namely, cytosolic Ca2+ and pH were buffered at 2 mm and 5.8, respectively.

In A. tricolor, large-conductance (up to 200 pS) HAACs exhibit complex regulation by cytosolic Ca2+ and ATP (Terry et al., 1991). In isolated membrane patches, channel activation was independent of ATP; however, when observed in the whole-cell configuration of the patch clamp technique, ATP inactivated the channel, indicating that regulation by ATP is probably via an indirect step (e.g. via a protein kinase that is not an intrinsic part of the channel). In addition, elevating Ca2+cyt to 10 µm increased channel activity where as decreasing Ca2+cyt resulted in the loss of voltage dependence. The physiological significance of this complex control is unclear, although it may reflect a strict control over channel opening to prevent excessive solute loss (Terry et al., 1991).

In contrast to most HAACs, the large-conductance (150 pS in symmetrical 150 mm Cl) HAAC in barley suspension cultured cells (Amtmann et al., 1997) and the HAAC in A. thaliana callus cells (Lew, 1991) are weakly voltage dependent. In these cases the I–V relationship of the open channel exhibits strong inward rectification consistent with a blockade of the channel pore at voltages positive of ECl.

Smaller conductance HAACs have also been recorded in Chara inflata (7–44 pS; see Tyerman, 1992) and cells from the coats of developing bean (Phaseolus vulgaris) seeds (18 pS in 108 mm Clcyt: 30 mm Clext; Zhang et al., 2004b). Both these channel types are inhibited by extracellular La3+ and are activated by Ca2+cyt elevations (Biskup et al., 1999; Zhang et al., 2004b).

3. Mechanosensitve and light-activated anion channels

Stretch-activated anion-selective channels have been described in the plasma membrane of V. faba guard cells (Cosgrove & Hedrich, 1991), tobacco suspension cultured cells (Falke et al., 1988) and mesophyll cells from A. thaliana leaves (Qi et al., 2004). These anion channels are likely to play roles in osmoregulation during osmotic stress and cell expansion. However, significant differences in their biophysical properties suggest that they have distinct roles in osmoregulation. In outside-out patches from tobacco cells, anion-selective mechanosensitive (MS) channels exhibit a large unitary conductance (97 pS in 220 mm Clext: 25 mm Clcyt) and long open time durations, suggesting a role in mediating large (turgor resetting) anion fluxes. In contrast, the MS channels in V. faba guard cells are characterized by rapid flickering between the open and closed states and a small unitary conductance (27 pS with 154 mm Clcyt: 84 mm Clext), possibly reflecting a role in osmosensing. MS channels in A. thaliana mesophyll cells also have a small unitary conductance but, in contrast to their counterparts in tobacco cells and V. faba guard cells, channels were activated by positive pressure applied to the patch pipette (Qi et al., 2004). Using trinitrophenol (TNP), which preferentially inserts in the outer leaf of the plasma membrane, Qi et al. (2004) demonstrated anion channel activation in the absence of positive pressure, suggesting that channel activation was dependent on the increased convex curvature of the membrane. This is similar to that reported for the large mechanosensitive channel in Escherichia coli (Pivetti et al., 2003).

Blue and white light-activated anion channels in A. thaliana hypocotyl (Cho & Spalding, 1996) and pea leaf mesophyll (Elzenga & van Volkenburgh, 1997) cells, respectively, exhibit small unitary conductances (ranging from 23 to 46 pS). Although their activities are strongly dependent on Ca2+cyt, light reception is not dependent on Ca2+ signalling (see Spalding, 2000 for further details). The complete inhibition of the blue light-activated channel in A. thaliana by 20 µm NPPB and 10 mm La3+ has been used to demonstrate its role in mediating membrane depolarizations associated with blue light inhibition of hypocotyl elongation (Spalding, 2000).

III. Anion channels in higher plant roots

Electrophysiological studies have identified a variety of anion channel types in the plasma membrane of higher plant root cells (Table 1). They display distinct and varied biophysical properties that are consistent with them having diverse physiological functions. However, based on their localization within the root, anion selectivity and regulation, these channels can be broadly arranged into two distinct groups: (1) those with likely roles in regulating inorganic anion transport across the root, and (2) those mediating organic anion exudation from the root with likely roles in enhancing phosphate bioavailability or reducing the bioavailability of toxic metals in the root rhizosphere. The biophysical properties and the possible physiological relevance of these channels are discussed.

1. Inorganic anion transport across higher plant roots

The roots of higher plants are responsible for the transport of nutrient ions from the soil solution to the xylem vessels and, via the transpiration stream, to the shoots. The root is composed of many specialized cell types, although, with respect to trans-root ion transport, the root cell types can be generalized as peripheral cells (i.e. epidermal and cortical cells), which mediate the net uptake of ions from the soil solution, and stelar cells, which mediate ion release into the xylem vessels. The specialized structure of the root ensures that mineral ion transport across the root to the xylem vessels involves the uptake and efflux of ions across the plasma membrane of peripheral and stelar cells, respectively. Therefore membrane proteins mediating ion transport across the plasma membrane of root cells represent important points which control the movement of nutrients across the root. K+ transport across the root has been extensively studied and, as a result, the molecular mechanisms mediating and controlling trans-root K+ movements are well characterized (see reviews by DeBoer, 1999; Roberts & Snowman, 2000; Tester & Leigh, 2001). A less coherent picture is currently available for trans-root anion transport, although it is inevitable that cation transport is tightly coupled to anion transport (Lin, 1981; Kochian et al., 1985) for reasons of overall charge neutrality.

Channels mediating anion influx in the root periphery Biophysical properties  OR-DAACs mediating anion influx have been reported in the cortex of wheat (Triticum aestivum) roots (Skerret & Tyerman, 1994; Garrill et al., 1996), the maize root apex (Pineros & Kochia, 2001; Pineros et al., 2002), root epidermal cells of A. thaliana (Diatloff et al., 2004) and Lupinus albus (Zhang et al., 2004a). Typically, whole-cell anion OR currents display rapid activation/deactivation kinetics and marked outward rectification resulting from a strong voltage dependence. The exception is OR anion currents in L. albus, which exhibit slow (seconds) kinetics and weak voltage dependence (Zhang et al., 2004a). The wheat root OR currents have been most extensively investigated (Skerret & Tyerman, 1994); these currents exhibit a selectivity sequence of NO3 > Cl > I (determined from the magnitude of anion influx currents following exchange of extracellular Cl with test anion) and are partially blocked by extracellular DIDS (69% using 200 µm), perchlorate (75% using 10 mm), niflumate (40% using 100 µm) and NPPB (40% using 100 µm) (Skerret & Tyerman, 1994; Garrill et al., 1996). L. albus OR currents are completely blocked by 100 µm A-9-C (Zhang et al., 2004a). The unitary conductance of root OR-DAACs is small; for example, 4 pS (15 mm Clcyt: 300 mm Clext) and 9.2 pS (104 mm Clcyt: 115 mm Clext) in wheat and maize roots, respectively.

Physiological relevance  Cytosolic Cl and NO3 concentrations in plant cells are approximately 10 and 4 mm, respectively (Felle, 1994; Miller & Smith, 1996; Lorenzen et al., 2004) and, in most soil conditions, low millimolar to micromolar concentrations of Cl and NO3 persist (Marschner, 2002). Thus, the equilibrium potential for anions is, under most conditions, more positive than the transmembrane voltage across the plant cell plasma membrane; consequently, anion uptake by roots is mediated via ‘active’ transport (probably via H+ symport: Ullrich & Novacky, 1990; Felle, 1994). For example, even in arable soil which can contain NO3 concentrations up to 3.1 mm (e.g. table 13.3 in Marschner, 2002), the equilibrium potential for NO3 will be approximately +6 mV. This voltage is significantly more positive than the root cell plasma membrane voltage (which can be more negative than −225 mV; Hirsch et al., 1998), and thus NO3 uptake cannot be passive via an ion channel. However, in saline (NaCl) soils, in which Na+ and Cl contents can exceed 30 mm (Xu et al., 2000), excessive Na+ uptake by plant cells (which is associated with significant plasma membrane depolarization) is accompanied by passive Cl uptake (Elzam & Rains Epstein, 1964; Cram & Laties, 1971; Kochian et al., 1985). Cl uptake is thought to maintain electroneutral Na+ uptake, thereby preventing excessive (and deleterious) Na+-induced plasma membrane depolarizations. Skerret & Tyerman (1994) proposed that Na+-induced membrane depolarization activated OR-DAACs, resulting in passive Cl influx in saline conditions. Indirect support for this theory is provided by a recent study using a recombinant fluorescent indicator CLOMELEON to monitor [Cl]cyt in A. thaliana roots (Lorenzen et al., 2004). Lorenzen et al. showed that 100 mm extracellular NaCl induced significant C1 influx and that this was strongly inhibited by 10 mm extracellular Ca2+ or Mg2+ or 2 mm La3+. As Cl influx via anion OR channels (Skerret & Tyerman, 1994; Diatloff et al., 2004; Zhang et al., 2004a) and H+/Cl symporter (Felle, 1994) is insensitive to extracellular cations, the most likely explanation for the inhibition of Cl influx by divalent and trivalent cations is the inhibition of Na+ influx as a result of the blockade of nonselective (Na+-permeable) cation channels by trivalent and divalent cations (see Demidchik et al., 2002 for a review of nonselective channels): that is, prevention of Na+-induced plasma membrane depolarization and this failure to induce Cl influx via OR-DAAC activation. It would be interesting to test this further by the simultaneous measurement of plasma membrane voltages and Clcyt using electrophysiological techniques and CLOMELEON technology.

Channels mediating anion efflux from the root periphery Biophysical properties  Depolarization-activated anion efflux channels have been described in A. thaliana roots (Kiegle et al., 2000; Dauphin et al., 2001; Diatloff et al., 2004) and the root hairs of two bean species, Vigna unguiculata and P. vulgaris (Dauphin et al., 2001).

Anion efflux channels are highly expressed and have been well characterized in the plasma membranes of epidermal and root hair cells of A. thaliana roots [referred to as the Arabidopsis root anion channel (ARAC); Diatloff et al., 2004]. Interestingly, ARAC-like currents have also been reported in other cell types in A. thaliana roots (including pericycle and cortex; Kiegle et al., 2000), indicating that these channels may be ubiquitously expressed in A. thaliana roots. Whole-cell ARAC currents are typical R-type currents exhibiting fast activation/deactivation kinetics and strong voltage-dependent activation (Figs 1 and 2). ARAC also exhibits inactivation, although notably, in contrast to GCAC1 in guard cells, inactivation is partial (approx. 50%), enabling the ARAC to mediate sustained anion efflux. ARAC appears similar to R-type currents in A. thaliana hypocotyls, as: (i) it is significantly permeable for SO42– Cl and NO3 but impermeable to organic anions, malate and citrate; (ii) it has a small single channel conductance (12 pS; 25 mm SO42–cyt: 45 mm Clext); (iii) channel activity is modulated by extracellular anions; (iv) it has a comparable inhibitor profile, i.e. effective inhibition by niflumate (Ki = 61 µm; Fig. 2), although only partial inhibition of the ARAC is achieved using NPPB (66% block at 100 µm) and A-9-C (16% block at 200 µm) (S. K. Roberts and R. Brown, unpublished results), and (v) intracellular SO42– is a potent activator of the ARAC.

Figure 2.

Effects of niflumic acid (NA) on Arabidopsis root anion channel (ARAC) currents in Arabidopsis root epidermal cells. (a) Whole-cell currents resulting from voltage pulses from +64 mV to −216 mV in 10-mV steps. The holding voltage was −216 mV. (b) As (a), except that the bath solution contained 500 µm. (c) Current–voltage relationship of steady-state whole-cell currents with bath solution containing 0 (•), 10 (▴), 50 (▾), 100 (◆), 200 (▪) and 500 (★) µm NA. (d) Dose–response curve of the inhibition of ARAC currents by NA. Values are the mean of three independent experiments (± standard error of the mean). Data are fitted with I = INA(1/[1 + ([NA]/Ki)]) +Iins, where I is total current, INA is the NA-sensitive current, Iins is the NA-insensitive current, Ki is the constant for inhibition of the NA-sensitive current and [NA] is the extracellular NA concentration. Ki = 61 ± 33 µm, INA = 81.9 ± 11.7% and Iins = 17.5 ± 11.6%. The pipette solution was 25 mm Cs2SO4, 2 mm MgCl2, 10 mm N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES), 1 mm MgATP and 2 mm EGTA adjusted to pH 7.2 with Tris base and to 720 mOsm kg−1 using sorbitol. The bath solution was 5 mm LaCl3, 10 mm CaCl2, 5 mm MgCl2 and 10 mm 2-(N-Morpholino)ethanesulfonic acid (MES), adjusted to pH 6.0 with Tris base and adjusted to 700 mOsm kg−1 using sorbitol. Further experimental details can be obtained from Diatloff et al. (2004).

Dauphin et al. (2001) described large S-type anion currents in the plasma membrane of root hairs from A. thaliana, V. unguiculate and P. vulgaris plants which are strictly induced by severe drought (dessication) stress. The pathways signalling the activation of these S-type currents have not been investigated, although, analogously to S-type currents in A. thaliana guard cells, this could involve ABA and (Ca2+-dependent?) kinases (see discussion of S-type channels, above). Interestingly, Dauphin et al. (2001) did not report the presence of R-type currents in A. thaliana root hair cells.

Physiological significance Anion homeostasis. Significant inorganic anion efflux from the roots of higher plant is well documented and may play an important role in determining the net rate of anion uptake by roots. NO3 efflux has been most extensively investigated (see Forde & Clarkson, 1999 and references therein), although Cl efflux from the roots of A. thaliana (Lorenzen et al., 2004), barley (Jackson & Edwards, 1966) and carrot (Cram, 1983) and SO42– efflux from carrot (Cram, 1983) and tomato (Lycopersicon esculentum) (Lopez et al., 2002) roots have also been documented. In steady-state conditions, NO3 efflux increases with increasing intracellular content and extracellular NO3 (Jackson et al., 1976; Breteler & Nissen, 1982; Deane-Drummond & Glass, 1983), although efflux is also influenced by a variety of stresses which perturb growth, for example, defoliation (MacDuff & Jackson, 1992). Thus, anion efflux has been regarded as an overspill mechanism that is stimulated when there is an imbalance between uptake and demand. In extreme cases, an anion efflux system could prevent toxic accumulation of anions in the cytosol, although it must be rare in nature for a plant to experience an excessive (toxic) supply of NO3, especially given the large capacity for the vacuole to store NO3. However, anion efflux may be more common in dealing with excessive Cl accumulation (e.g. in saline conditions; Yamashita & Matsumoto, 1996) or preventing toxic accumulation of cytosolic SO42– which may result from the limited storage capacity of plant vacuoles for SO42– compared with that for NO3 and Cl (Cram, 1983). In less extreme conditions, regulated efflux of anions has been proposed to represent a mechanism to enable ‘fine-tuning’ of cytosolic anion concentrations (Deane-Drummond, 1985, 1986; Macduff & Jackson, 1992). Interestingly, the biophysical properties of the ARAC are well suited to preventing the toxic accumulation of anions; for example, the ARAC is sensitive to extracellular anions such that activity is enhanced with increasing extracellular anion concentration and the potent activation of the ARAC by SO42– is suited to preventing the accumulation of cytotoxic levels of intracellular SO42–.

Osmoregulation. Osmotic stress caused by drought (hyper-osmotic) or excess water (hypo-osmotic) is frequently encountered by root cells and represents a significant challenge for higher plants (Zhu, 2002). Adaptations to stress and the signal transduction events responsible for triggering osmoresponses are well documented, with many recent advances identifying the genes responsible for various aspects of transduction and response (Zhu, 2002). One immediate affect of hyperosmotic stress is a hyperpolarization of the plasma membrane of plant cells (Kinraide & Wyse, 1986; Reuveni et al., 1987; Lew, 1996; Teodoro et al., 1998; Shabala & Lew, 2002; Shepard et al., 2002). In contrast, plasma membrane depolarization has been reported in A. thaliana cells in response to hypo-osmotic stress (Lew, 1996; Zingarelli et al., 1999; Shepard et al., 2002). These changes in membrane voltage are effected by changes in the net transmembrane fluxes of a variety of ions including K+, Cl, and H+ (e.g. Lew, 1998; Teodoro et al., 1998; Zingarelli et al., 1999; Shabala et al., 2000; Shepard et al., 2002). Interestingly, whereas osmotic stress modulates both K+ efflux (e.g. Zingarelli et al., 1999) and K+ influx (e.g. Shabala & Lew, 2002), most evidence suggests that only Cl efflux across the plasma membrane of plant cells is modulated by osmotic stress. This indicates a key role for channel-mediated Cl efflux across the plasma membrane in response to osmotic stress. For example, stimulation of net Cl influx in A. thaliana in response to hyperosmotic stress is the result of a dramatic reduction in Cl efflux (Pennarum & Maillott, 1988; Lew, 1998; Teodoro et al., 1998; Lorenzen et al., 2004). It is also noteworthy that hyperosmotic stress in A. thaliana induces (up to) a 40-mV hyperpolarization of the plasma membrane resulting in membrane voltages in excess of −200 mV (Zingarelli et al., 1999; Shabala & Lew, 2002); this would be sufficient to close the ARAC and reduce Cl efflux (see Fig. 2c). Conversely, membrane depolarizations induced by hypo-osmotic stress in A. thaliana cells results from enhanced Cl efflux (Teodora et al., 1998; Lorenza et al., 2004) and this could reflect the activation of ARAC. Consistent with this, hypo-osmotic induced Cl efflux is partially inhibited by the anion channel blockers A-9-C and NPPB (Teodora et al., 1998).

It has been suggested that osmotic stress in plant roots is perceived as a change in cell turgor pressure and that mechanosensitive anion channels act as ‘turgor sensors’ mediating Cl efflux (Teodoro et al., 1998; Zingarelli et al., 1999). Indeed, it is well established that Charaphytes (arguably the best understood system for turgor regulation and ion channel control) employ mechanosensitive channels in the perception of osmotic stress (for a review, see Shepard et al., 2002). However, there are no reports of mechanosensitive channels in higher plant root cells, and experiments in which oil microinjection was used to directly modulate turgor pressure within the growing root hairs of A. thaliana (Lew, 1996) failed to find evidence for a turgor-sensing mechanism. Despite this, it would be interesting to determine the effect of membrane stretch on ARAC activity (in analogy to the modulation of voltage-gated plant anion channel activity by membrane stretch; Barbara et al., 1994; Dioudonnéet al., 1997), to investigate a possible role for the ARAC in turgor sensing in A. thaliana roots.

It is unclear what role desiccation-activated S-type channels play in osmoregulation (Dauphin et al., 2001). The stimulation of this channel in response to a hyperosmotic shock favours enhanced Cl efflux; this is counterintuitive to the Cl flux expected for a role in osmotic adjustment during hyperosmotic stress. However, these S-type channels may play roles in signalling and/or modulating the response to osmotic stress. For example, Shabala & Lew (2002) reported oscillations in net K+ and H+ fluxes, which are suggested to ‘fine-tune’ the osmotic potential of the cell and contribute to the efficiency of cell osmotic adjustment. The S-type channel could provide an analogous mechanism allowing Cl fluxes to oscillate as part of a fine-tuning of Cl fluxes and membrane voltage during osmotic stress.

Boron efflux. Anion channels in the root periphery may play a central role in boron (B) resistance in barley plants (Hayes & Reid, 2004). B toxicity is a significant global problem in agricultural regions (Nable et al., 1997) where high levels of soil B limit growth and yield of many crop species. Based on ether : water partition coefficients for B, B influx is likely to be passive across the lipid bilayer. Hayes & Reid (2004) recently established that the lower B accumulation by B-tolerant barley (Sahara) compared with a B-sensitive cultivar (Schooner) probably resulted from enhanced borate anion [B(OH)4] efflux from the cytosol. Two mechanistic models have been proposed by Hayes & Reid (2004) for borate efflux from barley root cells. One is based on a mechanism previously proposed by Frommer & von Wirén (2002) for BOR1 (a putative borate transporter involved in xylem loading of borate; Takano et al., 2002) in which borate efflux occurs by anion (HCO3?) exchange. The alternative model proposes borate efflux from the cytosol via plasma membrane anion channels. This is based on the steep electrochemical gradient favouring borate efflux from the cytosol and the finding that borate efflux appears to be sensitive to the anion channel blockers DPC and niflumate (Hayes & Reid, 2004). However, the latter model is undermined by the absence of reports showing anion channel permeability to borate. An investigation of plant anion channel permeability to borate would be timely and may provide useful insights into B tolerance mechanisms in plants.

Channels mediating anion efflux from the root stele Biophysical properties  Physical and enzymatic manipulations have been developed to release and identify protoplasts from the root stele (Wegner & Raschke, 1994; Roberts & Tester, 1995; Maathuis et al., 1998). As a consequence of these developments, anion channels have been described in the xylem parenchyma cells (XPCs) of barley roots (Wegner & Raschke, 1994; Kohler & Raschke, 2000; Kohler et al., 2002), root stelar cells from maize (Gilliham & Tester, 2005) and A. thaliana root pericycle cells (Kiegle et al., 2000).

Three distinct anion channel conductances have been identified in barley xylem parenchyma cells (Kohler & Raschke, 2000): an inwardly rectifying anion channel (X-IRAC), a quickly activiating anion conductance (X-QUAC) and a slowly activating anion conductance (X-SLAC) (Fig. 3). X-IRAC- and X-QUAC-like activities have also been reported in maize root stelar cells, which display remarkably similar properties to their barley counterparts (Gilliham & Tester, 2005). In both maize and barley, X-IRAC occurs in a small proportion of cells (17 and 11%, respectively) and at low abundance. X-IRAC exhibits characteristics typical of HAACs; namely, strong inward rectification with negative going voltages, slow channel gating characterized by long (up to seconds) open and closed times and large unitary conductance (for barley, 70–85 pS using 34 mm Clext and 124 mm Clcyt; for maize, 90 pS using 2 mm Clext and 100 mm Clcyt). However, although X-IRACs from maize and barley display strong inward rectification, the gating mechanisms appear fundamentally distinct. In maize, X-IRAC activity is strongly voltage dependent (i.e. the channel open probability increases with negative voltages) whereas its counterpart in barley is voltage independent, exhibiting an increase in unitary conductance with membrane hyperpolarization (e.g. 71 pS at −133 pS to 47 pS at −73 mV and 13 pS at −13 mV); this is reminiscent of HAACs described in barley suspension cultured cells (Amtmann et al., 1997). Despite these fundamental differences in gating mechanism, X-IRAC activity in both maize and barley increases with increasing Ca2+cyt.

Figure 3.

Representative current–voltage curves for the three plasma membrane anion conductances from barley (Hordeum vulgare) xylem parenchyma cells, adapted from Kohler & Raschke (2000). X-SLAC, slowly activating anion conductance; X-QUAC, quickly activating anion conductance; X-IRAC, inwardly rectifying anion conductance.

X-QUAC exhibits rapid (ms) activation kinetics and partial (up to 50%) inactivation at hyperpolarized voltages. The I–V relationships of the whole-cell X-QUAC currents are complex and reveal unique voltage gating; typical currents exhibit large inward rectification at voltages negative of −40 mV, large outward rectification at currents positive of the equilibrium potential of the permeant anion and a negative slope conductance between approximately 0 and −50 mV, resulting in local maximum and minimum currents (reminiscent of the U-shaped I–V relationship displayed by R-type currents; compare Figs 1a and 3). X-QUAC currents are completely and irreversibly blocked by 100 µm DIDS and IAA-94 in barley and 100 µm niflumate (Ki = 17.5 µm) in maize. In contrast to X-IRAC, X-QUAC is reduced in both maize and barley by Ca2+cyt increases. Furthermore, in maize, X-QUAC activity is also reduced by ABA and water stress, prompting the proposal that this current is regulated in response to water stress via an ABA-induced Ca2+-dependent signalling mechanism (Gilliham & Tester, 2005). Selectivity sequences for X-QUAC have been determined by substituting extracellular Cl with a variety of test anions and deriving the relative permeability (PA/PCl) from current reversal potentials: in barley, NO3 (3) > Cl (1) >> malate2– (0.28) (Kohler & Raschke, 2000) and in maize, NO3 (1) = Cl (1) > I (0.59) > malate2– (0.18) > SO42– (0.12) > citrate3– (0.1) (Gilliham & Tester, 2005). However, with the exception of NO3 efflux via X-QUAC in barley (Kohler et al., 2002), the efflux of these test anions via X-QUAC has not been directly demonstrated and thus the physiological significance of these selectivity sequences for xylem loading is uncertain.

X-SLAC is only rarely observed in barley XPCs (7% of cells; Kohler & Raschke, 2000). Its very slow (tens of seconds) kinetics and voltage gating properties resemble those of S-type currents. Typically, X-SLAC is inactivate at hyperpolarized membrane voltages and activated by membrane depolarizations positive of −120 mV, exhibiting a peak current magnitude at −20 mV (Fig. 3). This current is abolished by elevated Ca2+cyt.

It is interesting that the stelar cells of barley and maize have markedly similar types of ion channel activity, yet ARACs (which bear only limited resemblance to the X-QUAC) appear to be the prevalent anion channel type in the pericycle of the A. thaliana root stele (Kiegle et al., 2000). It will be interesting to conduct studies on a wider range of plant species to determine if differences in ion channel types observed in the stele of maize, barley and A. thaliana represent fundamental differences between monocotyledonous and dicotyledonous species or whether they merely reflect differences in experimental conditions and/or cell types.

Physiological significance  It is likely that salt release into the xylem vessels involves the passive efflux of cations and anions across the plasma membrane of root stelar cells (Wegner & Raschke, 1994). In barley and maize root steles, K+ efflux channels have been described (see Roberts & Snowman, 2000 for a review) and thus the coexistence of anion channels in these root cell types establishes the possibility of simultaneous passive release of anions and cations. Channel-mediated ion release from stelar cells depends on the resting voltage of cells, both as a regulating factor of channel activity and as part of the driving force for passive transport. Kohler & Raschke (2000) identified a window of membrane voltage (between −50 and 30 mV) positive of Ek and negative of Eanions, in which K+ efflux and anion efflux channels could be simultaneously open. Although the membrane voltage of stelar cells from intact transpiring plants has not been determined, several studies indicate that the plasma membrane of these cells is likely to be depolarized relative to that in peripheral root cells (Dunlop, 1982; DeBoer & Prins, 1985; Roberts & Snowman, 2000). Thus, although this scenario may be regarded as oversimplified in that it ignores the likely activity of a myriad other electrophoretic transporters, it is plausible and useful in providing a starting point to describe salt release into the xylem vessels.

All three anion conductances described in barley and maize root stele exhibit distinct voltage dependence, frequency of occurrence and control by Ca2+cyt (Table 1), indicating that they are likely to play distinct roles in salt release. From the I–V relationships, Kohler & Raschke (2000) estimated Cl fluxes relative to fresh weight to be 11, 8 and 5 µmol l−1 g−1 h−1 for X-QUAC, X-SLAC and X-IRAC, respectively. These numbers far exceed the measured values for Cl release from xylem vessels of intact barley roots using 36Cl tracer (from 0.2 to 4 µmol l−1 g−1 h−1; Pitman, 1982), indicating the potential of any one of these conductances to mediate the loading of anions into the xylem vessels. As X-QUAC is the most frequently observed in barley and maize steles, this channel is likely to be the most influential (Kohler & Raschke, 2000; Gilliham & Tester, 2005). Consistent with this, DIDS (a potent inhibitor of X-QUAC in barley; Kohler et al., 2002) significantly reduces anion loading into the xylem of intact excised barley roots (Kawachi et al., 2002). Furthermore, water stress and ABA treatment result in an accumulation of solutes in maize roots (Cram & Pitman, 1972; Sharp & Davies, 1979); this is entirely consistent with the inhibition of anion loading into the xylem vessels via the downregulation of X-QUAC in maize root stelar protoplasts by ABA and water stress (Gilliham & Tester, 2005).

X-SLAC and X-IRAC may contribute less to net anion loading of the xylem based on their lower abundance and lower transport capacity. However, they could play more prominent roles in the regulation of anion loading into the xylem. For example, based on the increase in conductance at hyperpolarized voltages, Kohler et al. (2002) suggest that X-IRAC may regulate plasma membrane voltage by providing an additional pathway for anion efflux to prevent excessive hyperpolarization (possibly resulting from imbalance in cation/anion efflux and/or H+ pumping).

2. Organic anion efflux from higher plants

The exudation of organic compounds from roots can alter rhizosphere chemistry, soil microbiological populations and plant growth, although enhanced organic anion (OA) efflux from roots is primarily associated with nutrient (particularly phosphorus) deficiency and exposure to toxic metals (particularly Al3+). In almost all soils, phosphorus (P) is the most limiting nutrient because of its low bioavailability. Although the P content of soil is often adequate, P has a low bioavailability as a result of low concentrations (1–5 µm) of soluble P, with most P being adsorbed to soil minerals or fixed into inaccessible organic forms. Thus, plants have evolved a range of ‘coping’ strategies to enhance P uptake (see Vance et al., 2003 for a review); one key adaptation is the secretion of OAs from roots which subsequently acts to desorb phosphate (Pi) from the soil by anion exchange (Table 2). Secretion of OAs can be localized to specific regions of the root system and can be transient or continue for most of the life of the plant. Most extensively studied is the enhanced release of citrate and malate from proteoid or cluster roots (specialized short lateral roots with a dense covering of root hairs) in the members of the Proteacae (Skene, 2003).

Table 2.  Some plant species in which phosphorous deficiency enhances organic acid exudation from the roots
Plant speciesPrimary anion secretedRelative increase in OA secretion (%)CommentsReference
  1. The relative increase in organic anion (OA) secretion is calculated as (OA secretion from P-depleted plants/OA secretion from P-replete plants) ×100.

Chick pea (Cicer arietinum)Citrate, malonate173Whole rootNeumann & Romheld (1999)
Lupinus albusMalate, citrate692Cluster rootsNeumann & Romheld (1999)
Rice (Oryza sativa)CitrateUp to 485 (cultivar dependent)Whole rootKirk et al. (1999)
Medicago (Medicago sativa)Citrate181Whole rootLipton et al. (1987)
Rape (Brassica napus)Citrate, malate450/3814Root tip/whole rootHoffland et al. (1989); Zhang et al. (1997)
Radish (Raghanus sativus)Tartarate4280 Zhang et al. (1997)
Phaseolus vulgarisCitrate, tartate, acetateUp to 235 (cultivar dependent)Whole rootShen et al. (2002)
Soybean (Glycine max)Oxalate300Whole rootDong et al. (2004)
Arabidopsis thalianaCitrate, malateUp to 354 (cultivar dependent)Whole rootNarang et al. (2000)
Cabbage (Brassica oleraceae)Citrate, succinate600Whole rootDechassa & Schenk (2004)
Potato (Solanum tuberosum)Succinate3500Whole rootDechassa & Schenk (2004)

Al3+ toxicity significantly limits crop production on acid soils worldwide, with many crop species being particularly sensitive to Al3+ stress (see Kochian et al., 2004 for a review). OA efflux correlates strongly with Al3+ tolerance in a number of plant species (see table 1 in Kochian et al., 2004) which is thought to reflect the ability of OAs to chelate Al3+ and reduce its bioavailability.

At the near-neutral pH of the cytosol, most organic acids exist predominantly in their dissociated form (for example 99% of cytosolic malic acid will occur in the malate2– form). In addition, the large inside negative plasma membrane voltage of most plant cells ensures that the electrochemical gradient for the OA greatly favours their passive movement out of the cells. Consequently, it is widely accepted that OA efflux is probably mediated via ion channels.

Channels mediating organic anion efflux from the root periphery Biophysical properties Anion channels regulated by phosphate supply. Pi availability regulates anion channel activity in root epidermal cells from A. thaliana (Diatloff et al., 2004) and L. albus (Zhang et al., 2004a). Significantly, these channels are highly permeable to OAs, displaying high selectivity for malate and citrate over Cl (the relative permeability for citrate over Cl is approximately 26). The phosphate-regulated Arabidopsis root anion channel (PR-ARAC) exhibits properties typical for R-type channels (Fig. 4); in contrast, whole-cell OA-selective currents from L. albus resemble HAACs, displaying inward rectification, rapid (< 20 ms) activation kinetics and a slow (seconds) partial (50%) inactivation. Both channels are upregulated by Pi depletion; however, PR-ARAC is only active following Pi starvation, indicating a strict dependence on the Pi status of the plant. In contrast, OA-selective currents in L. albus are present in P-replete plants, although current density increases 2-fold following P starvation. Surprisingly, these currents were present at equivalent levels of activity in the epidermis of cluster and noncluster L. albus roots, indicating that the two cell types have equal capacity to release OA, despite the fact that OA exudation from cluster roots is far greater than that from noncluster roots (Skene, 2003). This suggests that either the OA-selective channels are differentially regulated in these different root cell types in planta or that the driving force (i.e. cytosolic OA concentration and membrane voltage) for OA efflux is much greater in cluster root cells. Interestingly, Yan et al. (2002) showed that H+ ATPase activity is enhanced in cluster roots, consistent with cluster root cells being more hyperpolarized than noncluster root cells and therefore more likely to possess a more polarized membrane voltage favouring channel activity and a greater electrochemical gradient for OA efflux. Clearly, it will be interesting to directly record the membrane voltages of cluster and noncluster root cells using electrophysiological techniques.

Figure 4.

Whole-cell current–voltage relationship of malate efflux currents from epidermal root protoplasts isolated from phosphate-starved plants. Currents were recorded 2 (•) and 8 (▪) min after obtaining the whole-cell configuration of the patch clamp technique. Inset: example of whole-cell currents used to construct the current–voltage relationship. Currents were recorded 2 min after obtaining the whole-cell configuration and result from voltage pulses ranging from −216 to +54 mV in 10-mV steps (for clarity, only currents every 40-mV interval are shown) from a holding voltage of −16 mV. The pipette solution was 60 mm malate, 1 mm MgATP, 10 mm N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES), 5 mm EGTA, pH 7.2, and sorbitol to 720 mOsm kg−1. The bath solution was 5 mm LaCl3, 10 mm CaCl2, 5 mm MgCl2, 10 mm 2-(N-Morpholino) ethanesulfonic acid (MES), pH 6.0, and sorbitol to 700 mOsm kg−1. Further experimental details can be obtained from Diatloff et al. (2004).

Little is currently known about the mechanisms regulating the phosphate-regulated channels, although PR-ARAC-mediated currents undergo rapid rundown within 20 min of achieving the whole-cell configuration in a patch clamp experiment, consistent with channel activity being under tight posttranslational control by an unknown cytosolic factor(s) which is washed out of the cytosol by the pipette solution.

Al3+-activated anion channels. Al3+-activated anion channels (ALAACs) have been reported in the root tips of wheat (Ryan et al., 1997; Zhang et al., 2001) and maize (Kollmeier et al., 2001; Pineros & Kochian, 2001; Pineros et al., 2002). These channels exhibit many features common to HAACs, including rapid/instantaneous activation kinetics, slow partial inactivation and inward rectification. ALAACs were first demonstrated by Ryan et al. (1997) as inwardly rectifying C1 currents, although a subsequent study by the same group showed that this channel was highly permeable to malate (with a calculated Pmal/PCl > 2.6; Zhang et al., 2001). A key finding of these studies was the significantly greater ALAAC activity in protoplasts isolated from an Al-tolerant wheat line compared with those from a near-isogenic Al3+-sensitive line; therefore, this study identified the wheat ALAAC as a possible major Al-tolerance factor.

Activation of ALAAC in wheat root protoplasts is strictly dependent on the presence of extracellular Al3+, that is removal of Al3+ resulted in complete inactivation. The mechanism by which Al3+ activates the channel is unknown, although activation is not possible in isolated excised patches of membrane, indicating that Al3+ activation is unlikely to be an intrinsic property of the channel. Activation may involve phosphorylation as indicated by the inhibition of Al3+-dependent malate release from wheat roots by the kinase inhibitor K-252a (Osawa & Matsumoto, 2001). The wheat root ALAAC is partially (50–80%) blocked by 100 µm niflumate and 10 µm DPC (Zhang et al., 2001) and exhibits a unitary conductance of 66 pS (102 mm Clcyt: 20 mm Clext; Ryan et al., 1997).

Recently, a novel type of membrane transporter has been cloned (aluminum-activated malate transporter; ALMT1) which may encode ALAAC in wheat roots (Sasaki et al., 2004). ALMT1 was identified via a subtraction hybridization approach using the pair of near-isogenic wheat lines which were used in the patch clamp studies (Ryan et al., 1997). Notably, ALMT1 has been: (i) localised to the plasma membrane of wheat cells (Yamaguchi et al., 2005); (ii) shown to express more strongly in the root tip of the A1-tolerant wheat line (Sasaki et al., 2004); and (iii) shown to mediate A1-dependent malate efflux (Sasaki et al., 2004). Furthermore, ALMT1 expression in suspension cultured tobacco cells (Sasaki et al., 2005) and intact barley plants (Delhaize et al., 2004) conferred increased Al tolerance, indicating that ALMT1 is probably a major Al-tolerance gene.

ALAACs from root tip protoplasts from the Al-tolerant maize cultivar Cateto Colombia 96/71 are also strictly dependent on the presence of extracellular Al3+ for activity (Pineros & Kochian, 2001; Pineros et al., 2002); however, in contrast to its counterpart in wheat, Al3+ was effective in activating the channel in isolated membrane patches. Thus, Al3+ may bind the channel directly (or with a membrane receptor closely associated with the channel). In Cateto Colombia 96/71 maize, ALAACs are characterized by long open and closed times lasting several seconds and a small unitary conductance ranging from 18 to 27 pS (100 mm Clcyt: 11 mm Clext). Channel activity is also dependent on Clext, reminiscent of that described for R-type channels (see section on Rapidly activating anion efflux channels). Notably, the selectivity of the ALAAC (for citrate or malate) from Cateto Columbia 96/71 maize is currently unknown. In a separate study using protoplasts from the root tips of the Al-tolerant maize cultivar ATP-γ (Kollmeier et al., 2001), ALAACs have been shown to be permeable to malate and citrate, albeit to a lesser extent than reported for ALAACs in wheat [i.e. the selectivity sequence, based on the magnitude of anion efflux currents, is Cl (1) >> malate (0.28) > citrate (0.17)]. However, the ALAACs identified by Kollmeier et al. (2001) differ significantly from those from Cateto Columbia 96/71 maize; for example, the ALAAC from the maize cultivar ATP-γ (1) has a significantly larger unitary conductance (144 pS using 104 mm Clcyt: 22 mm Clext), (2) does not inactivate after the removal of Al3+, and (3) is insensitive to extracellular Cl. Thus, it remains a distinct possibility that there are several distinct types of ALAAC in maize roots which have distinct electrophysiological properties. Thus, there may be marked differences in the properties of ALAACs in different maize cultivars and these differences could correlate with differences in Al tolerance.

Control of organic anion efflux  Although OA efflux from the roots of some plant species can be closely correlated with intracellular OA content, much evidence now suggests that intracellular content is unlikely to be the primary factor regulating OA efflux (for further discussion, see Delhaize et al., 2001; Ryan et al., 2001; Ma & Furukawa, 2003). In contrast, it is widely accepted that transmembrane transporters exert primary control on many metabolic pathways and processes (Kunze et al., 2002); thus, the identification of OA-permeable channels in root cells represents an exciting opportunity to gain unique insights into the processes that may control OA efflux from roots. Particularly, an understanding of the molecular processes dictating OA-permeable channel regulation will be key in understanding the role of these ion channels in adaptation to P stress or Al toxicity. Our current understanding indicates that ALAACs are under strong posttranslational control, although the mechanisms and intermediate steps involved are unclear. In the case of Pi-regulated channels, much less is known and regulation may be at the transcriptional and/or posttranslational level.

IV. Conclusions and Prospects

Higher plant cells are equipped with a variety of anion channels co-residing in the plasma membrane which are permeable to a range of physiological anions. Most anion channels can be divided into several distinct classes based on their voltage dependence and kinetics. Depolarization-activated channels include R-type and S-type channels, which mediate anion efflux, and outwardly rectifying channels, which mediate anion influx. Hyperpolarization-activated channels mediate anion efflux and generally display higher unitary conductances than depolarization-activated anion channels. In addition, root plasma membrane anion channels display unique features that indicate specific functions in root physiology. Most notable are the organic anion-selective ALAACs and phosphate-regulated channels. ALAACs, although resembling HAACs, possess a unique Al3+-dependent gating mechanism which identifies them with a key role in Al tolerance. Similarly, the dependence of anion channel activity on the phosphate status of the plant identifies them with roles in enhancing the phosphate acquisition efficiency of the plant.

Over the past 5 years there has been considerable progress in identifying and characterizing channel activities in the roots of higher plants using electrophysiological techniques. Although much can still be gained by electrophysiological investigation, the major challenge for the next 5 years is the identification and characterization of anion channel genes in plants. The recent identification of ALMT1 represents an exciting major development towards achieving this goal. However, the diverse electrophysiological properties of plant anion channels, both within and between the classes of anion channel, suggest that several large and diverse multigene families probably encode plant plasma membrane channels. Indeed, it would not be surprising if there was little homology between the genes encoding the different anion channel types encountered in plant cells. For example, plant genomes contain at least two gene families with similarity to anion channel genes in animal systems: the chloride channel (CLC) family (Nilius & Droogman, 2003) and a subset of the ATP-binding cassette (ABC) protein superfamily (Theodoulou, 2000). Genes encoding CLCs have been cloned from A. thaliana and other plant species (Barbier-Brygoo et al., 2000); although localization studies have shown plant CLCs to be localized to intracellular membranes (Hechenberger et al., 1996; Geelen et al., 2000), this does not preclude the possibility that some family members might reside at the plasma membrane. Indeed, variable membrane locations among family members have been reported (e.g. natural resistance-associated macrophage protein; Nramp metal transporters; Portnoy et al., 2000). ABC transporters transport a wide range of organic and inorganic solutes and include (1) the ABC transporter pdr12p, which mediates OA efflux from Saccharomyces cerevisiae (Piper et al., 1998), and (2) CFTR, which is a Cl channel in mammalian cell plasma membranes (Nilius & Droogman, 2003). The presence of CFTR homologues in plants is indicated by the findings of pharmacological studies (Leonhardt et al., 1999), which showed that S-type currents are sensitive to a range of CFTR-specific inhibitors, and raises the possibility that S-type channels may be encoded by the ABC superfamily.

It is envisaged that the development of strategies designed to identify plant anion channel genes will rely on knowing the physiological and biophysical attributes of the functional ion channel. For example, several laboratories have developed screens to identify functional plant anion channels based on yeast complementation using plant cDNA libraries (e.g. Ryan et al., 2003). In my laboratory, screens were developed based on either the complementation of yeast mutants (e.g. pdr12Δ, which is defective in weak organic acid efflux and shows hypersensitive growth to benzoate and propionate) or the sensitivity of yeast growth to toxic organic anion analogues (e.g. fluorocitrate, which binds the aconitase in mitochondria and inhibits respiration). Unfortunately, these screens failed to isolate OA-selective channel genes from A. thaliana cDNA libraries. However, based on our recent electrophysiological characterization of PR-ARAC from A. thaliana roots (Diatloff et al., 2004), these screens would be best served by cDNA libraries derived from root epidermal cells of P-starved A. thaliana plants.

Once genes are identified, they will provide new opportunities to reveal the roles of anion channels in plant and root physiology. These include heterologous expression approaches in which ion channel activity can be analysed with voltage clamp techniques, studies using plants transformed with GFP:gene fusion constructs to identify tissue and membrane localization, and the generation and phenotyping of gene knockout mutants. Indeed, the identification of plant anion channel genes represents the major goal for a better understanding of their central role in plant biology, which should deliver many exciting opportunities for plant biotechnology, particulary with respect to improving nutrient acquisition and enhancing tolerance to a range of abiotic stresses.

Acknowledgements

I am grateful to the reviewers for helpful comments and suggestions and to the BBSRC for funding ion channel research in my laboratory.

Ancillary