S. Thomine, Institut des Sciences du Végétal, CNRS, Avenue de la Terrasse, 91 198 Gif-sur-Yvette, France Fax: +33 1 69 82 37 68 Tel: +33 1 69 82 37 93 E-mail: email@example.com
Plant anion channels allow the efflux of anions from cells. They are involved in turgor pressure control, changes in membrane potential, organic acid excretion, tolerance to salinity and inorganic anion nutrition. The recent molecular identification of anion channel genes in guard cells and in roots allows a better understanding of their function and of the mechanisms that control their activation.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
The role of anions in plant cells is clearly distinct from that encountered in animal cells. First, chloride is by far the most prevalent anion in animal cells whereas plant cells contain a complex mixture of diverse anions including, besides chloride, nitrate, sulfate, phosphate and organic anions, such as citrate, malate or oxalate in varying proportions . Second, anions are accumulated in plant cells, whereas in most animal cells the chloride gradient favors the influx. Third, in mature plant cells, the main reservoir of ions is not the cytosol but rather the central vacuole which occupies about 80% of the cell volume. These differences imply original functions for plant anion channels and probably account for the recent discovery of anion channel gene families unique to plants and not found in animal genomes.
Ion channels switch between open and closed states according to the factors that control their gating. When an ion channel is open, massive ion fluxes occur according to their electrochemical gradients. An important step in studying and understanding the function of ion channels was development of the patch clamp technique by Neher and Sakmann in the late 1970s . In contrast, the first anion channel structure was determined in bacteria only in 2002  and the first plant genome sequence was released in 2000 . Thus, the electrophysiological properties of plant cell membranes were thoroughly studied before the genes encoding anion channels were identified.
At the level of plant cell plasma membrane, both the membrane potential, which is highly electronegative (usually below −100 mV), and the intracellular anion accumulation dictate anion efflux through anion channels when they open . In addition, in the case of organic anions such as malate and citrate, which are carboxylic acids, the pH gradient (neutral inside, acidic outside) further favours their efflux through channels, as their protonation in the extracellular space maintains a steep gradient of the anionic species. The slightly negative electrical gradient across the vacuolar membrane (tonoplast) also drives anion ‘efflux’ from the cytosol to the vacuolar lumen. However, active transport, such as H+/anion co-transport, is clearly required for the high accumulation of certain anions in this compartment [6,7]. Paradoxically, active transport may also be necessary to release anions from the vacuole when the cell undergoes important changes in turgor pressure.
The diversity of anions in plant cells means that anion channels serve a wide range of functions. Whatever its anion selectivity, the opening of an anion channel in the plasma membrane shifts the membrane electrical potential towards the equilibrium potential of anions, i.e. it will lead to a depolarization [8,9]. Cell depolarization can induce signalling events or lead to the activation of voltage-gated ion channels. When anion efflux through anion channels is coupled to potassium efflux, anion channels act as major players in plant cell osmotic regulation. Depending on their selectivity, anion channels may play more specific roles. For example, chloride-selective channels may be involved in salt tolerance [10,11]; nitrate-selective channels in nitrogen homeostasis and organic-acid-selective channels in carbon metabolism (e.g. malate channels in CAM (crassulacean acid metabolism) plants) or pH regulation .
Besides their selectivity, another important feature of anion channels is their gating. It is unlikely that a plant cell can survive with constitutively open anion channels because this would lead to massive loss of ions and depolarization. This review will provide examples of anion channels gated by diverse mechanisms. Many of the anion channels described are voltage regulated, opening in response to membrane depolarization or hyperpolarization . Additionally, some channels display strong regulation by intracellular signalling events, such as phosphorylation, and intracellular or extracellular metabolites [14–17]. In addition to canonical anion channels characterized by open–closed transitions, the review will address transporters that allow anion fluxes along their electrochemical gradient (i.e. MATE or NAXT) but have not been characterized as channels stricto sensu. Such transporters obey similar biophysical constraints and lead to similar consequences for the cell and their detailed characterization may reveal in some cases that they function as channels. Particular attention will be given to the most recent progress on two systems in which anion channels have been intensively studied: the guard cells of stomata, which undergo fast reversible anion channel dependent change in turgor, and roots, which illustrate the functions of anion channels in anion excretion to the rhizosphere or to the xylem. For other aspects of plant anion channel biology, the reader is referred to other review articles [1,9,18–21].
Guard cell anion channels: from electrophysiological characterization to molecular structure
Stomata are small pores in the epidermis of plant leaves and stems, which control plant gas exchange. Each stomatal pore is surrounded by a pair of guard cells which control the pore size by swelling or shrinking as a response to changes in the surrounding environment or to intrinsic signals. This depends on the activation of ion channels, changes in the guard cell osmotic pressure and movement of water in or out of the guard cells. Adequate stomatal regulation ensures sufficient uptake of carbon dioxide with minimal loss of water. This is particularly important in situations where water resources are limited. Rapid stomatal closure also limits the entrance of pathogens  and air pollutants such as ozone [23,24].
The activation of guard cell anion currents was recognized already more than 20 years ago as one of the first steps in the induction of plant stomatal closure [25,26]; however, genes encoding guard cell anion channels were characterized only recently [24,27,28]. Activation of guard cell anion channels and release of anions is the critical step for induction of plant stomatal closure [29,30]. This depolarizes guard cell plasma membrane and triggers the activation of guard cell outwardly rectifying K+ channels . Overall the osmotic pressure inside the cell is reduced, water flows out, guard cells become flaccid and the stomatal pore closes.
Applying the patch clamp technique to the plasma membrane of Vicia faba guard cell protoplasts led to the characterization of two distinct, rapid-type (R-type) and slow-type (S-type), anion channels [13,32,33]. Activation of R-type anion channels is voltage dependent and they activate/deactivate within milliseconds. In V. faba guard cells, R-type currents also exhibit time-dependent inactivation within tens of seconds. The activation/deactivation time constants of S-type anion channels are in the range of 10 s, channel activity shows weak voltage dependence and S-type currents do not inactivate with time. Both channels participate in stomatal closure but not equally in all responses. Drought-induced plant hormone abscisic acid (ABA) activates both S- and R-type anion currents . In contrast, increases in CO2 partial pressure trigger consistent activation of S-type currents only, whereas R-type channels are either activated or inactivated .
Identification, regulation of SLAC1
A gene encoding the guard cell plasma membrane S-type anion channel was genetically isolated from independent Arabidopsis mutant screens for ozone sensitive rcd (radical induced cell death) and for carbon dioxide insensitive (cdi) mutants [35,36]. Both rcd3 and cdi3 turned out to carry mutations in the gene At1g12480 and were thus renamed slac1-1 (S456F)  and slac1-2 (G194D) , respectively. Guard cells extracted from slac1 mutants do not have detectable Ca2+- and ABA-induced S-type anion channel activity, whereas activation of R-type channels remains intact . Heterologous expression of SLAC1 alone in Xenopus oocytes does not generate clear anion currents. Therefore, as it could not be excluded that SLAC1 represented only a subunit of the S-type anion channel, the protein was initially named slow anion channel associated 1 . SLAC1 expression is highly specific for guard cells. Plants lacking functional SLAC1 display impaired stomatal closure in response to all major endogenous (ABA, Ca2+, NO) and environmental (CO2, darkness, humidity, ozone) stimuli. These results illustrate the crucial role of guard cell plasma membrane S-type anion current activation for the induction of stomatal closure [24,28,37].
The importance of regulatory processes such as phosphorylation and dephosphorylation for activation and inactivation of S-type anion channels was shown already by Schmidt et al. . On this basis, it was suggested that lack of regulatory proteins in Xenopus oocytes could be the reason why initial experiments trying to produce SLAC1-type anion currents in heterologous systems failed . This possibility was independently tested by two laboratories [15,38] which showed that co-expression of SLAC1 with a protein kinase consistently induced S-type anion channel currents in Xenopus oocytes confirming that SLAC1 forms the guard cell plasma membrane S-type anion channel per se. Functional expression of SLAC1 reconstituted anion currents with slow kinetics, a higher permeability for nitrate than for chloride and low permeability to malate, bicarbonate and sulfate, as observed in guard cells [15,39,40]. The large increases in malate and fumarate content in slac1-2 and slac1-3 guard cells are thus probably not due to lack of efflux of these organic anions through the slow anion channels but is associated instead with a profound alteration in guard cell function in these mutants .
Both Geiger et al.  and Lee et al.  showed that phosphorylation by the protein kinase OST1 (also known as SnRK2.6 or SRK2E) is critical for full activation of SLAC1-dependent S-type anion currents in Xenopus oocytes (Fig. 1). In addition, S-type anion channel activities were clearly reduced in guard cells isolated from the ost1-2 mutant illustrating the importance of OST1-dependent phosphorylation for SLAC1 activation in guard cells . Additional evidence for OST1-dependent activation of SLAC1 was provided by Vahisalu et al.  who showed that OST1 phosphorylates multiple serines of SLAC1 hydrophilic N-terminal fragment. Furthermore, plants carrying a mutation in one of the phosphorylated serines (S120) show impaired stomatal responses to ozone , similar to that of slac1 loss-of-function mutants , indicating that OST1-dependent phosphorylation at this site is physiologically relevant for induction of stomatal closure. The C-terminal tail of SLAC1 is also phosphorylated by OST1 in vitro . Nevertheless, SLAH1, a homologue of SLAC1 which lacks the N- and C-terminal tails phosphorylated in SLAC1, fully complements the stomatal phenotypes of slac1-2 when expressed in guard cells under control of the SLAC1 promoter .
Elevation of cytosolic Ca2+ also triggers the activation of guard cell slow anion currents . Analysis of S-type anion channel activation by Ca2+ in the context of ABA signaling showed that treatment by ABA or protein phosphatase inhibitors facilitates Ca2+-mediated activation of slow anion currents by lowering the intracellular Ca2+ concentration required to trigger their activation [42,43]. One way of translating Ca2+ signals goes through activation of Ca2+-dependent protein kinases (CPKs, Fig. 1). Guard cells from single and double mutants of Arabidopsis CPK3 and CPK6 have impaired ABA- and Ca2+-induced S-type anion currents . Recently Geiger et al.  showed physical interaction between CPK21, CPK23 and SLAC1. Furthermore, co-expression of these proteins with SLAC1 activates S-type anion currents in heterologous systems. However, only CPK21 kinase activity is Ca2+ sensitive.
Elevation of CO2 leads to the activation of S-type anion currents. Recent data demonstrated that the activation of SLAC1 currents is mediated by intracellular bicarbonate generated from CO2 by β-carbonic anhydrase, rather than by intracellular pH changes [46,47]. Bicarbonate-induced activation of S-type anion currents is positively controlled by OST1 kinase and negatively by HT1 kinase [47,48]. Interestingly, this activation requires elevated intracellular calcium levels suggesting the need for concomitant signals for guard cell CO2 response.
One of the major recent breakthroughs in plant biology was the discovery of the cytosolic ABA receptor PYR/PYL/RCAR proteins [49,50], which inhibit protein phosphatase 2Cs (PP2Cs), such as ABI1 and ABI2, by the formation of a ternary complex between ABA, PP2Cs and PYR/PYL proteins (Fig. 1). This mechanism also controls the OST1-dependent activation of SLAC1 as, in the absence of ABA, OST1 is kept inactive by PP2Cs. Upon ABA binding to the receptor complex, the activity of the PP2Cs is inhibited and OST1 is activated either by autophosphorylation  or by an upstream kinase [52,53]. In agreement with these results, protein phosphatases ABI1, ABI2  and PP2CA  abolish OST1-dependent activation of SLAC1-induced anion currents in Xenopus oocytes. Interestingly ABI1 and ABI2 also control CPK21- and CPK23-dependent activation of SLAC1 currents  suggesting that Ca2+-dependent SLAC1 activation is also controlled by ABA.
These findings provide an elegant signaling module for Ca2+-dependent and Ca2+-independent regulation of plant guard cell S-type anion channel SLAC1 (Fig. 1). However, the functional significance of CPK21 and CPK23 for stomatal regulation in planta is not yet fully understood. Lack of functional SLAC1 or OST1 causes impaired stomatal response to ABA [24,54], air humidity [24,55] CO2  and ozone , indicating that OST1-dependent activation of SLAC1 is required for stomatal closure in response to these stimuli. In contrast, plants with impaired CPK23 gene do not exhibit any stomatal phenotypes even though their Ca2+-induced activation of SLAC1 is reduced by 70% . Similarly, stomatal responses to major environmental stimuli seem to be intact in plants with impaired CPK21 gene (E Merilo & H Kollist, unpublished). Collectively, current data suggest that, although three different protein kinases can activate SLAC1, OST1-dependent phosphorylation is the main prerequisite for SLAC1 activation in guard cells.
Structure of SLAC1 and its homologues
Recently, the 3D structure of HiTehA, a bacterial homologue of SLAC1, was resolved at high resolution (Fig. 2). The bacterial TehA does not possess a soluble C-terminal domain comparable with that found in SLAC1 protein. Except for this domain, a model of the SLAC1 3D structure could be constructed with high confidence based on the 3D structure of TehA [39,56]. This revealed three critical features of this channel: its trimeric structure, the properties of its pore and an essential gating mechanism. First, TehA forms homo trimers (Fig. 2C). This feature may account for the cooperative opening of this channel which was noted already in early single channel recordings . In addition, a multimeric structure suggests a possible mode of activation by phosphorylation at the interface between the subunits . Second, the structure of TehA sheds light on the structure of the anion-permeable pore of SLAC1. Each subunit forms an independent pore (Fig. 2B, C, E). The pore has a remarkably constant diameter of about 5 Å through the membrane and is lined by hydrophobic or hydroxyl residues (Fig. 2B, D, E); it does not show any distinct binding site for anion in contrast to the anion permeation pathway through chloride channel (CLC) proteins . Such weak interactions between the pore and the permeating anions account for the observed selectivity sequence of different anions, which follows the energetic cost of their dehydration. Third, the structure of TehA revealed the presence of a phenyl ring blocking the anion permeation pathway (Fig. 2B, D). This phenylalanine residue (F450 in SLAC1) is conserved through the entire SLAC1 protein family. Mutational analysis demonstrated that this phenyl ring gates the pore of both TehA and SLAC1 . The ability of the channel to switch between closed and open states thus relies on the ability to move the phenyl ring away from the anion permeation pathway, probably by phosphorylation-induced conformational change of the protein. Thus, a central question will be to understand how phosphorylation of SLAC1 is coupled to the positioning of the phenyl ring in the permeating pore.
In addition to SLAC1, the Arabidopsis genome encodes four SLAC1 homologues, SLAH1–4 [28,39]. At the subcellular level, SLAH1–3 are located in the plasma membrane . Transformation with SLAH1 and SLAH3 under control of the SLAC1 promoter rescued slac1 mutant phenotypes, indicating that SLAH1 and SLAH3 are also capable of forming S-type anion channels. SLAH1 and SLAH2 are expressed in roots and SLAH3 is expressed in whole plant. Initial characterization indicated that SLAH1–3 were not expressed in guard cells . However, a recent study revealed that, in different growth conditions, SLAH3 is expressed in guard cells where it functions as an S-type anion channel . Similar to SLAC1, SLAH3 activation is regulated by phosphorylation by CPK21, which is controlled by the ABA receptor–phosphatase complex (Fig. 1). In contrast to SLAC1, SLAH3 voltage dependence is modulated by extracellular which facilitates SLAH3 activation by shifting its activation threshold towards guard cell resting membrane potential . The function of SLAH3 and its regulation by nitrate in guard cell movements remains to be investigated.
Identification of AtALMT12/QUAC1 – a component of guard cell R-type anion channel activation required for plant stomatal closure in response to major endogenous and environmental stimuli
Recently two parallel studies characterized Arabidopsis AtALMT12 (aluminum activated malate transporter12), which on the basis of sequence is similar to ALMT1 (see below) and showed that this protein is required for full stomatal closure induced by various stimuli [27,59]. AtALMT12 is preferentially localized to the plasma membrane of guard cells; moreover detailed electrophysiological studies of almt12 guard cells and Xenopus oocytes expressing the protein revealed that AtALMT12 forms the malate-sensitive R-type anion channel (Fig. 1) . AtALMT12 activates when plasma membrane is depolarized . The maximum current of R-type anion channels is around −100 mV in Arabidopsis thaliana guard cells [25,27,60]. R-type anion currents generated upon expression of AtALMT12 in Xenopus oocytes display a peak around −100 mV which is shifted to more negative values when extracellular malate concentration is increased (Fig. 1) . Contrary to what its name implies, extracellular Al3+ treatment does not stimulate AtALMT12-dependent anion currents [27,59]; thus it was suggested to rename the protein quickly activating anion channel 1, QUAC1 . However, it should be noted that R-type anion currents are only reduced by 40% in almt12 mutant guard cells and that, in the absence of extracellular malate, R-type currents are indistinguishable from those observed in wild-type [27,59]. This implies that other proteins also play a role in the formation of guard cell R-type anion channels. Of the 14 Arabidopsis ALMTs  clear guard cell specific localization is only shown for AtALMT12/QUAC1; however, AtALMT13 and AtALMT14 are good candidates based on their high homology to AtALMT12/QUAC1 .
What is the relative contribution of S-type and R-type anion channels in stomatal movements?
It is well established that the activation of S-type anion currents requires phosphorylation by protein kinases [15,38,41,45]. In contrast, no direct intracellular signaling pathway is known for R-type anion channel activation or deactivation. R-type anion channel activity is tightly regulated by membrane potential [27,59,60]. It is therefore possible that R-type anion channel activation does not require any intracellular activation pathway but is merely triggered by membrane depolarization. The hypothesis that R-type channel does not require any intracellular activator is indirectly supported by the recent finding that AtALMT12/QUAC1 is fully functional in Xenopus oocytes  without the need of other plant proteins to activate it, as opposed to SLAC1 [15,38,45]. In open stomata, which have a strongly hyperpolarized membrane potential around −150 mV, the guard cell R-type anion channels are inactive (Fig. 1). Depolarization of plasma membrane can be achieved by activation of S-type anion channel [26,30,32], inhibition of H+-ATPases [62,63], activation of Ca2+ channel [60,64] or a combination of these processes. In addition, increases in extracellular malate concentration shift R-type outward anion current activation threshold to more negative voltages [16,27]. Thus the activation of R-type anion channels can be triggered either by membrane depolarization or by an increase in extracellular malate concentration. This suggests the presence of a feed-forward regulation for R-type anion channel activation where a slight membrane depolarization would trigger a slight activation and the release of malate which in turn would lead to enhanced activation due to the shift of the activation threshold to more hyperpolarized potentials.
Stomatal opening is initiated by plasma membrane H+-ATPase driven proton efflux from guard cells. This shifts the plasma membrane electrical potential towards hyperpolarized values, which in turn triggers activation of voltage-gated potassium uptake channels . Concomitant accumulation of positive charges inside the guard cells has to be balanced by anions. Classical studies from the 1970s by Raschke and Outlaw and their colleagues [66,67] showed that guard cells are capable of anion uptake from the extracellular space. In the absence of extracellular anions, stomatal opening is achieved by biosynthesis of organic anions inside the guard cells. Recently, it was shown that the Arabidopsis ABC transporter AtABCB14 mediates malate and possibly fumarate uptake from the apoplast to the guard cells during stomatal opening . Conversely, anion efflux channels such as SLAC1 and QUAC have to be closed to allow for guard cell turgor build-up during stomatal opening. The membrane hyperpolarization induced by proton pump activation during stomatal opening is probably sufficient to close R-type/QUAC channels, due to their steep voltage dependence. As described above, SLAC1 is activated by phosphorylation. An intriguing question is whether the same PP2Cs that inhibit OST1 activity also function in SLAC1 inactivation by dephosphorylation. It was shown that PP2CA physically interacts with SLAC1, and inhibits SLAC1 activity in Xenopus oocytes . In addition, other PP2Cs are able to dephosphorylate the SLAC1 N-terminal domain in vitro (Kollist, unpublished results) suggesting that ABA-PYR/PYL-PP2C signaling module might also control SLAC1 inactivation during stomatal opening.
Anion fluxes through the vacuolar membrane of guard cells
During stomatal movements, guard cells undergo great dynamics changes in vacuole morphology. These changes, associated with ion fluxes across the tonoplast, are essential for stomatal movements, highlighting the crucial function of the vacuole during this process . Vacuolar chloride channel currents regulated by a calcium-dependent protein kinase were identified in V. faba guard cells . Apart from this report, little is known about anion channels at the tonoplast of guard cells.
AtMRP5 belongs to the ATP-binding cassette (ABC) transporter family and is expressed in guard cells [71,72]. Disruption of AtMRP5 leads to an ABA insensitivity of stomatal closure. This is in agreement with reports showing that ABC transporter modulators affect guard cell anion currents and stomata aperture [73,74]. It had been proposed that AtMRP5 is a subunit or acts as a regulator of guard cell S-type anion channels . Recently, it was shown that AtMPR5 is an inositol-hexakisphosphate transporter presumably localized at the tonoplast . The implication of inositol-hexakisphosphate in calcium mobilization and inhibition of inward rectifying K+ conductance in guard cells may account for the stomatal phenotype of atmrp5 [72,76].
Recently, AtCLCc, a member of the CLC family was shown to be targeted to the tonoplast and implicated in stomatal movements. Plants lacking functional AtCLCc display light and ABA insensitivity of stomatal movements associated with a dramatic decrease in guard cell chloride content . The role of AtCLCc in chloride homeostasis provided the first evidence of the importance of coordination between anion transport at the plasma and vacuolar membranes during stomatal movements. As the CLC family comprises channels and transporters, electrophysiological studies will be necessary to establish which of these two transport mechanisms is used by AtCLCc. Three other AtCLC members are located in the tonoplast (AtCLCa, AtCLCb and AtCLCg); however, only one of them, AtCLCa, is expressed in guard cells . AtCLCa encodes an /H+ antiporter involved in homeostasis . The high expression of AtCLCa in guard cells suggests that it may also be implicated in stomata movements . This raises the possibility that two vacuolar CLC family members, AtCLCc and AtCLCa, with distinct preferences for chloride or nitrate, respectively, could cooperate to mediate anion transport across guard cell tonoplast depending on the availability of these two anions. (See note added in proof.)
Root anion channels: anion excretion and loading into the xylem
Anion channels have been described in all root cell types investigated . Besides the ubiquitous relevance of anion fluxes in any plant cell, anion channels fulfill several root-specific functions. These functions include anion loading to the xylem and anion excretion to the rhizosphere. Xylem loading allows anion translocation to the shoots; it is especially relevant for nitrate which is taken up by the root but mostly reduced to be assimilated in amino acids in the leaves . Xylem loading of organic anions such as citrate is important for the translocation of metal cations that move from the root to the shoot as complexes with organic acids [78,79]. Anion excretion to the rhizosphere also serves diverse functions. It regulates the uptake rate of some mineral nutrients through futile cycles, or counterbalances the efflux of positive charges . The best documented anion efflux in root peripheral cells is the excretion of organic acids to the rhizosphere. This is part of a process in which plants release 30% of the carbon fixed by photosynthesis . Release of organic anions serves several functions. The best established mechanism is the chelation of aluminum in acidic soils but it is also implicated in phosphate mobilization (see below). In addition, together with many other compounds released in the rhizosphere by plant roots, excreted organic anions may be used as carbon sources by bacteria and fungi living in the rhizophere and thus participate in the control of the microorganism populations . In the context of symbiosis, organic acids are excreted to intracellular symbiosomes to provide a carbon source to the nitrogen fixing bacteria. An organic anion efflux system belonging to the peptide transporter (PTR) family putatively involved in this function has been identified in alder nodules colonized by Frankia .
Anion efflux to the rhizosphere
Inorganic anion uptake is mediated by high and low affinity root cell transporters specific for various nutrients such as nitrate, phosphate and sulfate. Anion efflux is also an important process in root peripheral cells where it occurs along their electrochemical gradient and is probably mediated by anion channels or other passive transport mechanisms . Inorganic anion efflux to the rhizosphere may be necessary to regulate root cell pH by electrically counterbalancing the efflux of protons or to regulate whole plant inorganic anion uptake under stressful conditions. In addition, anion channels are important to control the plasma membrane electrical potential, which is a key parameter for nutrient acquisition . The recent molecular identification of the first nitrate efflux transporter from root peripheral cells, NAXT1, allowed the physiological function of root nitrate efflux to be tested.
NAXT1 (nitrate excretion transporter 1) belongs to the large NRT1/PTR family  and was identified by a biochemical strategy performed on A. thaliana suspension cells . NAXT1 is targeted to the plasma membrane and mainly expressed in cortical and epidermis cells of mature roots where it is responsible for passive excretion induced by medium acidification. Root cell acidification occurs during anoxia in flooded soils. Acid loading leads to a prolonged efflux associated with a decrease of root content. These responses are abolished in the naxt1 mutants leading to the hypothesis that NAXT1 participates in excretion to counterbalance H+ excretion by H+-ATPase required to attenuate cytosol acidification [80,85]. The NAXT subfamily contains seven members whose physiological functions mostly remain to be explored. The discovery of NAXT1 suggests a diversity of transport mechanisms within the PTR/NRT family. The best characterized member of this family, NRT1.1, is a proton nitrate symporter . Although the transport mechanism used by NAXT1 has not yet been fully characterized, it transports nitrate passively along its electrochemical gradient, as a nitrate-permeable channel would.
Root peripheral cells also harbor R-type anion channels with high nitrate and sulfate permeability [87,88]. These channels, although not yet identified at the molecular level, could serve functions similar to those proposed for NAXT1: they could regulate the uptake of nitrate and sulfate or counterbalance protons [80,89].
Aluminum toxicity is a serious problem on acidic soils, which represent about 30% of arable land, worldwide. In many species, Al3+ tolerance is associated with increased excretion of organic acids, such as citrate, malate or oxalate in Al3+-tolerant cultivars [90–92]. The organic anion secretion occurs at the root tip, which is most sensitive to Al3+ stress. Application of citrate or oxalate can mitigate the toxic effect of Al3+ on root growth. The importance of organic acid excretion for Al3+ tolerance prompted several laboratories to conduct patch clamp studies on protoplasts isolated from root tips of Al3+-sensitive or Al3+-tolerant cultivars of wheat and maize [93–95]. Wheat and maize excrete mostly malate or citrate, respectively. In studies on both species, Al3+-activated anion conductances with channel properties were recorded in the plasma membrane of root tip protoplasts. The channels are permeable to organic acids with different selectivities. The Al3+-activated channels of wheat root tip cells are more permeable to malate than to chloride . In maize, they are also permeable to citrate [93,97]. The Al3+-activated currents occur more frequently and are more strongly activated in protoplasts from Al3+-tolerant cultivars of wheat or maize compared with sensitive ones. The availability of cultivars with contrasting Al3+ tolerance allowed the molecular identification of the channels/transporters responsible for the higher organic anion efflux. In the case of wheat, cDNA library subtraction between two near-isogenic lines identified TaALMT1 . TaALMT1 encodes a transmembrane protein which defines a new protein family unique to plants. Homologues of TaALMT1 have been characterized in Arabidopsis, barley and maize and are present in all sequenced plant genomes . ALMT1 is currently one of the best characterized organic anion efflux channels, although other systems have been identified that bring a major contribution to Al3+ tolerance in other species (see below).
ALMT1 was shown to be an Al3+-activated malate efflux protein able to confer Al3+ tolerance to plants (Fig. 3) . Subsequent studies performed on Xenopus oocytes  and in tobacco cells  further elucidated TaALMT1 transport mechanism, selectivity and regulation. TaALMT1 is a malate-selective channel generating low basal currents in the absence of Al3+, whereas in the presence of Al3+ currents are strongly enhanced. In tobacco cells, the permeability of TaALMT1 is about 20-fold higher for malate than for Cl− or . When expressed in Xenopus oocytes, TaALMT1 is permeable not only to malate but also to inorganic ions Cl−, or when external anion concentrations are high . These electrophysiological results were confirmed in vivo in transgenic barley, wheat and Arabidopsis where TaALMT1 expression enhances malate efflux and thus Al3+ resistance [92,98,101]. Nevertheless, not all ALMT1 homologues are able to transport malate or are implicated in Al3+ tolerance. When ZmALMT1 is expressed in Xenopus oocytes, the Al3+-activated currents are small and the selectivity for organic acids (malate, citrate) over several inorganic anions is poor . This led to the hypothesis that ZmALMT1 is rather involved in anion homeostasis and mineral nutrition.
The analysis of TaALMT1 topology revealed that the TaALMT1 polypeptide forms six transmembrane α-helices with an N-terminal domain and a long C-terminal domain both facing the extracellular side of the plasma membrane (Fig. 3) . Recent studies focusing on the mechanism of Al3+-induced activation of ALMT1 revealed a crucial importance for its C-terminal domain. Al3+ enhances the activity of most ALMT1 homologues identified in plants [98–100,104] but the exact mechanism remains unclear. Al3+ activation is observed when ALMT1 is expressed in Xenopus oocytes, suggesting that it is an intrinsic property of ALMT1 protein . Truncation of the C-terminal domain of TaALMT1 leads to a loss of basal and Al3+-activated transport activity, which is rescued by grafting the Arabidopsis ALMT1 C-terminal domain . This domain is thus essential for the function of TaALMT1 and its homologues. By mutating acidic residues in the C-terminal domain, three residues (E274, D275 and E284) were identified, which are specifically required for activation of ALMT1 transport activity by Al3+ without affecting its basal activity. This suggests that these residues participate in Al3+ binding domains (Fig. 3) .
Several lines of evidence suggest that TaALMT1 activity is regulated by phosphorylation. Application of K252a, a protein kinase inhibitor, on wheat or Arabidopsis roots reduces Al3+-activated malate efflux [106,107]. The role of phosphorylation was also studied by Ligaba et al.  on Xenopus oocytes expressing TaALMT1. In this system, the application of protein kinase antagonists inhibits basal and Al3+-activated malate efflux. Moreover, the addition of a protein kinase C (PKC) activator enhances TaALMT1 mediated currents. In an attempt to identify phosphorylated residues, the authors mutated six putative PKC phosphorylation sites located in the C-terminal domain of TaALMT1. Mutation of threonine 323 to alanine results in a significant increase of TaALMT1 activity; in contrast, substitution of serine 384 to alanine greatly reduces TaALMT1 activity . This could indicate that serine 384 needs to be phosphorylated before Al3+ can activate AtALMT1 (Fig. 3). However, no direct evidence that this site is phosphorylated in vivo exists. Moreover, the phosphorylation sites were identified on the C-terminal domain localized outside of the cell and extracellular protein kinases have not yet been identified in plants. TaALMT1 topology is now subject to controversy [12,92,103].
Whereas TaALMT1 locus is clearly important for Al3+ tolerance in wheat, quantitative trait loci accounting for Al3+ tolerance in sorghum, barley and maize identified genes coding organic anion transporters of the multidrug and toxic compound exudation (MATE) family [109–111]. In contrast to ALMT1 transporters, plant MATE involved in Al3+ tolerance show a substrate preference for citrate rather than malate. Although canonic MATE involved in toxic compound efflux were shown to function as a proton coupled anion efflux pump, expression of ZmMATE1 from maize in Xenopus oocytes triggers inward currents that are compatible with organic anion efflux current through channels [109,111]. More detailed characterization is required to determine whether these MATE function as proton coupled transporters, facilitators or constitute a new class of anion channels in plants.
In addition to Al3+ tolerance, organic anion efflux to the rhizosphere has also been implicated in phosphate nutrition . Organic anion efflux is especially relevant for plant families that do not form mycorrhiza and grow on soils in which phosphate is poorly available. Organic acid secretion increases the availability of phosphate tightly bound to soil particles through ligand exchange. Specialized roots involved in organic acid efflux were first described in Proteaceae. However, the preferred experimental system to analyze this process has been the cluster roots of white lupin . Both the development of cluster roots and their organic acid excretion are enhanced under phosphate deficiency. Patch clamp analysis of protoplasts from white lupin cluster roots revealed the presence of citrate permeable channels . However, these channels are not restricted to cells from cluster roots. Another patch clamp study performed on A. thaliana epidermal cells identified citrate permeable channels that are found only in protoplasts isolated from phosphate starved roots . This suggests that all root epidermal cells may excrete citrate through citrate permeable channels and cluster roots may merely represent a way to increase the root–soil interface at which efflux occurs. The molecular identity of the anion channels mediating organic efflux under phosphate starvation is not known yet. It will be interesting to mine microarray or proteomic data from phosphate starved roots for homologues of ALMT or MATE, which are good candidates for this function.
Anion efflux to the xylem
Mineral nutrients are mostly translocated from roots to shoots through the xylem. Translocation thus requires the loading of mineral ions into the xylem. Given their electrochemical gradient, many ions may be excreted to the xylem sap through channel-mediated mechanisms. In the case of potassium, a combination of electrophysiological analyses and Arabidopsis molecular genetics has demonstrated that SKOR, an outward rectifying potassium channel of the Shaker family, participates in potassium loading into the xylem . The electrochemical gradient for anions, such as nitrate, is even more favorable to their loading to the xylem through anion channels.
Several electrophysiological studies relying on the ability to isolate and recognize protoplasts from stele cells have identified anion channels in this cell type in maize and barley [20,115–117]. These channels are highly permeable to nitrate and are good candidates to load this nutrient into the xylem. In addition, a proton-coupled nitrate transporter of the PTR family, NRT1.5, was shown to participate in the loading of nitrate to the xylem . Nitrate transfer to the shoots is reduced, but not completely abolished, in an nrt1-5 mutant. NRT1.5-mediated and anion-channel-mediated nitrate loading systems may thus coexist. Interestingly, one class of anion channel, similar to R-type currents, present in stele cells is downregulated by the phytohormone ABA . This suggests that under drought conditions, when the xylem flux is slowed down, coordinated downregulation of anion loading to the xylem occurs. The anion channels described by the patch clamp technique in stele cells still await molecular identification. The discovery of the genes encoding these channels will allow their importance for anion loading in the xylem to be tested.
The major anions in the xylem sap are nitrate, chloride, sulfate and phosphate. Citrate and malate, however, are present at sub-millimolar concentrations (0.1–0.5 mm). Based on physicochemical considerations and the low pH of the xylem sap, it was proposed that some metal cations, such as iron or aluminum, travel through the xylem as complexes with citrate.
Mutations in FRD3 (ferric reductase deficient 3) were recovered in screens for mutants that constitutively activate iron-deficiency responses . FRD3 encodes a MATE transporter showing high homology to SbMATE, AtMATE, HvAACT1 and ZmMATE that are involved in citrate excretion in sorghum, Arabidopsis, barley and maize, respectively . The characterization of FRD3 both identified a pathway for citrate excretion from the pericycle cells to the xylem and supported the importance of iron–citrate complexes for iron translocation from the roots to the shoots  (see note added in proof). The closest homologue of AtFRD3 in rice, OsFRDL1, is also involved in iron translocation from roots to shoots [79,122]. AtFRD3 and OsFRDL1 are expressed in root pericycle cells. When expressed in Xenopus oocytes, both AtFRD3 and OsFRDL1 facilitate citrate efflux and both frdl1 and frd3 have decreased xylem sap citrate concentrations. Interestingly, when ectopically overexpressed in Arabidopsis, FRD3 confers increased Al3+ tolerance . These results indicate that similar transport systems, the MATEs, are involved in citrate excretion to the rhizosphere or to the xylem sap. Depending on their expression pattern, in epidermal or in pericycle cells, the citrate transporting MATE are involved in Al3+ tolerance or iron translocation, respectively. Although both processes involve citrate fluxes along its electrochemical gradient and the fluxes are further favored by the pH gradient between the cytosol and the extracellular medium, it remains to be determined whether the citrate transporting MATE function as anion channels.
The knowledge in the field of plant anion channels has grown very quickly in the last few years, with the identification of the genes underlying the rapid (R-type) and the slow (S-type) anion channels of guard cells, the analysis of their regulation and the determination of the 3D structure of the slow anion channel, SLAC1. It will be interesting to determine whether some plant anion channels are encoded by other gene families. So far, the best characterized member of the plant CLC family, AtCLCa, works as a proton-coupled nitrate transporter . It is possible, however, that other plant CLCs function as anion channels . In addition, this review describes membrane proteins of the MATE family and of the PTR family (NAXT1) that mediate passive anion fluxes and could also function as anion channels [80,109,111].
The guard cell has proved to be a very efficient model to identify plant ion channels . Nevertheless, using electrophysiological techniques, anion channel activity can be recorded from virtually any cell type investigated. These channel activities most often resemble S-type or R-type guard cell anion currents. The identification of two anion channel gene families, the SLAC1 family and the ALMT1 family, should open the way to the identification of the specific proteins or protein combinations underlying most anion channel activities encountered in plant cells. Mutations in the corresponding genes will allow a more complete investigation of the physiological role of anion channels in plant cells. Notably, it has been proposed that anion channels participate in the control of osmotic pressure and growth, in the massive anion efflux triggered in response to pathogen attack or in the generation of long-distance electrical signals [123,124]. It will soon be possible to test these hypotheses using powerful genetic tools.
ALMT1, CLC and SLAC1 homologues are present in genomes of all sequenced species . With the development of next generation high throughput sequencing, homologous anion channel genes will be identified in many more species. In particular, it should allow the study of anion channels and their regulation in plants where electrical signaling and fast movements have been well characterized, such as Mimosa pudica and carnivorous plants.
Finally, anion channels are clearly at the crossroad of signaling, metabolism, nutrition and turgor regulation. One of the future prospects will be to integrate their function in the many networks in which they participate. How do guard cell anion channels integrate with the activity of other ion channels in this cell type to achieve adequate turgor regulation under fluctuating conditions? How do nitrate-permeable channels integrate with nitrate transporters and nitrate assimilation pathways to control nitrogen homeostasis? How do organic acid permeable channels integrate with organic acid transporters, the primary carbon metabolism and the control of cellular pH? Answering these questions will require a detailed knowledge of anion channel genes and proteins and their regulation and the building of quantitative models integrating transmembrane and metabolic fluxes together with cellular ion concentrations.
The work of M.J. and S.T. was supported by the Centre National de la Recherche Scientifique (CNRS) and the Agence Nationale pour la Recherche (ANR-Nitrapool: grant number ANR-08-BLAN-0008-02). The work of H.K. and K.L. was supported by ESF grant 7763, targeted funding theme SF0180071S07 and by European Regional Fund (the Center of Excellence in Environmental Adaptation).
Note added in proof
After this review was accepted, two important studies were published. Meyer et al. characterized a malate permeable channel belonging to the ALMT1 family in the vacuolar membrane of guard cells. Roschzttardtz et al. characterized the function of FRD3 malate efflux transporter throughout plant development. [Meyer S, Scholz-Starke J, De Angeli A, Kovermann P, Burla B, Gambale F & Martinoia E (2011) Malate transport by the vacuolar AtALMT6 channel in guard cells is subject to multiple regulation. Plant J67, 247–257. Roschzttardtz H, Séguéla-Arnaud M, Briat JF, Vert G & Curie C (2011) The FRD3 citrate effluxer promotes iron nutrition between symplastically disconnected tissues throughout Arabidopsis development. Plant Cell23, 2725–2737.]