Mechanisms of Cl- transport contributing to salt tolerance



    Corresponding author
    1. Centre for Ecohydrology and Future Farm Industries CRC, School of Plant Biology, The University of Western Australia, Crawley, Western Australia 6009, Australia and
    Search for more papers by this author

    1. School of Agriculture, Food and Wine, University of Adelaide, Adelaide, South Australia 5064, Australia
    Search for more papers by this author

N. L. Teakle. Fax: +61 8 6488 1108; e-mail:


Mechanisms of Cl- transport in plants are poorly understood, despite the importance of minimizing Cl- toxicity for salt tolerance. This review summarizes Cl- transport processes in plants that contribute to genotypic differences in salt tolerance, identifying key traits from the cellular to whole-plant level. Key aspects of Cl- transport that contribute to salt tolerance in some species include reduced net xylem loading, intracellular compartmentation and greater efflux of Cl- from roots. We also provide an update on the biophysics of anion transport in plant cells and address issues of charge balance, selectivity and energy expenditure relevant to Cl- transport mechanisms. Examples are given of anion transport systems where electrophysiology has revealed possible interactions with salinity. Finally, candidate genes for anion transporters are identified that may be contributing to Cl- movement within plants during salinity. This review integrates current knowledge of Cl- transport mechanisms to identify future pathways for improving salt tolerance.


Over 800 million ha worldwide is predicted to be salt affected (FAO 2008), representing a significant loss to agricultural production. For most plants to tolerate salinity, Na+ and Cl- uptake must be restricted while maintaining uptake of macronutrients such as K+, NO3- and Ca2+. There has been extensive research and reviews into the mechanisms of Na+ transport in plants under salt stress (Amtmann & Sanders 1999; Blumwald, Aharon & Apse 2000; Tester & Davenport 2003; Horie & Schroeder 2004; Apse & Blumwald 2007) and K+ (Maathuis & Amtmann 1999; Tester & Leigh 2001; Maser, Gierth & Schroeder 2002; Rodriguez-Navarro & Rubio 2006). The mechanisms of Cl- transport are less well understood than that of cation transport in general, both for plant nutrition and under high NaCl concentrations, despite its predominance as the anion in most saline soils. Cl- is considered an essential micronutrient that regulates enzyme activities in the cytoplasm, is an essential co-factor in photosynthesis, acts as a counter anion to stabilize membrane potential and is involved in turgor and pH regulation (Tyerman 1992; Marschner 1995; Teodoro, Zingarelli & Lado 1998; Xu et al. 2000; White & Broadley 2001). However, Cl- can be toxic to plants at high concentrations, with critical toxicity estimated to be 4–7 mg g−1 for Cl- sensitive species and 15–50 mg g−1 dry weight (DW) for Cl- tolerant species (Xu et al. 2000). Both Na+ and Cl- are toxic to plants at high concentrations, but some species can control Na+ transport better than Cl- and vice versa (Munns & Tester 2008). Control of Cl- transport and Cl-‘exclusion’ from shoots is correlated with salt tolerance in many species, particularly for legumes, such as Trifolium (Winter 1982; Rogers, Noble & Pederick 1997a), Medicago (Sibole et al. 2003), Glycine (Luo, Bingjun & Liu 2005) and Lotus (Teakle, Real & Colmer 2006; Teakle et al. 2007); as well as for woody species, for example, Pinus banksiana (Franklin & Zwiazek 2004), Citrus and Vitis (Sykes 1992; Romero-Aranda et al. 1998; Moya et al. 2003).

Recent information from functional characterization of anion transporters reveals some surprises in terms of selectivity and energy coupling that is relevant to salinity. Selectivity between Na+ and K+ to maintain favourable K+/Na+ ratios has been a focus of salt tolerance research. This selectivity has been related to the selectivity of specific transport systems (Schachtman, Tyerman & Terry 1991; Gassmann, Rubio & Schroeder 1996; Hua et al. 2003; Volkov & Amtmann 2006), or through combinations of transporters at different membranes and in different cells along the pathway from roots to shoots (Sunarpi et al. 2005; Apse & Blumwald 2007; Byrt et al. 2007). A similar situation may exist for Cl- with respect to major macronutrient anions (NO3-, SO42−, Pi), and organic anions, yet exclusion of Cl- related to selectivity of nutrient anion transporters and their location in the pathway from roots to shoots has not really been addressed. There are well-documented interactions between macronutrient anions and salinity that may be linked to transporter selectivity (Grattan & Grieve 1999). In particular, it is important to consider NO3- in our focus on Cl- and salinity tolerance because it is the more common univalent anion in plants under non-saline conditions (e.g. Fricke, Leigh & Tomos 1994; Frachisse et al. 1999) and many anion channels are more selective to NO3- than to Cl- (Roberts 2006). Cl- concentration in leaves can be reduced by the presence of NO3- in the growth medium (Abdolzadeh et al. 2008; Gimeno et al. 2009; Song et al. 2009). NO3-/Cl- interactions are also analogous to the focus on K+/Na+ interactions and selectivity when considering Na+ exclusion mechanisms.

Charge balance is an issue often not dealt with adequately, linked to selectivity and relevant to the relative roles of Cl- and Na+ in salt tolerance. The net movement of either ion must be balanced with the flux of other ions so that in each compartment there is charge equivalence to within a very small difference. From a Cl- perspective, uptake of Cl- could be balanced by the uptake of another cation, for example, Na+, or by the loss of another anion. Those anions that could be lost from a cell may not be readily expendable because they are linked to nutrition (e.g. NO3-) or carbon balance (e.g. malate). From a Na+ perspective, apart from the reciprocal of the above with Cl-, K+ may be lost from the cell (Shabala et al. 2006), which may be acceptable if Na+ replaces the osmotic role of K+ in the vacuole and K+ concentration is maintained within acceptable limits in the cytoplasm (Carden et al. 2003). What ultimately determines charge balance is the changes in voltage across membranes in the pathway because of minute differences in net charge movement, which then influences the driving force on counter ions, effectively as a self-regulating system. However, it is important that voltage excursions are not too extreme and the voltage remains negative (cytoplasm with respect to outside), otherwise other transport processes may be disrupted and membrane integrity may be compromised. The nature of the charge balance (cation or anion and the direction) depends on the types and selectivity of the transporters that are present at key membrane barriers and their response to changes in membrane voltage. While our review will focus on mechanisms of Cl- transport, the importance of charge balance previously described implies that anions and cations should of course not be studied in isolation. This is also important in considering energy requirements for transport (Britto & Kronzucker 2009). For more detail on Na+ transport mechanisms, we refer the reader to Plett & Møller (2009, this volume).

Recent reviews have covered aspects of anion transport (e.g. Roberts 2006; De Angeli et al. 2009), but not since 2001 (White & Broadley 2001) have the mechanisms of Cl- transport in relation to salt tolerance of plants been covered in any detail. The aim of this review is to ‘update’ current knowledge on Cl- transport under salt stress, identify mechanisms of Cl- regulation that contribute to salt tolerance in different plant species and finally to propose future directions for research to advance our current (lack of!) understanding of Cl- transport mechanisms. While our review will focus on aspects of Cl- transport associated with salt tolerance, it is important to note that regulating Cl- uptake and transport within the plant is just one aspect contributing to the complicated nature of salt tolerance. Several different traits are likely to be contributing to salt tolerance depending on the species, such as tolerance to osmotic stress, accumulation of compatible solutes and oxidative stress. For further information on other traits associated with salt tolerance that are beyond the scope of this review, we refer the reader to recent reviews on mechanisms of salt tolerance (Bartels & Sunkar 2005; Munns 2005; Flowers & Colmer 2008; Munns & Tester 2008).


Considering that NaCl is generally the dominant salt in saline soils, and that clearly NaCl = Na+and Cl-, it is quite surprising that most research on salt tolerance (except in species such as citrus and grapevines) focus on Na+ and not both Na+ and Cl-. Recent reviews have had a much stronger emphasis on Na+ rather than Cl-, reflecting the general research focus within the body of literature (Flowers & Colmer 2008; Munns & Tester 2008). The focus on Na+ in reviews on salt tolerance is a reflection of the significant number of experimental papers that only measure Na+ (and K+) but not Cl- in salt-stressed tissues. However, both Na+ and Cl- are metabolically toxic to plants if accumulated at high concentrations in the cytoplasm.

For some species Na+, and not Cl-, shoot or root concentrations are negatively correlated with salt tolerance (e.g. rice, Lin & Kao 2001; wheat, Kinraide 1999; Husain, von Caemmerer & Munns 2004; Plett & Møller 2009 this volume), while the reverse is true for other species. For example in soybean, leaf Cl- was negatively correlated with salt tolerance (R2 = 0.9), but there was no significant correlation for leaf Na+ (Luo et al. 2005). Other examples of species for which control of Cl- transport and Cl-‘exclusion’1 from shoots is correlated with salt tolerance are shown in Fig. 1. For several species, the ‘exclusion’ of both Na+ and Cl- is important for salt tolerance, for example, in Hordeum marinum (Islam et al. 2007) and Medicago truncatula (Aydi, Sassi & Abdelly 2008). We argue that salt tolerance is related to the ability of individual genotypes to regulate Cl- and Na+ transport to avoid toxicity of both ions, as either ion can be toxic depending on the ability of plants to minimize concentrations in the cytoplasm.

Figure 1.

Low shoot Cl- is correlated with salt tolerance for some species. Salt tolerance is measured by shoot DW under saline conditions relative to control, except for Glycine, which was estimated from % of seedlings not injured by salinity. NaCl treatments and correlation coefficients are: Glycine, 150 mm, R2 = 0.90 (Luo et al. 2005); Lotus, 200 mm, R2 = 0.89 (Teakle et al. 2006); Hordeum/Triticum, 300 mm, R2 = 0.99 (Islam et al. 2007); Trifolium, 40 mm, R2 = 0.95 (Rogers et al. 1997b); Eucalyptus, 100 mm, R2 = 0.92 (Marcar 1993).

Some genotypes are more effective at regulating Na+ or Cl- transport (or both); the differences probably reflecting different mechanisms of charge balance within the plant. For example, experiments using iso-osmotic solutions of 150 mm of Cl- (plus K+, Mg2+, Ca2+, NH4+) or 150 mm Na+ (plus NO3-, SO42−, H2PO4-), compared to 150 mm NaCl, found that genotypes of Glycine max were more sensitive to Cl-, while G. soja genotypes were more sensitive to Na+ (Luo et al. 2005). As described in the introduction, charge balance is important for understanding Cl- and Na+ transport processes under salt stress. Therefore, it is not surprising that Cl- and Na+ uptake are often linked. For example, in separate studies of salt tolerance in 40 species of Trifolium (Rogers et al. 2009) and 40 genotypes of Lotus tenuis (N.L. Teakle et al. unpublished data) shoot Na+ and shoot Cl- showed a strong positive correlation (R2 of 0.82 and 0.84, respectively; Fig. 2). However, the inheritance of Na+ exclusion was found to be independent of Cl- exclusion in citrus (Sykes 1992).

Figure 2.

The relationship between shoot Cl- and shoot Na+ for Trifolium species and genotypes of Lotus tenuis. (a) Forty Trifolium species were grown at 160 mm NaCl for 27 d (Rogers et al. 2009). (b) Forty genotypes of Lotus tenuis were grown at 300 mm NaCl for 35 d (N.L. Teakle et al. unpublished data).

In summary, it is clear that both Na+ and Cl- transport processes are involved in plant responses to salinity. Advances have been made by focussing on one of the pair (mostly Na+), when exclusion of that ion is correlated with salt tolerance. However, it is likely that further advances and potential improvement of salt tolerance will be made when Na+ and Cl- transport are studied in parallel.


This section describes processes of Cl- uptake, transport and accumulation, from the cellular to the whole plant level, which are associated with salt tolerance. It is likely that in salt tolerant species, several processes will be operating, and the key trait responsible for lower Cl- accumulation in shoots may change depending on the external NaCl concentration. For example, at low to moderate NaCl (∼85 to 170 mm), tolerant Glycine genotypes had lower leaf Cl- than sensitive genotypes (Pantalone et al. 1997); yet at higher NaCl (∼250 mm), the tolerant genotypes had up to twofold higher leaf Cl-. This indicates that at low NaCl, salt tolerance was correlated with Cl-‘exclusion’ from shoots, but at higher NaCl, tissue tolerance (presumably via accumulation of Cl- in vacuoles) was important for tolerance, in addition to exclusion. For each trait discussed next, examples are given from genotypes that differ in salt tolerance.

Reduced net Cl- uptake by roots

Controlling the initial uptake of Cl- (and Na+) at the root-soil interface would limit the total amount of these ions entering the plant and subsequently transported to shoots. For halophytes, turgor can be generated by accumulating high Cl- concentrations (340 to 375 mm; White & Broadley 2001), yet for most species to tolerate high salt concentrations, Cl- uptake by roots must be restricted. In contrast to Na+, there have been few studies on the unidirectional uptake of Cl- at high NaCl concentrations. Interestingly, a Cl- excluding grapevine rootstock (Paulsen) actually had a higher 36Cl- influx rate than K51-40, a Cl- accumulating rootstock (Abbaspour 2008). Cl- influx under steady state conditions does not always reflect eventual total Cl- accumulation by roots because ‘net’ uptake of Cl- depends on both influx and efflux by roots. NaCl-induced efflux of Cl- has been observed in several species, for example, in the halophyte Diplachne fusca (Bhatti & Wieneke 1984); barley (Yamashita & Matsumoto 1996; Britto et al. 2004); sorghum (Boursier & Lauchli 1989) and Arabidopsis (Lorenzen, Aberle & Plieth 2004). Genotypic differences in Cl- efflux have also been observed. In a recent study using scanning ion-selective electrodes, roots of salt tolerant Populus euphratica treated with 100 mm NaCl for 15 d exhibited a significant net efflux of Cl- at 400 to 1200 µm from the apex (Sun, Chen & Dai 2009). Under the same conditions, no efflux of Cl- was observed in the salt-sensitive Populus popularisand responses were NaCl-specific (and not a result of hyper-osmotic stress). The conclusion from their study was that Cl- (and Na+) extrusion from P. euphratica roots contributed to ion homoeostasis under saline conditions (Sun et al. 2009). While it is difficult to draw conclusions with such limited data available, these results indicate that greater efflux of Cl- from roots may contribute significantly to genotypic differences in reduced net Cl- uptake by roots, and warrants further investigation

Reduced net xylem loading of Cl-

Despite the importance of xylem ion transport for regulating shoot ion concentrations under salt stress, and plant nutrition in general, accurate measurements of ions (including Cl-) in the xylem have rarely been made. This is most likely because collecting xylem sap from intact transpiring plants is difficult because of the large negative pressures in xylem vessels and tracheids (Steudle 2000; de Boer & Volkov 2003). Destructive methods can yield useful comparative data within an experiment, but the ion concentrations measured may not accurately reflect those of the transpiration stream of intact plants (Watson, Pritchard & Malone 2001; Malone, Herron & Morales 2002). For example, Cl- fluxes to the xylem for intact transpiring plants are predicted to be over an order of magnitude lower than those measured using applied pressure (Pitman 1982), raising doubts about the equivalence of data from intact plants versus excised roots. Xylem-feeding insects, such as meadow spittlebugs (Philaenus spumarius) native to England, have been successfully used to collect xylem sap (insect excreta) from intact transpiring plants and the concentrations of ions measured using ion chromatography (Watson et al. 2001; Gong, Randall & Flowers 2006; Hall et al. 2006). This technique has been used to measure differences in Cl-‘exclusion’ from the xylem in two Lotus species that differ in salt tolerance (Teakle et al. 2007). However, the use of the xylem-feeding spittlebugs is limited by host specificity and seasonal availability of the insect.

Because of these difficulties in accurately collecting xylem sap, only a few genotypic comparisons of Cl- concentrations in the xylem are available. Minimizing xylem loading of Cl- to the shoot is critical to minimize shoot concentrations; and is a likely factor contributing to differences in shoot Cl- and hence salt tolerance in some species (Fig. 1). In one of the few studies to measure xylem Cl- from intact, transpiring plants under saline conditions, it was found that the more salt-sensitive species Lotus corniculatus had twice the xylem Cl- of the relatively more tolerant Lotus tenuis, which corresponded to the measured differences in shoot Cl- (Teakle et al. 2007). Similarly, a more salt-tolerant line of durum wheat had only 8 mm Cl- in the xylem, compared with the more salt-sensitive genotype with 21 mm after 10 d in 50 mm NaCl (measured using X-ray microanalysis; Lauchli et al. 2008). Such concentrations even for the salt-tolerant line, if loaded exclusively into the relatively small volume of the leaf apoplast, would build to osmotically prohibitive concentrations. Thus, uptake into the symplast of the leaf, and partitioning between and within leaf cells must occur (see below). Based on calculated xylem concentrations, reduced xylem loading of Cl- also contributed to salt tolerance in citrus seedlings (Moya et al. 2003). In contrast, Cl- in the xylem was estimated at 4.7 mm for barley growing at 50 mm NaCl (91% of Cl- excluded), which was higher than estimates for the more salt-sensitive bread and durum wheat species at 97 to 98% Cl-‘exclusion’ (see table 2 in Munns 2005). This suggests that factors other than reduced xylem loading of Cl- contribute to salt tolerance in barley. Anion channels in the xylem and possible candidate genes for xylem transport of Cl- are discussed below.

Radial flow of ions to the xylem may also occur through the apoplasmic path, also referred to as by-pass flow. Differences in salinity tolerance between genotypes can be accounted for by differences in this component of radial flow (Yadav, Flowers & Yeo 1996). Root pressure probe measurements show that the reflection coefficient for the root cylinder for salts is less than unity, compared with cortical cells with values not different from unity and indicating that ion leakage via the apoplast may occur (Azaizeh, Gunse & Steudle 1992). Suberin lamellae are considered to be primary barriers to by-pass flow in the apoplasm. Arabidopsis-enhanced suberin 1 (esb1) mutant accumulates more suberin in the root (Baxter et al. 2009). This corresponds to low transpiration and decreased accumulation in the shoot of Ca, Mn and Zn, but counterintuitively increased accumulation of other elements including Na (anions not measured). It would be interesting to examine these mutants under salinity to determine if Cl- is also increased.

Intercellular compartmentation of Cl-

Partitioning of Cl- between different cell types in both roots and shoots contributes to salt tolerance in some species. Within leaves, there is some evidence that Cl- is preferentially accumulated in the epidermis, reducing Cl- toxicity in mesophyll cells that are more important for photosynthesis. A salt-tolerant cultivar of barley was found to be more effective in excluding Cl- from mesophyll cells, compared with a more sensitive barley cultivar (Huang & Van Steveninck 1989). However, a more quantitative analysis using single-cell-sampling techniques for barley found that both epidermal and mesophyll Cl- increased with increasing NaCl, up to 563 and 167 mm, respectively, at 150 mm NaCl (Fricke, Leigh & Tomos 1996). Their study also found that net rates of photosynthesis at 150 mm NaCl were only mildly affected (83% of control), thus raising doubts that accumulation of Cl- in epidermal cells is associated with salt tolerance. Interestingly, in their study, epidermal Cl- was always at least threefold higher than mesophyll Cl-, yet concentrations of Na+ and K+ did not differ significantly between the two cell types (Fricke et al. 1996). This apparent charge imbalance could be overcome by the accumulation of organic anions in mesophyll cells, although this was not measured. In another study comparing a salt-tolerant barley cultivar with a salt-sensitive durum wheat, Cl- was preferentially accumulated in the epidermis compared with the mesophyll, but to a similar extent in both species (James, Davenport & Munns 2006); thus, allocation of Cl- to particular cell types might not be a large factor contributing to salt tolerance. Further work with other genotypes that have similar shoot Cl- concentrations, but differs in salt tolerance, could help determine if partitioning of Cl- in epidermal cells, away from the mesophyll, contributes to salt tolerance.

Another form of intercellular compartmentation of Cl- in leaves is the accumulation of Cl- in salt glands or bladders. These highly specialized cell structures are unique to some halophytes, which can accumulate Cl- (and Na+) in salt glands on leaf surfaces to lower internal leaf ion concentrations. Cl- secretion via salt glands can be significant, with approximately 20% of leaf Cl- excreted from salt glands of Leptochloa fusca at 100 mm NaCl (Jeschke et al. 1995). It is beyond the scope of this review to cover Cl- transport in salt glands, as this adaptation is limited to only some halophytes. We refer the reader to recent papers studying salt glands in the halophytes Bienertia sinuspersici (Park, Okita & Edwards 2009) and Limonium sinense (Ding et al. 2009). Interestingly, a cation-chloride cotransporter (CCC) was localized to leaf trichomes and hydathodes in Arabidopsis (Colmenero-Flores et al. 2007). Although the exact role of these transporters in Cl- transport is still unknown (see discussion next), it would be interesting to investigate if CCCs are also localized to salt glands of halophytes for possible efflux of Cl- (and Na+) from leaves.

At the root level, Cl- transport across different cell types from the cortex to the xylem could affect the total flux of Cl- to the shoot. For example, Storey, Schachtman & Thomas (2003) found that vacuolar Cl- was about 50% lower in the endodermis compared with the inner cortex and pericycle for grapevine roots treated with 25 mm NaCl. They also found that the salt-tolerant genotype (low shoot Cl-) had 20% higher vacuolar Cl- in pericycle cells, compared with the salt-sensitive genotype (high shoot Cl-), suggesting that the pericycle could play a role in restricting Cl- transport to the shoot. The key data missing from their study were a quantitative assessment of xylem parenchyma and xylem vessel ion concentrations. For lucerne, salt tolerance has also been associated with retention of Cl- (and Na+) in the vacuoles of the epidermis and outer cortex (Anderson & Van Steveninck 1987), presumably to minimize transport to the xylem for loading to the shoot. By contrast, Cl- concentrations across the root profile (epidermis → cortex → xylem) were the same for two durum wheat genotypes that differ in salt tolerance when grown at 50 mm NaCl (Lauchli et al. 2008). Surprisingly, the Cl- concentrations measured in their study were very low (less than 10 mm) in all root cells except the epidermis (30 mm), and were far less than the values measured for Na+. These types of studies are useful for understanding root ion accumulation in specific cell types, but caution is needed in interpreting the results as it does not always account for the actual fluxes of the ions involved.

Intracellular compartmentation of Cl-

Low shoot concentrations of Cl- and Na+ are not always correlated with salt tolerance. Within a genus, and even within a species, there are genotypes that can tolerate high Cl- concentrations, genotypes that have greater ‘exclusion’ of Cl- and likely genotypes that do both; thus, there is often not always a correlation between shoot Cl- and salt tolerance (Fig. 3). Most species can ‘exclude’ Cl- and Na+ up to a point (e.g. 90 to 98%; c.f. table 2 in Munns 2005), but ultimately salt tolerance is going to be improved by efficient sequestration of Cl- and Na+ in vacuoles to prevent accumulation in the cytoplasm to toxic levels. Halophytes have controlled uptake of Cl- (and other ions) to support turgor-driven growth; this relies on the effective sequestration of Cl- into vacuoles as even halophytes cannot tolerate high cytoplasmic Cl- (Flowers, Troke & Yeo 1977; Glenn, Brown & Blumwald 1999).

Figure 3.

The relationship between shoot Cl- and salt tolerance for Trifolium species and genotypes of Lotus tenuis. (a) Forty Trifolium species were grown at 160 mm NaCl for 27 d (Rogers et al. 2009). (b) Forty genotypes of Lotus tenuis were grown at 300 mm NaCl for 35 d (N.L. Teakle et al. unpublished data). Correlations were not significant.

Unfortunately, it is technically difficult to obtain direct experimental measurements of Cl- fluxes and concentrations in vacuoles of intact plants, although some estimates have been made using X-ray microanalysis (Hajibagheri & Flowers 1989), intracellular ion-sensitive microelectrodes (Felle 1994), tracer compartmental analysis (Britto et al. 2004) or Cl- sensitive fluorescent probes (Lorenzen et al. 2004). It has been estimated that plant cell vacuoles can accumulate up to 500 mm Cl- (Cram 1973). In the halophyte Suaeda maritima growing at 200 mm NaCl, Cl- concentrations ranged from 86 to 95 mm (cytoplasm), 430 to 465 mm (vacuole) and 111 to 130 mm (cell wall), with similar values found for Na+ (Hajibagheri & Flowers 1989). Using the fluorescent dye lucigenin, initial Cl- transport into tonoplast vesicles from the salt-tolerant ice plant (Mesembryanthemum crystallinum) followed saturation kinetics, with a Km of about 17 mm (Wissing & Smith 2000). This Km is significantly higher than those found for less salt-tolerant species, for example, red beet (6.5 mm, Pope & Leigh 1988); corn (4.3 mm, Bennett & Spanswick 1983); and oats (2.3–2.6 mm, Kaestner & Sze 1987; Pope & Leigh 1987). These results suggest that vacuolar Cl- transport in the halophytic ice plant saturates at higher Cl- concentrations than for less salt-tolerant species (Wissing & Smith 2000). For another halophyte, Atriplex gmelini, Cl- concentrations of isolated vacuoles were 260 mm, which was almost the same as that measured in protoplasts, suggesting that for this halophyte, most of the leaf Cl- (and Na+) is sequestered into vacuoles (Matoh, Watanabe & Takahashi 1987).

For some species, there is indirect evidence that more efficient vacuolar sequestration of Cl- is associated with salt tolerance. For example, salt-tolerant genotypes of citrus, grapevine and Lotus with low shoot Cl-, actually have higher root Cl- concentrations compared with the more sensitive genotypes (Storey & Walker 1999; Storey et al. 2003; N.L. Teakle et al. unpublished data), suggesting more efficient compartmentation of Cl- in root vacuoles for the tolerant genotypes. Higher leaf Cl- concentrations are also associated with salt tolerance in some avocado rootstocks (Xu et al. 2000) and lupin cultivars (Van Steveninck et al. 1982), which may indicate efficient sequestration of Cl- in leaf vacuoles. There have also been some more direct genotypic comparisons demonstrating that efficient intracellular compartmentation of Cl- is associated with salt tolerance. In a comparison between two maize genotypes that differ in salt tolerance, the more salt-sensitive genotype had consistently higher root cytoplasmic Cl- concentrations than the tolerant genotype (based on 36Cl flux analysis and electron microscopy; Hajibagheri et al. 1989). The estimates of cytoplasmic Cl- concentrations were surprisingly high, with approximately 563 mm in the sensitive genotype and 360 mm for the tolerant genotype (measured at 100 mm, Hajibagheri et al. 1989). These values are consistent with other published estimates, as cytoplasmic Cl- was also predicted to be about 350 mm for barley at 100 mm NaCl (Britto et al. 2004). In another study comparing two barley cultivars that differ in salt tolerance, X-ray microanalysis revealed that Cl- (and Na+) in the cytoplasm and vacuole were similar between the two genotypes; however, the tolerant cultivar had about half as much Cl- (and Na+) in the cell wall compared with the sensitive cultivar (Flowers & Hajibagheri 2001). High ion concentrations in the cell wall may result in reduced turgor, leading to reduced shoot growth in the more sensitive genotype. Further studies that directly measure intracellular Cl- for genotypes that differ in salt tolerance, but have the same leaf or root Cl- concentrations, are needed to confirm vacuolar sequestration is directly responsible for differences in salt tolerance.

Phloem re-circulation and translocation within the plant

The ability to effectively transport Cl- away from sensitive parts of the plant (e.g. expanding leaves) could be an important factor contributing to salt tolerance in some species. For halophytes, similar Cl- concentrations are measured in older versus younger leaves, whereas restricting Cl- accumulation in younger leaves is often associated with salt tolerance for glycophytes (Greenway & Munns 1980; White & Broadley 2001). In a study comparing two tree medics that differ in salt tolerance at 50–200 mm NaCl, the more sensitive Medicago arborea had significantly higher Cl- concentrations in younger leaves (leaf 2) than the more tolerant Medicago citrina, despite no difference between the genotypes in older expanded leaves (leaf 8) (Sibole et al. 2003). Increased partitioning of Cl- in sheaths away from leaf blades has been found to contribute to salt tolerance in some grasses. For example, in the moderately salt-sensitive maize, Cl- partitioning in the sheaf only occurred at up to 50 mm NaCl, whereas the moderately salt-tolerant sorghum partitioned twice as much Cl- in the sheaths versus blades up to 130 mm NaCl (Boursier et al. 1987). Possibly, this partitioning keeps potentially toxic levels of Cl- away from the primary sites of photosynthesis in the leaf, and is proposed to be important to salt tolerance of sorghum (Boursier et al. 1987). Partitioning of Cl- in the sheath versus laminae was also found for the salt-tolerant grass Leptochloa fusca (Klagges et al. 1993) and for barley (Greenway 1962). For other species, such as grapevines, petioles can also accumulate high Cl- concentrations to protect the laminae (Downton 1977). These results suggest that efficient transport of Cl-within the plant, away from sensitive tissues, contributes to salt tolerance in some species.

There is some debate around the role of phloem transport of Cl- and Na+ in salt tolerance. Often, there is little re-translocation of Cl- in the phloem, particularly in salt-tolerant species, supposedly to ensure that Cl- is not exported to growing tissues of the shoot (Munns 2002). In barley, Cl- flux in the phloem was calculated to be only 10% of xylem flux, suggesting that the phloem does not play a key role in regulating shoot Cl- concentrations (Munns, Fisher & Tonnet 1986). For salt-sensitive species, phloem concentrations of Cl- can increase to quite high levels under salt stress, for example, 47 mm in lupins (Jeschke, Pate & Atkins 1986) and 32 mm in maize (Lohaus et al. 2000). In castor bean treated with 128 mm NaCl, phloem Cl- concentration was twice that of Na+, but still insignificant relative to xylem Cl- (Jeschke & Pate 1991). In a detailed study of Cl- regulation in Hordeum vulgare, no evidence was found for Cl- circulation from the shoot and back to the external medium, based on 36Cl flux analysis and total Cl- contents in different tissues over time (Greenway & Thomas 1965). However, significant retranslocation of Cl-within the shoot was found, with young leaves receiving up to 30% of their total Cl- from older leaves (Greenway & Thomas 1965). Although limited data are available for direct measurements of phloem Cl- in genotypes that differ in salt tolerance, it seems that the proportion of Cl- that is recirculated via the xylem is the same for salt-tolerant and salt-sensitive species (White & Broadley 2001).


Passive and active transport

With normal concentration gradients of Cl- (and NO3-), the highly negative plasma membrane potential of plant cells requires that anion transport into a cell is active. This appears to be via co-transport with protons (Sanders 1980; Beilby & Walker 1981; Felle 1994). The opposite applies for anion efflux where transport out of the cell can be passive. However, this thermodynamic requirement for anion transport under normal conditions does not mean that passive transporters do not exist for influx. There seem to be transporters for many contingencies in plants. Thus, in roots there is an anion channel2 that opens to allow passive influx of anions (Skerret & Tyerman 1994; Diatloff et al. 2004; Zhang, Ryan & Tyerman 2004). This channel is also observed in other cell types (Roberts 2006). Passive influx can only occur when the membrane potential becomes more positive than the equilibrium potential for the anion, a situation that can occur when the external concentration of the anion is high relative to the internal concentration or when membrane potential is depolarized. Na+ entry to the cytoplasm would depolarize the membrane potential (Cakirlar & Bowling 1981; Schachtman et al. 1991; Blumwald et al. 2000), which could drive the membrane potential positive of the equilibrium potential for Cl-. The passive entry of Cl- could initially prevent excessive depolarization by clamping the membrane potential at the equilibrium potential for Cl-, provided this is reasonably negative.

Since the discovery of an anion channel in wheat roots that could allow anion influx under saline conditions (Skerret & Tyerman 1994), there has been some debate about the likelihood that Cl- could enter the cell passively. There is evidence in favour of this proposition, for example, net Cl- influx to the cytoplasm of Arabidopsis roots (measured using a fluorescent probe) was independent of external pH, inhibited by Ca2+ and Mg2+ (indicating a link to cation uptake), and inhibited by La3+ (Lorenzen et al. 2004). Analysis of the older literature led several authors to conclude that Cl- influx could be passive at high external Cl- (Skerret & Tyerman 1994; Tyerman & Skerret 1999; White & Broadley 2001; Roberts 2006). However, it seems that cytoplasmic Cl- concentration rises rapidly when root cells are exposed to high external Cl- (Lorenzen et al. 2004) and this is supported by compartmental analysis on barley roots (Britto et al. 2004). Taking the data of Lorenzen et al. (2004), the cytosolic Cl- concentration goes to about 50 mm after 1.5 h when the roots are exposed to 100 mm external NaCl. The equilibrium potential for Cl- therefore declines to about −15 mV, having begun initially at about −50 mV when NaCl was first applied. This loading of cytosolic Cl- would oppose the passive entry, requiring that the membrane potential become significantly depolarized. This was illustrated for barley roots from estimates of cytosolic Cl- concentrations showing that the equilibrium potential for Cl- drops to about +40 mV at an external Cl- concentration of 100 mm, as a result of increased cytoplasmic Cl- (Britto et al. 2004). It is unlikely that the membrane potential of viable cells would be anywhere near such a positive potential to allow passive influx. With an up-shock of external NaCl, it is possible for Cl- entry to be passive initially, particularly if the entry of Na+ depolarizes the membrane potential, but after the cytoplasmic Cl- concentration has stabilized at higher levels, influx would have to be active.

Anion flux to the vacuole can be passive, but with normal tonoplast membrane potentials, a concentration gradient for a univalent anion of only about threefold higher in the vacuole can be established. Thus, it is not so surprising that a proton antiporter for NO3- has been discovered (AtClCa), located on the tonoplast that transports with a coupling ratio of 2 NO3-: 1H+, allowing a much higher accumulation of NO3- in the vacuole (De Angeli et al. 2006). This transporter is a member of the ClC family of anion transporters for which there are anion channel and anion/proton antiporter members (Chen 2005). Proton antiport across the tonoplast for NO3-, but not for Cl-, was proposed earlier based on dissipation of membrane potential and pH gradients in tonoplast vesicles (Blumwald & Poole 1985) and from microelectrode measurements of nitrate gradients and tonoplast membrane potentials (Miller & Smith 1992). Channel-mediated anion transport across the tonoplast has been characterized electrophysiologically for malate (Cerana, Giromini & Colombo 1995; Hafke et al. 2003) and Cl- (Tyerman & Findlay 1989; Plant, Gelli & Blumwald 1994; Pei et al. 1996).

It is evident that knowledge of the Cl- concentration gradients across membranes will be very important to determine the requirement for active or passive influx or efflux depending on membrane potential. Figure 4 shows the Nernst potential for combinations of internal and external anion concentration and illustrating when active transport is required, or passive transport can occur. Cytoplasmic Cl- concentration is normally measured in the range of 5–20 mm under non-saline conditions. At the lower end of cytoplasm concentration and with an external concentration of 100 mm, membrane potentials more positive than about −70 mV would be required for passive influx to the cytoplasm. If membrane potentials were in this range, and for this concentration gradient to be sustained, there would also have to be active efflux from the cytoplasm (e.g. ClCa see below and Fig. 6).

Figure 4.

Cl− equilibrium potential (ECl) surface calculated using the Nernst equation for combinations of internal and external concentrations of Cl−. Blue shading indicates when influx is more likely to occur because the ECl is more negative and Vm could be positive of ECl when [Cl−]cyt is low and [Cl−]out is high.

Figure 6.

Indicative voltage range of activity for common anion channels observed from electrophysiology. The horizontal axis indicates the range of Vm and the labelled boxes indicate the approximate range of Vm where the particular class of anion channels activate. Heavier shading indicates a maximum in activity. ECl is normally positive but salinity will shift ECl negative (dashed arrow). If Vm is positive of ECl, some channels may allow anion influx. Channel names vary but here we use those for xylem parenchyma cells and root cortical or epidermal cells. IRAC = inward rectifier anion channel activated by hyperpolarization (anion efflux, there may be more than one channel accounting for this, White & Broadley 2001). R-Type and X-QUAC, rapid/quick activation anion channel, by depolarization and normally accommodating anion efflux. S-Type or X-SLAC, slow activation anion channel, by depolarization and normally accommodating efflux. OR-DAAC, outward rectifying depolarization activated anion channel, only able to accommodate anion influx and activation closely follows ECl. ALAAC, aluminium activated anion channel. This has some features in common with X-SLAC and X-QUAC which both may allow outward currents (anion influx). Note that X-QUAC in xylem parenchyma can also show high inward currents (anion efflux) at more hyperpolarized voltages contrasting to typical R-type channels.

Concentration kinetics, high and low affinity transport systems

Another important feature of anion influx, common to most ion influx systems, is that both high and low affinity systems occur. There are numerous examples of this for Cl- in the older literature, with the low affinity transport showing saturation or linear kinetics (White & Broadley 2001). In some cases, the high and low affinity Cl- influx can be separated by inhibitors (Kochian, Jiao & Lucas 1985). Taking NO3- transport via nitrate transporters (NRTs) as an example, specific NRTs transport exclusively in the high affinity range, low affinity range, or can have dual affinity (Tsay et al. 2007). In addition, some are constitutively active while others are induced by low nitrate status of the plant.

Uptake across the plasma membrane by low affinity proton symport transporters (nH+:mCl- where n > m) can strongly depolarize the plasma membrane (Felle 1994). Thus, high transport capacity and the ability to change the voltage is not just a property of channel-mediated transport, an excellent example being the NO3- efflux (NAXT1) transporter which has a high Vmax related to a very high density in the membrane, as judged by vesicle studies (Segonzac et al. 2007). Ion channels are generally considered to exhibit low affinity transport with first order concentration-dependent kinetics, but some do saturate (Dietrich & Hedrich 1998).

Channels can be highly selective by effectively transporting just one ion, or poorly selective, and this can depend on the absolute concentration and concentration gradients of the anions. An example of this is the complex selectivity of the Al3+-activated malate transporter (TaALMT1) anion channel discussed further below (Pineros, Cancado & Kochian 2008b). Secondary active transporters are usually taken to be more selective than channel-mediated transport, but often selectivity is not examined. It is possible that some low affinity secondary active transporters may not be so selective, for example, AtClCa appears to transport anions other than NO3-, including Cl- (De Angeli et al. 2009). Unfortunately, selectivity of anion transporters at high external Cl- is generally not measured.

Voltage dependence

Voltage dependence of transport is often associated with ion channels and is used in the classification of various anion channels observed with electrophysiology. This manifests as a change in conductance of the channel with change in membrane potential and is seen as a time-dependent change in conductance (either rapid or slow) when the voltage is altered during voltage clamp protocols. It can be very strong, such that a channel only opens for a range of potentials that essentially would allow transport in only one direction. This rectifier behaviour is particularly the case if the voltage sensitivity is closely tuned to the equilibrium potential for the transported ion. Thus, ion channels can change the membrane voltage by passing ion currents and at the same time are regulated by membrane voltage. Such voltage dependency is vital for control of channels; otherwise, they could dominate ion fluxes and dissipate ion gradients. Voltage dependency of channels, combined with other signalling, can explain some phenomena with respect to charge balance such as osmotic efflux of KCl (Blatt 2000; de Boer & Volkov 2003), Na+/K+ exchange during salinity induced Na+ uptake (Qi & Spalding 2004; Shabala & Cuin 2008) and control of pH gradients (Tyerman 1992; Colcombet et al. 2005).

Like channels, transporters can show voltage and time-dependent kinetics (Lu & Hilgemann 1999). AtCLCa shows voltage and time dependence, giving outward rectification (relative to the cytoplasm), which in this case (2NO3-:1H+ antiport) is equivalent to a preference for NO3- flux across the tonoplast out of the vacuole. Similarly, the proton-coupled phosphate cotransporter, HvPht1;6, shows voltage and time-dependent behaviour giving inward rectification when expressed in Xenopus oocytes (Preuss et al. 2009). The Km for transport can also be voltage dependent, as demonstrated for nitrate and histidine transport via BnNRT1;2 when expressed in Xenopus (Zhou et al. 1998). Some whole cell patch clamp records have been observed that are similar to the currents attributed to AtClCa and PhT1;6 in different systems, and there may have been cases where these currents were incorrectly assigned to anion channels.

Futile cycling and energy considerations

It has been stated that: ‘The prevention of Na+ uptake from the apoplast is more demanding in terms of ion selectivity and energy costs than the prevention of Cl- uptake’ (Munns 2005). This warrants further comment and a comparison of basic transport circuits ( Fig. 5). The statement by Munns assumes that exclusion of Cl- automatically occurs through the opposing effect of the negative membrane potential, while Na+ is not excluded in the first instance because it will passively diffuse down an electrochemical potential gradient. It is true that under most circumstances, Cl- uptake into the symplasm is active, most probably secondarily active via an anion:H+ symport. However, this symport can display low affinity transports (LATs), that is, continued and increased influx rate as the external Cl- concentration increases with a high Km (White & Broadley 2001). In addition, initial influx with a sudden increase in external NaCl could be via an anion channel as discussed previously. Net Cl- influx rates can be as high as Na+ influx under salinity (e.g. maize; Hajibagheri, Harvey & Flowers 1987). Therefore, just because Cl- influx is active in the steady state, it does not imply a low rate of transport. The high rate of influx via LATs places an acid load on the cytoplasm and could depolarize the membrane potential.

Figure 5.

Hypothetical transport circuits that may occur across the plasma membrane to account for influx of Na+ and Cl- and apparent futile cycling of Na+ and Cl-. For each case, the H+-ATPase establishes an inward H+ gradient and negative Vm. (a) Na+ influx via non-selective cation channels driven by the negative Vm and sustained by Na+:H+ antiporter. Charge balance for net influx of Na+ could be achieved by efflux of K+ or influx of Cl-. (b) Cl- influx via an anion channel and 2H+:Cl- symport when Vm is positive of ECl (see Fig. 4). Charge balance for net influx could be achieved by Na+ influx or anion efflux (e.g. NO3-). (c) Cl- influx/efflux at steady state after the cytoplasmic concentration of Cl- has increased because of initial influx and ECl has become more positive. Cl- efflux can occur via an anion channel. (d) Possible circuit incorporating the cation Cl- cotransporter (CCC). In this case, influx of Cl-, Na+ and K+ are electroneutral and dependent on the gradient established by the difference in [Na+] × [K+] × [Cl-]2. For high external NaCl concentrations, and because of the exponent in [Cl-], the direction of transport would be inward. Proton efflux is coupled with Cl- efflux via the H+-ATPase and an anion channel could maintain low cytosolic [Cl-]. Na+ efflux via H+ antiport would maintain low cytoplasmic [Na+]. Cycling of Na+ and Cl- would still occur, but influx of Cl- would appear to be strongly coupled to influx of cations (e.g. Lorenzen et al. 2004).

The second point to make is that the net transport or rate of accumulation of any ion depends on the balance between unidirectional influx and efflux. The efflux of Cl- is more likely to be passive because there is a gradient for passive efflux of anions from the cytoplasm to the external medium. Therefore, accumulation of Cl- will depend on the unidirectional active influx and the unidirectional passive efflux. If the Cl- LATs continues unabated, then to prevent net Cl- uptake the passive efflux becomes critical and will need to be increased. Generally, it has been observed that Cl- efflux is stimulated by increased concentration of external Cl- (Britto et al. 2004; Sun et al. 2009), and Cl- efflux can approach 90% of the influx in barley (Britto et al. 2004). This stimulation could occur via acidification of the cytoplasm and alkalization of the apoplast (Colcombet et al. 2005). Compare this with Na+, where the net accumulation is dependent on the unidirectional passive influx and the unidirectional active efflux. Is there actually a difference between these scenarios in terms of energy expenditure? Na+ enters passively and may depolarize the membrane, similar to the LATs Cl- influx. The antiporter dissipates the proton gradient, also placing an acid load on the cytoplasm requiring increased proton pump activity. One can see from the simple transport circuit in Fig. 5a,c that in both cases proton pumping is required for the two ions assuming a steady state in cytoplasmic ion concentration and constant pH. In both cases, we see an apparent futile cycling of ions, expending energy.

The energy expenditure in terms of oxygen consumed can be calculated in the same way that Britto & Kronzucker (2009) have done for Na+, except for a different stoichiometry of 2:1 for H+:Cl- influx (H+:Na+ antiport was taken as 1:1), 1:1 for H+ pumped per ATP and 5:1 for ATP synthesized per O2 produced. This demonstrates that similarly high expenditures of O2 are observed for Cl- fluxes as for Na+ fluxes. Taking fluxes measured for roots of intact barley plants in 100 mm NaCl with Cl- influx/efflux of 50–60 µmol g−1 h−1 (Britto et al. 2004) and Na+ influx/efflux of 70–80 µmol g−1 h−1 (Kronzucker et al. 2008), the oxygen consumed would be slightly higher for Cl- at 20–24 µmol O2 g−1 h−1, compared with 14–16 µmol O2 g−1 h−1 for Na+. This is comparable with the total respiratory O2 consumption for barley roots (approximately 25–30 µmol O2 g−1 h−1) (Bloom, Sukrapanna & Warner 1992; Clarkson et al. 1992). Although these rudimentary calculations serve to illustrate similar energy use for Cl- and Na+ based on a proton economy, they are too high to be realistic and have prompted reconsideration of the transport circuits illustrated in Fig. 5a,c (Britto & Kronzucker 2009). Alternative possibilities are coupled Na+:K+:2Cl- cotransport (Fig. 5d), which supposedly occurs with the CCC transporters that are electroneutral in animal cells (Gamba 2005) and also appear to be electroneutral in plants (N.L. Teakle et al. unpublished results). In this case, the flux proceeds in the direction from the greater product of the concentrations for each ion (see Eqn 1 in the latter part of the paper). Figure 7 illustrates the direction of transport of a CCC with various scenarios of concentration gradients.

Figure 7.

Examination of the possible role of cation-coupled Cl- (CCC) transport in removing Cl- and cations from the xylem, assuming a stoichiometry of 1K+:1Na+:2Cl-. We have used the xylem concentrations (mm) for Cl-, K+ and Na+ measured by Teakle et al. (2007) in Lotus corniculatus after 28 d in 200 mm external NaCl. Lotus tenuis had about half this xylem [Cl-] at the same stage and shows higher expression of the CCC ortholog. Values refer to concentrations in mm. Note that cytosolic Na+ and Cl- concentrations would have to be quite low to allow influx from the xylem with these measured concentrations and assuming that [K+] in the cytoplasm is about 100 mm. Taking more realistic concentrations for cytoplasmic Cl- and Na+ of about 20 mm, it is likely that transport would be inward to the xylem if the transporter resides on the plasma membrane.

Finally, there is the issue of selectivity. The Munns argument seems to imply that just because influx is active, it is likely to be more selective. There are increasing examples, when tested, of surprisingly non-selectivity between secondary active transport systems. For passive Cl- efflux, most anion channels are rather non-selective and are likely to release NO3- or even SO4- over Cl-, based on the selectivity sequences so far measured for the various anion channels (White & Broadley 2001; Roberts 2006).


Anion channels: electrophysiology

There are surprisingly few electrophysiological studies of anion channels in roots, particularly in the context of Cl- transport under salinity. When referring to named channels below, it should be kept in mind that these names are for the electrophysiological manifestation of some underlying transport system, that most probably functions as a channel. Rarely have the genes and proteins been identified for anion channels, unlike the situation for K+ channels and K+ transporters in general (Very & Sentenac 2003), although there are some promising recent discoveries for a putative gene that may code for the slow-anion channel in guard cells (Negi et al. 2008; Vahisalu, Kollist & Wang 2008). The critical evidence for naming these transports as ‘channels’ is if single channel events can be shown to underlie the ensemble whole membrane currents that are more often (and more easily) measured. Various characteristics are used to name channels, for example, voltage dependency (depolarization activated, or hyperpolarization activated), direction of transport (outward rectifier = anion influx, or inward rectifier = anion efflux), rate of activation (slowly activating or S-type, and rapidly activating or R-type) or what activated the channel [e.g. aluminium activated anion channels (ALAAC); Roberts 2006]. Figure 6 shows a generalized summary of the voltage range over which various anion channels activate, the direction of the flux that occurs and some common names of the channels. Note that there may be more than one type of protein that accounts for the particular electrophysiologically observed channel.

Before addressing plant root cells specifically, it is important to consider the information that has been obtained from the extensive electrophysiological studies on Charophytes that differ in salt tolerance (Beilby & Shepherd 2006). Charophyte membrane transport has been an excellent model guiding many higher plant membrane transport studies. For Charophytes, elevated external Na+ and changes in turgor are key elements in the response to salinity. Maintenance of the negative membrane potential via activation of the proton pump and reduced activation of ‘leak’ conductances enables cell viability under salinity (Shepherd et al. 2008). Depolarization to the threshold for action potentials is a critical point in salt sensitivity. Calcium-dependent anion channels are a major component of the action potential. Furthermore, mechanosensitive channels when activated under salinity interact with the depolarization and action potential in a Ca-dependent manner and hasten cell death (Shepherd et al. 2008). Recently, Beilby & Al Khazaaly (2009) have identified OH- permeable channels as the most likely component of the ‘leak’ conductance activated by salinity, a surprising observation given that previously these channels were only observed at high pH (Beilby & Bisson 1992). This is highly significant for higher plant roots because there are similarities with Charophytes in H+/OH- circulation at the root tip (Raven 1991), and there is evidence for their existence from combined patch clamp and microelectrode ion flux estimation (Tyerman et al. 2001).

Anion channels in root cells

Patch clamp studies on protoplasts made from various cell types from roots have demonstrated several different types of anion channels. These have been extensively reviewed recently (Roberts 2006) and for anion channels in other cell types, see Barbier-Brygoo et al. (2000). Next, we discuss the anion channels that have been identified by electrophysiology in specific root cells and we attempt to emphasize the potential role of these channels in Cl- fluxes under salinity; however, only a few studies have examined these channels in the context of high salinity.

Epidermal cells and cortical cells

Protoplasts from epidermis and cortical cells have shown the presence of outward rectifying depolarization-activated anion channels (OR-DAACS) in wheat, maize, Arabidopsis and lupin (Skerret & Tyerman 1994; Pineros & Kochian 2001; Diatloff et al. 2004; Zhang et al. 2004). These channels show similar features between species but differ in rate of activation and degree of rectification. Only a couple of studies have examined these channels specifically in the context of high salinity (Skerret & Tyerman 1994; Tyerman et al. 1997). These have already been discussed previously as a potential pathway for passive entry of Cl- at high salinity, but it is worthwhile noting other important features of these channels. They are stimulated by high cytosolic Ca2+ and by hyperosmotic treatment (Skerrett 1995), so that in addition to their potential role in preventing excessive depolarization during initial exposure to high external Cl-, it is possible that they are also involved in osmoregulation. They are about twice as permeable to NO3- than to Cl-, but at high Cl- concentrations, the currents through the channels can be very high and comparable with the magnitudes of currents through the outward rectifying K+ channel (Skerret & Tyerman 1994). This high capacity for transport, despite a low single channel current, means that the channel when open is able to dominate the membrane conductance and probably would tend to clamp the membrane potential close to the equilibrium potential for Cl-. If this equilibrium potential is more negative than the equilibrium potential for K+, then Na+ influx would be balanced by Cl- influx. On the other hand, if the K+ equilibrium potential was more negative than the Cl- equilibrium potential and the K+ outward rectifier channel was not inhibited by the high NaCl concentrations, then Na+ influx would be balanced by K+ efflux. These scenarios depend on maintaining normal cytoplasmic concentrations of K+ and Cl-; however, high salinity reduces K+ in the cytosol and increases Cl-, both of which cause a shift towards positive voltages of their respective equilibrium potentials.

Another class of channel observed in Arabidopsis protoplasts, derived from root epidermal cells, is the depolarization depolarisation-activated anion channel that mediates anion efflux. These are not activated at very negative membrane potentials, but upon depolarization to membrane potentials in the range of −120 to −150 mV, they open to carry large anion efflux currents (inward currents). They rapidly activate (R-Type) and slowly inactivate at constant voltage (slowly turn off), but not completely, because about 50% of the initial current is maintained (Diatloff et al. 2004). Further depolarization causes less activation so that there is virtually no activity positive of −40 mV, and well negative of the equilibrium potential for Cl- under non-saline conditions. These channels run down unless SO42− is present in the cytoplasm, and the channel is also permeable to SO42− as well as to NO3- and Cl- (Frachisse et al. 1999; Diatloff et al. 2004). These R-type depolarization-activated anion channels are relatively common in other cell types, the best characterized are guard cells and hypocotyl cells (Dietrich & Hedrich 1998; Frachisse et al. 1999), but there are some differences between cell types (Roberts 2006). Increasing extracellular anion concentration shifts the voltage activation of the R-type channel more negative, the extent depending on the anion involved and interactions with nucleotides on the cytoplasmic side (Colcombet et al. 2001). For Arabidopsis hypocotyl cells, the half activation potential shifts to above −170 mV at 100 mm external Cl- when the MgATP concentration is 1 mm in the cytosol (Colcombet et al. 2001). High cytosolic MgATP counteracts this effect. Thus, in high external Cl- concentrations, the activation potential could be so negative that the depolarization activation would not be observed and the channel would appear to be hyperpolarization activated, that is, going from less negative to more negative membrane potentials would activate the channel. The physiological significance of this channel under salinity is not clear. However, it has been proposed to be involved in elicitor responses from pathogens and seems to be a component of the signalling pathway leading to reactive oxygen species (ROS) production (Colcombet et al. 2009). As salinity generally increases ROS production, which has effects on other transport processes (e.g. aquaporins; Boursiac et al. 2008), it would be interesting to examine the role of the R-Type anion channel in ROS signalling under salinity.

Another depolarization-activated channel is the slowly activating anion channel (S-type). This has distinctly different selectivity and pharmacological properties to that of the R-type (Frachisse et al. 1999, 2000). For example, in Arabidopsis hypocotyl protoplasts, the S-type channel has a very high permeability to NO3- relative to Cl-, contrasting to the R-type channel which is more permeable to SO42− than to NO3- and Cl-. The S-type channel is generally considered to provide the capacity for sustained anion efflux upon membrane depolarization negative of the anion equilibrium potential. The range of voltage activation is more depolarized than that of the R-type channel and in hypocotyl cells tends to outward rectification (Frachisse et al. 2000), and therefore is likely to allow anion influx at depolarized membrane potentials and high external anion concentrations. This channel is regulated by pH and protein phosphorylation (Frachisse et al. 2000). Cytosolic acidification and extracellular alkalinization increase S-type currents independently of voltage in Arabidopsis hypocotyl protoplasts (Colcombet et al. 2005). This characteristic would allow the S-type channel to function in pH regulation by acting as an electrical shunt for the proton pump to allow higher pump activity. The S-type channel is a candidate for elevated anion efflux that occurs under salinity, particularly because high salinity will tend to acidify the cytoplasm (Martinez & Lauchli 1993; Felle 1994). This channel has also been identified in root hair cells after desiccation which depolarized the membrane potential (Dauphin et al. 2001).

A citrate- and malate-permeable channel with voltage characteristics similar to the R-type channel has been identified in Arabidopsis epidermal cells. This channel became evident when Arabidopsis was starved of Pi and was suggested to account for the efflux of citrate from Arabidopsis roots under Pi starvation (Diatloff et al. 2004). A citrate-permeable channel was also identified from Lupin roots. This channel is of the IRAC type and had a very high relative permeability for citrate over Cl- (26×) and was more evident from roots grown in low Pi (Zhang et al. 2004). Its function was proposed to be as a citrate effluxer to mobilize external Pi in conjunction with the proton pump that would activate the channel by hyperpolarization. It is unlikely that this channel would be important under salinity given its high apparent selectivity for citrate over Cl-, but it is possible that activation of the channel would be impeded at high salinities because of membrane depolarization. The other channel identified in Lupin was classed as an outward rectifier and is similar in voltage dependence to the R-type channel observed in Arabidopsis. This channel was more permeable to Cl- than to citrate (3.7×) and its frequency was not changed by growth in low Pi (Zhang et al. 2004).

Aluminium-activated anion channels have been studied in context of Al tolerance at low pH afforded by stimulation of malate or citrate efflux from root tips that complexes Al3+ and protects the roots from the toxic effects of Al3+ (Ryan, Delhaize & Jones 2001; Delhaize, Gruber & Ryan 2007). An Al3+-activated channel permeable to Cl- was initially discovered in a patch clamp study of protoplasts isolated from root tips of near isogenic lines of wheat differing in Al3+ tolerance (Ryan et al. 1997) and was later demonstrated to be more permeable to malate than to Cl- (Zhang et al. 2001). Anion channels were identified in protoplasts from maize roots that are activated by Al3+ (Pineros & Kochian 2001; Pineros et al. 2002), and which were permeable to citrate and malate although apparently less so than Cl- (Kollmeier et al. 2001). This channel could allow both Cl- efflux and influx (Pineros et al. 2002). Interestingly, the stele of maize roots exuded citrate upon Al3+ addition (Pineros et al. 2002). The ability of the maize channel to allow Cl- influx has implications for saline conditions combined with Al stress, as high cytosolic Cl- may compete with the efflux of organic acids, or the gating of the channel may be disrupted by the high external Cl- concentration resulting in reduced organic anion release. Although the wheat channel appeared to be more permeable to malate than to Cl- in the patch clamp analysis (more than 18-fold; Zhang et al. 2008), substantial Cl- currents can be observed. The wheat channel has not been demonstrated to carry anion influx in protoplasts derived from wheat roots, but this might be a consequence of the low external anion concentrations used in the experiments rather than a real rectification by channel gating, because the channel can outwardly rectify (Pineros et al. 2008b; Zhang et al. 2008).

In the context of water relations and changes in turgor pressure, we need to mention the stretch-activated channels, or mechanosensitive channels for which there are anion channel representatives (White & Broadley 2001; Roberts 2006; Zhang, Patrick & Tyerman 2007). There is only one electrophysiological characterization of stretch-activated anion permeable channels in roots (Haswell et al. 2008). Here, two mutants of the MscS-like proteins were characterized in Arabidposis root cortical cells by patch clamp. Wild-type plants have a 10 pA conductance channel that is activated by increased pressure (5–10 mmHg) in the whole cell configuration. This channel is more permeable to Cl- than to Ca2+, but neither the selectivity to other ions nor the voltage dependency is known, making it difficult at this stage to speculate on their role in salinity tolerance.

Xylem parenchyma cells

Several patch-clamp studies have been carried out on protoplasts isolated from cells of the stele to examine the processes for salt release to the xylem through coordinated cation (Wegner & Raschke 1994; Roberts & Tester 1995, 1997; Wegner & De Boer 1997) and anion efflux (Köhler & Raschke 2000; Gilliham & Tester 2005). The net release of K+ and anion is via efflux channels, with the molecular identity of a K+ channel (SKOR) being identified as a member of the SHAKER family (Gaymard et al. 1998). Three types of anion channels have been identified in barley xylem parenchyma protoplasts: an inwardly rectifying channel (X-IRAC), a quick activating anion channel (X-QUAC) and a slow activating anion channel (X-SLAC). X-QUAC and X-SLAC are similar to the R-type and S-type anion channels, respectively, as discussed previously, although X-QUAC appears to accommodate large inward currents (anion efflux) at hyperpolarized potentials that differs from typical R-type channels (Köhler & Raschke 2000, cf with Diatloff et al. 2004). Both X-IRAC and X-QUAC have been identified in maize stellar-derived protoplasts (Gilliham & Tester 2005). The X-IRAC activates at quite negative hyperpolarizations and seems equally permeable to both NO3- and Cl- (Köhler & Raschke 2000). This channel is unlikely to be important in anion efflux to the xylem because of its more negative activation, out of the range that corresponds to activation of the K+ outward rectifier channel (Köhler & Raschke 2000). X-QUAC seems to be the main candidate for accommodating anion efflux in conjunction with the outward cation rectifiers. X-QUAC is more selective for NO3- than for Cl- in both maize and barley (permeability ratios from reversal potentials: maize 1.7:1, barley 3:1) (Köhler & Raschke 2000; Gilliham & Tester 2005).

The actual currents that may occur through the channel under different anion loads in the cytoplasm is likely to be more complex than is suggested from these ratios, because extracellular anions shift the peak current as well as the reversal potential (Köhler & Raschke 2000). Increased extracellular NO3-, but not Cl- or malate, also regulates the channel to provide positive feedback so that similar efflux of NO3- would occur even when the gradient for efflux decreases (Kohler et al. 2002). A limitation of these results is that the conductances were not measured at high external NaCl concentrations. It was suggested that other plants might have higher selectivity for NO3- over Cl- than for barley, on account of the fact that barley accumulates Cl- in the shoot under salinity (Kohler et al. 2002). This appears not to have been rigorously examined, but the maize permeability ratio suggests that there may not be a large difference, and perhaps Cl- re-absorption from the xylem may differentiate between plants that exclude Cl- to different degrees (see next at section on CCCs and Fig. 7). The complex nature of the gating of the X-QUAC, combined with properties of X-SLAC (pH regulation), means that the response to salinity could be complicated, because of decreased NO3- in the cytoplasm and apoplast (from reduced uptake), and the effect of water stress discussed next. The control of Cl- transport to the shoot under saline conditions may be a result of reduced loading of Cl- via X-QUAC; however, the molecular identity of these xylem channels remains to be identified (see next).

An outwardly rectifying non-selective cation channel (NORC) was also identified in xylem parenchyma cells from barley roots (Wegner & Raschke 1994). The channel is about twice more permeable to cations than to anions, but probably allows permeation of organic anions in addition to Cl- and NO3-. This channel does not activate until the membrane potential becomes quite depolarized (–25 mV when the equilibrium potential for K+ is very negative, and normally nearer zero mV). The same channel is observed more frequently in protoplasts that have higher cytosolic Ca2+ concentrations (Wegner & De Boer 1997). NORC may be important in the context of high NaCl concentrations because of its non-selective transport, that is, it can transport Na+, K+ and Cl-, which incidentally is an interesting link to the cation-Cl- cotransporter in Arabidopsis (AtCCC) that transports K+, Na+ and Cl- (Colmenero-Flores et al. 2007 and see next).

Water stress and ABA have been known to promote accumulation of solutes in the roots in order for roots to continue growing and exploring the soil for water (Pitman 1982). Both the outward rectifier K+ channel and X-QUAC are regulated by ABA such that salt efflux to the xylem would be reduced under water stress. This is likely to also occur under salinity as roots certainly accumulate Cl- under these conditions. ABA regulation appears to be via transcriptional and post-translational regulation for both SKOR (Gaymard et al. 1998; Roberts 1998; Roberts & Snowman 2000) and X-QUAC (Gilliham & Tester 2005). Elevated cytosolic Ca2+ also seems to be important in the signalling pathway for both channels (Gilliham & Tester 2005).

Links with water flow

In general, there is a relationship between root water flow and xylem loading of K+ (Wegner & Zimmermann 2009), which must also apply for anions on the basis of charge balance. Flux of K+ to the xylem and root hydraulic conductivity were positively correlated with water flow across the root, but were not correlated with xylem pressure (Wegner & Zimmermann 2009). A correlation between root hydraulic conductance and transpiration was also observed for grapevine (Vandeleur et al. 2009). Wegner & Zimmermann (2009) argued that the flow-dependent xylem loading of K+ was via increased K+ accumulation within xylem parenchyma cells at higher flows, providing greater driving force for efflux via the K+ outward rectifier. A close correlation has been observed between Cl- uptake and water flow into two citrus genotypes differing in salt sensitivity (Moya et al. 2003). Although both genotypes showed a linear relationship of Cl- uptake versus total water absorbed, the salt-tolerant genotype showed a lower slope, implying that per unit of water absorbed there was less Cl- transport to the shoot. Transport via the apoplast across the root may explain these results, but a comparison between Vitis genotypes differing in salt tolerance found that bypass flow, that is, uptake across the root via the apoplast, did not account for the differences in Cl- transport to the shoot (Abbaspour 2008). Furthermore, flux analysis showed substantial differences in flux rates across both tonoplast and plasma membrane in the root cells.

There is another interesting link between water transport of roots and anion transport. Roots respond rapidly to NO3- in the external medium by increasing root hydraulic conductivity (Gloser et al. 2007). This effect is linked to intracellular NO3- concentration, rather than products of NO3- assimilation, stimulating aquaporin-mediated water transport across the plasma membrane of cortex cells (Gorska et al. 2008). Neither SO42− nor Pi anions stimulate water transport. It would be interesting to see what effect high salinity and Cl- has on this regulatory system, which is proposed to aid accumulation of NO3- adjacent to the root by increasing water flow to the root.

Anion transporters: the genes

It is likely that several transporter and channel genes (and associated proteins) are involved in Cl- exclusion from the shoot. This may include transporter/channel genes that have indirect effects on Cl- exclusion (e.g. aquaporins linked to water flow in roots) or control of anion channels [e.g. ATP binding cassette (ABC) transporters]. While previously Cl-‘exclusion’ was considered to be controlled by a single dominant gene (Abel 1969; Sykes 1992), it seems several mechanisms may be contributing to the amount of Cl- loaded to shoots (described above). In addition, Rogers et al. (1997b) found heritability of shoot Cl- to be only moderate in the first cycle of selection (maximum of 0.37) and very low (0.1) in the second selection cycle. Recently, a transcriptome analysis of two citrus genotypes differing in Cl- exclusion under elevated Cl- demonstrated the possible involvement of several anion transporter families, but differential expression between the genotypes was also detected for genes involved in energy and carbon metabolism and stress-related hormones (Brumos et al. 2009).

There are several candidate transporter gene families revealed by transcriptome analysis, mutant analysis and from functional analysis linked to anion transport, which may be involved in Cl- homeostasis in plants under salinity. We survey those gene families based on associations within the family to Cl- transport (either directly or indirectly), transcriptomic links to salinity, or by demonstrated Cl- transport activity. These include bacterial mechanosensitive channels of small conductance (MscS)-like (MSL), voltage dependent anion channels (VDAC) Porins, the CLC anion channels and anion/H+ antiporters, the NRT (NOD)s nitrate and peptide transporter family, the ABC transporter family, aluminium activated malate channels (ALMT), CCC and the recently identified slow anion channel associated protein (SLAC1) linked to anion channels in guard cells. It is likely that there may be other gene families or members of already identified families that function in Cl- transport and these may be revealed by either QTL analysis, mutant analysis or functional analysis where Cl- permeability is tested for the transporter. Families such as the multidrug and toxic compound extrusion (MATE), sulphate and phosphate transporters may be involved. Recently, a low affinity HPO42− transporter (HvPht1.6) was characterized in Xenopus oocytes (Preuss et al. 2009). This transports other oxy-anions equally well or better than HPO42− and also shows some Cl- transport. Unfortunately, often when anion transporters are functionally characterized, selectivity between different anions is not tested. Examples where selectivity was tested are discussed next.


The CLC ‘Chloride Channel’ gene family is well characterized in animals (Jentsch et al. 2002) and plant genes have been cloned from Arabidopsis (Hechenberger et al. 1996), rice (Nakamura et al. 2006), tobacco (Lurin et al. 1996) and soybean (Li et al. 2006) (reviewed in Barbier-Brygoo et al. 2000; De Angeli et al. 2007, 2009). In animals, they form two-pore Cl- channels (CLC0 and CLC1) or Cl-:H+ antiporters. The CLCs form homodimers with each monomer forming an anion transport pathway (Jentsch et al. 2002; Miller 2006). The three-dimensional crystal structure has been obtained for bacterial CLC proteins, allowing for characterization of amino acid residues important for gating and selectivity (Dutzler et al. 2002). There are seven members of this family in Arabidopsis and rice, and they are expressed (generally) in all tissues in various endomembranes (AtCLCa, tonoplast; AtCLCd, trans Golgi; AtCLCf, cis Golgi; AtCLCe, chloroplast thylakoid), but there is no evidence yet for a plasma membrane location, though they have been predicted to target the plasma membrane (Diedhiou & Golldack 2006). AtCLCa, c, e and f have been genetically linked to NO3- accumulation in Arabidopsis (Geelen et al. 2000; Harada et al. 2004).

There has been mixed success in functional characterization of CLCs, with only the tobacco CLC-Nt1 able to be expressed in Xenopus oocytes (Lurin et al. 1996). In their study, negative going voltage pulses activated a slow time-dependent inward current that was blocked completely by 10 mm Ca2+ (with 88 mm external Cl-) and external anions appeared to gate the transporter. Selectivity between anions was obtained from outward current tails observed during deactivation, and showed higher transport for NO3-, but there were still appreciable currents for Cl- (about 33% lower than for NO3-). SO42−, malate and glutamate were less than 33% of NO3-. Properties of the Arabidopsis AtCLCa were examined by comparing tonoplast anion currents (whole cell patch-clamp) between wild type and AtCLCa knockout mutants (De Angeli et al. 2006). There were time and voltage-dependent currents present in wild type that were completely absent in the knock out lines. These tended to outwardly rectify (i.e. favouring anion transport from vacuole to cytoplasm), and Cl- transport was much less than NO3-, but higher than for SO42−. From reversal potentials and pH dependence, it was deduced that the transporter worked as an antiporter with a stoichiometry of 2NO3-:1H+. Thus, under non-saline conditions, this transporter would accumulate NO3- into the vacuole.

As a single amino acid substitution can change selectivity from NO3- to Cl- (Bergsdorf, Zdebik & Jentsch 2009), this raises the possibility that other members of the CLC family might actually be Cl-/H+ antiporters. There is some indication that CLC members might be involved in Cl- transport. Comparison of rice genotypes that exclude Cl- (Pokkali) or accumulate Cl- (IR29) showed that there were differences in expression of OsCLC1 (Diedhiou & Golldack 2006). Under salt stress, IR29 had repressed expression of OsCLC1, while Pokkali showed induction in leaves and roots. The root induction was particularly strong and expression was located to the xylem parenchyma and phloem (Diedhiou & Golldack 2006). This could of course be a response to perturbed NO3- homeostasis rather than to high Cl- concentration. OsCLC1 and 2 are located on the tonoplast (Nakamura et al. 2006). In soybean, GmCLC1 and GmNHX1 (Na+/H+ antiporter) are both localized to the tonoplast and transcripts are increased by NaCl (125 mm) treatment and dehydration (Li et al. 2006). When both GmNHX1 and GmCLC1 were expressed separately in tobacco BY-2 cells, the cells were more salt tolerant (100 mm NaCl) but not dehydration tolerant. The GmCLC1 cells accumulated more Cl- in the vacuole (Li et al. 2006). In contrast to rice and soybean, the ortholog of Arabidopsis CLCd in Citrus leaves was not differentially expressed between the Cl- accumulating and excluding genotype under saline conditions (Brumos et al. 2009). Identification of the putative vacuolar Cl-/H+ antiporter is needed and its function in Cl- accumulation into vacuoles verified.


A gene encoding an Al3+-activated malate transporter (TaALMT1) was isolated from wheat (Sasaki et al. 2004) and encodes a protein with six putative membrane spanning regions (Motoda et al. 2007) that localizes to root plasma membranes (Yamaguchi et al. 2005). TaALMT1 is expressed constitutively and correlates with a major QTL for Al3+ resistance (Raman, Zhang & Cakir 2005). The expression of TaALMT1 in rice (Oryza sativa), cultured tobacco cells and barley gives Al3+-activated efflux of malate, which enhances the Al3+ resistance of the transgenic cells (Delhaize et al. 2004; Sasaki et al. 2004). TaALMT1 has homologs in Arabidopsis (AtALMT1 – AtALMT9), Brassica napus (BnALMT1) and maize (ZmALMT1), and encodes Al3+-activated malate transport proteins (Hoekenga, Maron & Pineros 2006; Ligaba et al. 2006; Kovermann et al. 2007). Functions of ALMT1 homologs other than conferring Al tolerance have been indicated. For example, ALMT1 homologs may function in nutrition and ion homeostasis (maize; Pineros et al. 2008a); malate transport (Arabidopsis; Kovermann et al. 2007), recruitment of beneficial bacteria to the rhizosphere (Arabidopsis; Rudrappa et al. 2008), the transport of malate, fumarate or citrate and possibly pH regulation (barley; B.O. Gruber et al.unpublished data).

Heterologous expression of TaALMT1 in Xenopus oocytes and tobacco-cultured cells has demonstrated that a basal level of anion transport is evident without Al stimulation and anion influx could be accommodated by TaALMT1 (Pineros et al. 2008b). Our studies on Xenopus oocytes expressing TaAMLT1 have revealed that this is more apparent at higher pH (pH 7.5) than at lower pH, particularly the ability to carry anion influx (S. Ramesh et al. unpublished data). It would be interesting to examine Cl- accumulation of the Al-tolerant wheat varieties under salinity, because if TaALMT1 is partially active without Al, it may contribute to Cl- fluxes across the plasma membrane. Furthermore, the selectivity of TaALMT1 for malate over Cl- is rather complex. At low external Cl-, the transport is more selective for malate than for Cl- (30:1), while at high external Cl- the channel becomes less selective (1:1) (Pineros et al. 2008b). This has implications for salinity effects, because at high external Cl-, the reversal potential will shift negative and the channel would potentially become equivalent to the OR_DAAC observed in wheat roots, and capable of transient inward flux of Cl-. If Al stress co-occurred with salinity, the ability of the root tips to excrete malate may be reduced because of the shift in reversal potential to more negative potentials combined with the depolarization under salinity, and the possible competition between Cl- and malate at the cytoplasmic face of the channel.


Recently, it has been predicted that xylem ion concentrations may also be controlled by active retrieval of Cl- from the xylem stream (Colmenero-Flores et al. 2007; Munns & Tester 2008). A possible candidate gene for the active retrieval of Cl- from xylem vessels is a CCC. CCCs are secondary active (but electroneutral) transporters that have been well characterized in the animal field (reviewed by Gamba 2005), and recently a Na+:K+:2Cl- co-transporter has been identified in Arabidopsis (AtCCC, Colmenero-Flores et al. 2007). Evidence for the role of AtCCC in controlling Cl- loading/unloading at the xylem/symplast boundary was based on: (1) AtCCC expression was found in xylem parenchyma cells using GUS constructs; and (2) Cl- concentrations of an Arabidopsis ccc mutant were 40% higher in shoots and 33% lower in roots, compared with wild-type plants at 50 mm NaCl for 15 d (Colmenero-Flores et al. 2007). Previous studies have suggested that Cl- transport in the xylem is an electroneutral process (Köhler & Raschke 2000) that is accompanied by permeable cations (Lorenzen et al. 2004), which suggests that CCCs could be a feasible candidate gene for xylem retrieval of Cl- under salt stress. A comparison of gene expression in leaves between two Citrus genotypes that differ in Cl- exclusion in response to elevated Cl- revealed that the Citrus AtCCC ortholog had increased expression over time (12 weeks) in the Cl--accumulating genotype (Carrizo citrange), but not in the Cl--excluding genotype (Cleopatra mandarin) (Brumos et al. 2009).

Based on the AtCCC expression in Xenopus where Na+, K+ (Rb+) and Cl- appear to be all transported and probably not coupled to proton transport (Colmenero-Flores et al. 2007), and our unpublished work indicating that the transport is electroneutral, one could hypothesize that the transport is consistent with that observed in some animal CCCs where the stoichiometry is Na+:K+:2Cl- (Russel 2000). For no net charge transfer, the voltage gradient does not affect the gradient for net movement and flux reversal occurs when the following relationship is satisfied:


Using this relationship and measured and acceptable concentration gradients, Fig. 7 shows possible gradients that would be required to drive transport into a xylem parenchyma cell via Na+:K+:2Cl- symport, based on Lotus xylem concentrations measured using spittle bugs (Teakle et al. 2007) in the presence of 200 mm external NaCl. Because of the squared term for Cl-, small changes in Cl- concentration have a large effect on the direction of transport. This figure suggests that only when the cytoplasmic concentration of Cl- and Na+ are very low will there be an influx from the xylem to the xylem parenchyma cytoplasm. The cytoplasmic concentrations required (about 2 mm Cl-) are much lower than those measured so far in root cells; however, xylem parenchyma cells may be atypical. Alternatively, the concentration in the xylem at more distal positions in the root may be much higher. Note that in Lotus corniculatus, which had lower expression of CCC, there is a higher xylem Cl- concentration, but Na+ and K+ are almost identical, perhaps accounted for by other transporters such as HKTs (Møller et al. 2009). Notwithstanding the charge balance issue, it is possible that this could occur, particularly if there is active transport at the tonoplast driving both Na+ and Cl- into the vacuole to keep Na+ and Cl- concentrations low in the cytoplasm, and depending on the cytosolic K+ concentration, which is unlikely to be less than about 100 mm.

Another possibility is that the transporter occurs in the tonoplast, driving electroneutral Cl- uptake into the vacuoles down K+ and Na+ concentration gradients. This might be important in the situation where large water flows across the root cause high concentrations of convected ions (Na+, Cl- and K+) in the cytosol, as has been recently proposed for K+ (Wegner & Zimmermann 2009). This could result in the same phenotype observed for the Arabidopsis ccc mutants and the difference found between the Lotus species, suggesting a link with transpiration. Transport of SO42− across the tonoplast in xylem parenchyma of Arabidopsis roots via AtSULTR4;1 and AtSULTR4;2 seems to account for differences in transport of sulphate to shoots. Further work is clearly needed to determine the localization of CCCs in plants, which will likely provide more clues to the function of these transporters under salinity.


The NRT and OPT or oligopeptide transporters (PTR) form a large gene family in higher plants (Tsay et al. 2007). The NRTs can be subdivided into NRT1 (low affinity) and NRT2 (high affinity or dual affinity). They have been functionally characterized as proton symporters with the exception of NAXT1, which belongs to the NRT1 subclass and is probably a passive uniporter (Segonzac et al. 2007). Often when functionally characterized, their selectivity to other anions is not tested. There is some evidence that members of the NRTs may be involved in Cl- homeostasis or that they transport Cl-. Passive NO3- fluxes were characterized in plasma membrane vesicles isolated from maize roots to show a similar acidic optimum to the proton pump (Pouliquin et al. 2000). The activation energy was high as may be expected for a transporter, and with a Km of 3 mm indicative of low affinity. The selectivity of this transport showed that Cl- transport was about 20% of that for NO3-. Considering the similarities to NAXT1 transport properties, this suggests that perhaps other members of the NAXT subclass may be permeable to Cl-. Another suggested involvement of NRTs comes from the comparison of transport genes expressed in the leaves of two Citrus genotypes after treatment with 50 mm KCl or NaCl; an NRT1 gene had constant expression over 7 weeks in the Cl- excluding genotype, but was down-regulated in the Cl- accumulating genotype (Brumos et al. 2009). However, this may be related to the effects of high Cl- on nitrate uptake.

ABC transporters

We include members of the ABC transporter family in our survey because there are indications that an ABC transporter controls the guard cell S-type anion channel and this channel or a close homolog is thought to be responsible for anion loading of the xylem in roots. The ABC transporters are a very large family in Arabidopsis and rice, with diverse roles including auxin transport and xenobiotic and metal detoxification (Davies & Coleman 2000; Rea 2007). They are located to either the plasma membrane or tonoplast membrane. There are 13 subfamilies among which are the multidrug resistance-associated proteins (MRP) and multidrug resistance (MDR) proteins (Rea 2007). A MDR member in Arabidopsis (AtABCB14 previously AtMDR12; Rea 2007) located to the plasma membrane has been shown to be involved in malate uptake by guard cells (Lee et al. 2008). Interestingly, the presence of external Cl significantly stimulates malate uptake when this protein is expressed in Escherishia coli cells in contrast to malate and fumarate, which competitively inhibit (Lee et al. 2008). There are 16 and 17 members of the MRP sub-family in Arabidopsis and rice, respectively (Rea 2007). Two of the MRPs (4 and 5) have been studied in detail because of their expression in guard cells. AtMRP5 is present on the guard cell plasma membrane and MRP5 mutants had reduced activation of the S-type anion channel by ABA and cytosolic Ca (Suh et al. 2007). AtMRP5 is also strongly expressed in the root and leaf vasculature (Gaedeke, Klein & Kolukisaoglu 2001). The MRP4 protein is located on the plasma membrane and is also expressed in primary roots (Klein, Geisler & Suh 2004). These studies on MRP4 and 5 indicate that ABC transporters play a central role in control of ion channels in guard cells, and given that they are also expressed in roots, and potentially alongside homologs of the S-type guard cell anion channel and outward rectifier K+ channel (SKOR), they may also be involved in regulation of anion release to the xylem.


A new family of possible anion transporters or anion channels was recently discovered in Arabidopis in an effort to uncover the gene for the S-type anion channel in guard cells (Negi et al. 2008; Vahisalu et al. 2008). Named SLAC1, the protein has homology to the C4-dicarboxylate transporter family and Mae1 of Schizosaccharomyces pombe. SpMae1 transports malate into the cell possibly co-transported with H+, but this was not evident when Mae1 was expressed in Saccharomyces cerivicea (Camarasa et al. 2001). SLAC1 is expressed strongly in guard cells and the protein, consisting of some 556–557 amino acids with predicted 10 membrane spanning helixes, was located to the plasma membrane (Negi et al. 2008; Vahisalu et al. 2008). Protoplasts isolated from guard cells showed that SLAC1 mutants had increased accumulation of malate, fumarate Cl- and K+ consistent with a deficiency in anion release from the guard cells (note anion release activates K+ release). The most direct indication of a close link of the gene with the S-type anion channel was that patch-clamped protoplasts from mutants (including a T-DNA mutant) had no activity of the S-type channel when this would be normally activated by ABA or Ca (Vahisalu et al. 2008). On the other hand, neither the R-type channel nor Ca channel was affected (Vahisalu et al. 2008). Unfortunately, SLAC1 could not be heterologously expressed for better functional characterization, but there is a strong possibility that this gene codes for the S-type anion channel. Significantly, three SLAC1 homologs (SLAHs) have been identified in Arabidopsis and there are nine orthologs in rice. Only AtSLAH2 complemented the AtSLAC1-2 mutant when expressed in guard cells. SLAH1, 2 and 3 were expressed in roots, with SLAH1 showing strong expression in the root vascular cylinder, making it a candidate for the S-type anion channels in root xylem parenchyma (Negi et al. 2008).

VDAC Porins

VDAC are beta-barrel porins that can show high conductance and facilitate the permeation of large molecules. They are located to the outer membrane of mitochondria (Clausen et al. 2004; Lee et al. 2009), and expression of VDACs can be affected by drought, salinity and pathogens. They are implicated in ABA signalling (Yan et al. 2009), and are a component of tRNA import to mitochondria (Salinas et al. 2006). A VDAC is located to small vesicles in the periphery of infected cells in nodules, suggesting that they may have even more diverse roles in plants (Wandrey et al. 2004). A salt stress-inducible VDAC was identified in Pennisetum glaucum (Desai et al. 2006). VDAC transcripts were also up-regulated by desiccation, cold treatment and salicylic acid, but not by ABA. Expression in rice at low levels seemed to impart salinity tolerance. A VDAC porin from the outer mitochondrial membrane was down-regulated in leaves of the Cl- accumulating Citrus genotype (Carrizo) compared with constant expression in the excluder (Cleopatra) (Brumos et al. 2009). However, given the diverse roles of VDACs in plants, it would be unlikely these genes would have a direct role in intracellular Cl- homeostasis, and are more likely to have indirect roles as a signal for cellular ion homeostasis (Desai et al. 2006)


There are 10 members of the MSL in Arabidopsis (Haswell 2007). Two members (MSL9 and MSL10) show expression in the root and are located to the plasma membrane. These appear to be Cl- permeable and account for the 10 pA stretch-activated channel that can be observed in Arabidopsis root cortex cells (Haswell et al. 2008). Growth of a quintupple mutant of MSL4, MSL5, MSL6, MSL9 and MSL10 has been examined at high salinity (up to 250 mm NaCl) as well as recovery from shock after transfer from high salt. No differences to wild-type plants were evident, suggesting that at least for high salinity there is no direct involvement of these channels (Haswell et al. 2008).


This review has highlighted the lack of information on Cl- transport in general, and particularly during saline conditions. Na+ is perhaps more widely studied in regards to salt tolerance mainly because more is known about Na+ transport mechanisms compared with Cl-. We have demonstrated that Cl- transport is complex and strongly linked to cation transport, therefore reaffirming the importance of studying anion and cation transport in parallel. Studies such as by Li et al. (2006), which examined both Na+ and Cl- vacuolar transporters in parallel, are more likely to advance our understanding of ion transport mechanisms under salinity.

Plant responses to salinity and Cl- transport processes associated with salt tolerance will vary depending on the species, and even the genotype within a species, as we have shown for Trifolium and L. tenuis (Fig. 3). This highlights the need for genotypic comparisons of ion transport processes at the whole plant, organ and cellular level. More accurate measurements of Cl- concentrations in the vacuole versus cytoplasm, xylem vessels versus xylem parenchyma cells, mesophyll versus epidermal cells, etc., in genotypes that vary in salt tolerance may help identify Cl- transport processes important for tolerance. Understanding which ion transport processes are contributing significantly to salt tolerance will help determine the transport mechanisms of interest for identifying candidate genes for Cl- transport via electrophysiology and molecular biology.

The examples given in this review show that almost anything is possible for Cl- transport mechanisms under saline conditions. Channels can show saturation, and transporters can have high Vmax and can be passive. Even within a transporter family, different members can show passive or active transport. By re-examining some of the dogmas associated with ion transport and measuring selectivity at high external NaCl concentration, it is likely that advances will be made in our understanding of Cl- transport mechanisms under salt stress.

While electrophysiological studies of anion channels are numerous, in contrast, discovery of the corresponding genes lags far behind. This review has proposed several candidate genes for Cl- transporters and channels that may be involved in salt tolerance. Once candidate genes for Cl- transport mechanisms are identified, functional characterization and localization will be critical for determining the physiological role of these genes in salt tolerance.


The authors wish to thank Matt Gilliham and Hank Greenway for helpful comments on the manuscript. The Future Farm Industries CRC provided funding to N.T., and the Australian Research Council provided funding to S.D.T. for this review.


  • 1

    Cl-‘exclusion’ refers to the ability of plants to prevent root uptake of Cl- from the soil and subsequent transport in the xylem to the shoot. In this review, we refer to Cl-‘excluding’ genotypes as those that have relatively low shoot Cl- concentrations, and Cl-‘accumulating’ genotypes that have relatively higher shoot Cl- concentrations. Note that all genotypes must exclude Cl- to some extent.

  • 2

    Channels have high transport rates per functional unit. Transport is passive and when observed with patch clamp, they show discrete opening and closing events corresponding to gated opening and closing of the pore. Transporters refer to carrier-mediated transport that generally have much lower transport rates per active unit compared with channels. Transport may be active or passive. The distinction between a transporter and a channel may be rather grey and there are examples of ion transport that occurs at the boundary between these and members of a gene family that show both types of transport (e.g. CLCs).