Thellungiella halophila, a salt-tolerant relative of Arabidopsis thaliana, has specific root ion-channel features supporting K+/Na+ homeostasis under salinity stress


  • Vadim Volkov,

    Corresponding author
    1. Plant Science Group, IBLS, University of Glasgow, Glasgow G12 8QQ, UK, and
    2. Division of Biological Sciences, University of Paisley, Paisley PA1 2BE, UK
      *(fax +44 141 330 4447; e-mail
    Search for more papers by this author
  • Anna Amtmann

    1. Plant Science Group, IBLS, University of Glasgow, Glasgow G12 8QQ, UK, and
    Search for more papers by this author

*(fax +44 141 330 4447; e-mail


Thellungiella halophila is a salt-tolerant relative of Arabidopsis thaliana with high genetic and morphological similarity. In a saline environment, T. halophila accumulates less sodium and retains more potassium than A. thaliana. Detailed electrophysiological comparison of ion currents in roots of both species showed that, unlike A. thaliana, T. halophila exhibits high potassium/sodium selectivity of the instantaneous current. This current differs in its pharmacological profile from the current through inward- and outward-rectifying K+ channels insofar as it is insensitive to Cs+ and TEA+, but resembles voltage-independent channels of glycophytes as it is inhibited by external Ca2+. Addition of Cs+ and TEA+ to the growth medium confirmed the key role of the instantaneous current in whole-plant sodium accumulation. A negative shift in the reversal potential of the instantaneous current under high-salt conditions was essential for decreasing sodium influx to twofold lower than the corresponding value in A. thaliana. The lower overall sodium permeability of the T. halophila root plasma membrane resulted in a smaller membrane depolarization during salt exposure, thus allowing the cells to maintain their driving force for potassium uptake. Our data provide quantitative evidence that specific features of ion channels lead to superior sodium/potassium homeostasis in a halophyte compared with a closely related glycophyte.


High levels of salinity in the soil negatively affect the growth and development of crops and therefore cause serious problems for world food production (Munns, 2005). However, a diverse range of non-crop plant species (so-called halophytes) can survive and even thrive in highly saline environments. Salt-tolerance strategies in halophytes range from physiological and morphological adaptations to changes in metabolism and protein structure (Flowers et al., 1977; Kirst, 1990; Tester and Davenport, 2003). Unfortunately, most halophytic plants are not amenable to molecular genetics and therefore very few mechanisms leading to salt tolerance in halophytes have been identified at the molecular level. The best-characterized glycophytic model species to date is Arabidopsis thaliana, and several genes linked to salt-stress responses have been identified in A. thaliana (Amtmann et al., 2004; Zhu, 2000). However, because A. thaliana is salt-sensitive, its usefulness for the study of salt tolerance is limited. Over recent years, a close relative of A. thaliana, Thellungiella halophila (synonymous to T. salsuginea; Al-Shebaz et al., 1999) has been developed as a model species for the study of plant adaptation to abiotic stress (Gong et al., 2005; Inan et al., 2004; Taji et al., 2004; Vera-Estrella et al., 2005; Volkov et al., 2004; Wang et al., 2006; Wong et al., 2005). Although highly similar to A. thaliana both in its morphology and in its genome sequence, T. halophila is an ‘extremophile’ that can survive low temperatures, drought and high salt (including salt-shock treatments; Bressan et al., 2001; Inan et al., 2004). Furthermore, it can be easily transformed and has several other characteristics such as short generation time and copious seed production that make it a good candidate for functional genomics analysis, in particular in comparison with A. thaliana (Amtmann et al., 2005; Zhu, 2000).

One of the biggest challenges for a plant growing on high salt is to ensure the uptake of essential mineral nutrients such as K+ while restricting the accumulation of potentially toxic ions such as Na+. We have carried out an extensive analysis of K+ and Na+ levels in A. thaliana and T. halophila under low- and high-salt conditions (Volkov et al., 2004; Wang et al., 2006), and found that (i) during the first 72 h of salt treatment (100 mm), root K+ is lost much more rapidly in A. thaliana than in T. halophila, and (ii) net uptake of Na+ after addition of salt is considerably lower in T. halophila than in A. thaliana both in the short (25–72 h) and long term (6 weeks). We conclude that K+/Na+ homeostasis is indeed a crucial factor for salt tolerance in T. halophila. Three pivotal questions arise from the data. (i) Why does T. halophila accumulate less Na+ than A. thaliana, because of reduced Na+ influx or because of increased Na+ efflux? (ii) Why does A. thaliana lose K+ more rapidly than T. halophila? (iii) Which ion transporters are responsible for K+ and Na+ transport in the two species?

In a recent study (Wang et al., 2006), we used 22Na tracer flux analysis to answer the first question. We found that uni-directional Na+ influx into root cells of T. halophila was less than half the uni-directional Na+ influx into root cells of A. thaliana. Quantitative comparison of uni-directional and net Na+ influx showed that both plant species export approximately 70% of the Na+ taken up back into the external medium. Due to the higher Na+ influx rates in A. thaliana, this implies that the absolute amount of exported Na+ is higher in A. thaliana than in T. halophila, and indeed 22Na efflux experiments confirmed this hypothesis. Much of the recent research into salt tolerance has focused on Na+/H+ antiport systems, in particular the A. thaliana plasma membrane transporter SOS1 (Shi et al., 2000, 2003). Although our results support a vital role of such Na+ export systems for growth of A. thaliana under salt stress, they clearly show that salt tolerance in its halophytic relative is achieved through restriction of Na+ influx rather than enhanced export.

The present study is concerned with the other two questions. Initial whole-cell patch-clamp experiments had revealed that T. halophila root cells exhibit similar types of plasma membrane cation currents as in A. thaliana and other plant species. Thus voltage steps to hyperpolarizing and depolarizing potentials induce time-dependent inward and outward currents respectively, as well as instantaneous currents (Volkov et al., 2004). Interestingly, all of these currents appeared to have a higher K+/Na+ selectivity in T. halophila than A. thaliana, indicating that structural features of root ion channels might underlie differential ion accumulation in the two species. However, as this view is based on data obtained by various laboratories using different experimental approaches (e.g. reversal potential versus relative currents), this issue required more detailed analysis. Furthermore, the relative importance of the different current types for K+ and Na+ uptake in T. halophila remained unknown. In this paper, we present a detailed analysis of all three types of current in T. halophila with respect to their relative selectivity for a wide range of cations, voltage dependence for activation, and sensitivity to inhibitors. Important features of instantaneous currents were measured in direct comparison between T. halophila and A. thaliana. Combination of these data with data on the inhibitor sensitivity of whole-plant K+ and Na+ contents and membrane potentials allowed us to identify the currents underlying K+ and Na+ transport and to carry out a quantitative evaluation of the impact of species-specific current–voltage relationships on whole-plant Na+ uptake and K+ loss.


Time-dependent inward currents

Hyperpolarization of the T. halophila root plasma membrane activated time-dependent inward currents indicative of the presence of inward-rectifying channels (Figure 1). In symmetric 100 mm KCl/100 mm KCl (pipette/bath) solutions, a time-dependent inward current was detected in 76% of protoplasts at −200 mV (n = 25). Activation kinetics could be satisfactorily fitted with a single exponential with τ = 320 ± 31 msec for −200 mV and τ = 303 ± 33 msec for −180 mV (n = 4, mean ± SE). From the voltage dependence of the conductance, a gating charge, z, of 1.9 ± 0.2, and a half-activation potential, V1/2, of −191 ± 1 mV were determined (Figure 1b). Time-dependent inward currents at −160 mV disappeared when the bath solution was changed from 100 mm KCl to 100 mm NaCl (n = 33), RbCl (n = 5), LiCl (n = 5) or CsCl (n = 5). At −200 mV, none of the protoplasts (n = 9) exhibited time-dependent inward current in 100 mm NaCl, whereas currents were still observed in 37.5% of protoplasts (n = 8) with 10 mm KCl in the bath. The difference in the appearance of current in the two solutions was significant (P < 0.05). This indicates that inward-rectifying channels in T. halophila root cells have a >10-fold higher permeability for K+ than for Na+. The current was completely blocked by adding 5 mm CsCl to the bath solution (100 mm KCl, n = 4, Figure 1c) and reinstated after Cs+ washout.

Figure 1.

 Activation and inhibition of time-dependent inward currents in T. halophila root protoplasts.
(a) Control.
(b) Conductance–voltage relationship.
(c) Block by Cs+.
(d) No inhibition by cytoplasmic Na+.
Concentrations of K+, Cs+ and Na+ in the pipette/bath solutions were as follows (other ions as given in Experimental procedures): 100 mm KCl/100 mm KCl (a, b), 100 mm KCl/95 mm KCl + 5 mm CsCl (c), 100 mm KCl + 10 mm NaCl/50 mm KCl + 50 mm NaCl (d). Protoplast diameter was 12 μm (a, c) or 17 μm (d). The holding potential was −70 mV (a, c) or −80 mV (d). The voltage was clamped in steps of 20 mV between +80 and −200 mV. Currents shown in (d) were recorded 1 h after obtaining the whole cell-configuration. Data points in (b) are the means of four measurements, error bars are SE. The dotted curve is a fit of the Bolzmann equation G/Gmax = 1/(1 + exp ((Vm − V1/2)/S)).

It was recently reported that inward-rectifying currents in roots of A. thaliana seedlings are inhibited by cytoplasmic Na+ (Qi and Spalding, 2004), and this inhibition was proposed to be the reason for the salt-induced K+-deficient phenotype of SOS1 mutants, which lack the plasma membrane Na+/H+ efflux system (Wu et al., 1996). In T. halophila root protoplasts, time-dependent inward currents were still considerable when 10 mm NaCl was present at the cytoplasmic side (n = 17, Figure 1d). We were interested in whether this could be an important difference between T. halophila and A. thaliana, and repeated the experiments with A. thaliana. Surprisingly, protoplasts isolated from the roots of mature A. thaliana plants did not show inhibition of the K+ inward rectifier by cytoplasmic Na+. A literature search revealed that inward-rectifying K+ currents were also recorded after heterologous expression of AKT1 in insect cells, although considerable amounts of Na+ were present in the pipette (Gaymard et al., 1996). We conclude that (i) Na+ inhibition of inward-rectifying K+ currents is not a direct effect of Na+ on AKT1, and (ii) Na+ inhibition of inward-rectifying K+ currents does not usually occur in mature roots of A. thaliana or T. halophila, and therefore might be specific for the seedling stage.

Time-dependent outward currents

Depolarization of the T. halophila root plasma membrane activated time-dependent outward currents indicative of outward-rectifying channels (Figures 1 and 2). Time-dependent outward currents (at +60 mV) were observed in 95% of the protoplasts explored (n = 93). The activation kinetics of the time-dependent outward current were voltage-dependent and also depended on the cation concentration in the bath. Activation became faster with decreasing K+ concentrations or replacement of K+ by Na+ and Li+ (Figure 2a). The half-activation time T1/2 in Na+ and Li+ solutions was 2–3 times shorter than T1/2 in the presence of larger ions such as K+, Rb+ and Cs+ (Figure 2c).

Figure 2.

 Activation and inhibition of time-dependent outward currents in T. halophila root protoplasts.
(a) Fast activation with 100 mm external Na+.
(b) Block by TEA+.
(c) Half-activation times of time-dependent outward currents in various bath solutions. Concentrations of K+, Na+ and TEA+ in the pipette/bath solutions were as follows (other ions as given in Experimental procedures): (a)100 mm KCl/100 mm NaCl; (b) 100 mm KCl/80 mm NaCl + 20 mm TEACl; (c) As indicated in the figure. Protoplast diameter was 17 μm. The holding potential was −70 mV. The voltage was clamped in steps of 20 mV between +80 and −60 mV. Half times (T1/2) of activation are presented for a range of clamped voltages. For all ions except Li+, T1/2 values are the means of four experiments ± SE. With LiCl in the bath, current activation was so fast that T1/2 could only be resolved in one of four experiments.

The time-dependent outward current was blocked by TEA+ (Figure 2b). Inhibition by TEA+ was weakly voltage-dependent, being 98 ± 1% at 20 mV and 90 ± 4% at 80 mV (20 mm TEACl, means ± SE, n = 6).

Time-dependent outward currents were highly selective for K+ over Cl. When 100 mm KCl in the bath was replaced by 10 mm KCl (n = 11), the reversal potential shifted by −48 ± 1 mV, indicating a Cl/K+ selectivity of <0.015. Selectivity of the outward-rectifying current for K+ over monovalent cations was also high and followed a permeability sequence of K+ > Rb+ > Cs+ > NHinline image > Li+ > Na+ (Table 1). For small cations (Li+ and Na+), inactivation was very fast and therefore tail currents became difficult to resolve. The values given in Table 1 are the minimal estimates for K+/cation selectivity. Addition of K+ slowed down the inactivation, which allowed us to confirm the K+/Na+ selectivity given in Table 1 by using various combinations of K+ and Na+ in the bath.

Table 1.   Relative selectivity (K+/ion) of instantaneous and outward-rectifying currents in root protoplasts of T. halophila
CurrentMethodLi+Na+Rb+Cs+NHinline imageTEA+
  1. Data are given as mean ± SE (for reversal potential, n = 4 for Rb+ and TEA+, n = 5 for other ions; for conductance, n = 3 for Li+ and n = 4 for other ions). Erev, reversal potential; Grel, relative conductance at V < −80 mV. ND, not determined.

Instantaneous currentErev9.1 ± 3.06.7 ± 0.81.5 ± 0.12.3 ± 0.16.7 ± 2.78.9 ± 1.8
Grel1.8 ± 0.21.7 ± 0.31.7 ± 0.51.4 ± 0.2ND3.4 ± 0.9
Time-dependent outward currentErev9.7 ± 0.913.1 ± 4.01.9 ± 0.13.8 ± 0.47.6 ± 1.4ND

Activity of the outward rectifier at voltages more negative than the reversal potential Erev has been reported for A. thaliana root cells (Maathuis and Sanders, 1995). We found that, in T. halophila, the activation potential (Eact) depended on the external cation. When K+ was present in the bath solution, time-dependent outward currents activated at voltages approximately 60 mV below Erev, thus allowing a net influx of cations between Eact and Erev, which is apparent as the remainder of the steady-state inward current after subtracting the instantaneous current from the total steady-state current (Figure 3a). Reducing the external K+ concentration resulted in a negative shift of Eact, which always remained negative with respect to Erev, although the amount of inward current decreased (Figure 3b). However, when external K+ was replaced with other cations, the Eact of the time-dependent outward current was close to Erev, thus preventing a net influx of cations into the cell. For example, no steady-state inward current was observed with 100 mm external Rb+ (Figure 3b), although the reversal potential for 100 mm Rb+/100 mm K+ is more positive than that for 30 mm K+/100 mm K+, and the respective inward current through a given number of open channels should be greater than the one produced by 30 mm K+ (Table 1). A solution of 100 mm Na+ also failed to exhibit a steady-state inward current through the outward rectifier (Figure 3a). Because of the low permeability of the outward rectifier for Na+ (Table 1), we further investigated this issue at the level of single channels.

Figure 3.

 Current–voltage curves of inward currents through outward-rectifying channels in root protoplasts of T. halophila in various external (bath) solutions.
(a) Whole-cell I–V curves obtained with 100 mm KCl or 100 mm NaCl in the bath. Data are given as means ± SE (n = 7). Note the appearance (absence) of an inward current in the presence of K+ (Na+).
(b) Whole-cell I–V curves obtained with 100 mm RbCl or 30 mm KCl in the bath. Data are given as means ± SE (n = 5). Note the appearance (absence) of an inward current in the presence of K+ (Rb+). The pipette solution was always 100 mm KCl. Instantaneous current was subtracted.
(c) Single-channel I–V plots for outward-rectifying channels in various external (bath) solutions. The pipette solution was always 100 mm KCl. Data points are means of 12 independent measurements (n = 7 for 10 mm KCl) from three protoplasts (two for 10 mm KCl). Error bars are SD.

Single-channel recordings of outward-rectifying channels in T. halophila in outside-out patches had a unitary conductance of 54 pS in symmetric 100/100 mm KCl (Figure 3c), a value that agrees with that reported for A. thaliana root cells (Maathuis and Sanders, 1995), and confirmed inward movement of K+ through typically flickering openings of outward-rectifying channels (Figure 4a). Channel openings below Erev were observed with 10 mm external K+ (Figures 3c and 4b) but not with 100 mm external Na+ although the respective open channel currents should be similar (Figures 3c and 4c).

Figure 4.

 Single-channel recordings in outside-out patches of T. halophila root protoplasts.
The pipette solution was 100 mm KCl. The bath solution contained 100 mm KCl (a), 10 mm KCl (b) or 100 mm NaCl (c). c, current level with no open channels; o, current levels of single-channel openings. Spiky openings of outward-rectifying channels allowing inward current are indicated with asterisks. Openings of a second type of channel are indicated with crosses.

Instantaneous currents

Instantaneous currents were measured at 20–40 msec after application of the voltage pulse. This type of current is likely to be mediated by channels that are voltage-independent in their gating. Selectivity of the instantaneous current for K+ over Cl was high (over 50) according to a shift of reversal potential by −41 ± 2 mV when 100 mm KCl in bath was substituted with 10 mm KCl (n = 11). Current–voltage (I–V) curves of the instantaneous current in 100 mm KCl/100 mm KCl and 100 mm KCl/100 mm NaCl (pipette/bath) are shown in Figure 5. Changing the bath solution from 100 mm KCl to 100 mm NaCl resulted in a shift of the reversal potential by −48 ± 4 mV (n = 13), corresponding to a Na+/K+ permeability ratio of 0.16 ± 0.02. Surprisingly, the shift in the reversal potential was not mirrored by a corresponding change in conductance. The relative Na+/K+ conductance (measured between −80 and −140 mV) was 0.78  ± 0.09 (n = 13). Permeability ratios and the relative conductance of the instantaneous current for a number of monovalent cations are given in Table 1. Interestingly, the conductance of the channel was very similar for all cations (except TEA+) but the voltage required to allow inward movement of the ions (Erev) differed. As published data for the cation permeability of voltage-independent channels in A. thaliana are based on relative conductance only (Demidchik and Tester, 2002), we repeated our experiments with root protoplasts from A. thaliana. Under the same conditions as used for T. halophila, the relative Na+/K+ conductance of instantaneous currents in A. thaliana protoplasts was 0.72 ± 0.19. This value agrees with published data (Demidchik and Tester, 2002), and is close to the respective value in T. halophila. However, in A. thaliana root protoplasts, we observed only a very small shift in the reversal potential of instantaneous currents when K+ in the bath was replaced with Na+ (−16 ± 2 mV, n = 6), corresponding to a Na+/K+ permeability ratio of 0.54 ± 0.03 (Figure 7). Hence, the discrepancy between the shift in the reversal potential and relative conductance was specific for T. halophila.

Figure 5.

 Whole-cell current–voltage curves for instantaneous currents in root protoplasts of T. halophila.
The bath solution contained 100 KCl (open circles) or 100 NaCl (closed circles). The pipette solution was always 100 mm KCl. Data are given as means ± SE (n = 13).

Figure 7.

 Comparison of current–voltage curves for whole-cell instantaneous currents in root protoplasts of A. thaliana (open symbols) and T. halophila (closed symbols).
Currents are normalized to protoplast surface. The pipette solution was always 100 mm KCl. The bath solution was 100 KCl (a) or 100 NaCl (b). Data are given as means ± SE (n = 6 for A. thaliana, n = 13 for T. halophila).

We were concerned that fast activation of the outward rectifier in external Na+ (see above) contributed to the instantaneous current and led to the observed shift in Erev. We tested this by adding TEA+ to the external Na+ solution, which completely blocked the time-dependent outward current (Figure 6). The treatment did indeed lead to a decrease in the instantaneous outward current, but the reversal potential remained negative (Figure 6). Selectivity for K+ over Na+ was thus confirmed as a distinctive feature of the instantaneous current of T. halophila root cells.

Figure 6.

 .Current–voltage curves for instantaneous and total currents in T. halophila root protoplast with and without TEA+. Concentrations of K+, Na+ and TEA+ in the bath solutions are given in the figure (other ions as given in Experimental procedures). The pipette solution was always 100 mm KCl. Filled symbols are for total steady-state currents, open symbols indicate the instantaneous current component only. Inset shows magnified part of the current-voltage curves.

Absolute instantaneous whole-cell currents were small (typically <50 pA at −160 mV). Instantaneous current normalized to membrane surface was compared between root protoplasts of A. thaliana and T. halophila. The plants were grown to similar sizes, and protoplasts were isolated using the same procedure. When the bath solution was 100 mm KCl, the overall conductance of the instantaneous current was more than twice as high in A. thaliana as in T. halophila (Figure 7a). This could be due to higher channel abundance or to higher single-channel conductance. After changing the bath to 100 mm NaCl, the difference in conductance was smaller than expected from the difference in Erev (see above), but, due to the negative shift of Erev in T. halophila, the inward current was still much smaller than in A. thaliana at any given voltage (Figure 7b).

Several inhibitors and regulators of cation channels in A. thaliana were tested on instantaneous currents in T. halophila root protoplasts (Table 2). We failed to find reproducible effects of glutamate (1 mm, n = 9) or 8-Br-cGMP, a membrane-permeable analogue of cGMP (50 μm, n = 6). Sometimes a decrease in current was observed after adding the latter, but this coincided with a general current rundown in the course of the experiment and was not reversed after washout. The effect of TEA+ on the instantaneous outward current has been described above (Figure 6). The effect of TEA+ on the inward component of the instantaneous current depended on the permeating ion. When TEA+ was added to a background of K+, the instantaneous inward current decreased slightly (probably due to some contribution of the outward rectifier to the current), but no inhibition occurred when TEA+ was added to a background of Na+ (see Figure 6, inset). Other monovalent cations, in particular Cs+, which strongly inhibited the inward rectifier, did not inhibit the instantaneous current.

Table 2.   Inhibitory profiles of instantaneous currents in A. thaliana and T. halophila
 A. thalianaT. halophila
  1. aDemidchik and Tester (2002).

  2. bMaathuis and Sanders (2001).

  3. cDemidchik et al. (2004).

TEA+No effectaNo effect
Cs+No effectaNo effect
8-Br-cGMPInhibitionbNo effect
GlutamateStimulationcNo effect
Zn2+ (1 mm)Complete blockaInhibition
Ba2+ (1 mm)Complete blockaInhibition
Ca2+ (0.1–1 mm)InhibitionaInhibition

We also tested the effect of divalent ions on the instantaneous current. Addition of 1 mm ZnCl2 to the bath solution resulted in a slight decrease in the instantaneous current (n = 5). Washout led to partial or complete recovery of the current. Addition of 1–2 mm BaCl2 also inhibited the instantaneous current (n = 3). Lowering external Ca2+ from 1 to 0.3 mm in a background of 100 mm KCl resulted in an increase of instantaneous current by 50–100% (n = 4). However, within several minutes after lowering the external Ca2+ concentration, the current dropped to initial or even lower values. Subsequent re-supply of external Ca2+ (1 mm) further reduced the current. In 0.1 mm external Ca2+, instantaneous currents were larger than in 0.3 mm Ca2+, but patches became very unstable. Thus the instantaneous current in T. halophila is clearly inhibited by external Ca2+, but current rundown and instability of patches at low Ca2+ prevented us from quantifying this effect.

We conclude that the major part of the instantaneous current in T. halophila is created by a K+-selective Ca2+-sensitive pathway that can be clearly distinguished from voltage-gated inward- and outward-rectifying channels by its lack of sensitivity to TEA+ and Cs+.

Regulatory profile of root and shoot ion concentrations

To identify which of the measured currents is responsible for net uptake of K+ and Na+, we measured the effect of channel inhibitors on root and shoot K+ and Na+ concentrations in T. halophila. When plants were exposed to 100 mm NaCl for 2 days, K+ concentrations of the roots decreased (Figure 8a). Addition of TEA+ (20 mm) or Cs+ (5 mm) reduced root K+ concentrations both under control and saline conditions, with Cs+ having a stronger effect than TEA+. K+ concentrations of the shoots also decreased in saline medium, but no significant effect of TEA+ or Cs+ was detected (Figure 8b). Na+ concentrations in both roots and shoots strongly increased after addition of NaCl to the medium (Figure 8c,d). Neither TEA+ nor Cs+ inhibited Na+ accumulation (Figure 8c,d). In salt-treated roots, Na+ concentrations even increased when 5 mm Cs+ was added to the medium (Figure 8c).

Figure 8.

 Effect of channel blockers on K+ and Na+ contents in roots and shoots of T. halophila.
(a, b) K+ contents in roots (a) and shoots (b). (c, d) Na+ contents in roots (c) and shoots (d). Plants were incubated for 2 days in control medium without or with 100 mm NaCl added (‘control’, ‘+Na+’). 20 mm TEACl or 5 mm CsCl were added to control medium (‘+TEA’, ‘+Cs’) or control medium with 100 mm NaCl (‘+Na + TEA’, ‘+Na + Cs’). Data are given as means ± SE (n = 4). Different letters indicate significant differences at P < 0.05.

Membrane potential

To relate the measured currents to the in vivo Na+ influx, we measured free-running membrane potentials in roots of A. thaliana and T. halophila. In a 1 mm KCl solution, root membrane potentials were not significantly different between the two species (Table 3), ranging between −60 and −205 mV in T. halophila and between −77 and −193 mV in A. thaliana. In both species, addition of 100 mm NaCl led to membrane depolarization, but the shift of membrane voltage was significantly smaller in T. halophila than in A. thaliana (37 ± 17 mV in T. halophila and 63 ± 27 mV in A. thaliana, n = 6, Table 3). During successive treatments with 100 mm NaCl, T. halophila cells recovered membrane potentials close to those recorded under low-salt conditions. By contrast, A. thaliana cells continued to depolarize and did not recover after subsequent washout of NaCl (Table 3).

Table 3.   Membrane potentials in roots cells of T. halophila and A. thaliana before and after application of 100 mm NaCl in a background of 1 mm KCl. Data are given as mean ± SD
 A. thalianaT. halophila
  1. *Significant difference at P < 0.05.

  2. **Significant difference at P < 0.01.

Membrane potential (in mV) before salt treatment−105 ± 44 (n = 6)−119 ± 56 (n = 6)
Membrane potential (in mV) after salt treatment−42 ± 51 (n = 6)−82 ± 46 (n = 6)
Depolarization (in mV) in response to salt treatment63 ± 26 (n = 6)*37 ± 17 (n = 6)*
Membrane potential (in mV) after several successive salt treatments−25 ± 13 (n = 5)**−138 ± 62 (n = 6)**


Identification of transport pathways mediating K+ and Na+ uptake in T. halophila

Our patch-clamp experiments identified three distinct types of ion currents in T. halophila root cells: time-dependent inward currents, which were highly selective for K+ over Na+ and blocked by Cs+, time-dependent outward currents, which were moderately selective for K+ over Na+ and inhibited by TEA+, and instantaneous currents, which were moderately selective for K+ over Na+ and not inhibited by Cs+ or TEA+. All three currents mediated the uptake of K+, albeit over different voltage ranges. Inward-rectifying channels allow increasingly large K+ influx at membrane potentials more negative than −140 mV (Figure 1). Instantaneous currents facilitate moderate K+ influx at any voltage more negative than EK (Figure 7), whereas outward-rectifying channels provide a pathway for low K+ influx within a limited range of voltages just below EK (Figure 3). Both outward-rectifying and instantaneous currents mediate efflux of K+ at voltages above EK.

All three current types in T. halophila root cells provide a very limited pathway for Na+ uptake. Selectivity for K+ over Na+, a typical feature of voltage-dependent K+ channels, was also found for the instantaneous current, which constitutes the only pathway for Na+ uptake at physiological membrane potentials. Na+ currents through inward-rectifying channels are negligible even at very negative voltages, and Na+ uptake through outward-rectifying channels is prevented at voltages below EK due to the specific gating properties of this channel type. We found that both activation potential and activation kinetics (Eact) of the outward rectifier differed for K+ and Na+. K+-dependent gating had previously been reported for the outward rectifier of Vicia faba guard cells, and was explained with an allosteric model assuming a binding site for external K+ that is distinct from the pore (Blatt and Gradmann, 1997). Our results indicate that, in T. halophila, the binding site differs in its relative cation selectivity.

Measurements of K+ and Na+ contents under low- and high-salt conditions showed that K+ accumulation in roots of T. halophila was reduced by Cs+, whereas Na+ contents were not decreased by Cs+ or TEA+ (the observed increase of Na+ uptake in the presence of Cs+ could be due to membrane hyperpolarization caused by inhibition of the K+ inward rectifier). From comparison of the inhibitory profiles of K+ and Na+ concentrations with those of the whole-cell cation currents, we conclude that K+ uptake is predominantly through inward-rectifying channels, whereas Na+ uptake is mediated by the instantaneous current. The same conclusion has been reached for A. thaliana (Demidchik and Tester, 2002; Hirsch et al., 1998; Maathuis and Sanders, 2001) and other plant species (Amtmann and Sanders, 1999; Davenport and Tester, 2000; Tyerman and Skerrett, 1999; White, 1997).

Instantaneous currents in T. halophila and A. thaliana differ in two important aspects. Firstly, the overall amount of current (per surface) is smaller in T. halophila than in A. thaliana. From whole-cell experiments, it cannot be decided whether this is due to lower abundance or lower conductance of the underlying individual channels. Secondly, instantaneous currents in T. halophila are selective for K+ over other cations, especially Na+. This marks an important difference, not just compared with A. thaliana, but also compared with several glycophytic crop species including wheat, barley and maize, where the current is characterized by a general lack of selectivity among cations (Maathuis and Sanders, 2001; Roberts and Tester, 1997; Tyerman et al., 1997). Single-channel measurements in the glycophytic plants have revealed that the instantaneous current is carried by ion channels, which are voltage-independent in their gating (‘voltage-independent channels’). Excised patches of T. halophila showed channel openings that clearly differed in their gating properties from both inward- and outward-rectifying channels (see, for example, Figure 4a). The activity of these channels disappeared quickly, indicating the requirement for an unidentified positive regulator, which is lost during patch excision. Due to this ‘rundown’, it was not possible to establish the ion selectivity of this channel. Nevertheless, the similarity of the pharmacological profiles of the instantaneous currents in T. halophila and the glycophytic species, e.g. insensitivity to Cs+ and TEA+ and inhibition by external Ca2+, suggests that similar types of channels produce these currents in the two species. None of the other transporters that are potential Na+ uptake pathways in plant root cells (e.g. HKT1 or members of the KUP/HAK family) have been reported to be Ca2+-sensitive. Our finding that uni-directional Na+ fluxes into roots of both T. halophila and A. thaliana are inhibited by external Ca2+ (Essah et al., 2003; Wang et al., 2006) therefore supports a role of voltage-independent channels in Na+ uptake in both species. Differential abundance (or conductance) and selectivity of voltage-independent channels therefore emerge as prime factors for explaining the differences in Na+ accumulation and salt tolerance between the two species (Inan et al., 2004; Volkov et al., 2004; Wang et al., 2006).

K+/Na+ homeostasis in A. thaliana and T. halophila

Instantaneous currents in T. halophila showed a complex I–V relationship in conditions of high cytoplasmic K+ and high external Na+ concentrations. Replacement of external K+ with equimolar external Na+ shifted the current reversal potential towards the negative, indicating selectivity of the channel for K+ over Na+. However, the inward conductance did not decrease to the extent expected from the Goldman–Hodgkin–Katz (GHK) equation. This finding indicates that ion movement through the T. halophila channel is not independent and therefore does not obey the GHK equation. Interestingly, ‘background’ currents in the plasma membrane of the salt-tolerant alga Lamprothamnium papulosum exhibit the opposite behaviour; upon increasing external salinity, they increase in conductance without a changing reversal potential (Beilby and Sheperd, 2001). Inter-dependency of permeating ions has been described for several channel types, and can even occur in very large pores (Alcaraz et al., 2004). The simplest interpretation of our data is that the voltage-independent channel in T. halophila has two protein–ion interaction sites that differ in their relative affinity for different cations (single-file multi-ion model, Hille, 2001), the first one discriminating only weakly between the cations, the second having a much higher affinity for K+, Rb+ and Cs+ than for Na+, Li+ or NHinline image (‘selectivity filter’). Only if additional driving force is applied in the form of a concentration gradient or a negative membrane potential can the latter cations out-compete K+ at the second site and pass through the channel. From a physiological point of view, the relatively high conductance for Na+ means that the observed difference in Erev of the instantaneous current between T. halophila and A. thaliana does not automatically imply that T. halophila takes up less Na+ than A. thaliana. Determination of root cell membrane potentials has provided us with the ‘missing link’ required for quantitative assessment of the impact of root cell current–voltage relationships on root Na+ uptake. After addition of 100 mm NaCl, membrane potentials depolarized to values around −80 mV in T. halophila and −40 mV in A. thaliana. The fact that the Na+-induced depolarization was smaller in T. halophila than in A. thaliana is indicative of lower Na+ permeability of the root plasma membrane in T. halophila, thus confirming the results obtained by patch clamp. A more negative membrane potential allows root cells of T. halophila to maintain the driving force for electrogenic nutrient uptake and to prevent the extensive loss of K+ that occurs in A. thaliana (Wang et al., 2006). ‘Unfortunately’, it also increases the driving force for Na+ influx. However, excessive Na+ uptake in T. halophila in high salt is prevented by the observed negative shift of the current reversal potential. According to the I–V curves shown in Figure 7(b), the instantaneous Na+ inward current in 100 mm external NaCl amounts to 14 mA m−2 at −80 mV in T. halophila, and is therefore only half the size of the Na+ inward current at −40 mV in A. thaliana (28 mA m−2). This finding agrees qualitatively with previous results from 22Na tracer fluxes, which showed that uni-directional influx of Na+ into the roots of salt-treated plants was twofold greater in A. thaliana than in T. halophila (Wang et al., 2006). Taking into account the average size of root protoplasts, the current observed in T. halophila at −80 mV (14 mA m−2) also corresponds quantitatively to the measured fluxes (0.1 μmol Na+/min.g FW) for a similar external Ca concentration; Wang et al., 2006) if 3% of the total root cell surface contributes to Na+ uptake, which is a reasonable assumption (see Experimental procedures for calculations).

This exercise shows that, although additional layers of complexity are likely to be involved in the plant responses to salt stress, the intrinsic electrical properties of the root cell plasma membrane are sufficient to explain the observed differences in Na+ influx between the two species. Moreover, we have previously shown that the differences in uni-directional Na+ influx are sufficient to explain the differences in whole-plant Na+ accumulation between the two species (Wang et al., 2006). Hence, a direct link between root ion currents and whole-plant ion accumulation has been established (Figure 9). Any additional complexity (e.g. transcriptional and post-translational regulation of ion transporters) is likely to concern the fine-tuning rather than the basic mechanism for K+/Na+ homeostasis in T. halophila.

Figure 9.

 Diagram summarizing the properties of voltage-independent channels in A. thaliana and T. halophila and their effects on membrane potential, uni-directional fluxes in root cells and whole-plant K+/Na+ homeostasis.
Data supporting the upper part of the scheme are provided by this study; data for the lower part are provided by Wang et al. (2006).

It is clear that the above considerations based on average values do not fully account for the variation between different cell types in the root tissue. Membrane potentials varied considerably between the individual experiments, and in many cases were too positive to allow a function of inward-rectifying channels in K+ uptake. This observation provides a possible explanation for the physiological role of voltage-independent channels. Although whole-plant K+ uptake is to a large extent mediated by inward-rectifying channels, as reflected in its sensitivity to Cs+, it is likely that voltage-independent channels also contribute to K+ influx. Their role in K+ uptake might be particularly important when cells are depolarized to voltages at which the inward rectifier is inactive, e.g. in saline conditions. We conclude that the combination of voltage-independent gating and selectivity for K+ over Na+ of this particular class of root cation channels in T. halophila supports K+ uptake while restricting Na+ uptake in a saline environment.

Conclusions and outlook

The data presented here provide experimental proof that specific features of ion currents across the plasma membranes of root cells lead to superior K+/Na+ homeostasis in a salt-tolerant species. We were able to obtain this proof by combining results from patch-clamp experiments with information on membrane potentials, whole-plant Na+ uptake and uni-directional 22Na fluxes in direct comparison between T. halophila and A. thaliana, two closely related species that differ in their salt tolerance. These findings not only put Na+ uptake pathways back into the spotlight of salt tolerance research (Amtmann and Sanders, 1999; Tyerman and Skerrett, 1999; White, 1999), but also have important implications for future studies into the genetic basis of toxic Na+ accumulation in glycophytic species. Due to its low background of root Na+ uptake, T. halophila can now be used as an experimental system to screen for increased Na+ uptake after expression of putative Na+ transporters as well as regulatory proteins from glycophytic species. This approach is particularly promising as heterologous expression systems such as Xenopus oocytes have often failed to produce functional channels from transcripts encoding cyclic nucleotide-gated channels (CNGCs) and glutamate receptor-like channels, which are prime candidates for Na+ uptake channels in A. thaliana (Davenport, 2002; Maathuis and Sanders, 2001). Furthermore, gene swap between A. thaliana and T. halophila allows functional analysis of potential salt tolerance determinants in a whole-plant context. Finally, future structure–function analysis of orthologous channel genes from T. halophila and A. thaliana based on sequence comparison and site-directed mutagenesis will increase our understanding of structural features underlying differential K+/Na+ selectivity in plant ion channels.

Experimental procedures

Plant material

T. halophila (Shandong) and A. thaliana (Columbia) plants were grown under hydroponic conditions at 24°C on a controlled 14 h/10 h or 10 h/14 h (150 or 200 μE m−2 sec−1 light) day/night cycle, respectively (Volkov et al., 2004). These conditions produce plants that are comparable in their developmental programme, i.e. exhibiting large leaf rosettes and delayed flowering. The basic nutrient solution contained 1.25 mm KNO3, 0.5 mm Ca(NO3)2, 0.5 mm MgSO4, 0.625 mm KH2PO4 and micronutrients (Arteca and Arteca, 2000). For salt treatment, 100 mm NaCl was added to the medium.

Protoplast isolation

Root protoplasts were isolated from 6 to 8-week-old plants following a procedure developed for A. thaliana roots (Demidchik and Tester, 2002). About 200 mg of chopped root pieces (1–2 mm long) were digested in 2 ml of enzyme solution containing 1.5% cellulase Onozuka RS (Yakult Honsha, Tokyo, Japan), 1% cellulysin (CN Biosciences, Nottingham, UK), 0.1% pectolyase Y-23 (Kikkoman Co., Noda City, Japan), 0.1% bovine serum albumin (Sigma, St. Louis, MO, USA), 10 mm KCl, 10 mm CaCl2, 2 mm MgCl2 and 2 mm MES (2-(N-morpholino)-ethanesulfonic acid). The pH was 5.7 (adjusted using Tris-base), and the osmolarity was adjusted with sorbitol to 350 mOsm. Roots were incubated on a shaker (60 rpm) for 30–50 min at 28°C. The digested tissue was gently washed several times with ice-cold ‘storage solution’ containing 10 mm KCl, 2 mm CaCl2, 2 mm MgCl2, 2 mm MES, pH 5.7 (Tris), 350 mOsm (sorbitol). Protoplasts were extracted by squeezing the digested tissue through a 20 μm mesh into 2–3 ml of ‘storage solution’. This method produced two populations of protoplasts differing in size and morphology. Comparison with T. halophila root cross-sections (see for example Figure 9 in Inan et al., 2004) suggests that the larger protoplasts (13–20 μm in diameter) with a central vacuole originate from the root cortex and epidermis, whereas the smaller (<10 μm), denser ones originate from the stele. Only the former type of protoplast was used in this study.

Patch clamp

Protoplasts were patch-clamped using standard techniques (Amtmann et al., 1997). Patch-clamp pipettes were pulled on a vertical electrode puller (PP-83, Narishige, Tokyo, Japan) from glass capillaries (Kimax 51, Kimble Products, Vineland, NJ, USA). Pipettes were filled with pipette solution containing 100 mm KCl (8–10 mm NaCl was added in some experiments), 1 mm CaCl2, 1 mm MgCl2, 2–2.5 mm EGTA, 2 mm Mg-ATP, 2 mm HEPES, pH 7.0 (Tris). The free calcium concentration was 200–500 nm. Final pipette resistances were around 10 MΩ. The sealing solution contained 20 mm CaCl2, 2 mm MES, pH 5.7 (Tris), 350 mOsm (sorbitol). After seal formation, the sealing solution in the bath was replaced by experimental solutions containing 1 mm CaCl2, 1 mm MgCl2, 2 mm MES, pH 5.7 (Tris), 350 mOsm (sorbitol) and various amounts of Cl salts of K+, Na+ and other cations. All solutions were sterile-filtered. Most experiments were carried out in the whole-cell mode, where the bath represented the external solution and the pipette solution the cytoplasmic compartment. The reference agar bridge contained 100 mm KCl. Liquid junction potentials (which were not more than 10 mV) were measured and corrected for as described by Amtmann and Sanders (1997). Experiments were carried out at room temperature (22–25°C) with a bath perfusion rate of 0.14 ml min−1. Currents were recorded and processed using a standard patch-clamp amplifier L/M-EPC7 (Heka Elektronik, Lambrecht/Pfalz, Germany), ITC-18 digitizer (Computer Interface Instrutech Corporation, Port Washington, NY, USA) and ‘Pulse + PulseFit’ software, version 8.53 (Heka Elektronik). Data were online low-pass-filtered at 0.2–1 kHz using an eight-pole Butterworth filter (KEMO Ltd, Beckenham, Kent, UK) and sampled at 5–10 kHz. Holding potentials were −70 to −100 mV. Statistical analysis was performed using standard software packages (EXCEL 2000 for Windows, Microsoft and SigmaPlot for Windows, SPSS Science, Chicago, IL, USA) as well as Panda II and Henry II (Y-Science, Glasgow, UK; Selectivity ratios were determined from reversal potentials in the various bath solutions and from the relative conductance. For kinetic analysis, a linear regression fit of the instantaneous current was subtracted from the total current.

Membrane potential

The apical 3–4 cm of the main root of a 4-week-old T. halophila or A. thaliana plant were fixed at both ends to the bottom of a purpose-built chamber using 1.5% agar. The remaining major part of the root system was placed into a Petri dish containing the nutrient solution used for plant growth. The fixed root section was bathed in a solution with 1 mm CaCl2, 1 mm MES, buffered by Tris to pH 5.7, supplemented with 0.1 mm, 1 mm or 10 mm KCl. For salt treatment, 100 mm NaCl was added to the 1 mm KCl solution. After a root cell was impaled and stable voltage readings had been recorded, bath solutions with various KCl concentrations were applied to measure the response of the membrane potential to external K+. Subsequently salt treatment was given for 20–30 min, followed by a return to bath solution without NaCl. The cycle was repeated several times. All operations were carried out on a Zeiss Axiovert microscope (Zeiss, Jena, Germany) with 63× LWD DIC (long working distance, differential interference contrast) optics. Measurements were carried out in a continuously flowing solution at around 20 chamber volumes min−1 (Grabov and Blatt, 1998). Recordings were obtained with a microelectrode filled with 200 mm K+ acetate, pH 7.5, to minimize Cl leakage (Blatt and Armstrong, 1993). Connection to the amplifier headstage was via 1 M KCl Ag/AgCl half cells, and a matching half cell and 0.1 M KCl–agar bridge served as the reference electrode

Determination of tissue ion contents

To investigate the effect of channel blockers on tissue ion contents, 4-week-old T. halophila plants were transferred to nutrient solution supplemented with 5 mm CsCl or 20 mm TEACl without or in the presence of 100 mm NaCl. Plants were subjected to the experimental treatments for 2 days. Ion contents in the roots and shoots of T. halophila plants were determined by inductively coupled plasma optical emission spectroscopy (ICP-OES). Ions were extracted as described previously (Volkov et al., 2004). Four biological replicates were sampled individually.


To translate ion currents (mA m−2 of protoplast surface) into ion fluxes (μmol Na+/(min.g FW), the following conversions were made. A current of 1 mA corresponds to a flux of 1 mC/sec, which for a monovalent ion is equal to (1/Faraday constant) 0.0104 μmol/sec or 0.624 μmol/min. The average protoplast diameter D was 16 μm. Assuming a spherical shape of the protoplast, its surface S = π·D2 was 0.8 nm2, and its volume V = 4/3·π·(D/2)3 was 0.214 × 10−14 m3 or 2.14 pl (picolitres). Taking root and protoplast density as 1 g cm−3, 1 g of root FW accommodates 1 ml/2.14 pl = 46.7 × 107 protoplasts with a total surface area (St) of 0.38 m2. Hence, a Na+ current of 1 mA m−2 may be considered equivalent to a flux of 0.624 × 0.38 = 0.24 μmol Na+/min g root FW if the entire root plasma membrane surface participates in the measured flux. Assuming that a proportion of 3% of the total root cell surface contributes to Na+ uptake, the measured current of 14 mA m−2 compares to a flux of 0.1 μmol Na+/min g root FW (see Discussion).


We would like to thank Mike Blatt, Adrian Hills, Sergey Sokolovski, Ufo Sutter, Bo Wang (University of Glasgow), Wieland Fricke (University of Paisley), Frans Maathuis (University of York), Vadim Demidchik (University of Cambridge) and Alexei Redkozubov (Institute of Physiologically Active Substances, Russian Academy of Sciences), for fruitful discussion throughout the project. Financial support for this study came from the Bower Fire Fund.