Sodium Influx and Accumulation in Arabidopsis

Abstract

Arabidopsis is frequently used as a genetic model in plant salt tolerance studies, however, its physiological responses to salinity remain poorly characterized. This study presents a characterization of initial Na⁺ entry and the effects of Ca²⁺ on plant growth and net Na⁺ accumulation in saline conditions. Unidirectional Na⁺ influx was measured carefully using very short influx times in roots of 12-d-old seedlings. Influx showed three components with distinct sensitivities to Ca²⁺, diethylpyrocarbonate, and osmotic pretreatment. Pharmacological agents and known mutants were used to test the contribution of different transport pathways to Na⁺ uptake. Influx was stimulated by 4-aminobutyric acid and glutamic acid; was inhibited by flufenamate, quinine, and cGMP; and was insensitive to modulators of K⁺ and Ca²⁺ channels. Influx did not differ from wild type in akt1 and hkt1 insertional mutants. These data suggested that influx was mediated by several different types of nonselective cation channels. Na⁺ accumulation in plants grown in 50 mm NaCl was strongly reduced by increasing Ca²⁺ activity (from 0.05-3.0 mm), and plant survival was improved. However, plant biomass was not affected by shoot Na⁺ concentration, suggesting that in Arabidopsis Na⁺ toxicity is not dependent on shoot Na⁺ accumulation. These data suggest that Arabidopsis is a good model for investigation of Na⁺ transport, but may be of limited utility as a model for the study of Na⁺ toxicity.

Soil salinity reduces yields in a wide variety of crops and affects almost 1,000 million ha of agricultural land globally (Szabolcs, 1994). Increasingly, Arabidopsis is being used as a model plant to dissect the molecular bases for growth inhibition by elevated soil Na⁺. The genetic and molecular power of this plant has provided a series of recent insights into molecular processes involved in the responses of Arabidopsis to high soil salinity (for review, see Zhu, 2000, 2002). However, some basic characterizations of the responses of this plant to elevated salinity have not been performed to assess its utility as a model for other plants, notably the cereals such as wheat (Triticum aestivum), maize (Zea mays), and rice (Oryza sativa). In particular, the relationship between Na⁺ influx, Na⁺ accumulation, and Na⁺ tolerance has not been examined in detail.

In many plants, a large component of tolerance to long-term exposure to Na⁺ can be attributed to the ability of plants to exclude Na⁺ from the shoot (Munns, 2002; Tester and Davenport, 2003). The first step in the movement of Na⁺ from the soil solution to the shoot is the initial entry of Na⁺ into the cells of the root epidermis and cortex. This is termed (unidirectional) Na⁺ influx, and it is distinct from net influx (or accumulation), which is the end result of the processes of both influx and efflux.

It is proposed that the initial influx step is a key determinant of overall shoot Na⁺ accumulation (Schubert and Läuchli, 1990; Davenport and Tester, 2000). In particular, it is observed in wheat (and many other plants) that Ca²⁺ inhibition of initial Na⁺ influx is directly correlated with the Ca²⁺-induced decrease in shoot Na⁺ accumulation, and these are correlated with a reduction in growth inhibition by Na⁺ upon addition of Ca²⁺ (Cramer, 2002).

The pathways for initial Na⁺ influx are still not identified. Tester and Davenport (2003) proposed three pathways for Na⁺ influx. Two protein-mediated pathways can be distinguished by their sensitivity to the addition of extracellular Ca²⁺, and a third pathway appears to be due to “leakage” into the root via the apoplast. The relative contribution of each of these pathways varies with species and growth conditions—for example, most Na⁺ influx into rice appears to be mediated by an apoplastic pathway (Yeo et al., 1987, 1999; Yadav et al., 1996). Davenport and Tester (2000) proposed that the partial sensitivity to Ca²⁺ of a nonselective cation channel in wheat roots provided evidence consistent with these channels being the primary pathway for both Ca²⁺-sensitive and -insensitive components of protein-mediated Na⁺ influx. However, the molecular identity of nonselective cation channels remains obscure (Demidchik et al., 2002), with likely candidates including cyclic nucleotide-gated channels (Maathuis and Sanders, 2001; Leng et al., 2002) and Glu receptors (Lacombe et al., 2001a; Davenport, 2002). There are several other candidate pathways for initial Na⁺ influx, reviewed by Tester and Davenport (2003), notably HKT1 (Rubio et al., 1995; Rus et al., 2001) and LCT1 (Schachtman et al., 1997; Amtmann et al., 2001).

Interestingly, the relationship observed in many plants between (Ca²⁺-sensitive) Na⁺ accumulation and (Ca²⁺-stimulated) tolerance does not appear to hold for maize or rice. In maize, although Ca²⁺ alleviates Na⁺ toxicity, growth inhibition is correlated with the sensitivity of leaf extension to salt-stimulated increases in abscisic acid (ABA) rather than the extent of Na⁺ accumulation (Cramer and Quarrie, 2002). Thus, in maize, the Ca²⁺ alleviation of Na⁺ toxicity is apparently due to effects of Ca²⁺ on ABA accumulation, rather than on Na⁺ accumulation. In rice, Na⁺ accumulation and sensitivity are correlated, but they are insensitive to Ca²⁺ (Yeo et al., 1987, 1999; Yadav et al., 1996). In Arabidopsis, it is notable that growth of salt-sensitive mutants (Zhu et al., 1998) and a salt-accumulating mutant (Nublat et al., 2001) does not appear to be closely related to shoot levels of Na⁺. For example, the salt-hypersensitive mutant of Arabidopsis, sos1, has a lower shoot Na⁺ content and lower Na⁺ influx than wild-type plants (at least in the presence of low K⁺; Ding and Zhu, 1997). This behavior, which is more reminiscent of the Na⁺ responses of maize than of wheat, rice, and many other plants, also appears to be reflected in the lack of correlation found by some workers between tolerance and the degree of Na⁺ accumulation in Brassica spp. (He and Cramer, 1993; Porcelli et al., 1995). Unusual responses of some Brassica spp. to the addition of Ca²⁺ is also noteworthy, with some workers finding that addition of Ca²⁺ in saline conditions did not reduce shoot Na⁺ concentration but improved growth (Huang and Redmann, 1995) or did not improve growth at all (Schmidt et al., 1993), at least in some species (Ashraf and Naqvi, 1992). Thus, to maximize the utility of Arabidopsis as a model system for the identification of genes encoding Na⁺ influx pathways and Na⁺ tolerance mechanisms, a thorough characterization of Na⁺ influx and its relationship to accumulation and tolerance was necessary.

The aims of this study were (a) to characterize rigorously the unidirectional influx of Na⁺ with a view to underpinning future molecular studies identifying the pathways for Na⁺ influx, and (b) to characterize the Na⁺ sensitivity of growth of Arabidopsis in transpiring conditions and to relate this to shoot Na⁺ accumulation, with a view to comparing this response to those of other Brassica spp. and the cereals.

RESULTS

Time Course of Na⁺ Influx

The amount of ²²Na⁺ in root tissue increased very rapidly, with uptake being linear for the first 2 min of exposure to the uptake solution and slowing greatly within 5 min (Fig. 1). Thus, an influx time of 2 min was used for subsequent experiments, because longer times would result in an underestimation of unidirectional influx. This influx is unlikely to be due simply to wall binding because influxes occurred after pre-equilibration of wall binding sites for times that were much longer than the influx and rinse times. Furthermore, simple wall binding cannot easily be reconciled with both the elimination of influx in roots that had been boiled (Table I), nor with the significant and repeatable inhibition of influx by submillimolar concentrations of a range of organic compounds (Tables I, II, and III).

Figure 1.

Increase in Na⁺ content (micromoles per gram) of excised roots of Arabidopsis after transfer into 200 mm NaCl plus 0.2 mm Ca²⁺ activity labeled with ²²Na⁺. a, Time course for up to 20 min; b, more detailed measures over the first 2 min, fitted with a line y = 3.31x + 0.16, r² = 0.96. Data represent means, and error bars the se (n = 4).

Table I.

The effect of pretreatment for 55 min with DEPC, a protein modifying reagent, or 10 min of boiling in de-ionized water on Na⁺ influx into 12-d-old excised roots of Arabidopsis

Values represent mean ± se (n = 18 for controls, n = 6 for all other treatments). FM, Fresh mass. -, not measured.

Treatment	Na+ Influx
	0.05 mM Ca2+	0.2 mM Ca2+	3.0 mM Ca2+
	μmol g−1 FM min−1
Control	3.00 ± 0.26	1.88 ± 0.12	1.34 ± 0.19
0.05 mM DEPC	−	1.17 ± 0.20b	−
0.5 mM DEPC	0.75 ± 0.08a	0.82 ± 0.06b	0.66 ± 0.04c
Boiling	−	0.04 ± 0.04b	−

^a,b,cWhen statistically significantly different to the respective controls, P < 0.05.

Table II.

The effects of flufenamate and quinine on Na⁺ influx into 12-d-old excised roots of Arabidopsis

Values represent mean ± se (n = 18 for the controls, n = 6 for all other treatments). FM, Fresh mass.

Treatment	Na+ Influx
	0.05 mM Ca2+	0.2 mM Ca2+	3.0 mM Ca2+
	μmol g−1 FM min−1
Control	3.00 ± 0.26	1.88 ± 0.12	1.34 ± 0.19
0.1 mm Flufenamate	2.49 ± 0.26	1.30 ± 0.17b	1.27 ± 0.13
1 mm Quinine	1.0 ± 0.09a	1.40 ± 0.14b	0.65 ± 0.08c

^a,b,c When statistically significantly different to the respective controls, P < 0.05.

Table III.

The effect of various biochemical modulators of cation channels on Na⁺ influx into 12-d-old excised roots of Arabidopsis

Values represent mean ± se (n = 18 for the general control, n = 10 for the glutamate control and 2 mm glutamate treatment, n = 6 for all other treatments). FM, Fresh mass.

Treatment	Na+ Influx	Percentage Change Relative to Controla
	μmol g−1 FM min−1
Control (0.2 mM Ca2+)	1.88 ± 0.12
10 mM H2O2 (10 min)	2.34 ± 0.07	+25
10 mM H2O2 (50 min)	1.45 ± 0.08	−23
40 mM H2O2 (10 min)	0.98 ± 0.03	−48
0.5 mM 8-Br-c-AMP	1.51 ± 0.15
0.5 mM 8-Br-c-GMP	1.34 ± 0.15	−29
1 mM BABA	2.19 ± 0.32
10 mM BABA	1.56 ± 0.33
1 mM GABA	2.16 ± 0.20
10 mM GABA	2.66 ± 0.19	+42
Control (for glutamate)	1.78 ± 0.19
2 mM Sodium glutamate	2.23 ± 0.24	+25
10 mM Sodium glutamate	2.08 ± 0.19

^a When statistically significantly different to the respective controls, P < 0.05.

A likely consequence of this rapid influx would be a rapid increase of Na⁺ in root tissue upon changes in external Na⁺ concentrations. This can be confirmed by directly measuring changes in tissue Na⁺ concentrations upon changes in external Na⁺. When roots of Arabidopsis grown and treated in exactly the same way as for ²²Na⁺ influx experiments were placed in 100 mm NaCl plus 3 mm Ca²⁺ activity for 10 min and then rinsed for 5 min in ice-cold 10 mm CaCl₂, tissue Na⁺ concentration was measured to be 16.5 ± 1.7 μmol g^-1 fresh mass (n = 4). Roots analyzed immediately before such a Na⁺ influx treatment contained 1.6 ± 0.2 μmol Na⁺ g^-1 fresh mass (n = 4), suggesting a net influx into the excised roots of 1.5 μmol g^-1 fresh mass min^-1, notably similar to that measured using ²²Na⁺. Thus, it is very likely that the influxes measured in this work are not only truly unidirectional but are also occurring across the plasma membrane into living cells within the root.

Varying Extracellular Na⁺ Concentration and Ca²⁺ Activity

Sodium influx increased linearly with increasing external Na⁺ concentration, at least in the range tested, of 1 to 200 mm NaCl (Fig. 2). These fluxes were conducted with a constant Ca²⁺ activity of 0.2 mm, requiring an increase in Ca²⁺ concentration from 0.25 to 0.68 mm, due to the effect of NaCl on Ca²⁺ activity as NaCl increased from 1 to 200 mm (see figure 4 in Tester and Davenport, 2003).

Figure 2.

Increasing unidirectional influx of Na⁺ (in micromoles per gram per minute) into excised roots of Arabidopsis with increasing concentrations of external Na⁺ and a constant Ca²⁺ activity of 0.2 mm, fitted with a line y = 0.023x + 0.15, r² = 0.97. Data represent means, and error bars the se (n = 4).

Figure 4.

The effect of 10-min or 3-h pretreatment in sorbitol or PEG (that are iso-osmotic with 50 mm NaCl) on influx of Na⁺ from 1 mm NaCl. All solutions contained 1 mm NaCl and 0.05 or 1.0 mm Ca²⁺ activity in addition to added osmotica. Influx was measured over 2 min into excised roots of Arabidopsis. Na⁺ influx was calculated as a percentage relative to control plants pretreated for 10 min in just 1 mm NaCl plus 0.05 or 1.0 mm Ca²⁺ activity. Data represent means, and error bars the se (n = 4). Numbers below the graph represent the percentage of total Na⁺ influx that was sensitive to a change in external Ca²⁺ activity from 0.05 to 1 mm, calculated according to the formula: [(Na⁺ influx [%] at 0.05 mm Ca²⁺) - (Na⁺ influx [%] at 1 mm Ca²⁺)]. Control fluxes were 74.9 ± 13.2 and 43.4 ± 5.5 nmol ^g-1 min^-1 at 0.05 and 1.0 Ca²⁺ activities, respectively.

The effect of Ca²⁺ activity on Na⁺ influx was more complicated (Fig. 3). When roots were not pretreated in sorbitol before uptake, then Na⁺ influx was significantly inhibited (up to 70%) by increasing extra-cellular Ca²⁺ activity up to 1 mm. At activities above 1 mm, Ca²⁺ had no further effect on Na⁺ influx, i.e. the remaining Na⁺ influx was Ca²⁺ insensitive. It should be noted that Ca²⁺ is being used in this study simply as a pharmacological agent, and no claims are being made to its relevance to Na⁺/Ca²⁺ interactions in the soil solution.

Figure 3.

Inhibition of unidirectional influx of Na⁺ (in micromoles per gram per minute) into excised roots of Arabidopsis from 50 mm NaCl with increasing Ca²⁺ activity. Roots were either pretreated for 3 h in sorbitol before influx (white circles), or fluxes were performed after only 10 min of equilibration in non-radioactive influx solution (black circles). Data represent means, and error bars the se (n = 4 with sorbitol, n = 8 without sorbitol).

In experiments where excised roots were pretreated in sorbitol for 3 h, however, Na⁺ influx was not affected by differences in external Ca²⁺ activity (Fig. 3). Pretreatment with sorbitol reduced Na⁺ influx relative to untreated roots only at low Ca²⁺ activities (below 1 mm Ca²⁺ activity), that is, sorbitol pretreatment appeared to affect only Ca²⁺ sensitivity of influx. Na⁺ influx at low Na⁺ concentrations is also Ca²⁺ sensitive (Essah, 2002; see also control fluxes in Fig. 4), but this sensitivity to Ca²⁺ was not affected by pretreatment with low concentrations of sorbitol (data not shown). Ca²⁺ sensitivity of Na⁺ influx decreased with increasing concentrations of external sorbitol and Na⁺. The fact that sorbitol pretreatment affected only the Ca²⁺-sensitive component of Na⁺ influx suggests that there are two independent pathways for Na⁺ influx, one that is Ca²⁺ sensitive and one that is Ca²⁺ insensitive.

Effect of Osmotica on Na⁺ Influx

To investigate further the inhibition of the Ca²⁺-sensitive component of Na⁺ influx by sorbitol, roots were pretreated with an iso-osmotic concentration of another osmoticum, PEG 8000 (60 mg mL^-1), having the same osmotic potential (97 mmol kg^-1) as 50 mm NaCl. The results obtained showed that pretreating roots in iso-osmotic PEG led to a significant decrease in the Ca²⁺-sensitive component of Na⁺ influx relative to the control (Fig. 5), suggesting that the decrease may be due to the effects of osmotic potential. However, pretreating roots for 3 h in 50 mm NaCl did not reduce the Ca²⁺ sensitivity of Na⁺ influx, but rather increased both the Ca²⁺-sensitive and -insensitive components of Na⁺ influx by 28% and 38%, respectively (Fig. 5). On the other hand, growing plants on NaCl (phytagel plates supplemented with 50 mm NaCl and 0.2 mm Ca²⁺ activity) before influx significantly reduced influx compared with control plants, but did not significantly affect the Ca²⁺ sensitivity of Na⁺ influx.

Figure 5.

The effect of a 3-h pretreatment in iso-osmotic sorbitol, polyethylene glycol (PEG), or 50 mm NaCl (Na pretr) or 12 d of growth in 50 mm NaCl (Na-grown), on unidirectional influx of Na⁺ into excised roots of Arabidopsis. Data represent means, and error bars the se (n = 4). Numbers below the graph represent the percentage of total influx that was sensitive to external Ca²⁺, calculated according to the formula: [(influx at 0.05 mm Ca²⁺ - influx at 1 mm Ca²⁺)/influx at 0.05 mm Ca²⁺] * 100.

To investigate further whether the effects of 3-h pretreatments in iso-osmotic sorbitol and PEG on the Ca²⁺ sensitivity of Na⁺ influx were attributable to the osmotic effects of these solutes, Na⁺ influx was measured in 1 mm NaCl in roots pretreated with sorbitol or PEG at concentrations iso-osmotic with 50 mm NaCl. Sorbitol or PEG was included in the uptake medium, and length of pretreatment was either 10 min or 3 h. Short-term exposure to high concentrations of sorbitol or PEG did not alter Na⁺ influx relative to controls (10-min pretreatment in 1 mm NaCl and 0.05 or 1 mm Ca²⁺ activity) (Fig. 4). However, longer exposure (3 h) to sorbitol clearly reduced Ca²⁺ sensitivity of influx (similar to the level observed in 50 mm NaCl; Fig. 5). This last result is in contrast to the lack of effect of low concentrations of sorbitol on Ca²⁺ sensitivity of influx at low concentrations of Na⁺ (reported above). This suggests that influx at lower concentrations of NaCl (1-10 mm) is similar in its characteristics to that at 50 mm and above and that the effect of sorbitol on Ca²⁺ sensitivity of influx depends on prolonged exposure to fairly high concentrations of sorbitol.

Pharmacology

For this study, Na⁺ influx for 2 min from a 50 mm NaCl solution was conducted without the 3-h sorbitol pretreatment and with 0.2 mm Ca²⁺ activity to enable measurement of the effects of the pharmacological agents on both the Ca²⁺-sensitive and -insensitive components of Na⁺ influx.

Diethylpyrocarbonate (DEPC), a reagent that modifies His and Tyr residues in proteins (Mankelow and Henderson, 2001; Row and Gray, 2001) significantly inhibited Na⁺ influx at the low (0.05 mm DEPC) and high (0.5 mm) concentrations tested (Table I). It could be inferred from Na⁺ influx over the range of Ca²⁺ activities tested that both the Ca²⁺-sensitive and -insensitive components of Na⁺ influx were inhibited by DEPC. However, DEPC did not completely inhibit Na⁺ influx, suggesting either that DEPC did not inhibit all protein-mediated influx or that DEPC penetration of the roots was partial. The latter suggestion is unlikely, because an experiment using the very high concentration of 5.0 mm DEPC did not inhibit influx any more than did 0.5 mm DEPC. The DEPC-insensitive component of Na⁺ influx was not due to apoplastic binding of Na⁺, as demonstrated by the lack of Na⁺ uptake into roots killed by boiling in de-ionized water for 10 min before influx (Table I). The rinse conditions used in influx experiments were apparently sufficient to displace all apoplastically bound ²²Na⁺.

The partial effects of Ca²⁺ and DEPC on Na⁺ influx suggests that there were three components (although not necessarily three separate mechanisms) of Na⁺ influx: a Ca²⁺-sensitive, DEPC-sensitive component; a Ca²⁺-insensitive, DEPC-sensitive component; and a component that was insensitive to both Ca²⁺ and DEPC. The mechanisms underlying these components were investigated using biochemical modifiers of known transport mechanisms. In the conditions used for these Na⁺ influx assays (no sorbitol pretreatment and 0.2 mm Ca²⁺ activity), the three components of Na⁺ influx were of approximately equal magnitudes (approximately 0.6 μmol g^-1 min^-1 each).

The monovalent cations Cs⁺ and tetraethyl-ammonium-Cl (TEA-Cl; K⁺ channel blockers) did not have any significant effect on influx (Table IV), suggesting that Na⁺ influx was not via TEA⁺ and Cs⁺-sensitive K⁺-selective channels. Among the divalent cations tested, Ba²⁺ (and Ca²⁺) significantly inhibited Na⁺ influx, whereas Zn²⁺ caused a nonsignificant reduction in Na⁺ influx. Although the effects of these divalent cations suggest that there could be a charge effect involved in the inhibition of Na⁺ influx into the roots, this is unlikely given the effects of the trivalent cations, Gd³⁺ and La³⁺. Gd³⁺ had no effect on Na⁺ influx, irrespective of the concentration used (0.01- 1.0 mm). La³⁺ inhibited influx at low concentrations (0.01 and 0.1 mm: data not shown), but significantly increased influx when higher concentrations were applied (Table IV). Amiloride (an inhibitor of Na⁺/H⁺ antiporters) and verapamil (a Ca²⁺ channel blocker) had no significant effects on Na⁺ influx. Taken together, these results suggest that Na⁺ uptake into the roots was not via K⁺ channels, Na⁺/H⁺ antiporters, and/or Ca²⁺ channels. The stimulation of Na⁺ influx by high concentrations of La³⁺ may be an indirect effect due to the cytotoxicity of these concentrations. Evidence consistent with cytotoxicity of millimolar concentrations of La³⁺ has been observed in Arabidopsis (J. Love, A.N. Dodd, and A.A.R. Webb, unpublished data), and the membrane depolarization of Neurospora crassa by 1 mm La³⁺ observed by Corzo and Sanders (1992) is also consistent with increased Na⁺ influx (see results with H₂O₂, below).

Table IV.

The effect of various inorganic and organic inhibitors of cation channels on Na⁺ influx into 12-d-old excised roots of Arabidopsis

Values represent mean ± se (n = 18 for controls, n = 6 for all other treatments). FM, Fresh mass.

Treatment	Na+ Influx	Percentage Change Relative to Controla
	μmol g−1 FM min−1
Control (0.2 mM Ca2+)	1.88 ± 0.12
5 mM CsCl	1.91 ± 0.28
10 mM TEA-Cl	1.70 ± 0.10
1 mM BaCl2	1.36 ± 0.05	−28
1 mM ZnCl2	1.53 ± 0.13
1 mM GdCl3	1.60 ± 0.16
1 mM LaCl3	2.51 ± 0.16	+34
0.1 mM Verapamil	1.53 ± 0.15
0.1 mM Amiloride	1.85 ± 0.17

^a When statistically significantly different to the respective controls, P < 0.05.

Flufenamate (an inhibitor of nonselective cation channels in animals) and quinine (a non-specific cation channel blocker) significantly reduced the influx of Na⁺ with 0.2 mm Ca²⁺ (Table II). Influxes were therefore conducted at low (0.05 mm) and high (3.0 mm) Ca²⁺ activities to further investigate the effects of these compounds. The results indicated that flufenamate was (partially) inhibiting only the Ca²⁺-sensitive portion of Na⁺ influx, whereas quinine inhibited both the Ca²⁺-sensitive and -insensitive components (Table II).

The application of H₂O₂ (an agonist of plant plasma membrane Ca²⁺-permeable cation channels; Demidchik et al., 2002) yielded a dual pattern of results (Table III). Addition of 10 mm with a short pretreatment (10 min) enhanced the rate of Na⁺ influx relative to the control, but a longer pretreatment (50 min) resulted in an inhibition of influx to below that of the control. Application of a higher concentration of H₂O₂ (40 mm) resulted in inhibition of influx after only 10 min pretreatment, consistent with an effect that is determined by cytosolic H₂O₂ concentration. The effects of flufenamate, quinine, and H₂O₂ are all consistent with the proposal that Na⁺ influx was at least partially via nonselective cation channels.

To investigate further what type of nonselective cation channels could be operating in Na⁺ uptake, the effects of bromocyclic monophosphates and aminobutyric acids on Na⁺ influx were tested. The bromocyclic monophosphates (Na⁺ salts) tested reduced influx, although only the inhibition by 8-bromoguanine 3′,5′-cyclic monophosphate was statistically significant at P < 0.05 (Table III). The rapid reduction in Na⁺ influx by external application of membrane-permeable analogs of cGMP and cAMP (to a lesser extent) suggests a down-regulation of the channel proteins responsible for the uptake, probably through direct binding.

Addition of l and 10 mm dl-β-aminoisobutyric acid (BABA, a partial agonist of Gly receptors in animals [Schmieden and Betz, 1995] and observed to affect some plant processes [Zimmerli et al., 2001]) caused an increase and a decrease in influx, respectively, although these effects were not significant. Addition of 4-aminobutyric acid (GABA, a major inhibitory neurotransmitter in the brain and a precursor in the plant Pro synthesis pathway [Shelp et al., 1999]), however, caused a progressive increase in influx as its concentration was raised from 1 (to 3 and then 7 mm) to 10 mm. The effects on Na⁺ influx of GABA at 7 and 10 mm were statistically significant (P < 0.05). These results suggest that BABA and GABA could be regulators of the protein(s) that serve as a pathway for Na⁺ influx into Arabidopsis roots. This stimulation cannot be attributed to a weak acid effect of GABA causing cytosolic acidification, because GABA has a pKa of 10.56, so it would remain fully protonated at both extracellular and cytosolic pHs.

The effects of Glu, an agonist of neuronal nonselective cation channels, were also tested. This required slight modifications to the general procedure for influx assays, in that there was just a 5-min pretreatment in Glu-containing solutions, and both the pretreatment and the labeled influx solutions contained 2 mm MES. To this was added 2 or 10 mm Na-Glu, and the pH was adjusted to 5.6 with Tris base. The concentration of NaCl was reduced to 40 mm in solutions containing 10 mm Na-Glu. Under these conditions, the influx of Na⁺ into excised roots was increased by 25% in 2 mm Na-Glu and by 17% in 10 mm Na-Glu (Table III).

Effects of Defined Mutations on Na⁺ Influx

Plants with reduced activity of K⁺ channels involved in K⁺ influx into roots (akt1; Hirsch et al., 1998; Spalding et al., 1999) and K⁺ loading into the xylem (skor-1; Gaymard et al., 1998) showed no significant difference in Na⁺ influx when compared with their respective wild-type controls (Table V).

Table V.

A comparison of Na⁺ influx into excised roots from seedlings of control plants and those with defined mutations

The controls were Columbia (for akt1), Ws (for skor-1 and scr-2), and Columbia gl1 (for sos3-1, hkt1-3, and sos3 hkt1). Seedlings were 12 d old, except for scr-2 seedlings, which were 19 d old, to allow the mutant plants to reach a comparable size with control plants. Values represent mean ± se (n = 6). FM, Fresh mass.

Mutant Name	Na+ Influx
	Control	Mutant
	μmol g−1 FM min−1
akt1	1.95 ± 0.09	2.14 ± 0.28
skor-1	1.82 ± 0.10	1.50 ± 0.14
sos3-1	2.48 ± 0.17	2.40 ± 0.13
hkt1-3	2.48 ± 0.17	2.47 ± 0.12
sos3-1 hkt1-3	2.48 ± 0.17	2.38 ± 0.10
scr-2	1.82 ± 0.10	2.10 ± 0.09a

^a When statistically significantly different to the respective controls, P < 0.05.

Na⁺ influx was also measured in the salt-sensitive sos3-1 mutant (Liu and Zhu, 1997), the knockout mutant of a gene encoding a putative Na⁺ transporter, hkt1-3, and the double mutant sos3-1 hkt1-3 (allelic to the sos3-1 hkt1-1 and sos3-1 hkt1-2 double mutants described by Rus et al. [2001]). These mutants are in the Columbia gl1 background, which showed higher Na⁺ influx than that measured in other ecotypes in this study (Table V). There were no significant differences in Na⁺ influx between any of these mutant plants and control plants (Table V).

The scr2 mutant, which shows altered root development and endodermal function (Scheres et al., 1995; Di Laurenzio et al., 1996), had significantly higher influx than its wild type (Table V).

The rss (for reduced salt sensitivity; Werner and Finkelstein, 1995) and sañ (from “salobreño,” the Spanish word for “salty land”; Quesada et al., 2000) recessive mutants are capable of germinating under saline conditions (up to 200 mm NaCl, and iso-osmotic concentrations of sorbitol and mannitol). No effects of the rss and sañ mutations were observed on Na⁺ influx into the roots (Essah, 2002).

Effects of Na⁺ Concentration and Ca²⁺ Activity on Growth and Tissue Ion Concentrations of Arabidopsis Plants

Growing plants in 50 mm NaCl inhibited plant growth, reducing the fresh and dry mass by 30% to 50% when compared with control plants grown in 1 mm NaCl (data not shown). Increasing the Ca²⁺ activity from 0.05 to 0.2 or 3 mm improved survival of plants. Only 14 of 24 plants survived 3 weeks of 50 mm NaCl treatment at 0.05 mm Ca²⁺ activity, compared with 24 of 24 and 21 of 24 plants surviving at 0.2 and 3 mm Ca²⁺ activities, respectively. However, there were only very small differences in the fresh and dry mass of the surviving plants, and their overall appearance was also very similar. Increasing the Ca²⁺ activity from 0.05 to 0.2 or 3 mm had a small effect on fresh and dry shoot masses of the surviving 50 mm NaCl-treated plants, but this effect was not statistically significant (Fig. 6a). The data in Fig. 6a were multiplied by a “survival factor” to take into account the differential survival in the three treatments and tend in fact to exaggerate the effect of Ca²⁺ in enhancing growth. The water content as well as fresh and dry root mass ratios (i.e. root mass divided by total plant mass) did not differ significantly for control and 50 mm NaCl-treated plants, and their values remained constant with higher Ca²⁺ (Essah, 2002). In contrast, increasing Ca²⁺ significantly decreased shoot Na⁺ concentrations (Fig. 6b), along with increasing [K⁺]:[Na⁺] (Fig. 6c) and decreasing [Na⁺]:[Ca²⁺] (Fig. 6d). Results for the roots showed the same trends as shoots, but the absolute effect of Ca²⁺ was smaller. Considering net uptake of Na⁺ into the whole plant, raising Ca²⁺ activity from 0.05 to 3.0 mm reduced the amount of Na⁺ accumulated by 83%, or 74% if expressed as a total amount accumulated per unit fresh mass of roots (Table VI). Similar but more severe effects were observed when plants were grown in 100 and 200 mm Na⁺ (data not shown).

Figure 6.

Effects of Ca²⁺ on growth and cation accumulation in 6-week-old Arabidopsis plants grown hydroponically in 50 mm NaCl after the 3rd week of planting. a, Shoot dry mass; b, shoot Na⁺ concentration; c, ratio of shoot K⁺ to Na⁺ concentration; d, ratio of shoot Na⁺ to Ca²⁺ concentration. Data represent means, and error bars the se (n = 24). Measurements were only made on surviving plants. To incorporate into the data the different mortalities with different treatments, shoot mass data (a) were multiplied by a “survival factor” of 0.58 for 0.05 mm Ca²⁺ activity treatment (14 of 24 plants survived the 50 mm NaCl treatment), 1 for 0.2 mm Ca²⁺ treatment (24 of 24 plants survived), and 0.875 for the 3.0 mm Ca²⁺ treatment (21 of 24 plants survived). Shoot mass includes the amount of Na⁺ in the tissue, but this accounted for 15.6%, 7.4%, and 3.0% of the total dry mass in the three Ca²⁺ activities used, having no significant effect on the overall trends. Ion concentrations in b to d were measured only in surviving plants and expressed on a dry weight basis for those plants, with no correction for survival.

Table VI.

Effect of Ca²⁺ on total Na⁺ accumulation per plant and on the amount of Na⁺ accumulated per unit mass of root, measured in 6-week-old plants grown hydroponically as described in “Materials and Methods,” and supplemented with 50 mm NaCl from 3 weeks after planting

Data represent mean ± se (n = 14 at 0.05 mm Ca²⁺ activity, n = 24 at 0.2 mm Ca²⁺, and n = 21 at 3.0 mm Ca²⁺ activity). FM, Fresh mass.

Ca2+ Activity	Na+ Content	Na+ Content per Unit Mass of Root
mm	μmol plant−1	μmol g−1 root FM plant−1
0.05	225 ± 15	1281 ± 163
0.2	155 ± 11	853 ± 79
3.0	39 ± 7	335 ± 68

DISCUSSION

Relationship of Influx to Net Accumulation

Unidirectional influx of ²²Na⁺ into roots of Arabidopsis plants could only be measured over a few minutes before the cytosol became fully labeled with ²²Na and significant efflux contributed to the measurements. In 50 mm external NaCl, influx was around 1.8 μmol g^-1 root fresh mass min^-1. This is within the range measured in wheat (Davenport and Tester, 2000; R. Munns, R. Davenport, and M. Tester, unpublished data) and rice (L. Wang, M. Tester, and R. Davenport, unpublished), as well as in model cells such as charophyte algae (e.g. Hope and Walker, 1975; Table VI), where physical separation of cellular contents from the wall and attached solution allow an unequivocal measure of influx.

The suggestion from our data that cytosolic Na⁺ turns over at very high rates is strongly supported by the thorough work of Cheeseman (1982), who measured a cytosolic half-time for exchange of Na⁺ in maize roots of approximately 5 min. Thus, we suggest that almost all published data measuring ²²Na⁺ accumulation into higher plant tissues are made over time periods that are too long to ensure they are not significantly reduced by efflux of ²²Na⁺. This would result not only in an underestimation of rates of unidirectional influx across the plasma membrane, but also cause confusion between influx and efflux pathways, which could result in misattribution of function to particular proteins or protein families.

The very high rates of exchange of Na⁺ between the external solution and the cytosol would also explain the difference between the values we measure using 50 mm Na⁺ (around 1.8 μmol g^-1 min^-1) and the smaller fluxes found in Arabidopsis roots by other workers (of 0.1-0.15 μmol g^-1 min^-1) whose measurements were made with either longer influx times (Maathuis and Sanders, 2001) or longer rinses (Elphick et al., 2001).

These rapid rates of influx imply high rates of Na⁺ efflux to the soil solution. Arabidopsis plants grown in 50 mm NaCl accumulated 335 to 1,281 μmol Na⁺ per gram of root fresh mass (depending on the Ca²⁺ activity) over 3 weeks (Table VI). These amounts would be accumulated within 3 to 12 h at measured rates of unidirectional influx (assuming a constant influx rate during growth). The disparity between unidirectional and net Na⁺ uptake implies high rates of Na⁺ efflux.

Three Components of Influx

Measurements of unidirectional influx into roots of Arabidopsis seedlings indicated that there were three components of influx: a DEPC-insensitive component (which was not affected by extracellular Ca²⁺), and two components, which were inhibited by DEPC and were sensitive or insensitive to inhibition by extracellular Ca²⁺. In the following discussion, we attempt to relate these components to transport mechanisms. It must be emphasized that the components we distinguish are purely phenomenological, and it is possible that all are due to the activity of a single transporter or that more than one transporter type contributes to each component.

Apoplastic Na⁺ Uptake

The influx experiments were designed to measure Na⁺ influx across the plasma membrane, with rinse conditions designed to displace apoplastic Na⁺. However, the discovery of a DEPC-insensitive component of Na⁺ influx raised the possibility that some of the Na⁺ influx was not protein-mediated and represented apoplastic Na⁺ binding. We think that the DEPC-insensitive component of influx was not due to apoplastic binding or apoplastic transport for several reasons. First, influx into boiled roots was negligible, indicating that rinse conditions were sufficient to displace most of the labeled extracellular Na⁺. Although boiled roots lack some of the complexity of intact roots, in particular the endodermal barrier, uptake of labeled Na⁺ into intact roots saturated rapidly, and so it was considered likely that any component of uptake due to leak of Na⁺ across the endodermis into the apoplast of the inner half of the root would exchange equally rapidly during rinsing. Second, DEPC only partially inhibited Na⁺ currents in isolated protoplasts from Arabidopsis roots, and was only effective in 56% of protoplasts (Demidchik and Tester, 2002). DEPC does not affect the activity of all proteins (e.g. Spires and Begenisich, 1990) and additionally may not have gained full access to membrane transport proteins present in the plasma membranes of cells in the outer half of the root. Therefore we think it is likely that the DEPC-insensitive, Ca²⁺-insensitive component of influx was protein mediated.

The scarecrow (scr) mutant has altered endodermal development and did show slightly higher influx, suggesting that in these plants, some of the influx measured was due to greater apoplastic leak across the endodermis (which could then result in higher uptake into stelar cells and so would be included in the influx measurement). This study did not attempt to measure apoplastic leak, except to attempt to exclude it from measurements, and the lack of transpiration would also reduce any apoplastic component. However, any Na⁺ that did leak across the endodermis and was taken up into stelar cells would be included in the measurement. The strong effect of Ca²⁺ in reducing net Na⁺ uptake suggests that bypass leak does not contribute significantly to uptake to the shoot (given that bypass flow, at least in rice, is Ca²⁺ insensitive; Yeo and Flowers, 1985, Yeo et al., 1987).

Na⁺ Influx Was Not via HKT1

AtHKT1 has been demonstrated to act as a Na⁺ uniporter when expressed in Xenopus sp. oocytes, and both AtHKT1 and the wheat homolog TaHKT1 have been implicated in plant Na⁺ uptake (Rus et al., 2001; Laurie et al., 2002). TaHKT1 has been reported to be insensitive to extracellular Ca²⁺ (Tyerman and Skerrett, 1999). However, hkt1-3 knockout mutants did not differ in Na⁺ influx from wild type. It has also been suggested that HKT1 is negatively regulated by the sos-signaling pathway (Zhu, 2002), however, neither sos3 nor sos3-hkt1-3 double mutants differed in Na⁺ influx from wild type (Table V). Similar results have also been obtained with the hkt1-1 allele (data not shown) and with sas2, which is a mutation in AtHKT1 (Berthomieu et al., 2003).

Na⁺ Influx Was Not via K⁺ Channels

Influx of K⁺ into Arabidopsis roots is partially via the K⁺ transporters, AKT1 and AtKC (Hirsch et al., 1998; Spalding et al., 1999). Inward K⁺ channels measured in root protoplasts had a relatively high permeability for Na⁺ and were partially inhibited by Ca²⁺ (Maathuis and Sanders, 1995). However, several lines of evidence excluded K⁺ channels as pathways for Na⁺ influx in Arabidopsis. Na⁺ influx was unaffected by TEA⁺ and Cs⁺, which block K⁺-selective currents in roots. Moreover the akt1 root K⁺ channel gene mutant did not differ in Na⁺ influx from wild type.

Contribution of Nonselective Cation Channels (NSCCs) to Na⁺ Influx

NSCCs are generally considered to constitute the major pathway for Na⁺ influx (Amtmann and Sanders, 1999; Demidchik et al., 2002), and Na⁺ currents in Arabidopsis root protoplasts were carried mainly via NSCCs (Demidchik and Tester, 2002). The Na⁺ influxes measured in the present study matched the characteristics of Na⁺ currents measured in protoplasts in their lack of sensitivity to TEA⁺, Cs⁺, and amiloride, and slight sensitivity to verapamil (which affected protoplast Na⁺ currents only at hyperpolarized voltages). In both cases Na⁺ influx depended on extracellular Na⁺ and was not saturated at 150 to 200 mm NaCl. Both root Na⁺ influx and protoplast Na⁺ currents were partially inhibited by submillimolar Ca²⁺ activities, although the proportion of Na⁺ transport inhibited was greater in protoplasts than intact roots. Similarly, quinine inhibited Na⁺ transport in both systems, but had a greater inhibitory effect on protoplast Na⁺ currents. By contrast, Na⁺ influx into roots was unaffected by Gd³⁺ and was stimulated by La³⁺, whereas Na⁺ currents in root protoplasts were strongly inhibited by these trivalent cations. This may reflect differences in Na⁺ transporters contributing to measured transport or may reflect the differences in voltages in these systems: In patch-clamped protoplasts, the membrane voltage could be held at hyperpolarized voltages, whereas membrane voltage was not fixed in intact roots. In wheat root protoplasts, La³⁺ block of cation currents required membrane hyperpolarization (S.D. Tyerman and R. Davenport, unpublished data). DEPC partially inhibited Na⁺ currents in root protoplasts and Na⁺ influx into root segments at 0.05 mm. The inhibition of Ca²⁺-sensitive Na⁺ transport was incomplete at 0.05 mm DEPC in both systems, but application of 0.5 mm DEPC completely inhibited Ca²⁺-sensitive Na⁺ influx into roots and at least partially inhibited Ca²⁺-insensitive Na⁺ influx.

The main NSCC candidates in Arabidopsis are cyclic nucleotide-gated channels and Glu receptors. Both cAMP and cGMP inhibited Na⁺ influx (although this was statistically significant only in the case of cGMP). This is in accordance with a cyclic nucleotide inhibition of cation currents in Arabidopsis root protoplasts (Maathuis and Sanders, 2001), but is in contrast to reports of cyclic nucleotide stimulation of K⁺, Na⁺, and Ca²⁺ currents via heterologously expressed cyclic nucleotide-gated channel subunits from Arabidopsis (Leng et al., 1999, 2002; Balagué et al., 2003). The contribution of these ion channels to native membrane currents remains unclear. Glu had a slight stimulatory effect on Na⁺ influx at 2 mm but not 10 mm. Again, this is difficult to relate to the minimal reports of constitutively active, Glu-insensitive plant Glu receptor activity in heterologous systems (Cheffings, 2001; Lacombe et al., 2001b). Interestingly, GABA, another neurotransmitter and plant metabolite, stimulated Na⁺ influx, suggesting a possible diagnostic test for transporters contributing to Na⁺ influx.

Recently, a number of hyperpolarization-activated Ca²⁺ conductances (HACCs) have been characterized in Arabidopsis, and in at least some cases, these appear to be carried by Ca²⁺-permeable NSCCs (Demidchik et al., 2002). In Arabidopsis, millimolar levels of H₂O₂ activated HACCs in protoplasts from guard cells (Pei et al., 2000) but not roots (Demidchik et al., 2003). In the present study, 10-min exposure to 10 mm H₂O₂ increased Na⁺ influx by 25%, whereas increasing the length (to 50 min) or the dosage (to 40 mm) of H₂O₂ inhibited Na⁺ influx (suggesting that intracellular accumulation of H₂O₂ was required for inhibition). This suggested that some Na⁺ influx occurred via H₂O₂-stimulated channels, or that Na⁺ influx was increased by indirect effects of H₂O₂, for example membrane depolarization. The subsequent inhibition of Na⁺ influx below control levels upon prolonged exposure to H₂O₂ suggested an indirect effect of H₂O₂ (although it is possible that unlike guard cell HACCs, H₂O₂-sensitive root NSCCs did not simply return to resting levels after activation but were inhibited).

It is not possible from the data presented here to attribute Na⁺ influx to a specific transport pathway. These data agree with other studies in identifying NSCCs as the most likely route of Na⁺ influx, however, much of this evidence is negative (based on lack of effect of modifiers of more selective channels). Plant NSCCs have proven difficult to characterize due to lack of specific blockers and effective agonists, and so we could not use rigorous diagnostic tests to confirm NSCC involvement. Moreover, our data suggested that several independent mechanisms probably operate in Na⁺ influx. For instance, the effect of sorbitol pretreatment in inhibiting only the Ca²⁺-sensitive portion of influx suggests that the Ca²⁺-sensitive pathway is discrete from the Ca²⁺-insensitive components. Alternatively, sorbitol pretreatment could cause modification of the properties of a single transporter type. At present, we do not have the tools to distinguish these possibilities.

Control of Na⁺ Influx

Sodium influx was responsive to osmotic conditions. Growth in NaCl reduced Na⁺ influx to approximately one-half of that of control plants (pretreated for only 10 min in NaCl before measurement) but did not affect Ca²⁺ sensitivity of influx, suggesting either a uniform down-regulation of multiple pathways or reduction in activity of a single transporter type with partial Ca²⁺ sensitivity. However, a short (3-h) exposure to 50 mm NaCl increased Na⁺ influx (again without affecting Ca²⁺ sensitivity), suggesting a transient up-regulation of transport activity (or some effect of, for example, a change in membrane potential). In contrast to the effect of short pretreatment in NaCl, in plants pretreated in osmotica with low membrane permeability (PEG and sorbitol) for 3 h, Na⁺ influx was reduced to the Ca²⁺-insensitive component of influx. One explanation of this difference is that NaCl and the other osmotica had opposite effects on the ability of the plants to adjust to the sudden change in osmotic potential of the solution. NaCl uptake would provide a rapid means of turgor adjustment, whereas (relatively impermeant) sorbitol and PEG may induce synthesis of organic intracellular osmotica and a reduction in activity of NSCCs (which could leak cellular solutes in the low-salt conditions in which sorbitol and PEG were applied). Interestingly, the incubation of roots in high osmolarity enzyme solutions for several hours during protoplasting did not seem to affect Ca²⁺ sensitivity of Na⁺ currents in Arabidopsis root protoplasts (Demidchik and Tester, 2002). These data suggest that Na⁺ influx pathways in Arabidopsis may be partially regulated by turgor.

Is Arabidopsis a Good Model for Na⁺ Transport in Crop Plants (Such as Wheat)?

In wheat, the characteristics of Na⁺ influx have been shown to correlate reasonably well with both net uptake and toxicity (Davenport and Tester, 2000). In Arabidopsis, the pattern of Na⁺ influx and accumulation appear similar to wheat. Unidirectional Na⁺ influx was of similar magnitude to wheat, and showed similar insensitivity to cation channel inhibitors. In Arabidopsis, both unidirectional and net Na⁺ fluxes were more sensitive to inhibition by Ca²⁺ than in wheat but appeared to be correlated, as in wheat.

Although transport characteristics in Arabidopsis and wheat appeared to be comparable, there was an interesting difference in the effect of Ca²⁺ on growth. Although increasing Ca²⁺ increased survival of Arabidopsis plants exposed to high NaCl concentrations, it had little effect on the growth of surviving plants. This was surprising because plants grown in 0.05 mm Ca²⁺ activity accumulated dramatically higher shoot Na concentrations than plants grown in 3.0 mm Ca²⁺ activity, yet there was no difference in biomass or water content between these plants. Liu and Zhu (1997) also found that increasing external Ca²⁺ from 0 to 3 mm concentration (equivalent to approximate Ca²⁺ activities of 0 to 1.25 mm, respectively) barely increased the growth of wild-type Arabidopsis seedlings treated with 50 mm (or 100 mm, although the data were not shown) NaCl, with growth being measured by root elongation. A consequence of these different effects of Ca²⁺ on Na⁺ accumulation and growth is that the extent of Na⁺ accumulation in the shoot did not appear to be related to growth. This has been also observed in the sos mutants of Arabidopsis (Ding and Zhu, 1997; Zhu et al., 1998).

These results with Arabidopsis generally accord with a range of observations made with several Brassica spp. (He and Cramer, 1993; Schmidt et al., 1993; Huang and Redmann, 1995; Porcelli et al., 1995), suggesting that this unusual behavior may be common in the Brassicaceae. The physiological basis for these responses are not known, although it is possible that the response of Brassica spp. to high Na⁺ is related more to ABA-mediated growth inhibition, as appears to occur in maize (Cramer and Quarrie, 2002), another plant in which the salt tolerance of different varieties is not related to the extent of Na⁺ accumulation in the shoot (Alberico and Cramer, 1993; Cramer et al., 1994; Al-Mansour, 1996).

We suggest that Arabidopsis provides a reasonable model for the study of Na⁺ transport processes in other salt-sensitive species, but it probably should not be used as a general model for studies of Na⁺ tolerance, because the bases of toxicity may differ between Arabidopsis and species such as wheat.

MATERIALS AND METHODS

Preparation of Seedlings

Seeds of Arabidopsis ecotype C 24 were surface-sterilized in 3% (v/v) sodium hypochlorite containing 0.02% Triton X-100 for 15 min and rinsed six times with sterile de-ionized water. Approximately 20 surface sterilized seeds were grown in vertical 10-cm-wide square sterile plates containing 50 mL of medium per plate. The medium was composed of Murashige and Skoog basal salts (Duchefa, Haarlem, Netherlands), 1% (w/v) Suc, and 0.25% phytagel (Sigma-Aldrich, St. Louis), and pH adjusted to 5.7 with KOH before autoclaving. After planting, seeds were vernalized for 2 d at 4°C to break any residual dormancy and to ensure uniform germination. Plates were then transferred to a growth room with a temperature of 22°C, an approximate photon flux density of 100 m mol m^-2 s^-1 supplied by 30-W Grolux and 30-W warm white fluorescent tubes, and a 16-h photoperiod. Seedlings were used 12 d after the end of the vernalization, when roots were 8 to 9 cm long.

General Procedure for Na⁺ Influx Experiments

In most experiments, entire root systems were excised from the shoot (to eliminate potential complications arising from transpiration) and were pretreated in 15 mL of unlabeled uptake solution for 10 min, with solutions changed after 5 min. This was to equilibrate cell wall-bound Ca²⁺ and Na⁺ with that of the external solution, eliminating binding of ²²Na⁺ to extracellular sites in a manner that could not be reversed with the rinses used. Channel blockers or modulators were also included during this pretreatment when tested. However, in some experiments, excised roots were pretreated for 3 h (instead of 10 min) in concentrations of sorbitol, PEG, or NaCl iso-osmotic with the uptake solution (plus 0.2 mm Ca²⁺ activity), to reduce the effects of osmotic shock on influx measurements. A longer recovery time after removal from the Phytagel and shoot excision was not used, to reduce long-term consequences arising from the removal of the photosynthate supply of the root. Recovery for up to 3 h had no effect on rates of influx, suggesting that any damage to roots arising from removal from the Phytagel did not have a large impact on influxes. In contrast, we found that removing roots from agar plates appeared to damage root function. Likewise, it was found that influx into roots excised from plants grown in solution culture was very similar to influx into roots grown on solid medium, suggesting that the damage to root hairs inevitable with the removal of plants from the solid medium did not significantly affect influx (P. Essah, unpublished data). As such, the most experimentally convenient protocol was selected.

Uptake was measured in 15 mL of unbuffered uptake solution containing various concentrations of NaCl and CaCl₂, 37 to 185 kBq L^-1 of ²²Na⁺, and channel blockers or modulators as indicated. Unless otherwise stated, influx solutions contained 50 mm NaCl and a Ca²⁺ activity of 0.2 mm, and influxes were measured over 2 min. Solutions were unbuffered, because the pH was found to remain unchanged after several 2-min uptake periods. There was no significant depletion of solutions during the course of up to 30 influx periods (monitored by measurements of the radioactivity of the solutions). To reduce effects of plant-to-plant variation and errors in both counting and weighing, 10 root systems were used for each influx assay.

At the end of the influx, roots were blotted and then transferred into 200 mL of ice-cold 200 mm NaCl plus 10 mm CaCl₂ for two successive rinses of 2 min and then 3 min. The aim of these rinses was to displace apoplastic ²²Na⁺ while inhibiting efflux from the root cells. Similar rates of influx were found if roots were rinsed in (iso-osmotic) 50 mm NaCl plus 10 mm CaCl₂, but the higher concentration was used to increase confidence that all apoplastic ²²Na⁺ was being removed in all treatments. Solutions were stirred on gently moving shakers (45 rpm). Roots were finally blotted gently, weighed rapidly, and transferred into plastic vials with 2.5 mL of scintillation cocktail (Optiphase Hisafe, Fisher Chemicals, Loughborough, UK). Samples were counted on a liquid scintillation counter (Beckman Instruments, Fullerton, CA). All chemicals used were analytical grade, except PEG 8000, which was ultra-pure and low in aluminum (catalog no. P-2139, Sigma-Aldrich). Ca²⁺ activities were calculated using GEOCHEM-PC v2.0 (Parker et al., 1995).

Transplantation and Application of Treatments in Solution Culture

For the growth experiment, 12-d-old seedlings (see “Preparation of Seedlings”) were transferred onto plastic supports with holes placed over hydroponic growth solution in deep trays. The nutrient solution contained 2.5 L of: 5.0 mm KNO₃, 0.25 mm KH₂PO₄, 2.0 mm MgSO₄, 0.1 mm Ca(NO₃)₂, 0.05 mm FeEDTA, plus micronutrients (70 μm H₃BO₃, 14 μm MnCl₂, 0.5 μm CuSO₄, 1 μm ZnSO₄, 0.2 μm Na₂MoO₄, 10 μm NaCl, and 0.01 μm CoCl₂). This was similar to Hoagland solution, but with the phosphate concentration reduced to one-tenth of that in the original Hoagland solution because phosphate toxicity can sometimes be observed in saline conditions (Grattan and Maas, 1984). In addition, one-tenth of the Ca²⁺ concentration was used, to enable effects of varying supplemental Ca²⁺ activities on Na⁺ accumulation to be seen.

An aquarium pump fitted with an air stone gently aerated the solution. A propagator cover was placed over the top, and its vents were opened after 2 d, by which time the plants should have largely recovered from the shock of their transplantation. Plants were grown in this solution for 4 d, after which the solution was changed, and treatments were applied. Treatments consisted of two NaCl concentrations of 1 and 50 mm and three Ca²⁺ activities (0.05, 0.2, and 3.0 mm, supplied as 0.1 mm Ca(NO₃)₂ and supplemental CaCl₂). For the 50 mm NaCl treatment, NaCl was supplied in two steps of 25 mm separated by a 2-d interval. This gradual increase enabled plants to adapt to the change in osmotic pressure. Solutions were topped up to 2.5 L daily with de-ionized water. Hydroponically grown plants were exposed to a temperature of 22°C, relative humidity of approximately 70%, and photoperiod of 10 h with a photon flux density of approximately 120 m mol m^-2 s^-1.

Fresh and Dry Mass Determination

Plants were harvested 4 weeks after transplanting (and hence 6 weeks after planting). For each plant, the root and base of the stem were rinsed in de-ionized water for a few seconds to reduce surface contamination by Na⁺ from the growth solution. Roots (i.e. all tissue below the hypocotyl) and shoots were separated and blotted, and fresh masses were determined quickly. Plant tissues were then oven-dried for 48 h at 70°C, and their dry masses determined. Water content (WC) of tissues was calculated as a percentage according to the following equation:

where FM and DM are the fresh and dry masses, respectively.

Tissue Na⁺ Concentration Determination

To determine tissue Na⁺ and K⁺ concentrations, the dried plant material was boiled for 20 min in 10 mL of 100 mm nitric acid, made up to volume with de-ionized water, and analyzed for Na⁺ and K⁺ using a flame photometer (M410, Corning, Palo Alto, CA). Tissue Ca²⁺ concentration was determined using an atomic absorption spectrophotometer (Baird Alpha 1, Cambridge, UK).