Separating multiple, short-term, deleterious effects of saline solutions on the growth of cowpea seedlings


  • Peter M. Kopittke,

    1. The University of Queensland, School of Land, Crop and Food Sciences, St. Lucia, Qld 4072, Australia
    2. Cooperative Research Centre for Contamination Assessment and Remediation of the Environment (CRC-CARE), The University of Queensland, St. Lucia, Qld 4072, Australia
    Search for more papers by this author
  • F. Pax C. Blamey,

    1. The University of Queensland, School of Land, Crop and Food Sciences, St. Lucia, Qld 4072, Australia
    2. Cooperative Research Centre for Contamination Assessment and Remediation of the Environment (CRC-CARE), The University of Queensland, St. Lucia, Qld 4072, Australia
    Search for more papers by this author
  • Thomas B. Kinraide,

    1. Agricultural Research Service, USA Department of Agriculture, Beaver, WV 25813–9423, USA
    Search for more papers by this author
  • Peng Wang,

    1. State Key Laboratory of Soil and Sustainable Agriculture, Institute of Soil Science, Chinese Academy of Sciences, Nanjing 210008, China
    2. Graduate School of Chinese Academy of Sciences, Beijing 100049, China
    Search for more papers by this author
  • Suzie M. Reichman,

    1. School of Civil, Environmental and Chemical Engineering, RMIT University, Melbourne, Vic. 3001, Australia
    Search for more papers by this author
  • Neal W. Menzies

    1. The University of Queensland, School of Land, Crop and Food Sciences, St. Lucia, Qld 4072, Australia
    2. Cooperative Research Centre for Contamination Assessment and Remediation of the Environment (CRC-CARE), The University of Queensland, St. Lucia, Qld 4072, Australia
    Search for more papers by this author

Author for correspondence:
Peter M. Kopittke
Tel: +61 7 3346 9149


  • Reductions in plant growth as a result of salinity are of global importance in natural and agricultural landscapes.
  • Short-term (48-h) solution culture experiments studied 404 treatments with seedlings of cowpea (Vigna unguiculata cv Caloona) to examine the multiple deleterious effects of calcium (Ca), magnesium (Mg), sodium (Na) or potassium (K).
  • Growth was poorly related to the ion activities in the bulk solution, but was closely related to the calculated activities at the outer surface of the plasma membrane, {Iz}0o. The addition of Mg, Na or K may induce Ca deficiency in roots by driving {Ca2+}0o to < 1.6 mM. Shoots were more sensitive than roots to osmolarity. Specific ion toxicities reduced root elongation in the order Ca2+ > Mg2+ > Na+ > K+. The addition of K and, to a lesser extent, Ca alleviated the toxic effects of Na. Thus, Ca is essential but may also be intoxicating or ameliorative.
  • The data demonstrate that the short-term growth of cowpea seedlings in saline solutions may be limited by Ca deficiency, osmotic effects and specific ion toxicities, and K and Ca alleviate Na toxicity. A multiple regression model related root growth to osmolarity and {Iz}0o (R2 = 0.924), allowing the quantification of their effects.


Saline soils are of global importance and, although naturally occurring, their formation is greatly accelerated by human activity. It has been estimated that between 20% and 50% of all irrigated lands world-wide are affected by excess salt, costing an estimated US$12 billion yr−1 in lost agricultural productivity (Pitman & Läuchli, 2002). Saline soils tend to be dominated by sodium (Na) salts, although some soils may contain excess potassium (K) or magnesium (Mg). For example, ultramafic (serpentine) soils, such as those in Scotland, Morocco or Australia, often contain high Mg, with soil solution Mg : Ca as high as 18 (Johnston & Proctor, 1981; Nagy & Proctor, 1997). High Mg concentrations also occur through the influence of seawater. For example, before vegetation is established, bauxite residue from aluminium mining is often neutralized by leaching with seawater, after which the soil solution of bauxite residue contains an excess of both Na and Mg (Menzies et al., 2009). Thus, there is a need to investigate not only the effects of high Na, but also of K and Mg, on the growth of plants.

Munns (2002) proposed that salinity reduces plant growth through three main mechanisms. First, an increase in the osmotic potential results in a decrease in growth because of water stress. In this instance, the magnitude of the growth reduction is directly related to the osmotic potential of the solution. Second, elevated levels of salts may induce calcium (Ca) deficiency or other nutritional imbalances. The development of Ca deficiency under saline conditions has long been established, even in the presence of external Ca concentrations generally regarded as adequate (Adams, 1966). This has been attributed to the displacement of cell membrane-associated Ca2+ by competitive cations, such as Na+, K+ and Mg2+ (Cramer & Läuchli, 1986; Cramer, 2002). Third, ions such as Na+ and Cl may be directly phytotoxic, accumulating within the cytoplasm to levels that inhibit normal cellular processes. The separation of these multiple toxic effects of salinity is not straightforward and comparatively few studies have provided a detailed examination of these three mechanisms in a single experimental system (most probably as a result of the extremely large number of treatments required to adequately separate these toxic effects).

Many studies have investigated the relationships between plant growth and the concentrations or activities of ions in the bulk solution. There is increasing evidence, however, that plant responses to ions (inhibition, alleviation of inhibition and transport) are dependent on their activities at the outer surface of the plasma membrane (PM) of root cells (for examples, see Table 5 of Kinraide & Wang (2010)). The PM surface carries negative charges, thus creating a negative surface potential (ψ0o) (Kinraide, 2006). The presence of these negative charges influences the distribution of ions at the PM surface, increasing the activity of cations, but decreasing that of anions (hereafter, the activity of an ion I with charge z in the bulk solution is designated as {Iz}b, whereas its activity at the outer surface of the PM is designated as {Iz}0o). Based on this electrostatic interaction, Kinraide (1999) examined the role of {Iz}0o on the toxic effects of excess Na and K in wheat (Triticum aestivum) seedlings, concluding that, although both root and shoot growth are reduced because of osmotic stress, Na is not toxic unless Ca is also deficient. A limited number of other studies, including those of Suhayda et al. (1990) and Reid & Smith (2000), have also considered the role of ψ0o in salinity – these authors concluded that ψ0o is likely to be of importance in influencing toxicity and ion transport in saline systems. Extending our knowledge regarding plant–ion interactions in saline systems is of importance in terms of both the breeding of plants that are better adapted to these environments and the efficient amelioration of saline lands.

The aim of the current study was to provide further interpretations of the short-term, phytotoxic effects of salinity, giving particular consideration to changes in ion activities at the outer surface of the PM of root cells. In addition to Na, the effects of excesses of Ca, Mg and K were also investigated, with a total of 404 treatments investigated in seven experiments.

Materials and Methods

General experimental methods

Short-term solution culture experiments were conducted as described by Kopittke et al. (2008) in a laboratory (140 μmol m−2 s−1 of photosynthetically active radiation) maintained at 25°C at The University of Queensland, St. Lucia, Australia. Briefly, 3-d-old seedlings of cowpea (Vigna unguiculata (L.) Walp. cv Caloona) were placed in Perspex strips (seven seedlings per strip) on top of 600-ml glass beakers filled to the brim (650 ml) with a basal solution of 1 mM CaCl2 and 5 μM H3BO3 at pH 5.6. Seedlings were grown for c. 18 h before transfer to the treatment solutions in different beakers for a further 48 h. All solutions were continuously aerated. Root and shoot lengths were determined using photography at the time of transfer (0 h) and at harvest 48 h later. Previous investigations using this experimental system have shown that the elongation rates of roots growing in Ca-sufficient and toxicant-free solutions are almost completely constant during this 48-h period (see Fig. 2 of Kopittke et al. (2009)).

Two replicates of all treatments were imposed with chloride salts of Ca, Mg, Na and K and with mannitol (Supporting Information Tables S1–S7). As a maximum of 66 experimental units (beakers) could be set up at any given time, the experiments described below were conducted over multiple months under similar conditions (the maximum number of experimental units was increased to 120 for Expt 5 onwards). Control treatments were included in all experiments to check for good growth under nonsaline conditions. Other than in Expt 1, all solutions contained 5 μM boron (B) as H3BO3 and a minimum of either 1 mM CaCl2 or a calculated 3.5 mM {Ca2+}0o (whichever was higher) to ensure that growth was not limited by Ca deficiency. Boron has a low phloem mobility and hence must be supplied continuously in the rooting medium (Brown & Shelp, 1997). In the current study, given that > 99.9% of B would have been present as undissociated (and uncharged) H3BO3, variations in ψ0o would not have influenced B nutrition.

Ion activities in the bulk solution were calculated using PhreeqcI 2.15 (Parkhurst, 2009) employing the default Phreeqc database. When preparing treatments, estimates of ψ0o and {Iz}0o were calculated using Eqns 1 and 11 of Kinraide (2006). Final calculations of ψ0o and {Iz}0o were performed using the Gouy–Chapman–Stern (GCS) model (computer program available from T. B. Kinraide) with measured pH values (measured at both 0 and 48 h, see Tables S1–S7). Osmolarity (mOsM) was calculated as:

image(Eqn 1)

where ϕ is the osmotic coefficient (being 1.86 for NaCl, 1.85 for KCl, 2.58 for MgCl2, 2.56 for CaCl2 and 1.04 for mannitol; Robinson & Stokes, 2002) and C is the concentration of solute i (mM).

Ca deficiency

Expt 1 was designed to determine the effects of salinity (excess Mg, Na or K) on root and shoot growth as affected by {Ca2+}b and {Ca2+}0o. These were varied both by changing the Ca concentration in the bulk solution and by adding Mg, Na or K, which marginally decreased {Ca2+}b, but also depolarized PM, thereby decreasing {Ca2+}0o. A total of 60 treatments was investigated in a 6 × 10 factorial design, containing 2 μM to 4 mM Ca and up to 20 mM Mg, 30 mM Na or 30 mM K. The osmolarity of all solutions was ≤ 66 mOsM.

Osmotic effects and specific ion toxicity

Expt 2 was designed to elucidate the osmotic effects of salinity on the growth of cowpea seedlings. The effects of excess Mg, Na or K with adequate Ca were compared with those of solutions containing only Ca, Ca plus d-mannitol (M1902; Sigma) or a mixture of salts in which Ca, Mg, Na and K contributed jointly to the osmolarity. The highest osmotic treatments were c. 400 mOsM (c.−1 MPa). There were 72 treatments in a 6 × 12 factorial design. The viscosity of the mannitol-containing solutions did not reduce the rate of solution aeration.

Expt 3 was designed to provide further information on specific ion toxicities. A total of 28 treatments was established with 14 treatments at 150 mOsM and 14 treatments at 250 mOsM. There were the following treatments: six in which Mg partially replaced Ca, six in which K replaced Na, one containing mannitol and Ca only, and one containing a mixture of salts in which Ca, Mg, Na and K contributed jointly to the osmolarity.

Expt 4 was designed to determine whether specific ion toxicities were related to {Iz}b or {Iz}0o. A total of 60 treatments was established, investigating the specific ion toxicities of Ca and Na. Solutions were established with increasing concentrations of Ca or Na, ranging up to c. 280 mOsM. There were four treatments at each Ca or Na concentration, established by the addition of small concentrations of H and/or aluminium (Al). The H+ activity used (32 μM, pH 4.50) and the Al concentrations used (≤ 10 μM) were sufficient to cause some depolarization of PM (and hence decrease in {Ca2+}0o or {Na+}0o), with negligible effects on ionic strength and {Iz}b and negligible direct effects on root elongation.

Alleviation of specific ion toxicities

Expt 5 was designed to investigate the alleviation of specific ion toxicities. Of the 47 treatments established, 17 investigated K alleviation of Ca, Mg or Na toxicity, 15 investigated Mg alleviation of Ca, Na or K toxicity and 15 investigated Na alleviation of Ca, Mg or K toxicity. Sufficient amounts of the intoxicating cation were added to all solutions to achieve an osmolarity of 250 mOsM, and the alleviating cation was added at five rates: 0, 0.5, 1, 2.5 or 7.5 mM (two additional K rates were included, 0.1 and 0.2 mM, to study K alleviation of Na toxicity). The concentrations of the intoxicating cation were not reduced, and hence the osmolarity increased slightly (≤ 270 mOsM) as the alleviating cation was added.

Expt 6 was designed to further investigate K alleviation of Na or Mg toxicity. A total of 77 treatments was established, with seven Na and four Mg concentrations each with seven rates of K addition. As with Expt 5, the concentrations of the intoxicating cation were not reduced and hence the osmolarity increased slightly as the alleviating cation was added.

Data analysis

The root elongation rate (RER) and shoot elongation rate (SER) were calculated and analysed using regression analysis, fitting either linear curves or curves of the general form

image(Eqn 2)

where b is the maximum growth rate in toxicant-free and Ca2+-sufficient solutions, c is a strength coefficient and increases with the strength of the toxicant, T is the toxicant intensity (i.e. osmotic strength (mOsM), {Iz}b (mM) or {Iz}0o (mM)) and h is a shape coefficient (Taylor et al., 1991; Kinraide, 1999). If acting by similar mechanisms (see Kinraide, 1999), Eqn 2 can be modified for multiple toxicant variables as

image(Eqn 3)

Regression analyses were conducted using SYSTAT 12 (Cranes Software International Ltd, Bangalore, India). Curves were plotted for which R2 ≥ 0.5, an arbitrary decision to ensure the meaningful determination of relationships between growth and toxicant variables. Unless otherwise stated, no coefficients are reported for which the 95% confidence interval encompassed zero.

Expt 7 was conducted to examine the ability of these regression relationships (as fitted for Expts 1–6) to predict plant growth across a wide range of solution compositions. A total of 60 treatments was established with osmolarities ranging from 3 to 420 mOsM. All solutions contained ≥ 3.5 mM {Ca2+}0o and Mg, Na and K at randomly allocated rates ≥ 0 mM. This gave solutions with two, three or four cations with a range of compositions.

Across this entire study, a total of 404 treatments was investigated, giving 808 experimental units (two replicates), 5656 seedlings (seven seedlings per replicate) and 15 204 length measurements (0 and 48 h for both roots and shoots). Shoot length was not measured in Expts 4–7. The calculated values of ψ0o ranged between −98.3 and 10.1 mV (Tables S1–S7). Across all treatments, solutions contained up to (mM) 135 Ca, 120 Mg, 200 Na and 200 K (although most treatments were substantially lower than these maxima, see Tables S1–S7). These values are similar to those sometimes found in nature, such as in solutions of saline soils (Eaton et al., 1960; Wolt, 1994) or in soils influenced by seawater (which contains c. 53 mM Mg and 470 mM Na; DOE, 1994). For example, Eaton et al. (1960) reported that solutions of saline soils from California (USA) contained up to 120 mM Ca and 110 mM Mg.


Ca deficiency

Poor relationships were found between {Ca2+}b and both RER (R2 = 0.482) and SER (R2 = 0.117) in Expt 1 (Fig. 1a,c). Indeed, the addition of Mg, Na or K influenced the {Ca2+}b required for maximum growth. Using root growth as an example, {Ca2+}b associated with a 50% reduction in RER varied by three orders of magnitude, from < 2 μM in solutions containing only Ca to c. 2 mM in solutions also containing 20 mM Mg.

Figure 1.

 Effects of the Ca2+ activity in the bulk solution {Ca2+}b (a, c) and at the outer plasma membrane surface {Ca2+}0o (b, d) on the elongation rate of roots (a, b) and shoots (c, d) of cowpea seedlings over 48 h (Expt 1). All solutions had an osmolarity of ≤ 66 mOsM to ensure that growth was not reduced by osmotic effects (Fig. 2). Regressions are plotted where R2 ≥ 0.5. The x-axes are plotted on a logarithmic scale to allow the presentation of values for {Ca2+}0o, which vary by c. four orders of magnitude, although elsewhere regression analyses are performed on a linear scale (for example, Eqn 6). The legend applies to all panels. Chloride salts were used throughout.

In contrast with {Ca2+}b, both root (RER, R2 = 0.904) and shoot (SER, R2 = 0.569) growth was significantly related to {Ca2+}0o irrespective of solution composition (Fig. 1b,d). A decrease in {Ca2+}0o had a greater adverse influence on root growth than on shoot growth; root growth was almost totally inhibited at low {Ca2+}0o, whereas SER decreased by only c. 40% (Fig. 1b,d). Critical values corresponding to a 10% reduction in elongation rate were calculated as 1.6 mM {Ca2+}0o for the roots and 0.79 mM {Ca2+}0o for the shoots.

Osmotic effects

The difficulty in distinguishing between nonspecific osmotic stress and that caused by specific ions is often addressed by the use of mannitol or other organic solutes. Growth in solutions containing mannitol was generally less than in iso-osmotic solutions of mixed salts (Ca, Mg, Na and K) (Fig. 2a,b). This reduction in growth, however, was not caused by a toxic effect of mannitol, but by combined effects of {Iz}0o and osmolarity (see section on Modelling growth reductions). Regardless, mixed salt solutions are suitable as indicators of osmotic stress (Fig. 2a,b). Shoot growth was generally more sensitive to osmotic stress than was root growth, as shown by the linear decrease in SER over the range studied (Fig. 2a,b). A 50% decrease in SER was evident at c. 180 mOsM and of RER at c. 310 mOsM (Fig. 2a,b). (Interestingly, high concentrations of mannitol caused no distortion of roots, as is evident in those grown at high osmolarity in mixed salt solutions (Fig. S1).)

Figure 2.

 Osmotic effects on the elongation rate of cowpea roots (a, c) and shoots (b, d) over 48 h (Expt 2). The ‘Ca’ solutions contained only Ca, whereas the other solutions contained both the ion of interest (Mg, Na, K or mannitol) and c. 4 mM {Ca2+}0o (Fig. 1). The ‘mixed’ solutions contained Ca, Mg, Na and K at concentrations which contributed jointly to the osmolarity. In (a) and (b), the regressions are fitted only to the mixed salt data (closed squares for mixed; open squares for mannitol) and plotted again in (c) and (d). For (a), y = 1.52/exp[(0.003x)4.67]; for (b) y = −0.00635x + 2.20. The treatment with the highest osmolarity is excluded from the linear regression for shoot data. The key applies to the two panels (c) and (d) in this figure.

Specific ion toxicity

Although an increase in osmolarity reduced growth, solutions dominated by a single salt tended to reduce growth more than did iso-osmotic mixed salt solutions (Fig. 2c,d). For example, a 200 mOsM mixed salt solution reduced root growth by only 10%, whereas a Ca-only solution at the same osmolarity reduced root growth by c. 65% (Fig. 2c). These specific ion toxicities were less pronounced for shoots than for roots (Fig. 2d).

Expt 3 investigated the effects of Mg substitution of Ca, and of K substitution of Na, at constant osmolarity. Results at 250 mOsM confirmed the greater toxicity of Ca than Mg in roots and shoots (Fig. 3a,c). The greater toxicity of Na than K was evident in roots (Fig. 3b), but not in shoots (Fig. 3d) (cf. Fig. 2c,d). Interestingly, a small substitution of K for Na increased RER to a similar value as that of mixed salts at 250 mOsM, providing evidence of K alleviation of Na toxicity (to be described further). Similar findings were evident at 150 mOsM (Fig. S2).

Figure 3.

 Effects of substituting Mg for Ca (a, c, circles) and K for Na (b, d, circles) at 250 mOsM on the elongation rate of cowpea roots (a, b) and shoots (c, d), compared with iso-osmotic mixed salt solutions of Ca, Mg, Na and K at the same osmolarity (solid line) (Expt 3). All solutions contained c.≥ 3.5 mM {Ca2+}0o to ensure that growth was not limited by Ca deficiency (Fig. 1b). Similar effects were evident at 150 mOsM (Supporting Information Fig. S2).

As was demonstrated for Ca deficiency (Fig. 1), specific ion toxicities were more closely related to {Iz}0o than to {Iz}b, especially for Na (Fig. 4). In Expt 4, {Iz}0o and {Iz}b were separated through the addition of low levels of H and Al, which were not sufficient to change the ionic strength (and hence bulk activity) appreciably, but resulted in some depolarization of PM and decrease in {Iz}0o (the addition of Al decreased the negativity of ψ0o by c. 10–30 mV at low ionic strengths, see Table S4). However, it was more difficult to separate {Ca2+}0o and {Ca2+}b because the extent of PM depolarization was substantially larger in Ca-toxic solutions than in Na-toxic solutions (Table S4). Hence, the depolarization resulting from H and Al additions (and consequent changes in {Ca2+}0o) was comparatively small.

Figure 4.

 Evaluation of the toxic effects of Na on the elongation rate of cowpea roots using data from Expt 4 and Treatments 85–91 of Expt 2 (see Supporting Information Table S2). Root elongation rate is related to the activity of Na in the bulk solution {Na+}b (a) or the activity of Na at the outer surface of the plasma membrane {Na+}0o (b). Additions of H and Al were made at nontoxic concentrations to depolarize the plasma membrane (and reduce {Na+}0o), with negligible influence on ionic strength (and consequently {Na+}b). All solutions contained c. ≥ 3.5 mM {Ca2+}0o to ensure that root growth was not limited by Ca deficiency (Fig. 1b).

Alleviation of specific ion toxicity

Because replacement of only a small proportion of Na by K in iso-osmotic solutions at 250 mOsM disproportionately improved root growth (Fig. 3b), Expt 5 further investigated this effect. A concentration of 0.5 mM K (corresponding to c. 0.5 mM {K+}0o in these solutions) markedly improved root growth in Na-toxic solutions at 250 mOsM (Fig. 5b). There appeared to be a slight alleviation of Mg toxicity (Fig. 5b), but this effect was not significant. Additions of K to Ca- or Mg-toxic solution did not improve root growth (Fig. 5b), nor did the addition of Mg (Fig. 5a) or Na (Fig. 5c) to any of the solutions.

Figure 5.

 Effects on the elongation rate of cowpea roots of adding Mg (a), K (b) or Na (c) to solutions containing c. 250 mOsM Ca, Mg, Na or K, compared with effects of iso-osmotic solutions of Ca, Mg, Na and K at the same osmolarity (Expt 5). Osmolarity increased as the cation was added to a maximum of c. 270 mOsM. All solutions contained c.≥ 3.5 mM {Ca2+}0o to ensure that root growth was not limited by Ca deficiency (Fig. 1b).

The alleviation of Na toxicity was further demonstrated by the results of Expt 6, in which K was added to solutions containing excess Na (Fig. 6a). Increasing osmolarity decreased RER, but K had an alleviatory effect from c. 100 to 300 mOsM. By contrast, there was no alleviation of Mg effects by increased {K+}0o (Fig. 6b).

Figure 6.

 Elongation rate of cowpea roots as affected by the addition of K (expressed as the activity of K+ at the outer surface of the plasma membrane, {K+}0o) to solutions containing excess Na (a) or excess Mg (b) (Expt 6). The bulk solution Na or Mg concentration was held constant in each series, and hence {Na+}0o or {Mg2+}0o decreased marginally and osmolarity increased marginally as K was added. All solutions contained c.≥ 3.5 mM {Ca2+}0o to ensure that root growth was not limited by Ca deficiency (Fig. 1b).

Modelling growth reductions

The development of a model to explain the short-term salinity effects of Ca, Mg, Na or K requires the assessment of Ca deficiency effects, osmotic effects, specific ion effects and the alleviation of specific ion effects. Thus, RER was assessed as a function of osmolarity (mOsM) and {Iz}b (mM), modifying Eqn 3 to include a term for Ca deficiency, (1 − 1/exp [c{Ca2+}b]) (Eqn 4). The first part of Eqn 4 provides a term to show that an increase in Ca increases RER (i.e. overcomes Ca deficiency), whereas, in the second part, an increase in osmolarity, Ca, Mg, Na or K decreases RER (i.e. induces toxicity). Initial estimates of the strength coefficients (i.e. ch) were calculated from individual analyses using Eqn 2 of relevant datasets. For example, ‘c’ was initially estimated from Expt 1, ‘d’ from the mixed salt treatments of Expt 2, etc. Fitting Eqn 4 resulted in R2 = 0.732 (= 344), although the coefficients for Mg and Na were not significant (i.e. Mg and Na did not have additional toxic effects when expressed as activities in the bulk solution), whereas the coefficient for K was negative (i.e. K improved growth).

image(Eqn 4)

The alternative approach used in this study is to relate RER to osmolarity and {Iz}0o by modifying Eqn 4 to

image(Eqn 5)

resulting in R2 = 0.882 (cf. the R2 value of 0.732 for osmolarity and {Iz}b). To include K alleviation of Na toxicity (Figs 3b, 5b), the term for Na was expanded to g0/(1 + g1{K+}0o){Na+}0o (Eqn 6). The alleviation of Na toxicity by the addition of Ca was not easy to investigate directly, given that all solutions contained basal Ca and that Ca causes substantial depolarization of PM (and hence the bulk Na concentration must also be increased to keep {Na+}0o constant, thereby also increasing the osmolarity). However, the addition of a further term for Ca alleviation of Na toxicity (thus expanding the term for Na to g0/(1 + g1{K+}0og2{Ca2+}0o){Na+}0o) was investigated (Eqn 6). It is noteworthy that, when a term was added to account for K alleviation of Mg toxicity, f0/(1 + f1{K+}0o){Mg2+}0o, the 95% confidence interval for f1 (−0.0040 to 0.022) encompassed zero, which is in keeping with the finding that K does not significantly alleviate Mg toxicity (Fig. 6b). Eqn 6 was fitted to the data:

image(Eqn 6)

resulting in R2 = 0.924; this equation related RER to the effects of osmolarity {Iz}0o for both Ca deficiency and specific ion toxicities, and the alleviation of Na toxicity by the addition of K and Ca (Fig. 7a, Table 1). It is also noteworthy that, when a term was added for mannitol, the 95% confidence interval (−0.00021 to 0.00014) encompassed zero, suggesting that mannitol had no direct toxic effects. Rather, the reduced growth observed in mannitol-containing solutions (Fig. 2a,b) can be explained entirely by changes in {Iz}0o (particularly {Ca2+}0o) and osmolarity. The data also suggest the possibility that increases in {Ca2+}0o marginally improve growth in the presence of toxic {K+}0o. This effect is only comparatively small, but further investigation is warranted.

Figure 7.

 Comparison between measured cowpea root elongation rate (RER) and predicted RER using Eqn 6, which takes into account osmotic effects, toxicities of Ca, Mg, Na and K (including the alleviation of Na toxicity by K and Ca) and Ca deficiency. In (a), data from Expts 1–6 (= 344) were fitted to Eqn 6. In (b), the coefficients determined from Expts 1–6 (Table 1) were validated using data from Expt 7 (= 60). Both linear regressions (black) and 1 : 1 lines (grey) are fitted. The R2 values correspond to the linear regression.

Table 1.   Estimates of the coefficients (and errors) for Eqn 6 when fitted to the 344 treatments from this study (Expts 1–6)
CoefficientEstimateStandard errorEstimate/standard error95% confidence interval
  1. The equation, root elongation rate (RER) = b[1 − 1/exp(c{Ca2+}0o)]/exp{[d(osmolarity) + e{Ca2+}0o + f{Mg2+}0o+g0/(1 + g1{K+}0o+ g2{Ca2+}0o){Na+}0o + h{K+}0o]i}, included terms for the effects of osmolarity {Iz}0o in relation to both Ca deficiency and specific ion toxicities, and alleviation of the specific ion toxicity of Na by the addition of K and Ca. Ion activities at the outer surface of the plasma membrane ({Iz}0o) are expressed as mM and osmolarity as mOsM.


There were no systematic deviations from the 1 : 1 line in Fig. 7(a), but a shortcoming exists in that the same data were used to determine and to test the parameters in Eqn 6. Therefore, Expt 7 used 60 random treatments of two, three or four ions to test the parameters calculated from the data in Expts 1–6 (Table 1). Linear regression analysis demonstrated a good relationship between the measured and predicted RER (R2 = 0.918, Fig. 7b), although the measured RER tended to be slightly lower than predicted. Thus, Eqn 6 (see also Table 1) appears to satisfactorily explain the multiple factors limiting short-term root growth of cowpea in saline solutions.

Finally, Cl toxicity was not assessed by anion substitution experiments. Nevertheless, an assessment was possible through the use of Eqn 6 because {Cl}0o is only weakly collinear with osmolarity (R2 = 0.551) and is even less correlated with other parameters in Eqn 6. Regression analysis of root elongation by Eqn 6, as it appears in Table 1, yielded R2 = 0.924. Deletion of d(osmolarity) from the equation resulted in R2 = 0.848. Replacement of d(osmolarity) with d1{Cl}0o resulted in d1 = 0.00169 (significant) and R2 = 0.873. Inclusion of both d(osmolarity) and d1{Cl}0o resulted in d = 0.00144 (significant), d1 = 0.000162 (not significant) and R2 = 0.924. Consequently, specific ion toxicity from Cl was at least 10-fold less than that for K+ and nearly 100-fold less than that for Na+, if it was toxic at all. Like all ions, Cl contributed to osmolarity.


Ca deficiency

Cowpea RER was decreased at ≤ 1.6 mM {Ca2+}0o, which is slightly higher than the value of 0.76 mM for wheat roots (Kinraide, 1999) and lower than the value of c. 6 mM for pea (Pisum sativum) and wheat (Wang et al., 2010) or c. 15 mM for melon (Cucumis melo) (Yermiyahu et al., 1997). Although all three cations (Mg, Na and K) reduced root elongation by inducing Ca deficiency, the effect was more pronounced for Mg than for Na or K. This is evident by comparing the {Ca2+}b required for maximal root elongation in solutions containing 30 mM Na with that in solutions containing 20 mM Mg (Fig. 1a). Although all cations reduced the negativity of ψ0o (and hence reduced {Ca2+}0o), the magnitude of the reduction was a function of ion charge and the strength of binding to the PM surface. For ions commonly of environmental and agricultural importance, the predicted effectiveness at reducing the negativity of ψ0o follows the order: Al3+ > H+ > Ca2+ ≈ Mg2+ > Na+ ≈ K+ (Kinraide, 2006; Wang et al., 2008). Therefore, the addition of Mg results in a larger decrease in the negativity of ψ0o (and hence a larger decrease in {Ca2+}0o) than Na or K, as is evident in modelled data (Fig. S3a). As a consequence, reductions in growth under saline conditions as a result of Ca deficiency are more likely to be induced by high Mg than high Na or K. As an example, if it is assumed that saline or sodic soils typically contain ≥ 0.5 mM Ca in the soil solution (Naidu et al., 1995; Chhabra, 1996), root growth would be expected to be limited by Ca deficiency (i.e. {Ca2+}0o ≤ 1.6 mM) with ≥ 4.2 mM Mg in the bulk solution, but only when Na or K are ≥ 100 mM (Fig. S3b). With a higher solution Ca concentration of 2 mM, root growth would be expected to be limited by Ca deficiency with ≥ 22 mM Mg, or with ≥ 400 mM (≥ 750 mOsM) Na or K (Fig. S3b). Given that most solutions of saline soils contain > 2 mM Ca, growth is more likely to be reduced by high osmolarity than by Ca deficiency when salts are dominated by Na or K.

The results of the current study confirm that RER is more closely related to {Ca2+}0o than to the concentration or activity of Ca, or to the ratio of other ions to Ca, in the bulk solution (Yermiyahu et al., 1997; Kinraide, 1999; Wang et al., 2010). We propose that the close relationship between {Ca2+}0o and root growth provides an advance in understanding the interactive effects of Ca and other cations in saline conditions. These effects have almost always been expressed either as the ratio of the concentrations of a single ion to that of Ca (for example, Na : Ca) or of total cations (TC) to Ca (TC : Ca) in the bulk solution (for reviews, see Grattan & Grieve (1992) or Grattan & Grieve (1999)). It has been reported that adverse effects are evident when Na : Ca is > c. 10–20 (Greenway & Munns, 1980; Cramer, 2002) or when TC : Ca is > c. 5–10 (Grattan & Grieve, 1992). The results of the current study, however, demonstrate that a reduction in growth is not closely related to cation ratios in bulk solution. Indeed, a 10% reduction in RER occurred with Na : Ca from 83 (1 mM Na and 0.012 mM Ca) to 32 (30 mM Na and 0.95 mM Ca); corresponding values for TC : Ca were from 83 (1 mM Na and 0.012 mM Ca) to c. 5 (20 mM Mg and 4 mM Ca). These data show that there is no single value, either of Na : Ca or TC : Ca, that can be used to predict short-term root growth of cowpea seedlings. Re-examination of published data also demonstrates the importance of {Ca2+}0o as a measure for predicting Ca deficiency (Notes S1).

In the current study, shoot growth was not as markedly decreased by Ca deficiency as was root growth, which had largely ceased at low {Ca2+}0o (Fig. 1b,d). This is consistent with some previous observations that shoot growth of seedlings is reduced less by Ca deficiency than is root growth [for example, see Tanaka & Woods (1972)]. It is likely that this results from the short-term nature of studies conducted in the current experimental system. Because Ca does not move towards the root tip, Ca must be present in the test solution to maintain structural and functional integrity. Indeed, it has been noted that root growth is reduced rapidly when placed in solutions lacking Ca (Burstrom, 1953; Kinraide, 1998; del Amor & Marcelis, 2003). It is noteworthy that a reduction in root elongation is more rapid in nutrient solutions free of Ca than in deionized water (Marschner, 1995). However, shoots can be supplied from the Ca reserves of the seed; hence, shoot growth did not respond greatly to low Ca in the test solutions, whereas root growth decreased rapidly (data not shown) and markedly. It is likely that larger variations in shoot growth would occur following depletion of the seed nutrient reserves in longer term experiments.

Osmotic stress

Using mixed salt solutions, shoot growth was generally more sensitive to osmotic stress than was root growth; a 50% reduction occurred at c. 310 mOsM in roots and at c. 180 mOsM in shoots (Fig. 2a,b). The magnitude of this growth reduction corresponds well with that of previous studies, although considerable variation exists among species (for example, see Fig. 1 of Munns & Tester (2008)). Further, as is evident in Fig. 1 of Kinraide (1999), for example, a threshold was observed for root growth, but not for shoot growth; although shoot growth was sensitive to any increase in osmolarity, increases in osmolarity in solutions < 200 mOsM caused only a small reduction in RER (Fig. 2a,b).

Specific ion toxicities and alleviation

The primary site of specific ion toxicity is generally considered to be the leaves, where salts accumulate to levels exceeding the ability of the cells to compartmentalize ions in the vacuole (Munns, 2002). However, in the current study, specific ion toxicities were observed primarily in the roots, with each of the ions Ca2+, Mg2+, Na+ and K+ reducing growth more than could be explained solely by osmotic effects (Figs 2, 3, S2). This apparent discrepancy regarding the site of specific ion toxicity (roots vs shoots) probably results from differences in the length of experimental periods in the various studies. In short-term experiments, there would be little deposition of salts in shoot tissues. In long-term experiments, it is likely that specific ion toxicities would be observed at lower solution concentrations, resulting not from toxicity at the site of uptake (roots), but from prolonged deposition from the transpiration stream (shoots). Regardless, the observation that solutions dominated by a single salt are more toxic than mixed salt solutions are consistent with previous reports, such as the study of Termaat & Munns (1986) with barley (Hordeum vulgare) and wheat.

As was evident with Ca deficiency, specific ion toxicities were related to {Iz}0o rather than {Iz}b, especially for Na (Fig. 4). For the roots, the coefficients for Ca, Mg, Na and K toxicity (Table 1) in Eqn 6 were all significant, thereby indicating that all four cations were toxic in addition to their osmotic effects. However, the extent to which these cations were toxic varied, with Ca being the most and K the least toxic to roots. The same level of growth inhibition as defined by {Ca2+}0o required 1.5 times {Mg2+}0o, 3.2 times {Na+}0o or 29 times {K+}0o (Table 1). However, care must be taken when extrapolating these findings to other experimental conditions and other species.

The addition of K improves growth of a range of plant species under Na-toxic conditions, possibly through the maintenance of a favourable Na : K ratio in the cytoplasm, as required for protein synthesis and enzymatic reactions (Leigh & Wyn Jones, 1984; Shabala & Cuin, 2008). It has been proposed also that increased Na concentrations decrease K uptake and increase K efflux (Shabala & Cuin, 2008). It is possible that any K efflux may be related to changes in ψ0o (more specifically, to the surface-to-surface transmembrane potential difference Em,surf, which is inversely related to ψ0o), although further work is required. The current study confirmed the beneficial effect of increased K on cowpea root growth under Na-toxic conditions (Figs 5b, 6a, S2b). There was a substantial improvement in RER with a small addition of K, with no clear extra benefit at higher K concentrations. Regression analysis of data from Expt 6 (Fig. 6a) using Eqn 7 yielded a negative value for e1, indicating that the magnitude of {K+}0o required for alleviation increases as {Na+}0o increases (b = 1.15, = 0.00598, = 4.21, = 0.345, = 6.52 and e1 = −0.00425, the 95% confidence intervals for e1 being −0.00091 to −0.00759). For example, c. 0.2 mM {K+}0o is required for 90% of maximum alleviation at 60 mM {Na+}0o, but c. 0.7 mM {K+}0o is required at 160 mM {Na+}0o.

image(Eqn 7)

It was difficult to demonstrate Ca alleviation of Na toxicity because Ca is an essential requirement for root growth even in short-term studies (Fig. 1b). However, the parameters of Eqn 7 (Table 1) indicated a small ameliorative effect of Ca, the effect being less pronounced for K. The coefficients (3.97 for K and 1.13 for Ca) showed that K was c. 3.5 times more effective than Ca at alleviating Na toxicity. Although Na appeared to possibly interfere with K metabolism, as evidenced by the alleviation if its toxic effects, the mechanisms by which the other cations are toxic are unknown. Indeed, the highly toxic, specific ion effects of Ca were unexpected.

Interactive effects of Ca in salinity toxicity

The preceding discussion suggests at least five effects of Ca in salinity toxicity. First, Ca2+ is essential for root elongation even in the absence of toxicants, but the addition of a PM-depolarizing solute may reduce {Ca2+}0o to growth-limiting activities. Indeed, the term (1 − 1/exp[c{Ca2+}0o]) in Eqn 6 demonstrates that growth is reduced as {Ca2+}0o approaches zero. Second, Ca2+ depolarizes the PM surface, thereby decreasing the PM surface activity of cationic toxicants (but increasing the PM surface activity of anionic toxicants). This effect is taken into account when toxicants are expressed as PM surface activities (such as {Na+}0o or {SeO42–}0o). Third, Ca2+ contributes to the reduction of water potential (osmolarity) and thereby contributes to toxicity. Fourth, Ca2+ at high PM surface activities is intoxicating to roots, causing specific ion toxicity. This effect is independent of osmotic effects and is expressed using the term e{Ca2+}0o in Eqn 6. Finally, Ca2+ specifically alleviates the effects of some toxicants (for example, Na+ toxicity in the current study, but also H+ toxicity (Kinraide, 1998)). This effect is independent of PM depolarization (the second effect), and is expressed by the expansion of the strength coefficient g in Eqn 6 to g0/(1 + g1{K+}0o + g2{Ca2+}0o).


This study set out to separate the multiple deleterious effects of salinity on the growth of cowpea seedlings based on ion activities at the outer surface of PM. This approach resulted in a model (Eqn 6) that relates RER to {Iz}0o for Ca deficiency, osmolarity, {Iz}0o for toxicities of Ca, Mg, Na or K, and for alleviation of Na toxicity by K and Ca. On a molar basis, Mg causes a substantially larger decrease in the negativity of ψ0o (with a corresponding decrease in {Ca2+}0o) than does Na or K. Therefore, reduction in growth resulting from Ca interactions is more likely for Mg-dominated salinity. High osmolarity reduced shoot growth more than root growth, which was reduced by 50% at c. 310 mOsM (cf. 180 mOsM for shoots). Specific ion toxicities were evident in the roots, with decreasing effects in the order of Ca, Mg, Na and K. The addition of K (and, to a lesser extent, Ca) alleviated the toxic effects of Na, suggesting that Na possibly interferes with the metabolic functions of these ions. This study demonstrates that the short-term growth of cowpea seedlings in saline solutions is related to {Iz}0o rather than {Iz}b, and that root growth is primarily limited by Ca deficiency, osmotic effects and specific ion toxicity, with Na toxicity alleviated at least in part by K and Ca.


The authors acknowledge the assistance of Dr B.A. McKenna and Dr U.T. Cobley. This research was funded through the Cooperative Research Centre for Contamination Assessment and Remediation of the Environment (CRC-CARE) Project 3-03-05-09/10. The support of the Environment Protection Authority (EPA) Victoria, Australia is also acknowledged.