Cell Membrane Surface Potential (ψ₀) Plays a Dominant Role in the Phytotoxicity of Copper and Arsenate^,

Abstract

Negative charges at cell membrane surfaces (CMS) create a surface electrical potential (ψ₀) that affects ion concentrations at the CMS and consequently affects the phytotoxicity of metallic cations and metalloid anions in different ways. The ζ potentials of root protoplasts of wheat (Triticum aestivum), as affected by the ionic environment of the solution, were measured and compared with the values of ψ₀ calculated with a Gouy-Chapman-Stern model. The mechanisms for the effects of cations (H⁺, Ca²⁺, Mg²⁺, Na⁺, and K⁺) on the acute toxicity of Cu²⁺ and As(V) to wheat were studied in terms of ψ₀. The order of effectiveness of the ions in reducing the negativity of ψ₀ was H⁺ > Ca²⁺ ≈ Mg²⁺ > Na⁺ ≈ K⁺. The calculated values of ψ₀ were proportional to the measured ζ potentials (r² = 0.93). Increasing Ca²⁺ or Mg²⁺ activities in bulk-phase media resulted in decreased CMS activities of Cu²⁺ ({Cu²⁺}₀) and increased CMS activities of As(V) ({As(V)}₀). The 48-h EA50{Cu²⁺}_b ({Cu²⁺} in bulk-phase media accounting for 50% inhibition of root elongation over 48 h) increased initially and then declined, whereas the 48-h EA50{As(V)}_b decreased linearly. However, the intrinsic toxicity of Cu²⁺ (toxicity expressed in terms of {Cu²⁺}₀) appeared to be enhanced as ψ₀ became less negative and the intrinsic toxicity of As(V) appeared to be reduced. The ψ₀ effects, rather than site-specific competitions among ions at the CMS (invoked by the biotic ligand model), may play the dominant role in the phytotoxicities of Cu²⁺ and As(V) to wheat.

Current environmental quality criteria and risk assessment procedures for metals and metalloids are predominantly based on total or dissolved metal concentrations (De Schamphelaere and Janssen, 2002). However, it is widely recognized that total or dissolved metal concentrations are sometimes poor predictors of bioavailability and toxicity. The physicochemical characteristics of soil and water, such as the contents of common cations and organic matter, have important effects on the bioavailability and toxicity of metals. Therefore, modifying factors for the bioavailability and toxicity should be taken into account in the regulatory frameworks (Peijnenburg et al., 1997).

The biotic ligand model (BLM; Di Toro et al., 2001; De Schamphelaere and Janssen, 2002), as a useful construct for assessing the effects of metals on organisms, has gained increasing attention from both academic scientists and regulators. Recently, the U.S. Environmental Protection Agency (EPA) has incorporated the BLM into its regulatory framework, and some other countries are considering the implications of following suit (Slaveykova and Wilkinson, 2005). The most important assumption of the BLM is that metal toxicity occurs as a result of the reaction of a free metal ion (or other reactive metal species) with binding sites (biotic ligand [BL]) at the cell membrane surface (CMS; Di Toro et al., 2001; Santore et al., 2001). The magnitude of the toxic effect is proportional to the concentration of metal-BL complex. However, the actual toxic lesion may not be the interaction of the toxicant and BL. For example, the BL may be a binding site in a transport channel such that influx of the toxicant is proportional to the number of sites occupied. The actual lesion may occur intracellularly. In fact, the mechanisms of metal intoxication are generally very poorly understood. Decades of investigation have not revealed the principal mechanisms by which Cu²⁺ and Al³⁺ inhibit root elongation, for example, despite several known effects ranging from the induced synthesis of reactive oxygen species to alterations of cell membrane structure (effects that often require greater concentrations than those required to inhibit root elongation; Murphy et al., 1999).

The BLM provides a possible mechanism of ionic alleviation of toxicity. The ameliorative effectiveness of major cations, such as H⁺, Ca²⁺, Mg²⁺, Na⁺, and K⁺, on the toxicity of metals has been viewed as a site-specific competition between toxic and ameliorative cations for binding sites at the CMS (Di Toro et al., 2001; Santore et al., 2001; De Schamphelaere and Janssen, 2002). However, the BLM fails to interpret the toxicity of metalloid anions, such as As(V) and selenate, or the enhancement of anion toxicity by the treatments that reduce the toxicity of cations (Yermiyahu and Kinraide, 2005; results presented in this article). Just as the toxic lesion itself may not be directly associated with the BL envisioned in the BLM (see above), the alleviation of cationic toxicity by ameliorative cations may not be related directly to site-specific competitions between toxicants and ameliorants (Kinraide, 2006)—a topic considered in this article.

Almost all cell surfaces are intrinsically negatively charged (Kinraide et al., 1998; Shomer et al., 2003; Hassler et al., 2004). These negative charges create negative potentials at the CMS (ψ₀), which play important roles in plant-ion interactions (Gimmler et al., 2001; Kinraide, 2001, 2006; Kinraide et al., 2004; Yermiyahu and Kinraide, 2005). The ψ₀ controls ion distribution between the CMS and the bulk-phase medium (BM): Negative values for ψ₀ concentrate cations and deplete anions at the CMS. For instance, when ψ₀ = −45.0 mV, mono-, di-, and trivalent cations will be concentrated, and anions of corresponding negative ions will be depleted 6-, 33-, and 191-fold, respectively. Although the cell wall is negatively charged, it has small effects upon ion activities at the CMS (Kinraide, 2004).

The surface potentials are also controlled by the ionic composition in the BM. Common cations, especially H⁺, Ca²⁺, and Mg²⁺, reduce the negativity of ψ₀ by ionic screening and binding. These cations are known to alleviate the uptake and biotic effects of toxic cations (commonly metals; Kinraide, 2006) and to have the opposite effects upon toxic anions (Yermiyahu and Kinraide, 2005). It is difficult to measure ψ₀ directly, but the electrical potential (ζ potential) at the hydrodynamic plane of shear at a small distance from the CMS can be determined by electrophoretic mobility (Delgado et al., 2007). In addition, a Gouy-Chapman-Stern (G-C-S) model is now available to calculate the ψ₀ of plant cell membranes in response to the solution ionic environment (Kinraide et al., 1998; Kinraide and Yermiyahu, 2007).

The BLM considers the competition of coexisting cations to alleviate the toxicity of toxic cations, but the BLM neglects ψ₀, which enriches toxic cations at the CMS. The BLM assumes that toxicity by cations requires them to bind to a hypothetical BL and the alleviation of toxicity by ameliorative cations requires them to bind to the same BL. The assumption might be unrealistic and, if true, difficult to verify. In fact, ψ₀ effects can give a false appearance of competition in cases where competition is weak or does not occur at all (Kinraide, 2006). Thus, ψ₀ effects should not be negligible and may be more important than the effects of site-specific competition.

Therefore, this study aims to (1) verify that a G-C-S model calculates values for ψ₀ of wheat (Triticum aestivum) root that are at least proportional to measured ζ potentials; (2) investigate the role of ψ₀ in the effects of common cations on the toxicity (inhibition of root elongation) of Cu²⁺ and As(V) and the probable mechanisms for the effects; and (3) establish the relationship between the toxicity threshold (EA50, activities producing 50% inhibition) for Cu²⁺ and As(V) and the ψ₀, a relationship that can be used for risk assessment of Cu²⁺ and arsenate.

RESULTS AND DISCUSSION

Effects of Common Cations on ζ Potentials of Root Protoplasts

Cations in BM reduce the negative potential of the CMS by ionic binding and charge screening (Tatulian, 1999). The ζ potential reflects the electrical potential at the hydrodynamic plane of shear at a small distance from the CMS and, consequently, it is of somewhat lower magnitude than the ψ₀. Measured ζ potentials of wheat root protoplasts as exposed with different concentrations of Ca²⁺, Mg²⁺, Na⁺, K⁺, and H⁺ are summarized in Supplemental Table S1. As shown in Figure 1, the negativity of ζ potentials declined significantly with increases of monovalent (Na⁺, K⁺, H⁺) and divalent (Ca²⁺, Mg²⁺) cations in the BM. The order of effectiveness for reducing the negativity of the ζ potential was H⁺ > Ca²⁺ ≈ Mg²⁺ >Na⁺ ≈ K⁺. This order is a function of ion charge and binding strength to the CMS (Kinraide and Yermiyahu, 2007). These trends can be well quantified with logarithmic equations and r² values are 1.00, 0.99, 0.76, 0.85, and 0.91 for Ca²⁺, Mg²⁺, Na⁺, K⁺, and H⁺, respectively. The ζ potentials were not significantly decreased by Cu²⁺ at concentrations ≤4 μm.

Figure 1.

ζ potentials of wheat root protoplasts as a function of the concentration of Ca²⁺ or Mg²⁺ (A), Na⁺ or K⁺ (B), or activity of H⁺ in the BM (C). Vertical bars represent the sds.

The G-C-S model, combining classical electrostatic theory (Gouy-Chapman) with ion binding, was developed to calculate the ψ₀ (Kinraide et al., 1998). Model parameters were derived or estimated from many studies (see further discussion below). The values of calculated ψ₀ were compared with the measured ζ potentials, and significant linear correlations (r² = 0.93) between the ζ potentials and ψ₀ were obtained in all datasets (Fig. 2). This indicates that the adopted model parameters of the G-C-S model are, to a great extent, capable of computing ψ₀ values for wheat root that are at least proportional to ζ potentials.

Figure 2.

Relationship between the ζ potentials and the CMS electrical potentials (ψ₀) calculated with the G-C-S model. Dashed lines indicate 95% confidence limits.

Ion Activities at the CMS

Negative ψ₀ enriches cation and depletes anion activities at the CMS. In the studies, assessing Ca-Cu interactions, increasing the bulk activity of Ca²⁺ ({Ca²⁺}_b) decreased ψ₀ negativity from −50.9 to −5.9 mV. As a result, the enrichment factor of Cu²⁺ activity ({Cu²⁺}₀/{Cu²⁺}_b) decreased from 51 to 1.6. With the increase of {Ca²⁺}_b, {Ca²⁺}₀ increased initially and then reached a plateau (Fig. 3A). In contrast, both {Mg²⁺}₀ and {H⁺}₀ declined markedly (Fig. 3, B and C) despite constant concentrations in the BM. In studies assessing K-Cu interactions, increasing {K⁺}_b from 0.08 to 8.9 mm decreased ψ₀ negativity from −50.3 to −44.5 mV, and consequently the corresponding enrichment factor of Cu²⁺ activity decreased from 44.6 to 29.7. Meanwhile, {K⁺}₀ increased considerably, but {Ca²⁺}₀, {Mg²⁺}₀, and {H⁺}₀ declined modestly (Fig. 3, D–F). As for As(V), its predominant species at the CMS were H₂AsO₄⁻ and HAsO₄²⁻, which accounted for 98.2% and 1.76%, respectively, at ψ₀ = −45 mV. Increased negativity of ψ₀ depleted the activity of HAsO₄²⁻ more than that of H₂AsO₄⁻ at the CMS.

Figure 3.

EA50{Cu²⁺} values and CMS activities of ions in response to Ca²⁺ or K⁺ additions to the rooting medium. Subscript b refers to variables in the BM. Subscript 0 refers to variables at the CMS. A to C, The {Ca²⁺}₀, {Mg²⁺}₀, and {H⁺}₀ in response to {Ca²⁺}_b (left column; Ca set) and the EA50{Cu²⁺}_b (A) and EA50{Cu²⁺}₀ (B and C) as a function of {Ca²⁺}_b in Ca set. D to F, The {K⁺}₀, {Ca²⁺}₀, and {H⁺}₀ in response to {K⁺}_b (right column; K set), and the EA50{Cu²⁺}_b (D) and EA50{Cu²⁺}₀ (E and F) as a function of {K⁺}_b in K set. Error bars indicate 95% confidence intervals.

Relative Root Elongation

Relative root elongations (RRLs) were fit with the equation RRL = 100/exp[(aT)^b] with toxicant (T) expressed as bulk-phase or CMS activities. Figure 4 presents the plots of RRL as a function of {Mg²⁺}_b together with [Cu²⁺]_b, {Cu²⁺}_b, or {Cu²⁺}₀. The correlation between root elongation and {Cu²⁺}₀ (Fig. 4C) is clearly superior to the correlation between root elongation and {Cu²⁺}_b (Fig. 4B) or [Cu²⁺]_b (Fig. 4A). Substitution of [Cu²⁺]_b, {Cu²⁺}_b, or {Cu²⁺}₀ into the equation yielded r² values of 0.81, 0.87, and 0.91, respectively. From Figure 4, A and B, it is clear that the RRL increased initially, but then RRL decreased slightly with increasing [Mg²⁺]_b (i.e. black triangles were lower than white circles). Similar results were observed for the Ca dataset (data not shown). The initial alleviation of Cu²⁺ toxicity could be explained by decreases in {Cu²⁺}₀ caused by a reduction in the negativity of ψ₀ that accompanied an increase in MgCl₂. The subsequent increase in toxicity may be due to the decrease of other competitive cations such as Ca²⁺ and H⁺ at the CMS (see discussion below), whereas {Mg²⁺}₀ remained constant (i.e. plateaued as did {Ca²⁺}₀ in Fig. 3A).

Figure 4.

RRL in response to Cu²⁺ and Mg²⁺. Mg²⁺ is expressed as concentration in the rooting medium and Cu²⁺ is expressed as concentration in the medium ([Cu²⁺]_b; A), activity in the medium ({Cu²⁺}_b; B), or activity at the CMS ({Cu²⁺}₀; C).

For the Cu studies, the 450 data points from all experiments were fit with RRL = 100/exp[(a{Cu²⁺}₀)^b]. The regression analysis yielded coefficients a = 0.0485 and b = 0.96; r² = 0.90. When RRL was assigned a value of 50%, a corresponding {Cu²⁺}₀ value of 14.1 μm was derived. This value is the computed activity of Cu²⁺ ion at the CMS that produces 50% inhibition. In the case of arsenate, root elongation was better correlated with {As(V)}₀ (r² = 0.93) than with {As(V)}_b (r² = 0.90) or [As(V)]_b (r² = 0.88; Fig. 5). Hormesis, a stimulation of response at low doses followed by inhibition at high doses, was observed in all sets in the As series. Therefore, a modified equation of RRL = 100/exp[(a({As(V)}₀ − c)^b] was used to fit RRL. The regression analysis yielded coefficients a = 2.26, b = 1.10, and c = 0.12; r² = 0.93 and n = 360. When RRL = 50%, a value of 0.434 μm {As(V)}₀ was determined. When the ψ₀ was considered, the activities of ionic toxicants at the CMS rather than their activities in the BM were better predictors of solution Cu²⁺ and arsenate toxicity.

Figure 5.

RRL in response to arsenate [As(V)] concentration in the medium ([As(V)]_b; A), activity in the medium ({As(V)}_b; B), or activity at the CMS ({As(V)}₀; C).

Effects of ψ₀ on the Toxicity of Copper and Arsenate

Treatments that reduce the negativity of ψ₀, such as increases in salt concentration or decreases in pH, reduce {Cu²⁺}₀ and increase {As(V)}₀. This alleviates Cu²⁺ toxicity (Fig. 3, A and D), but aggravates As toxicity (Fig. 6, A and D). With the increase of {Ca²⁺}_b, the 48-h EA50{Cu²⁺}_b was significantly increased initially and then reduced (Fig. 3A). This indicates that increasing {Ca²⁺}_b initially alleviated and then aggravated the Cu²⁺ toxicity. A similar phenomenon of initial alleviation and subsequent aggravation of toxicity was also observed with increasing {Mg²⁺}_b (Fig. 4, A and B, where black triangles were lower than white circles).

Figure 6.

EA50{As(V)} values as functions of Ca²⁺, Na⁺, and ψ₀. Note the references to the BM (indicated by subscript b) and the CMS (expressed by the subscript 0). A to C, The EA50{As(V)}_b as a function of {Ca²⁺}_b (A) and calculated CMS potentials (B), and the EA50{As(V)}₀ and the {Ca²⁺}₀ as a function of {Ca²⁺}_b in Ca set (C). D to F, The EA50{As(V)}_b as a function of {Na⁺}_b (D) and calculated CMS potentials (E), and the EA50{As(V)}₀ and {Na⁺}₀ as a function of {Na⁺}_b in Na set (F). Error bars indicate 95% confidence intervals.

The 48-h EA50{Cu²⁺}₀, instead of the usual EA50{Cu²⁺}_b, was introduced to possibly distinguish the ψ₀ effects from competition effects. In addition, the values of 48-h EA50{Cu²⁺}₀ were also regarded as the intrinsic toxicity of metals to plants. As seen in Figure 3A, {Ca²⁺}₀ increased initially and then reached a plateau at about 1.0 mm {Ca²⁺}_b as the {Ca²⁺}_b increased. However, the 48-h EA50{Cu²⁺}₀ declined at all times (Fig. 3, B and C), accompanied by decreases in the CMS activities of other cations (Mg²⁺, H⁺, Na⁺, and K⁺). This raises the possibility that the decreased EA50{Cu²⁺}₀ (Fig. 3, B and C) was due to the decreased competition from Mg²⁺ or some other cation at the CMS (discussed in the next section).

For K⁺, because it is less depolarizing than Ca²⁺ or Mg²⁺, the ψ₀ declined only slightly, from −50.3 to −44.5 mV; and {Ca²⁺}₀, {Mg²⁺}₀, and {H⁺}₀) declined slightly as {K⁺}_b increased from 0.08 to 8.9 mm (Fig. 3, E and F; {Mg²⁺}₀ not shown). The slightly decreased EA50{Cu²⁺}₀ (Fig. 3, E and F) may be related also to the slight decrease of Ca²⁺, Mg²⁺, and H⁺ at the CMS. The 48-h EA50{Cu²⁺}_b was expected to increase about 2-fold (the amount that {Cu²⁺}₀ was estimated to increase), but it did not. This expected increase in EA50{Cu²⁺}_b may have been offset by the weakened competition from Ca²⁺ and Mg²⁺ (discussed in the next section).

For arsenate bioassays, increases of Ca²⁺, Mg²⁺, Na⁺, K⁺, or a decrease of pH resulted in aggravated arsenate toxicity (only the effects of Ca²⁺ and Na⁺ are shown in Fig. 6). As shown in that figure, the 48-h EA50{As(V)}_b decreased linearly with increasing {Ca²⁺}_b (r² = 0.96) and {K⁺}_b (r² = 0.59). However, the 48-h EA50{As(V)}_b was better corrected with ψ₀ than with {Ca²⁺}_b or {K⁺}_b. The linear regression for EA50{As(V)}_b versus ψ₀ yielded r² = 1.00 for the Ca set and 0.77 for the K set (Fig. 6, B and E). The EA50{As(V)}₀ increased slightly with increasing {Ca²⁺}_b, {Mg²⁺}_b, or {H⁺}_b, whereas the values of EA50{As(V)}₀ did not change significantly as {Na⁺}_b or {K⁺}_b increased (Figs. 6, C and F, and 8A). These effects may be related to the different physiological mechanisms of these ions; the root length in control treatments of each set increased with the increasing Ca²⁺ or Mg²⁺, but did not change significantly with the increasing Na⁺ or K⁺ (data not shown).

Figure 8.

The EA50s expressed as the surface activity (EA50{As(V)}₀) (A) and ln of the bulk-phase activity (EA50{As(V)}_b) (B) as functions of ψ₀. The dashed line in A presents a predicted EA50{As(V)}₀ value of 0.434 μm; the full line in B presents the function ln(EA50{As(V)}_b) = −0.017ψ₀ + 0.041 (r² = 0.69, n = 20); and the dashed lines in B present 95% confidence limits.

Mechanisms for the Cation Effects on the Toxicity of Copper and Arsenate

The results of this study and previous studies indicate several mechanisms for ion-toxicant interactions. These are listed below. These mechanisms are intended to account for ionic alleviation or aggravation of intoxication by ionic toxicants. Some commonly studied cationic toxicants include H⁺, Cu²⁺, Pb²⁺, Hg²⁺ (and other heavy metals), trivalent metals (La³⁺, but especially Al³⁺), Na⁺ (at higher concentrations), and some anions (such as selenate and arsenate). The first mechanism listed below is the principal mechanism considered in our study, and we consider it to be very well substantiated now.

(1) Electrostatic interactions: The cationic components of salts commonly cause reductions in the negativity of CMS electrical potentials (ψ₀). The anionic components (commonly Cl⁻ or SO₄²⁻) generally have small effects because of weak binding to the CMS and small concentrations at the CMS because of electrostatic repulsion. The depolarizing effectiveness of some common cations follows the order Al³⁺ > H⁺ > Ca²⁺ ≈ Mg²⁺ > Na⁺ ≈ K⁺, depending upon charge and strength of binding to the CMS. The reduced negativity of ψ₀ causes reductions at the CMS of cationic toxicants and increases of anionic toxicants. In some cases, these effects appear to account almost entirely for ion-toxicant interactions (Yermiyahu and Kinraide, 2005; Kinraide, 2006).

(2) Site-specific competition: This mechanism accounts for ion-toxicant interactions according to the BLM, at least for cation-cation and anion-anion interactions. The operation of mechanism 1 above does not negate the operation of site-specific competition. Addition of Ca²⁺ to an intoxicating Cu²⁺ solution could both reduce {Cu²⁺}₀ and displace Cu²⁺ from surface ligands through direct competition. SO₄²⁻, although it would do little to enhance ψ₀ negativity, could compete with SeO₄²⁻ for binding sites, including transport sites. The electrostatic interactions and the site-specific competition models may be combined by incorporating into the equations of the BLM {I^Z}₀ instead of {I^Z}_b, where I^Z is any ion I of charge Z (Kinraide, 2006).

(3) Other mechanisms: One may envision several additional ion-toxicant interactions. Consider a case where one ion, say Ca²⁺, alleviates Cu²⁺ toxicity according to the two mechanisms above, but a third ion, say Mg²⁺, alleviates toxicity by an additional mechanism (perhaps through an especially great affinity for a site involved in Cu²⁺ uptake, but not a site to which Cu²⁺ binds). Additions of Ca²⁺ to solutions of constant {Cu²⁺}_b and {Mg²⁺}_b would alleviate toxicity by reducing {Cu²⁺}₀, but would enhance toxicity by reducing {Mg²⁺}₀.

The apparent enhancement of intrinsic Cu²⁺ toxicity illustrated in Figure 7A and the apparent alleviation of intrinsic arsenate toxicity illustrated in Figure 8A are problematic. A simple explanation is a systematic computational error caused by an overestimate of R⁻, which is the CMS density of negative charges in units μmol/m². The use of smaller values for R⁻ for the computation of ψ₀ reduce the slopes in both Figures 7A and 8A without reducing the slopes in Figures 7B and 8B. Alternative explanations for the apparent change in the intrinsic toxicities may be related to the changes in the activities of other influential ions, both positive and negative, as suggested in mechanism 3 above.

Figure 7.

The EA50s expressed as the surface activity (EA50{Cu²⁺}₀) (A) and ln of the bulk-phase activity (EA50{Cu²⁺}_b) (B) as functions of ψ₀. The dashed line in A presents a predicted EA50{Cu²⁺}₀ value of 14.1 μm; the full line in B presents the function ln(EA50{Cu²⁺}_b) = 0.027ψ₀ + 0.36 (r² = 0.78, n = 25); and the dashed lines in B present 95% confidence limits.

A slight enhancement of intrinsic Cu²⁺ toxicity by Ca²⁺ has been noted previously (Kinraide, 2006). Even more significant, and not subject to modeling error, is the Ca²⁺ enhancement of extrinsic toxicity (expressed in terms of {Cu²⁺}_b) noted by Lock et al. (2007a), as well as possible enhancements by Ca²⁺ of extrinsic Co²⁺ toxicity (Lock et al., 2007b) and extrinsic Ni²⁺ toxicity (Lock et al., 2007c). Despite these occasional oddities, which are now under investigation, a very large number of studies (for review, see Kinraide, 2006; Yermiyahu and Kinraide, 2005) demonstrate the superiority of expressing both the toxicity and uptake of cations and anions in terms of CMS activities rather than in terms of BM activities.

Prediction of Copper and Arsenate Toxicity

As shown in Figures 7B and 8B, the 48-h EA50s (expressed as cupric and arsenate ion activities in BM) can be predicted on the basis of ψ₀ calculated with the G-C-S model. For Cu²⁺ toxicity, the 48-h EA50 can be expressed as EA50{Cu²⁺}_b = exp(0.27ψ₀ + 0.36) (r² = 0.78; n = 24), and for arsenate toxicity, the 48-h EA50 can be expressed as EA50{As(V)}_b = exp(−0.017ψ₀ + 0.041) (r² = 0.69; n = 20). Almost all the observed EA50s were between 95% prediction limits.

The G-C-S Model

The G-C-S model combines classical electrostatic theory (Gouy-Chapman theory) with ion binding so that the electrical potential at the CMS (ψ₀) can be computed (Kinraide et al., 1998; Tatulian, 1999). The ψ₀ appears to be little influenced by the cell wall (Kinraide, 2004). Up-to-date parameters for the model are presented in Kinraide and Yermiyahu (2007). On the basis of several lines of evidence, including ζ potential measurements and adsorption measurements, we are quite confident of model parameters relating to ion binding strength at the CMS (Kinraide and Yermiyahu, 2007). Another critical parameter, the surface charge density of negative charges (R⁻ = 0.3074 μmol/m² is currently assumed in our model) appears to be more variable and may be influenced by species, tissue, and preparation in the case of vesicles and protoplasts. Consequently, the values for ψ₀, and the CMS ion activities computed from ψ₀ with the Nernst equation, may be only proportional to the actual values. (Figure 2 in Kinraide et al. [1998] illustrates this proportionality.) This limitation appears not to reduce seriously the great superiority of expressing plant-ion interactions (including uptake, toxicity, and the alleviation of toxicity) in terms of CMS ion activities rather than bulk-phase ion activities (Kinraide, 2001, 2006; Yermiyahu and Kinraide, 2005).

CONCLUSION

The parameters adopted in the G-C-S model for computing ψ₀ can be substantially verified by measured ζ potentials. The Ca²⁺ depolarizing effectiveness on reducing the negativity of ψ₀ was the same as that by Mg²⁺. The ψ₀ provides a new avenue for exploring the mechanisms of ion interactions and the biotic effects of toxic ions. In fact, ignoring the electrostatic effects (tantamount to assuming ψ₀ = 0) must surely lead to greater errors of interpretation than the small errors (relative to assuming ψ₀ = 0) in estimates of ψ₀. An increase in cations or a decrease in pH in BM reduces the negativity of ψ₀, reduces the surface activity of Cu²⁺, and increases the surface activity of arsenate by electrostatic mechanisms. As a result, the extrinsic toxicity of Cu²⁺ was alleviated, whereas the extrinsic toxicity of arsenate was aggravated. Effects upon the intrinsic toxicities of Cu²⁺ and arsenate may have been opposite to the effects upon extrinsic toxicity, but those effects are not certain and require further study. We also analyzed published studies (Lock et al., 2007a, 2007b, 2007c) and obtained compatible results. Exploring the mechanisms for an apparent enhanced intrinsic toxicity was under way. In addition, only cereals are discussed in this article, the results applying to measurements in dicots require further study.

MATERIALS AND METHODS

Experimental Design

In order to verify the reliability of the G-C-S model for calculating the ψ₀, and to investigate the independent effect of different cations on Cu²⁺ and arsenate toxicity, only one cation concentration varied at a time while all other cation concentrations were kept as nearly constant as possible. Seven sets of ζ potential measurements were determined, one each for Ca²⁺, Mg²⁺, Na⁺, K⁺, pH, Cu²⁺, and (Ca²⁺ + Mg²⁺) (Supplemental Table S1).

For Cu²⁺ toxicity tests (Cu series), five sets of Cu²⁺ bioassays were performed, one each for Ca²⁺, Mg²⁺, Na⁺, K⁺, and pH (Supplemental Table S2). In each medium, five Cu concentrations (0.25, 0.50, 1.0, 2.0, 4.0 μm) and a control were tested. For the arsenate toxicity tests, five sets of bioassays were performed, one each for Ca²⁺, Mg²⁺, Na⁺, K⁺, and pH (Supplemental Table S3). In each medium, five arsenate concentrations (0.67, 1.3, 3.3, 6.7, 26.7 μm) and a control were tested. The selected cation concentrations reflected the variability occurring in natural soil solutions (Schwab, 2000).

Preparation of the Test Solutions

In all experiments, the salts used were analytical grade and double-deionized water was used throughout. Except for pH tests, medium was buffered with 2.0 mm MES adjusted to pH 6 with NaOH (Lock et al., 2007a, 2007b). For each toxicity assay, CuCl₂ and NaH₂AsO₄ were added to the prepared test medium. All media were prepared 1 d before the start of tests to obtain near-equilibrium solutions. The chemical characteristics of different test media are summarized in Supplemental Tables S2 and S3. Solution pH was measured before and after each test. The concentrations of Cu were determined by graphite furnace atomic absorption spectrophotometry (Varian 220). The concentrations of arsenate and other cations (Ca, Mg, Na, and K) were determined by inductively coupled plasma-atomic emission spectroscopy (POEMS-II; TJA). Preliminary tests showed that the measured concentrations did not differ significantly from their nominal values and variability of concentrations before and after 24 h of test was found to be 10%.

Determination of ζ Potentials of Root Protoplasts

Two-day-old seedlings with uniform root length were transferred to the culture solution containing 0.25 mm CaCl₂, 0.25 mm MgCl₂, 0.5 mm NaCl, and 0.5 mm KCl. The medium was buffered with 2.0 mm MES adjusted to pH 6 with NaOH. Seedlings were grown at 25°C for 2 d. Wheat (Triticum aestivum) roots were collected, washed with deionized water, and cut into 1- to 2-mm sections. Three grams of root material were incubated for 12 h in 30 mL of enzyme solution containing 2% cellulase Onozuka R-10 (w/v), 1% pectinase (w/v), 1% bovine serum albumin (w/v), 1% polyvinylpyrrolidone (w/v), 0.5 mm mannitol, 2.0 mm MES, 0.25 mm CaCl₂ at pH 6.0. The mixture of protoplasts and enzyme solution was filtered through two-layer nylon net (0.1-mm pore diameter) and the filtrate was centrifuged at 800 rpm for 5 min. The supernatant was discarded and the protoplasts in the residue were suspended and washed by three successive additions of 15 mL of a protoplast washing solution containing 0.4 mm mannitol, 2.0 mm MES, and 25 μm CaCl₂ at pH 6.0. After washing, the protoplasts in the residue were suspended again in 10 mL of protoplast washing solution. Twenty milliliters of 20% Suc were injected into the bottom of the centrifuge tube with a long-needled syringe for density gradient centrifugation. The mixture was centrifuged at 800 rpm for 5 min and the pure protoplasts were collected from the two-phase solution interface. The collected protoplasts were suspended again and washed several times with protoplast washing solution. The final preparation contained at least 5,000 pure protoplasts/mL.

The protoplasts were suspended (at least 10⁵ protoplasts/mL) in the test medium (Supplemental Table S1) and incubated for 1 h at 25°C ± 1°C. The ζ potentials of the root protoplasts were measured using a JS94H microelectrophoresis meter (Powereach Instruments; Wang et al., 2008). The equilibrated suspension was agitated and transferred to the electrophoresis vessel after the electrode was wetted to avoid disturbance caused by air bubbles. The migration of the protoplast particles under a potential gradient of about 10 V/cm was observed on a computer screen through a camera and a multimedia system with 1,200-time amplification. An average of the electrophoretic velocity values was obtained by the computer through timing 10 particles, first in one direction and then in the opposite direction upon reversal of the polarity of the applied electrical field. The values of ζ potentials were calculated by computer software, using the Helmholtz-Smoluchowski equation, provided with Powereach equipment (see Delgado et al., 2007). This step was repeated at least 10 times and the mean value was reported in this article. The measurement error was found to be <2 mV and the temperature was kept at a constant 25°C ± 1°C.

Toxicity Assays

Forty-eight-hour root elongation toxicity tests with wheat were performed following ISO Guideline 11269-1 (International Organization for Standardization, 1993) and EPA Guideline 850-4200 (U.S. Environmental Protection Agency, 1996). The test seeds were provided by Nanjing Agriculture University. For acute toxicity tests, 2-d-old seedlings with uniform root length (1–2 cm) were cultured in darkness for 48 h at 25°C in an acid-washed polyethylene beaker containing 500 mL of test medium. Each treatment was performed with three replicates and six seedlings per replicate. At termination, the two longest roots of each seedling were measured and the mean value of 12 measured values per replicate was recorded. RRL was calculated using the formula RRL, % = 100 (RL_T − RL_S)/(RL_C − RL_S), in which RL_T represents the mean RL in the presence of toxicants (i.e. Cu²⁺ or arsenate), RL_C represents RL in the corresponding toxicant-free control, and RL_S represents RL in toxicant sufficient to saturate growth-inhibitory processes. The RL_S is nearly equal to RL at the time of seedling transfer to the test medium. Growth can be plotted against measures of toxicant intensity, such as the concentration ([T]_b), the activity in BM ({T}_b), or the activity at the CMS ({T}₀). The resulting curves are often negatively sigmoidal and have been expressed by a Weibull equation (Kinraide et al., 2004). If growth is limited by T, then RRL = 100/exp[(aT)^b], where a and b are curve-fitting parameters.

Computation of Ion Activities at the CMS

The activity of ion I^Z at the CMS ({I^Z}₀) was computed from the activity of I^Z in the BM ({I^Z}_b) according to the Nernst equation ({I^Z}₀ = {I^Z}_b exp[−ZFψ₀/(RT)]). The ψ₀ was calculated with a G-C-S model (Kinraide et al., 1998; Yermiyahu and Kinraide, 2005). F, R, and T are the Faraday constant, the gas constant, and temperature, respectively.

Data Treatment and Statistics

Ion species and activities were calculated using the visual MINTEQ (version 2.51) chemical equilibrium program (EPA). The equilibrium phases in the speciation calculation included atmospheric CO₂ (pCO₂ = 10^−3.5 atm). As for arsenate, H₂AsO₄⁻ (90.6%) and HAsO₄²⁻ (9.36%) are the predominant and thermodynamically stable species at pH 6.0. The 48-h EA50 values expressed as the activity of free cupric or arsenate ion in BM, denoted EA50{Cu²⁺}_b or EA50{AS(V)}_b, were then calculated from the observed root growth at each calculated free cupric or arsenate (sum of {HAsO₄²⁻}_b and {H₂AsO₄⁻}_b) ion activity in the BM ({Cu²⁺}_b or {As(V)}_b). The 48-h EA50 values expressed as the activity of free cupric or arsenate ion at the CMS, denoted EA50{Cu²⁺}₀ or EA50{As(V)}₀, were then calculated from EA50{Cu²⁺}_b or EA50{AS(V)}_b incorporated into the Nernst equation (see above). EA50s were calculated by fitting a sigmoid curve to the dose-effect relationships according to the model of Haanstra et al. (1985). RRL was related to toxicant intensity in different phases (e.g. [T]_b, {T}_b, {T}₀) by regression analysis (RRL = 100/exp[(aT)^b]) using Origin Professional 6.0. All the values for the regression coefficients in the Results and Discussion section are at a significance level of P < 0.05.

Supplemental Data

The following materials are available in the online version of this article.

• Supplemental Table S1. Measured ζ potentials of wheat root protoplasts as exposed with different ion environments.

• Supplemental Table S2. Copper toxicity assays.

• Supplemental Table S3. Arsenate toxicity assays.