Modelling metal–metal interactions and metal toxicity to lettuce Lactuca sativa following mixture exposure (Cu2+–Zn2+ and Cu2+–Ag+)

Environmental Pollution 176 (2013) 185e192

Contents lists available at SciVerse ScienceDirect

Environmental Pollution journal homepage: www.elsevier.com/locate/envpol

Modelling metalemetal interactions and metal toxicity to lettuce Lactuca sativa following mixture exposure (Cu2þeZn2þ and Cu2þeAgþ) T.T. Yen Le a, b, Martina G. Vijver c, Thomas B. Kinraide d, Willie J.G.M. Peijnenburg b, c, *, A. Jan Hendriks a a

Radboud University Nijmegen, Institute for Water and Wetland Research, Department of Environmental Sciences, 6500 GL Nijmegen, The Netherlands National Institute of Public Health and the Environment, Laboratory for Ecological Risk Assessment, PO Box 1, 3720 BA Bilthoven, The Netherlands c Leiden University, Institute of Environmental Sciences (CML), Department of Conservation Biology, 2300 RA Leiden, The Netherlands d Appalachian Farming Systems Research Center, Agricultural Research Service, United States Department of Agriculture, Beaver, WV 25813-9423, USA b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 30 August 2012 Received in revised form 22 November 2012 Accepted 15 January 2013

Metal toxicity to lettuce Lactuca sativa was determined following mixture exposure based on the concepts of concentration addition (CA) and response addition (RA). On the basis of conventional models assuming no interaction between mixture components, Agþ was the most toxic, followed by Cu2þ and Zn2þ. Furthermore, ioneion interactions were included in quantitatively estimating toxicity of interactive mixtures of Cu2þeZn2þ and Cu2þeAgþ by linearly expanding the CA and RA models. About 80e92% of the variability in the root growth could be explained by this approach. Estimates by the extended models indicate significant alleviative effects of Zn2þ on Cu2þ toxicity whereas Cu2þ did not significantly affect Zn2þ toxicity. According to the extended CA model, Cu2þ significantly reduced Agþ toxicity while Agþ enhanced Cu2þ toxicity. Similar effects were not found by the extended RA model. These interactions might result from their individual uptake mechanisms and toxic actions as published in literature. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: Mixture Toxicity Interaction Plant Metal

1. Introduction Metals are usually present in the environment in mixtures of varying composition. Toxicity of metal mixtures may vary widely while interactions in mixtures may deviate significantly from the biological actions of single metals (Norwood et al., 2003; Otitoloju, 2002; Manzo et al., 2010). Exposure to metal mixtures at concentrations below environmental quality guideline levels for individual components was reported to result in adverse effects, attributed to interactions between the constituents (Cooper et al., 2009). The reliability of toxicity estimations can thus be improved by taking into account interactions in mixtures (Otitoloju, 2002). Such interactions occur at different levels, i.e., in the environment, at the root surface, and within the plant (Kabata-Pendias and Pendias, 1984; Pahlsson, 1989). Interactions outside organisms determine the environmental availability of metals depending on the physicochemical conditions. Subsequently, in the toxicokinetic phase, interactions between different metals influence the uptake of metals by organisms. In the toxicodynamic phase, interactions at ligands within organisms affect their joint toxicity. While metale metal interactions in the environment have been predicted well

* Corresponding author. E-mail address: [email protected] (W.J.G.M. Peijnenburg). 0269-7491/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.envpol.2013.01.017

by speciation modelling, interactions at the toxicokinetic and toxicodynamic phases are usually excluded in available models for assessment of mixture toxicity, e.g., concentration addition (CA) and response addition (RA) (Bliss, 1939; Hewlett and Plackett, 1979). The conventional concept of these models is based on the assumption that the presence of one substance does not affect the biological action of the others in their mixture. Accordingly, deviations from the ideal behaviour of mixtures, which result from the interactions, cannot be quantified by these models based on the conventional concept. Backhaus et al. (2000a, 2000b) developed a modelling approach, which allowed accurately assessing toxicity of mixtures of similarly or dissimilarly acting organic compounds based on the concepts of CA and RA, respectively. However, in this approach, toxicity is examined for mixtures with a fixed ratio between the concentrations of the components, e.g., the ratio of their values of EC1 or EC50. This experimental design limits the application of the modelling approach of Backhaus et al. as in the environment metal mixtures exist with various ratios between the mixture constituents. Furthermore, in the assessment based on the concepts of CA and RA, toxicity of mixtures is mainly predicted from toxicological data for single substances only (Manzo et al., 2010; Sharma et al., 1999). However, joint toxicity of multiple chemicals should also be assessed by tests on mixtures (Frias-Espericueta et al., 2009). The present study aimed at modelling toxicity of metal mixtures taking into account potential interactions between their different

186

T.T.Y. Le et al. / Environmental Pollution 176 (2013) 185e192

components. Metal toxicity was experimentally assessed by tests on metal mixtures, instead of single exposure. Moreover, the ratio between the exposure levels of different mixture components was not fixed in the experimental set-up. Furthermore, because of this experimental design, the modelling approach that provided accurate estimations of toxicity of mixtures of organic compounds in the studies of Backhaus et al. (2000a, 2000b) is not suitable in the present study. Accordingly, a different method based on the concepts of CA and RA was used in the present study. Particularly, toxicity of metal mixtures was modelled by developing mathematical relationships expressing the metalemetal interactions and fitting these relationships to experimental data from the tests on metal mixtures, instead of based on ecotoxicological parameters for single metals. Specifically, toxicity of Cu2þ, Zn2þ, and Agþ in their non-interactive mixtures was assessed by fitting the empirical data to mathematical expressions of the conventional concept of CA and RA assuming no interactions between these ions. Furthermore, toxicity of interactive mixtures of Cu2þeZn2þ and Cu2þeAgþ was simulated through expansion of the conventional CA and RA models based on assumed linear interactions in the mixtures. In particular, expansion coefficients that describe interactive effects of one metal on the toxicity of the other in the mixture were included in estimating their joint toxicity. These coefficients distinguish extended models from conventional models. Metal toxicity was assessed in terms of inhibition of the root elongation of lettuce Lactuca sativa. 2. Methods 2.1. Test species and toxic endpoint Metal toxicity was assessed on lettuce, Lactuca sativa, in hydroponic exposures, in order to allow for controlled modifications of the test media. Lettuce was selected as test species in view of its high capacity to accumulate metals and the presence of a protocol by the Organisation for Economic Cooperation and Development (OECD) (McKenna et al., 1993; OECD, 2006; Le et al., 2012a). Root elongation was reported to be sensitive to metal exposure and has been widely used as toxic endpoint (Thakali et al., 2006; Lock et al., 2007; Voigt et al., 2006; Kinraide, 1999; Kinraide et al., 2004; Kopittke et al., 2011). Consequently, the root growth was used to evaluate metal toxicity in the present study. 2.2. Chemical measurements and speciation Free ion activities of Hþ, Cu2þ, and Agþ were measured by using hydrogen, copper, and silver sulfide ion-selective electrodes (Metrohm), respectively. These electrodes were calibrated by measurements at different concentrations of these ions in solution (Le et al., 2012a). Additionally, free Zn2þ activities in the exposure solutions were determined from total zinc concentrations by the speciation model Windermere Humic-Aqueous Model VI with Steiner solution as the default medium (Tipping, 1998). As Agþ, a component of the Steiner solution, is not included in this version of the WHAM model, a survey was carried out by using the Chemical Equilibria in Aquatic Systems (CHEAQS) model so as to investigate effects of Agþ on the activity of Cu2þ and Zn2þ (Verweij, 2004). Results from the survey indicate negligible effects of Agþ on the free ion activity of Cu2þ and Zn2þ. These findings were attributed to lower binding constants of Agþ to abiotic ligands in the solution, 2  2þ and Zn2þ as proved by the database e.g., NO 3 , SO4 , and H2PO4 , than those for Cu used for the CHEAQS model. The chemical composition of the Steiner solution used in the chemical speciation is given in Table S1.

higher level was kept below 2, except for the exposure solution at the background level of the Steiner solution in order to avoid uncertainties from the lack of replication in the present study. For mixtures of Cu2þeAgþ, Cu2þ was added to the Steiner solution first. The measurement of the free ion activity of Cu2þ was then carried out using the electrode as mentioned above before Agþ was put in the solution. This approach was used because Agþ did not have effects on the free ion activity of Cu2þ as presented in the previous section. For mixtures of Cu2þeZn2þ, Zn2þ was added to the Steiner solution at the concentration mentioned above before Cu2þ was put in the solution. This approach was used as it allowed including effects of Zn2þ in the measurement of the free ion activity of Cu2þ by the ion-selective electrode on the one hand and including the effects of Cu2þ on the free ion activity of Zn2þ determined by the WHAM model on the other hand. Moreover, unlike the experimental design in the studies of Backhaus et al. (2000a, 2000b), in the present study toxicity tests were not carried out at a fixed ratio between the exposure levels of different mixture components. On the one hand, the method applied in the present study thus overcomes the disadvantage of the approach of Backhaus et al. (2000a, 2000b) as mentioned in the Introduction. On the other hand, this set-up prevents the difficulties in controlling the free ion activities of metals in the exposure solution in order to have a fixed ratio between the free ion activities of two metals in the mixture. Solution pH was kept at 7.0 using 3-[N-morpholino] propane sulfonic acid at 0.75 g L1 and NaOH (Le et al., 2012a, 2012b). Exposure solutions were daily renewed. The solution pH was always checked before putting lettuce into the solution. 2.4. Toxicity assays Seeds of Lactuca sativa were germinated for 4 days at 15  C in the Steiner solution during a normal light cycle of 16:8 h light:dark. The germinated plants were then fixed in a parafilm strap with a surface area of around 30 cm2. The parafilm strap floated on the surface of a glass beaker (10 cm height and volume: 100 mL) with the roots immersed in the medium. Four plants were put in each beaker. The growth of lettuce (Growth, mm) exposed to a given exposure solution was calculated as the average of the increase in the root length of the 4 plants after 4 days of exposure. The replication by repeating toxicity tests on the same solution was not carried out in the present study. However, this lack of replicates was expected not to undermine the statistical significance of the toxicological data generated because of the experimental design applied in the present study: small gap between the exposure levels as described above and 4 plants grown in each solution. In total, 238 toxicity tests were carried out, including 122 tests performed with additions of Cu2þ and Zn2þ (but not Agþ) to the Steiner solution and 116 tests performed with additions of Cu2þ and Agþ (but not Zn2þ) to the Steiner solution. 2.5. Mathematical expression of metal toxicity 2.5.1. Toxicity of metals following single exposure The response of plants in terms of root growth (Growth; mm) after single exposure to metal ion Mnþ can be expressed in relation to its free ion activity in the solution {Mnþ} (mM) according to the following equation: Growth ¼

b h  d i exp c  Mnþ

(1)

where coefficient b (mm) is the growth of lettuce roots in the medium free of the metal ion (i.e., {Mnþ} ¼ 0); coefficient c (mM1) reflects the metal-specific toxicity strength. When {Mnþ} ¼ c1, Growth ¼ 36.8% b; and coefficient d (dimensionless) describes the slope of the adjacent curve. This exponential equation was found to be the most suitable to describe the root elongation following metal exposure and has been applied in a number of studies investigating metal toxicity to plants (Kinraide and Parker, 1989; Kinraide, 1999; Kinraide et al., 2004; Kopittke et al., 2011). The free metal ion activity at the 50% response level, i.e., median effective activity EC50M, could be calculated from the strength coefficient and the slope parameter in Eq. (1) as follows:

2.3. Preparation of the test solutions Steiner solution was used as test medium (Steiner, 1961). Cu2þ, Zn2þ, and Agþ were added into the Steiner solution in the form of nitrate salts. This form was used because of negligible interference of NO 3 on the measurement of ion-selective electrodes compared to other anions such as Cl. Preliminary tests were carried out at different concentrations of the metals from the background level in the Steiner solution to the level that causes serious effects on lettuce in order to determine the test exposure level. The exposure levels determined were 108e107 M for Agþ (expressed in free ion activity), 1010e106 M for Cu2þ (expressed in free ion activity), and 0e5  103 M (expressed in total concentration) or 106e103 M (expressed in free ion activity determined by WHAM model as presented below) for Zn2þ. These exposure levels were used to prepare test solutions. The ratio between one exposure level expressed as the free metal ion activity and the next

EC50M ¼

ðln 2Þ1=d c

(2)

The derivation of Eq. (2) is presented in the SA. 2.5.2. Toxicity of non-interactive mixtures If mixture components do not interact with each other, the growth of lettuce roots exposed to the mixture can be written according to the conventional concept of CA and RA assuming no interactions between mixture constituents. The CA model is based on the assumption that different substances in their mixture have the same modes of action (Bliss, 1939). Accordingly, the growth of lettuce roots (Growth, mm) following exposure to a non-interactive mixture of Cu2þ, Zn2þ, and Agþ can be written as

T.T.Y. Le et al. / Environmental Pollution 176 (2013) 185e192

Growth ¼

b h   d i    exp c1  Cu2þ þ c2  Zn2þ þ c3  Agþ

(3)

where b (mm) is the growth of lettuce roots in the medium free of Cu2þ, Zn2þ, and Agþ; coefficients c1, c2, and c3 (mM1) represent the strength of toxicity of Cu2þ, Zn2þ, and Agþ in their non-interactive mixtures, respectively; d (dimensionless) is the slope parameter describing toxicity of these metals in their non-interactive mixture; and {Cu2þ}, {Zn2þ}, and {Agþ} (mM) are the free ion activity of Cu2þ, Zn2þ, and Agþ in the solution, respectively (Kinraide, 1999). In the RA model, mixture components are supposed to have different modes of action of toxicity (Hewlett and Plackett, 1979). Therefore, based on the RA concept, the response of lettuce exposed to non-interactive mixtures can be expressed as a multiplicative function of the response of the plants following exposure to each constituent separately. For example, without interactions between Cu2þ, Agþ, and Zn2þ, the growth of lettuce roots exposed to a mixture of these metal ions can be expressed by the following equation Growth ¼

b h d1  d2  d3 i    exp c1  Cu2þ þ c2  Zn2þ þ c3  Agþ

(4)

where coefficients b, c1, c2, and c3 have the same meaning as in Eq. (3); d1, d2, and d3 (dimensionless) are slope parameters describing toxicity of Cu2þ, Agþ, and Zn2þ in their non-interactive mixtures, respectively; and {Cu2þ}, {Zn2þ}, and {Agþ} (mM) are the free ion activity of Cu2þ, Agþ, and Zn2þ in the exposure solution, respectively (Kinraide, 1999). Eqs. (3) and (4) are conventional expressions of CA and RA assuming that the presence of one metal does not affect the toxicity of the others in their mixtures. 2.5.3. Toxicity of interactive mixtures If metals in mixtures interact with each other, i.e., the presence of one metal affects the toxicity of the others in the mixtures, the interactions can be taken into account in quantifying toxicity of the mixtures by expanding the conventional CA and RA models. In particular, expansion coefficients representing the interactions can be incorporated in the strength coefficients in Eqs. (3) and (4) in two different ways as described in the SA (Kinraide, 1999; Kinraide et al., 2004). The expansion approach that results in higher statistical significance was selected as the best simulation of the interactions. The comparison of the expansion coefficient with zero determines whether one substance reduces or increases the toxicity of another (See SA). In addition, the interactive effect was considered statistically significant if the 95% confidence interval (CI) of the expansion coefficient does not encompass zero. A full description of equation derivation is presented in the SA. 2.5.3.1. Mixtures of Cu2þ and Zn2þ. An expansion coefficient c12 (mM1) representing interactive effects of Zn2þ on Cu2þ toxicity can be integrated into the strength coefficient of Cu2þ toxicity. Similarly, another expansion coefficient c21 (mM1) might be incorporated into the strength coefficient of Zn2þ toxicity to reflect effects of Cu2þ on Zn2þ toxicity. According to the CA model and based on the assumption of linear interactions, the response of lettuce exposed to Cu2þeZn2þ mixtures expressed as the root growth (Growth, mm) could be fitted well to Eq. (5) as all coefficients in this equation estimated by the regression analysis were statistically significant: Growth ¼

" exp

b ! #   n o d c1  Cu2þ 2þ  2þ  þ c2  Zn 1 þ c12  Zn

(5)

where b (mm) is the growth of lettuce roots in the medium free of Cu2þ and Zn2þ; c1 (mM1) is the strength coefficient of Cu2þ toxicity in the medium free of Zn2þ; c2 (mM1) is the strength coefficient of Zn2þ toxicity in the medium free of Cu2þ; c12 (mM1) is the expansion coefficient representing effects of Zn2þ on the toxicity of Cu2þ; d (dimensionless) reflects the slope of the adjacent curve; and {Cu2þ} and {Zn2þ} (mM) are the free ion activity of Cu2þ and Zn2þ in the solution, respectively. Based on the assumption of linear interactions, in the RA model the growth of lettuce roots (Growth, mm) in response to exposure to mixtures of Cu2þ and Zn2þ follows Eq. (6) as coefficients in this equation estimated by the regression analysis were statistically significant: Growth ¼

" exp

b #  !d1  d2   c1  Cu2þ þ c2  Zn2þ  2þ  1 þ c12  Zn

(6)

where coefficients b, c1, c2, and c12 have the same meaning as in Eq. (5); d1 and d2 (dimensionless) reflect the slope of the adjacent curve describing toxicity of Cu2þ and Zn2þ in their mixtures, respectively; and {Cu2þ} and {Zn2þ} (mM) are the free ion activity of Cu2þ and Zn2þ in the solution, respectively. 2þ

þ

þ

2þ

2.5.3.2. Mixtures of Cu and Ag . Interactive effects of Ag on Cu toxicity can be represented by an expansion coefficient c13 (mM1) that is incorporated into the

187

strength coefficient of Cu2þ toxicity. A similar expansion coefficient c31 (mM1) reflecting effects of Cu2þ on Agþ toxicity can be integrated into the strength coefficient of Agþ toxicity. The CA model can be extended as Eq. (7) to express joint toxicity of Cu2þ and Agþ, taking into account effects of their interactions, because coefficients in this equation estimated by the regression analysis were statistically significant: Growth ¼

" exp

b ! #    n o n o d c1  Cu2þ  þ  þ c3  1 þ c31  Cu2þ  Agþ 1 þ c13  Ag

(7)

where b (mm) is the growth of lettuce roots in the medium free of Cu2þ and Agþ; c1 (mM1) is the strength coefficient of Cu2þ toxicity in the medium free of Agþ; c3 (mM1) is the strength coefficient of Agþ toxicity in the medium free of Cu2þ; c13 (mM1) is the expansion coefficient representing effects of Agþ on the toxicity of Cu2þ; c31 (mM1) is the expansion coefficient describing effects of Cu2þ on the toxicity of Agþ; d (dimensionless) reflects the slope of the adjacent curve describing toxicity of Cu2þ and Agþ in their mixtures; and {Cu2þ} and {Agþ} (mM) are the free ion activity of Cu2þ and Agþ in the solution, respectively. By contrast, no expansion coefficient was found to be statistically significant to represent interactions between these metal ions according to the extended RA model. 2.6. Statistical analyses Coefficients in Eqs. (3)e(7) were determined by multiple regression analyses using the SYSTAT software. Coefficients are considered statistically significant if their 95% CI is statistically deviating from zero, i.e., not encompassing zero. The strength of the significance increases with increasing absolute value of the ratio between the estimate of the coefficient and the asymptotic standard error, i.e., parameter/ASE in the regression results. All toxicity data generated in the present study (from 238 tests) were used in the conventional models to assess toxicity of Cu2þ, Agþ, and Zn2þ in their non-interactive mixtures. All data were used in this analysis as in the conventional models for non-interactive mixtures it was assumed that the presence of one metal does not affect the biological action of the others in the mixtures. Among the 238 tests, data from 122 tests without additions of Agþ were used in the extended models to assess toxicity of interactive mixtures of Cu2þ and Zn2þ as Agþ was not present in the solutions. Moreover, toxicity of interactive mixtures of Cu2þ and Agþ was evaluated in the extended models using results from 116 remaining tests with no Zn2þ added to the Steiner solution, assuming negligible effects caused by Zn2þ at the background concentration in the default medium.

3. Results 3.1. Toxicity of Cu2þ, Agþ, and Zn2þ in non-interactive mixtures Differences between the strength of single metal toxicity predicted by the non-interactive mixture models based on the concepts of CA and of RA were small (Table 1). Estimates of all the coefficients and statistical parameters describing toxicity of Cu2þ, Agþ, and Zn2þ in non-interactive mixtures are given in Tables S2 and S3. There was no statistically significant difference between the estimates by the CA model and by the RA model in predicting Cu2þ and Zn2þ toxicity as shown by an overlap between the 95% CIs of the strength coefficients predicted by the two models for both metals (Table 1). An opposite observation was found in predicting toxicity of Agþ (Table 1). The difference between the strength coefficients of Agþ toxicity estimated by the CA model and by the RA model was small, but significant as shown by their non-overlapping 95% CIs. The assessment based on both models indicates that Zn2þ was far less toxic than Cu2þ and Agþ while there were negligible differences between the toxicities of Cu2þ and of Agþ (Table 1). These findings suggest that the CA and RA models did not yield substantial differences in the estimates of toxicity of these metals, based on the assumption of no interactions between them. These observations were also found based on the values of the median effective activity EC50M (Table 2). 3.2. Toxicity of binary interactive mixtures 3.2.1. Mixtures of Cu2þ and Zn2þ The extended CA (Eq. (5)) and RA (Eq. (6)) models estimated the toxicity of Cu2þeZn2þ mixtures at different exposure levels equally

188

T.T.Y. Le et al. / Environmental Pollution 176 (2013) 185e192

Table 1 Estimates of strength coefficients of toxicity of Cu2þ, Zn2þ, and Agþ: to Lactuca sativa following exposure to their non-interactive mixture found in the present study according to the conventional concept of concentration addition (CA) and response addition (RA) models; and to Vigna unguiculata following exposure to single metals reported in the study of Kopittke et al. (2011). 95% confidence intervals (CI) are provided. Source

Species

Strength coefficient ci (mM1)

Model

Present study

Lactuca sativa

Kopittke et al. (2011)

Vigna unguiculata

CA RA Single-metal exposure

well (R2 ¼ 0.92; Fig. 1A and B). However, the growth of lettuce roots exposed to mixtures of Cu2þ and Zn2þ at their low activities in solution was frequently underestimated (Fig. 1). According to both extended CA and extended RA models, Zn2þ significantly reduced toxicity of Cu2þ as the 95% CI of the expansion coefficient c12 were positive and statistically significantly deviating from zero (Tables 3 and 4). By contrast, Cu2þ did not have significant effects on Zn2þ toxicity, i.e., no statistically significant value of the expansion coefficient c21 was found to represent these impacts.

Agþ

Cu2þ

Zn2þ

c1 (95% CI)

c2 (95% CI)

c3 (95% CI)

2.92 (2.54e3.30) 3.13 (2.71e3.55) 2.00

6.15  103 (5.29  103e7.01  103) 6.31  103 (5.41  103e7.21  103) 4.26  102

2.93 (2.52e3.34) 3.79 (3.50e4.08) 25.9

values were substantially deviating from each other, i.e., about one order of magnitude, when lettuce was exposed to the single metals. Additionally, toxicity of Cu2þ following exposure to non-interactive mixtures of the metals was substantially lower than its toxicity

3.2.2. Mixtures of Cu2þ and Agþ Approximately 80% of the variability in the growth of lettuce roots exposed to mixtures of Cu2þ and Agþ at different free ion activities could be explained by the mathematical relationship expressed by Eq. (7) (n ¼ 116; R2 ¼ 0.80; Fig. 2). The 95% CIs of the expansion coefficients representing effects of Agþ on Cu2þ toxicity (c13) and effects of Cu2þ on Agþ toxicity (c31) deviated significantly from zero (Table 5). This indicates that Cu2þ and Agþ interacted with each other, significantly affecting their toxicities to lettuce. Particularly, Agþ significantly enhanced Cu2þ toxicity (c13 < 0) while Cu2þ significantly reduced Agþ toxicity (c31 < 0) (See the SA). 4. Discussion 4.1. Toxicity of Cu2þ, Zn2þ, and Agþ in non-interactive mixtures In the present study, based on the strength coefficients, toxicity to lettuce Lactuca sativa decreased in the sequence of Agþ > Cu2þ > Zn2þ, similar to the order reported by Kopittke et al. (2011) for cowpea Vigna unguiculata (Table 1). An inconsiderable difference was found in the toxic potency of Cu2þ to the two plant species (Table 1). Yet, the coefficients describing the strength of toxicity of Agþ and of Zn2þ to Vigna unguiculata were about one order of magnitude higher than those for Lactuca sativa (Table 1). Significant differences were found in the comparison of Cu2þ toxicity to Agþ toxicity between the two exposure conditions: exposure to the non-interactive mixtures in the present study and exposure to the single metals in a previous study of Le et al. (2012b) (Table 2). Particularly, the value of EC50Cu was similar to the value of EC50Ag when lettuce was exposed to metal mixtures while these Table 2 Toxicity of Cu2þ, Zn2þ, and Agþ expressed by the median effective activity EC50M (mM): following exposure to their non-interactive mixtures estimated by the concentration addition (CA) and response addition (RA) models assuming no interactions in the metal mixtures found in the present study; and following single exposure predicted in the study of Le et al. (2012b). For the results in the study of Le et al. (2012b), 95% confidence intervals are provided. Exposure

Models

Cu2þ

Zn2þ

Agþ

Source

Non-interactive mixtures Single exposure

CA RA

0.26 0.22 0.02e0.04

122.58 123.14 91.10e124.00

0.26 0.23 0.12e0.15

Present study Le et al., 2012b

Fig. 1. The response of lettuce roots expressed as Growth (mm) is plotted as a function of the exposure level in the solution: the free ion activity of Cu2þ ({Cu2þ}; mM) and the free ion activity of Zn2þ ({Zn2þ}; mM) according to the extended concentration addition (CA) model (A) and the extended response addition (RA) model (B). The surfaces describe the estimations by the models. The dotted points represent the experimental data.

T.T.Y. Le et al. / Environmental Pollution 176 (2013) 185e192

189

Table 3 Estimates of coefficients in Eq. (5) and statistical parameters determined by the regression analysis representing toxicity of interactive Cu2þeZn2þ mixtures as well as toxicological interactions between Cu2þ and Zn2þ according to the extended concentration addition model (n ¼ 122; R2 ¼ 0.92) (ASE stands for asymptotic standard error). Coefficient

Definition

Estimate

ASE

Parameter/ ASE

95% confidence interval Lower

Upper

b (mm) c1 (mM1) c12 (mM1) c2 (mM1) d (/)

Control growth Strength of Cu2þ toxicity Effects of Zn2þ on Cu2þ toxicity Strength of Zn2þ toxicity Slope

49.46 3.88 25.28  103 6.53  103 1.26

1.14 0.40 8.23  103 0.36  103 0.10

43.42 9.65 3.07 18.29 12.07

47.20 3.09 8.98  103 5.82  103 1.05

51.72 4.68 41.57  103 7.24  103 1.46

following the single exposure as the value of EC50Cu when lettuce was exposed to the mixtures was about one order of magnitude higher than the corresponding figure when lettuce was exposed to Cu2þ singly. By contrast, there was no significant difference in the estimation of toxicity of Zn2þ to lettuce as the value of EC50Zn determined in mixture assessment was in the range of the 95% CI calculated in single assessment (Table 2). 4.2. Toxicity of interactive mixtures of Cu2þeAgþ and Cu2þeZn2þ Metalemetal interactions are complicated as single metals, e.g., Cu2þ, Agþ, and Zn2þ, separately may have toxic effects on plants via a number of mechanisms. These processes may contribute to the interactions in Cu2þeAgþ and Cu2þeZn2þ mixtures predicted in the present study. In general, modes of action of Cu2þ include increasing the membrane potential and subsequently affecting the membrane permeability, and blocking ion channels (Demidchik et al., 1997; Kiss and Osipenko, 1994; Salama et al., 1992; Gilly and Armstrong, 1982). Noticeably, Cu2þ does not affect the conductance of the Kþchannel (Demidchik et al., 1997). Cu2þ can competitively replace other cations, e.g., Zn2þ and Ca2þ, at their binding sites in plant cell protein and lipid compounds, disrupting the metabolism (Mierle and Stokes, 1976; Watkins and Ferguson, 1982; Lidon and Henriques, 1993). According to Coskun et al. (2012), Agþ inhibited Kþ influx by two different mechanisms: directly as a Kþ-channel blocker at lower concentrations and indirectly via membrane destruction, e.g., increased permeability, at higher concentrations. This is consistent with the observation by Hendrix and Higinbotham (1974) that in plants, Agþ was able to substitute Kþ in membranes, thus inhibiting the uptake of other cations by roots. Similar to Cu2þ, elevated concentrations of Zn2þ rapidly result in changes in the membrane potential, i.e., depolarisation, of the root cell (Kenderesova et al., 2012). In addition, both deficiency and excess of Zn2þ may increase membrane permeability (Michael and Krishnaswamy, 2011; Chen et al., 2009; Kaya and Higgs, 2000). Furthermore, the induction and expression of proteins, e.g., ZIP, at the plasma membrane involved in the transport of metals like Cu2þ and Zn2þ are increased under Zn2þ deficiency and inhibited under Zn2þ sufficiency (Grotz and Guerinot, 2006; Maser et al., 2001;

Eckhardt et al., 2001; Ramesh et al., 2003; Ishimaru et al., 2005; Eide et al., 1996; Desbrosses-Fonrouge et al., 2005). Based on the modes of action of Cu2þ, Agþ, and Zn2þ reported in previous studies as mentioned above, some interactions might occur between these metals following exposure to their mixtures and might contribute to the results obtained in the present study by the extended CA and RA models. If Cu2þ and Agþ act in the same ways, i.e., reducing the membrane potential or blocking the Kþchannel, they may interact with each other, possibly influencing their toxicities. Particularly, either of these two mechanisms might lead to the inhibition of the transport and subsequent toxicity Cu2þ and Agþ by each other. As such, Cu2þ might reduce Agþ toxicity. Other mechanisms might additionally contribute to the alleviative effects of Cu2þ on Agþ toxicity. In particular, Cu2þ-induced compounds may detoxify Agþ and this mechanism was found to contribute to the alleviative effects of Cu2þ on Agþ toxicity in the study of Howe and Merchant (1992). In addition, the decline in the induction of non-selective uptake mechanisms, e.g., polypeptides, due to the elevated exposure to Cu2þ might further reduce Agþ uptake and toxicity (Howe and Merchant, 1992). Additionally, Agþ might lead to damage to the cell membrane, possibly increasing Cu2þ uptake and subsequently enhancing Cu2þ toxicity. These interactions might, to some extent, account for the interactions predicted in the present study according to the CA model, i.e., Cu2þ reduced Agþ toxicity while Agþ enhanced Cu2þ toxicity. Considering the possibility that Cu2þ and Agþ have different modes of action (RA model), e.g., Agþ blocks the Kþ-channel while Cu2þ does not affect the conductance of the Kþ channel, it would be expected that Cu2þ and Agþ do not affect the toxicity of each other as predicted in the present study. The alleviative effects of Zn2þ on Cu2þ toxicity to lettuce Lactuca sativa predicted in the present study is consistent with results reported for cress Lepidium sativum, duckweed Lemna minor, and pigeon pea Cajanus cajan in other studies (Montvydiene and Marciulioniene, 2007; Ince et al., 1999; Dirilgen and Inel, 1994; Sresty and Rao, 1999). This common inhibition might be related to interactions between these metals as follows. Possessing similar physical properties as noted by Weast (1976), Cu2þ and Zn2þ share main modes of actions and transport mechanisms as described above and reported previously (Arguello, 2003; Rensing et al., 1999;

Table 4 Estimates of coefficients in Eq. (6) and statistical parameters determined by the regression analysis representing toxicity of interactive Cu2þeZn2þ mixtures as well as toxicological interactions between Cu2þ and Zn2þ according to the extended response addition model (n ¼ 122; R2 ¼ 0.92) (ASE stands for asymptotic standard error). Coefficient

Definition

Estimate

ASE

Parameter/ ASE

95% confidence interval Lower

Upper

b (mm) c1 (mM1) c12 (mM1) d1 (/) c2 (mM1) d2 (/)

Control growth Strength of Cu2þ toxicity Effects of Zn2þ on Cu2þ toxicity Slope of Cu2þ toxicity curve Strength of Zn2þ toxicity Slope of Zn2þ toxicity curve

49.28 3.54 14.31  103 1.08 6.68  103 1.36

1.10 0.40 6.14 0.12 0.36 0.13

44.78 8.76 2.33 9.11 18.57 10.47

47.10 2.74 2.14  103 0.84 5.97  103 1.10

51.46 4.33 26.47  103 1.31 7.39  103 1.62

190

T.T.Y. Le et al. / Environmental Pollution 176 (2013) 185e192

Fig. 2. The response of lettuce roots expressed as Growth (mm) is plotted as a function of the exposure level in the solution: the free ion activity of Cu2þ ({Cu2þ}; mM) and the free ion activity of Agþ ({Agþ}; mM) according to the extended concentration addition model. The surface describes the estimations by the model. The dotted points represent the experimental data.

Axelsen and Palmgren, 1998). Consequently, their uptake might be inhibited by each other due to their potential direct interactions at the plasma membrane (Bowen, 1969; Giordano et al., 1974; Kausar et al., 1976; Hawf and Schmid, 1967; Chaudhry and Loneragan, 1972). As such, elevated Zn2þ may reduce Cu2þ transport, potentially alleviating Cu2þ toxicity. Moreover, at high levels of Zn2þ, the decline in the synthesis and expression of transporter proteins as reviewed above might reduce Cu2þ uptake further. Zn2þ might be replaced by Cu2þ because of higher affinity of Cu2þ for exchangeable sites on the root cell walls compared to Zn2þ (Nishizono et al., 1987; Ernst et al., 1992; Franco et al., 2002). The inhibition of Zn2þ toxicity may be additionally caused by changes in the structure of transporter proteins, which are induced by binding of Cu2þ to amino acids in the proteins (Bal et al., 1993; Freedman et al., 1982; Lopez-Millan et al., 2004; Orfei et al., 2003; Ooi et al., 1996; Petris et al., 2003; Stephens et al., 2011). However, under Zn2þ deficiency, the enhanced induction of proteins associated with the transport of Zn2þ as mentioned above might increase Zn2þ uptake, possibly compensating for the inhibition of Zn2þ uptake caused by Table 5 Estimates of coefficients in Eq. (7) and statistic parameters determined by the regression analysis representing toxicity of interactive Cu2þeAgþ mixtures as well as toxicological interactions between Cu2þ and Agþ according to the extended concentration addition model (n ¼ 116; R2 ¼ 0.80) (ASE stands for asymptotic standard error). Coefficient

Definition

Estimate

ASE

Parameter/ ASE

95% confidence interval

b (mm) c1 (mM1)

Control growth Strength of Cu2þ toxicity Effects of Agþ on Cu2þ toxicity Strength of Agþ toxicity Effects of Cu2þ on Agþ toxicity Slope

50.23 4.12

1.77 0.45

28.41 9.22

46.72 3.23

53.73 5.01

2.14

0.45

4.78

3.03

1.26

3.53

0.26

13.71

3.02

4.04

8.16

2.38

3.43

12.88

3.44

1.76

0.25

6.96

1.26

2.27

Lower

c13 (mM1) c3 (mM

1

)

1

c31 (mM d (/)

)

Upper

elevated Cu2þ exposure levels. These mechanisms possibly contribute to the insignificant effects of Cu2þ on Zn2þ toxicity as well as alleviative effects of Zn2þ on Cu2þ toxicity predicted in the present study. Another mode of action that might be responsible for the interactions between Cu2þ and Zn2þ is their effects on decreasing the membrane potential, which may subsequently reduce uptake of these metals (Kinraide, 1999; Kinraide et al., 2004). As a result, Zn2þ may alleviate Cu2þ toxicity as found in the present study. The effects of these cations on the membrane potential vary, depending on their concentrations, affinity, and mode of action (Kenderesova et al., 2012). Cu2þ is more effective than Zn2þ in decreasing negative charge (Irving and Williams, 1948; Bowen, 1966; Isermann, 1979). However, the Cu2þ activities in the solution tested in the present study were some orders of magnitude lower than the Zn2þ activities, possibly contributing to insignificant effects of Cu2þ on the membrane potential as well as on Zn2þ uptake toxicity as reported in the present study. In short, because of the complicated biological actions of metals, different interactions might occur between the metals at both toxicodynamic and toxicokinetic phases. The interactive effect of one metal on the toxicity of the other in the mixture might result from a combination of these mechanisms. Although the pattern of the interactions between Cu2þ and Agþ and between Cu2þ and Zn2þ predicted in the present study might be related to the toxic actions of these single metals reported in literature, physiological studies investigating biological actions of metals at different plant cell compartments following mixture exposure are required to support this relationship and to provide better insight into the mechanism of metalemetal interactions. 4.3. Modelling metalemetal interactions by mathematical equations The results of the present study demonstrate that around 80e 92% of the variability in the toxicity of Cu2þeZn2þ and Cu2þeAgþ mixtures could be explained by the equations used. The underestimation of the root growth of lettuce exposed to mixtures of Cu2þ and Zn2þ at their low activities in the exposure solution might be related to their essentiality, i.e., these metals may be beneficial to the growth of lettuce at these low exposure levels. The different interactions in mixtures of Cu2þ and Agþ predicted by the extended CA model and by the RA model indicate that the conclusion about interactions is strongly influenced by the mathematical relation used, consistent with the observation of Hernandez and Blazer (2006). Linearity as applied in the present study has been widely used to express interactions between different chemicals (Kinraide et al., 2004; Preacher et al., 2006). Linear relationships provide a simple description of the data from the perspective that the contribution of each predictor is summarised in a single coefficient (Hastie and Tibshirani, 1990). However, the use of linearity as in the present study to interpret interactions between different metals may lead to particular uncertainties. For example, linear relationships imply an increase in toxic effects with increasing exposure levels that does not hold under conditions of deficiency of essential metals. In addition, the dependence of the interactions between different metal ions on their doses resulting from the variations in the biological actions of the metals with varying exposure level as demonstrated above is not included in the linear relationship. Besides the linearity, interactions between different substances may follow other patterns and accordingly be expressed by other mathematical relations (Hamm et al., 2005). The use of linearity only while excluding other relationships does not necessarily reflect the actual interaction between variables (Lubinski and Humphreys, 1990; Cohen, 1978; Birnbaum, 1973; Busemeyer and Jones, 1983). Particularly, the statistical significance of the

T.T.Y. Le et al. / Environmental Pollution 176 (2013) 185e192

interaction found by the linear regression may be mainly due to an overlap with unchecked, but significant, nonlinear relations (Cortina, 1993; Lubinski and Humphreys, 1990). The method of using mathematical relationships as applied in the present study demonstrates some advantages in estimating mixture toxicity. This approach shows good predictive power in quantifying toxicity of metal mixtures taking into account potential metalemetal interactions. Additionally, this modelling approach reveals a full doseeresponse curve, describing toxicity as a function of the free ion activity of all mixture components, instead of only one single value of the concentration or activity at a certain response level, e.g., 50%. Furthermore, the mathematical extension of the CA and RA models provide quantitative estimates of toxicity of interactive mixtures while the application of the conventional concept of the models only offer qualitative estimations (higher or lower than additive effects). In summary, the approach of using mathematical relationships might be considered as a potential method for modelling metale metal interactions and integrating these interactions in predicting metal toxicity. Based on the conventional concept of assuming no interactions in mixtures, negligible differences were found between the estimations of metal toxicity by the CA model and by the RA model. By contrast, different patterns of metalemetal interactions were found based on the extended CA model and on the extended RA model. These interactions might result from various actions of toxicity of the single metals. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.envpol.2013.01.017. References Arguello, J.M., 2003. Identification of ion selectivity determinants in heavy metal transport P1B-type ATPases. Journal of Membrane Biology 195, 93e108. Axelsen, K.B., Palmgren, M.G., 1998. Evolution of substrate specificities in the P-type ATPase superfamily. Journal of Molecular Evolution 46, 84e101. Backhaus, T., Altenburger, R., Boedeker, W., Faust, M., Scholze, M., Grimme, L.H., 2000a. Predictability of the toxicity of a multiple mixture of dissimilarly acting chemicals to Vibrio fischeri. Environmental Toxicology and Chemistry 19, 2348e2356. Backhaus, T., Scholze, M., Grimme, L.H., 2000b. The single substance and mixture toxicity of quinolones to the bioluminescent bacterium Vibrio fischeri. Aquatic Toxicology 49, 49e61. Bal, W., Kozlowski, H., Kupryszewski, G., Mackiewicz, Z., Pettit, L., Robbins, R., 1993. Complexes of Cu(II) with Asn-Ser-Phe-Arg-Tyr-NH2; an example of metal ionpromoted conformational organization which results in exceptionally high complex stability. Journal of Inorganic Biochemistry 52, 79e87. Birnbaum, M.H., 1973. The devil rides again: correlations as an index of fit. Psychological Bulletin 79, 239e242. Bliss, C.I., 1939. The toxicity of poisons applied jointly. Annals of Applied Biology 26, 585e615. Bowen, H.J.M., 1966. Trace Elements in Biochemistry. Academic Press, London. Bowen, J.E., 1969. Absorption of copper, zinc, and manganese by sugarcane leaf tissue. Plant Physiology 44, 255e261. Busemeyer, J.R., Jones, L.E., 1983. Analysis of multiplicative combination rules when the causal variables are measured with error. Psychological Bulletin 93, 549e 562. Chaudhry, F.M., Loneragan, J.F., 1972. Zinc absorption by wheat seedlings and the nature of its inhibition by alkaline earth cations. Journal of Experimental Botany 23, 552e560. Chen, W.R., He, Z.L., Yang, X.E., Feng, Y., 2009. Zinc efficiency is correlated with root morphology, ultrastructure, and antioxidative enzymes in rice. Journal of Plant Nutrition 32, 287e305. Cohen, J., 1978. Partialled products are interactions: partialled powers are curve components. Psychological Bulletin 85, 858e866. Cooper, N.L., Bidwell, J.R., Kumar, A., 2009. Toxicity of copper, lead, and zinc mixtures to Ceriodaphnia dubia and Daphnia carinata. Ecotoxicology and Environmental Safety 72, 1523e1528. Cortina, J.M., 1993. Interaction, nonlinearity, and multicollinearity: implications for multiple regression. Journal of Management 19, 915e922. Coskun, D., Britto, D.T., Jean, Y.-K., Schulze, L.M., Becker, A., Kronzucker, H.J., 2012. Silver ions disrupt Kþ homeostasis and cellular integrity in intact barley (Hordeum vulgare L.) roots. Journal of Experimental Botany 63, 151e162.

191

Demidchik, V., Sokolik, A., Yurin, V., 1997. The effect of Cu2þ on ion transport systems of the plant cell plasmalemma. Plant Physiology 114, 1313e1325. Desbrosses-Fonrouge, A.-G., Voigt, K., Schroder, A., Arrivault, S., Thomine, S., Kramer, U., 2005. Arabidopsis thaliana MTP1 is a Zn transporter in the vacuolar membrane which mediates Zn detoxification and drives leaf Zn accumulation. FEBS Letters 579, 4165e4174. Dirilgen, N., Inel, Y., 1994. Effects of Zn and Cu on growth and metal accumulation in duckweed, Lemna minor. Bulletin of Environmental Contamination and Toxicology 53, 442e449. Eckhardt, U., Marques, A.M., Buckhout, T.J., 2001. Two iron-regulated cation transporters from tomato complement metal uptake-deficient yeast mutants. Plant Molecular Biology 45, 437e448. Eide, D., Broderius, M., Fett, J., Guerinot, M.L., 1996. A novel iron-regulated metal transporter from plants identified by functional expression in yeast. Proceedings of the National Academy of Sciences of the United States of America 93, 5624e5628. Ernst, W.H.O., Verkleij, J.A.C., Schat, H., 1992. Metal tolerance in plants. Acta Botanica Neerlandica 41, 229e248. Franco, C.R., Chagas, A.P., Jorge, R.A., 2002. Ion-exchange equilibria with aluminium pectinates. Colloids and Surfaces A-Physicochemical and Engineering Aspects 204, 183e192. Freedman, J.H., Pickart, L., Weinstein, B., Mims, W.B., Peisach, J., 1982. Structure of the glycyl-L-histidyl-L-lysineecopper(II) complex in solution. Biochemistry 21, 4540e4544. Frias-Espericueta, M.G., Voltolina, D., Osuna-López, I., Izaguirre-Fierro, G., 2009. Toxicity of metal mixtures to the Pacific white shrimp Litopenaeus vannamei postlarvae. Marine Environmental Research 68, 223e226. Gilly, W.M.F., Armstrong, C.M., 1982. Divalent cations and the activation kinetics of potassium channels in squid giant axons. Journal of General Physiology 79, 965e996. Giordano, P.M., Noggle, C.J., Mortvedt, J.J., 1974. Zinc uptake by rice, as affected by metabolic inhibitors and competing cations. Plant and Soil 41, 637e646. Grotz, N., Guerinot, M.L., 2006. Molecular aspects of Cu, Fe and Zn homeostasis in plants. Biochimica et Biophysica Acta-Molecular Cell Research 1763, 595e608. Hamm, A.K., Carter Jr., H.W., Gennings, C., 2005. Analysis of an interaction threshold in a mixture of drugs and/or chemicals. Statistics in Medicine 24, 2493e2507. Hastie, T.J., Tibshirani, R.J.,1990. Generalized Additive Models. Chapman & Hall, London. Hawf, L.R., Schmid, W.E., 1967. Uptake and translocation of zinc by intact plants. Plant and Soil 27, 249e260. Hendrix, D.L., Higinbotham, N., 1974. Heavy metals and sulphhydryl reagents as probes of ion uptake in pea stem. In: Membrane Transport in Plants. Springe Verlag, Berlin, p. 412. Hernandez, L.M., Blazer, D.G., 2006. Genes, Behavior, and the Social Environment: Moving Beyond the Nature/Nurture Debate. In: Committee on Assessing Interactions Among Social, Behavioral, and Genetic Factors in Health, p. 384. Hewlett, P.S., Plackett, R.L., 1979. An Introduction to the Interpretation of Quantal Responses in Biology. Edward Arnold, London, UK, p. 82. Howe, G., Merchant, S., 1992. Heavy metal-activated synthesis of peptides in Chlamydomonas reinhardtii. Plant Physiology 98, 127e136. Ince, N.H., Dirilgen, N., Apykyan, I.G., Tezcanli, G., Ustun, B., 1999. Assessment of toxic interactions of heavy metals in binary mixtures: a statistical approach. Archives of Environmental Contamination and Toxicology 36, 365e372. Irving, H., Williams, R.J.P., 1948. Order of stability of metal complexes. Nature 162, 746e747. Isermann, K., 1979. Neure Erkenntnisse der Pflanzenernahrung. Chemie in Unserer Zeit 13, 97e141. Ishimaru, Y., Suzuki, M., Kobayashi, T., Takahashi, M., Nakanishi, H., Mori, S., Nishizawa, N.K., 2005. OsZIP4, a novel zinc-regulated zinc transporter in rice. Journal of Experimental Botany 56, 3207e3214. Kabata-Pendias, A., Pendias, H., 1984. Trace Elements in Soils and Plants. CRC Press, Boca Raton, FL, USA. Kausar, M.A., Chaudhry, F.M., Rashid, A., Latif, A., Alam, S.M., 1976. Micronutrient availability to cereals from calcareous soils. Plant and Soil 45, 397e410. Kaya, C., Higgs, D., 2000. Short-term relationships between membrane permeability and growth parameters in three tomato cultivars grown at low and high zinc. Journal of Plant Nutrition 23, 1373e1383. Kenderesova, L., Stanova, S., Pavlovkin, J., Durisova, E., Nadubinska, M., Ciamporova, M., Ovecka, M., 2012. Early Zn2þ-induced effects on membrane potential account for primary heavy metal susceptibility in tolerant and sensitive Arabidopsis species. Annals of Botany 110, 445e459. Kinraide, T.B., Parker, D.R., 1989. Assessing the phytotoxicity of mononuclear hydroxy-aluminum. Plant Cell and Environment 12, 479e487. Kinraide, T.B., Pedler, J.F., Parker, D.R., 2004. Relative effectiveness of calcium and magnesium in the alleviation of rhizotoxicity in wheat induced by copper, zinc, aluminium, sodium, and low pH. Plant and Soil 259, 201e208. Kinraide, T.B., 1999. Interactions among Ca2þ, Naþ and Kþ in salinity toxicity: quantitative resolution of multiple toxic and ameliorative effects. Journal of Experimental Botany 50, 1495e1505. Kiss, T., Osipenko, O.N., 1994. Toxic effects of heavy metals on ionic channels. Pharmacological Reviews 46, 245e267. Kopittke, P.M., Blamey, F.P.C., McKenna, B.A., Wang, P., Menzies, N., 2011. Toxicity of metals to roots of cowpea in relation to their binding strength. Environmental Toxicology and Chemistry 30, 1827e1833. Le, T.T.Y., Peijnenburg, W.J.G.M., Hendriks, A.J., Vijver, M.G., 2012a. Predicting effects of cations on copper toxicity to lettuce (Lactuca sativa) by the biotic ligand model. Environmental Toxicology and Chemistry 2, 355e359.

192

T.T.Y. Le et al. / Environmental Pollution 176 (2013) 185e192

Le, T.T.Y., Vijver, M.G., Hendriks, A.J., Peijnenburg, W.J.G.M., 2012b. Modeling toxicity of binary metal mixtures (Cu2þeAgþ, Cu2þeZn2þ) to lettuce, Lactuca sativa, with the Biotic Ligand Model. Environmental Toxicology and Chemistry. http:// dx.doi.org/10.1002/etc.2039. Lidon, F.C., Henriques, F.S., 1993. Effect of copper toxicity on growth and the translocation of metals in rice plants. Journal of Plant Nutrition 16, 1449e 1464. Lock, K., van Eeckhout, H., De Schamphelaere, K.A.C., Criel, P., Janssen, C.R., 2007. Development of a biotic ligand model (BLM) predicting nickel toxicity to barley (Hordeum vulgare). Chemosphere 66, 1346e1352. Lopez-Millan, A.-F., Ellis, D.R., Grusak, M.A., 2004. Identification and characterization of several new members of the zip family of metal ion transporters in Medicago truncatula. Plant Molecular Biology 54, 583e596. Lubinski, D., Humphreys, L.G., 1990. Assessing spurious “moderator effects”: illustrated substantively with the hypothesized (“synergistic”) relation between spatial and mathematical ability. Psychological Bulletin 107, 385e393. Manzo, S., Buono, S., Cremisini, C., 2010. Cadmium, lead and their mixtures with copper: Paracentrotus lividus embryotoxicity assessment, prediction, and offspring quality evaluation. Ecotoxicology 19, 1209e1223. Maser, P., Thomine, S., Schroeder, J.I., Ward, J.M., Hirsch, K., Sze, H., Talke, I.N., Amtmann, A., Maathuis, F.J.M., Sanders, D., Harper, J.F., Tchieu, J., Gribskov, M., Persans, M.W., Salt, D.E., Kim, S.A., Guerinot, M.L., 2001. Phylogenetic relationship within cation transporter families of Arabidopsis. Plant Physiology 126, 1646e1667. McKenna, I.M., Chaney, R.L., Williams, F.M., 1993. The effects of cadmium and zinc interactions on the accumulation and tissue distribution of zinc and cadmium in lettuce and spinach. Environmental Pollution 79, 113e120. Michael, P.I., Krishnaswamy, M., 2011. The effect of zinc stress combined with high irradiance stress on membrane damage and antioxidative response in bean seedlings. Environmental and Experimental Botany 74, 171e177. Mierle, G.M., Stokes, P.M., 1976. Heavy metal tolerance and metal accumulation by planktonic algae. In: Hemphill, D.D. (Ed.), Trace Substances and Environmental Health. University of Missouri, Columbia, pp. 113e122. Montvydiene, D., Marciulioniene, D., 2007. Assessment of toxic interaction of metals in binary mixtures using Lepidium sativum and Spirodela polyrrhiza. Polish Journal of Environmental Studies 16, 777e783. Nishizono, H., Ichikawa, H., Suziki, S., Ishii, F., 1987. The role of the root cell wall in the heavy metal tolerance of Athyrium yokoscense. Plant and Soil 101, 15e20. Norwood, W.P., Borgmann, U., Dixon, D.G., Wallace, A., 2003. Effects of metal mixtures on aquatic biota: a review of observations and methods. Human and Ecological Risk Assessment 4, 795e811. [OECD] Organisation for Economic Cooperation and Development, 2006. OECD Test Guideline 227: terrestrial plant test-vegetative vigour. In: OECD Guidelines for Testing Chemicals, Paris, France. Ooi, C.E., Rabinovich, E., Dancis, A., Bonifacino, J.S., Klausner, R.D., 1996. Copperdependent degradation of the Saccharomyces cerevisiae plasma membrane copper transporter Ctr1p in the apparent absence of endocytosis. EMBO Journal 15, 3515e3523.

Orfei, M., Alcaro, M.C., Marcon, G., Chelli, M., Ginanneschi, M., Kozlowski, H., Brasun, J., Messori, L., 2003. Modeling of copper(II) sites in proteins based on histidyl and glycyl residues. Journal of Inorganic Biochemistry 97, 299e307. Otitoloju, A.A., 2002. Evaluation of the joint-action toxicity of binary mixtures of heavy metals against the mangrove periwinkle Tympanotonus fuscatus var radula (L.). Ecotoxicology and Environmental Safety 53, 404e415. Pahlsson, A.-M.B., 1989. Toxicity of heavy metals (Zn, Cu, Cd, Pb) to vascular plants. A literature review. Water Air and Soil Pollution 47, 287e319. Petris, M.J., Smith, K., Lee, J., Thiele, D.J., 2003. Copper-stimulated endocytosis and degradation of the human copper transporter, hCtr1. Journal of Biological Chemistry 278, 9639e9646. Preacher, K.J., Curran, P.J., Bauer, D.J., 2006. Computational tools for probing interactions in multiple linear regression, multilevel modeling, and latent curve analysis. Journal of Educational and Behavioral Statistics 31, 437e448. Ramesh, S.A., Shin, R., Eide, D.J., Schachtman, D.P., 2003. Differential metal selectivity and gene expression of two zinc transporters from rice. Plant Physiology 133, 126e134. Rensing, C., Ghosh, M., Rosen, B.P., 1999. Families of soft metal-ion transporting ATPases. Journal of Bacteriology 181, 5891e5897. Salama, G., Abramson, J.J., Pike, J.K., 1992. Sulfhydryl reagents trigger Ca2þ release from the sarcoplasmic reticulum of skinned rabbit psoas fibres. Journal of Physiology (London) 454, 389e420. Sharma, S.S., Schat, H., Vooijs, R., Van Heerwaarden, L.M., 1999. Combination toxicology of copper, zinc, and cadmium in binary mixtures: concentrationdependent antagonistic, nonadditive, and synergistic effects on root growth in Silene vulgaris. Environmental Toxicology and Chemistry 18, 348e355. Sresty, T.V.S., Rao, K.V.M., 1999. Ultrastructural alterations in response to zinc and nickel stress in the root cells of pigeonpea. Environmental and Experimental Botany 41, 3e13. Steiner, A.A., 1961. A universal method for preparing nutrient solution of a certain desired composition. Plant and Soil 15, 134e154. Stephens, B.W., Cook, D.R., Grusak, M.A., 2011. Characterization of zinc transport by divalent metal transporters of the ZIP family from the model legume Medicago truncatula. Biometals 24, 51e58. Thakali, S., Allen, H.E., Di Toro, D.M., Ponizovsky, A.A., Rooney, C.P., Zhao, F.-J., McGrath, S.P., 2006. A terrestrial biotic ligand. 1. Development and application to Cu and Ni toxicities to barley root elongation in soils. Environmental Science & Technology 40, 7085e7093. Tipping, E., 1998. Humic ion-binding model VI: an improved description of the interactions of protons and metal ions with humic substances. Aquatic Geochemistry 4, 3e48. Verweij, W., 2004. Chemical Equilibria in Aquatic Systems (CHEAQS Pro). http:// home.tiscali.nl/cheaqs. Voigt, S., Hendershot, W.H., Sunahara, G.I., 2006. Rhizotoxicity of cadmium and copper in soil extracts. Environmental Toxicology and Chemistry 25, 692e701. Watkins, C.B., Ferguson, I.B., 1982. The interaction of copper and zinc with calcium in apple fruit. Scientia Horticulturae 17, 319e325. Weast, R., 1976. CRC Handbook of Chemistry and Physics. CRC Press, Cleveland.