Low magnesium status in plants enhances tolerance to cadmium exposure


  • Christian Hermans,

    1. Laboratory of Plant Physiology and Molecular Genetics, Université Libre de Bruxelles, Bd du Triomphe, 1050 Brussels, Belgium
    Search for more papers by this author
  • Jiugeng Chen,

    1. Laboratory of Plant Physiology and Molecular Genetics, Université Libre de Bruxelles, Bd du Triomphe, 1050 Brussels, Belgium
    Search for more papers by this author
  • Frederik Coppens,

    1. Department of Plant Systems Biology, VIB, B-9052 Ghent, Belgium
    2. Department of Plant Biotechnology and Bioinformatics, Ghent University, B-9052 Ghent, Belgium
    Search for more papers by this author
  • Dirk Inzé,

    1. Department of Plant Systems Biology, VIB, B-9052 Ghent, Belgium
    2. Department of Plant Biotechnology and Bioinformatics, Ghent University, B-9052 Ghent, Belgium
    Search for more papers by this author
  • Nathalie Verbruggen

    1. Laboratory of Plant Physiology and Molecular Genetics, Université Libre de Bruxelles, Bd du Triomphe, 1050 Brussels, Belgium
    Search for more papers by this author

Author for correspondence:
Christian Hermans
Tel: +32 2 6505417
Email: chermans@ulb.ac.be


  • In a transcriptomic study of magnesium (Mg) starvation in Arabidopsis, we identified several genes that were differentially regulated which are involved in the detoxification process of nonessential heavy metals such as cadmium (Cd).
  • We further tested the impact of low Mg status on Cd sensitivity in plants. Interestingly, a −Mg pretreatment of 7 d alleviated the bleaching of young leaves caused by Cd. No or little difference in Cd tissue concentration between the +Mg and −Mg plants was observed, suggesting that lower Cd toxicity was probably not attributable to modified root to shoot translocation.
  • Mg deficiency also promoted an increase in the iron (Fe) concentration (up to one-fourth) in Cd-treated leaves. Because high Fe concentrations have previously been reported to prevent the harmful effects of Cd, we explored whether Fe homeostasis plays a role in the Mg–Cd interaction. A protective effect of −Mg pretreatment was also observed on Fe starvation. However, Fe foliar spray partially alleviated Cd-induced chloroses, while it almost completely restored chlorophyll content in Fe-deficient leaves.
  • In conclusion, the protective effect of Mg against Cd toxicity could be attributable partly to the maintenance of Fe status but also to the increase in antioxidative capacity, detoxification and/or protection of the photosynthetic apparatus.


Mineral interactions have been demonstrated to be important for the manifestation of deficiency and toxicity in plants (Suzuki, 2005; Ward et al., 2008). The role of high magnesium (Mg) supply in alleviating aluminium (Al) (Bose et al., 2011) and cadmium (Cd) (Kashem & Kawai, 2007) toxicity has already been described. Here we report a positive effect of low Mg status on the alleviation of toxicity symptoms caused by Cd exposure in Arabidopsis.

Mg is an essential macronutrient that plants utilize in large quantities for their growth (Karley & White, 2009). In addition to being the central atom of the chlorophyll molecule, Mg is essential for the functioning of many enzymes and chelation to nucleotidyl phosphate forms (Shaul, 2002). Early responses to Mg deficiency include the enhancement of antioxidative mechanisms (Cakmak & Marschner, 1992; Cakmak & Kirkby, 2008; Hermans et al., 2010b) and accumulation of sugars in leaves, which occur before any noticeable effect on photosynthetic activity (Cakmak et al., 1994a,b; Hermans et al., 2004, 2005; Hermans & Verbruggen, 2005). In Arabidopsis, we have shown that Mg deficiency reduces the growth of young leaves more than that of the roots, as carbon allocation to the youngest leaves is affected more than that to the roots (Hermans & Verbruggen, 2005; Hermans et al., 2006). Those observations were consistent with the finding that the number of genes differentially regulated by Mg depletion in the roots was much lower than that in the leaves (Hermans et al., 2010a,b).

Cd is a nonessential element for living organisms, with the exception of marine diatoms (Morel, 2008). In plants, Cd can be taken up by roots through the same plasma membrane transporters as those used for other cations such as calcium (Ca), iron (Fe) and zinc (Zn) (Korshunova et al., 1999; Clemens, 2006; Nakanishi et al., 2006; Lux et al., 2011). Because Cd is chemically similar to those essential elements, it can deregulate their homeostasis or cause their displacement from proteins (reviewed in Verbruggen et al., 2009). A major source of Cd toxicity is its reactivity to sulphydryl groups. Cd can deplete the reduced glutathione pool and increase reactive oxygen species (ROS) production by impairing the mitochondrial and photosynthetic electron transfer chains (Heyno et al., 2008; Smeets et al., 2008; Bi et al., 2009; Sandalio et al., 2009). Furthermore, Cd can directly or indirectly damage the photosynthetic apparatus by replacing Mg in the chlorophyll structure (Küpper et al., 1998), by inhibiting photoactivation of photosystem II (Faller et al., 2005) or by affecting chlorophyll biosynthesis (Gadallah, 1995), the organization and assembly of light-harvesting complexes (Janik et al., 2010) and chloroplast structure and density (Krupa, 1988; Stoyanova & Tchakalova, 1997; Baryla et al., 2001; Carrier et al., 2003; Bi et al., 2009). Interestingly, plants exposed to high Cd and to low Fe share similar chlorotic symptoms. As high Cd content in the growth medium inhibits Fe uptake in a number of species (Kovács et al., 2010), some authors assert that Cd-induced Fe deficiency in the shoots is of particular importance for the development of chloroses (Siedlecka & Krupa, 1999; Solti et al., 2008). However, other studies in Brassica tend to refute this assertion, suggesting that chlorosis is not attributable to mineral deficiency (Carrier et al., 2003). Discrepancies between these reports may be attributable to variable species sensitivities and/or differences in the Cd concentrations applied.

In our recent global transcriptomic studies of Mg deficiency in Arabidopsis (Hermans et al., 2010a,b), we identified a category of genes involved in the process of detoxification of drugs and heavy metals, such as Cd. In this report, we show that a Mg starvation pretreatment can alleviate Cd toxicity in plants.

Materials and Methods

Hydroponics culture

Arabidopsis thaliana Heynh Columbia (Col-0) plants were grown hydroponically in short-day conditions (8 h light; 100 μmol photons m−2 s−1/16 h darkness). The growth conditions and the nutrient solution composition are described elsewhere in Hermans & Verbruggen (2005) and Hermans et al. (2010a).

Mineral content analysis

Plant tissues were harvested and dried at 60°C. Full mineral profiles were assayed by inductively coupled plasma mass spectrometry performed by Purdue Ionomics Information Management Systems (PIIMS, Purdue, IN, USA).

Chlorophyll content analysis

Pigments were extracted with 80% acetone and chlorophyll content was estimated using the method of Porra et al. (1989).

Photosynthesis measurements

The behaviour of photosystem II was assessed using chlorophyll a (Chla) fast fluorescence kinetics (direct induction with 660-nm exposure), recorded with the Handy Plant Efficiency Analyser fluorometer (Hansatech Instruments, Norfolk, UK). The OJIP transients were analysed according to the JIP test procedure, as previously described by Hermans et al. (2003); Hermans & Verbruggen (2005).

RNA extraction, reverse transcription and quantitative PCR (qPCR)

Total RNA was extracted from 100 mg of ground frozen root and leaf samples with TRIzol reagent (RNeasy; Qiagen, Venlo, NL, USA). Reverse transcription was performed starting from 1 μg of total RNA using the RevertAid H Minus First Strand cDNA Synthesis kit (Fermentas, St. Leon-Rot, Germany). The Janus Automated Workstation (PerkinElmer, Waltham, MA, USA) handled all pipetting for PCRs in a final volume of 5 μl using the LightCycler 480 SYBR Green I Master mix (Roche). qPCR reactions were performed with the LightCycler 480 (Roche), with preincubation at 95°C for 10 min, 45 cycles at 95°C for 10 s, 60°C for 15 s and 72°C for 15 s, and melting curves according to the manufacturer’s recommendations. Primers designed using NCBI Primer-Blast (Rozen et al., 2000) are listed in Supporting Information Table S2.

Statistical analysis

Statistical analyses for biological data were carried out using the spss statistical software (SPSS, Chicago, IL, USA). One-way analysis of variance (ANOVA) and subsequent multiple comparison tests (LSDs) were performed to calculate the significance level of differences between means. Significance was set at the 5% level (< 0.05).


Five-week-old Araridopsis (Col-0) plants grown hydroponically were subjected to Mg starvation (complete omission of Mg from the nutrient solution) for a period of 1 wk. At day 7 of treatment, half of the Mg-fully supplied (+Mg) and Mg-starved (−Mg) plant populations were subjected to mild Cd exposure by transfer to a nutrient solution contaminated with 2 μM CdSO4 or were fed with an Fe-depleted solution (Fig. 1a). The experimental design included treatments in which plants were fully supplied with Mg and not subjected to Cd exposure (+Mg-Cd; the control), subjected to Cd exposure (+Mg+Cd) or subjected to Fe depletion (+Mg-Fe); and treatments in which plants were Mg-starved and not subjected to Cd exposure (-Mg-Cd), subjected to Cd exposure (-Mg+Cd) or subjected to Fe depletion (-Mg-Fe). We considered 1 wk of Mg depletion to be sufficient before applying the +Cd and −Fe treatments because, while no visible symptoms in leaves (Hermans et al., 2010b) or decrease in biomass (Fig. S1) was perceptible, the Mg content had already decreased considerably in plant organs. Indeed, the mineral analysis revealed that, after 7 d of Mg starvation treatment, the concentration had decreased significantly (< 0.05) by 74% in roots and by 56, 67 and 77%, respectively, in old mature leaves (OL), young mature/expanding leaves (YL) and immature leaves/rosette buds (IL) compared with Mg-fully supplied controls (Fig. 2a).

Figure 1.

Mineral treatments and hydroponics culture of Arabidopsis. (a) Plants were grown for 5 wk. At day 0 (d0), plants were fed with a nutrient solution containing magnesium (+Mg) or depleted in Mg (−Mg) for 1 wk. At d7, different treatments were applied, which consisted of the addition of 2 μM cadmium (+Cd) to or the removal of iron (−Fe) from the nutrient solution. Colour pictures of representative rosettes were taken after 7 d of Mg starvation pretreatment (d7) followed by 3 d or 7 d of Cd exposure or Fe starvation indicated by (d7)+3 or (d7)+7. The treatments were as follows: +Mg−Cd, Mg fully supplied without Cd addition (control); +Mg+Cd, Mg fully supplied and Cd added; +Mg−Fe, Mg fully supplied and Fe depleted; −Mg−Cd, Mg depleted; −Mg+Cd, Mg depleted and Cd added; −Mg−Fe, Mg and Fe depleted. Bar, 5 cm. (b) Total chlorophyll content expressed per unit fresh weight measured in old mature leaves (OL) and expanding and young mature leaves (YL). = 3; ± SE. The photosynthetic performance index (PI) was measured in the same leaves. = 10; ± SE. The different letters denote values significantly different at < 0.05, by an LSD test (uppercase letters for comparisons between leaves for the same treatment; lowercase letters for comparisons between treatments for the same leaf stage).

Figure 2.

Mineral profile in organs of Arabidopsis subjected to variable magnesium (Mg), cadmium (Cd) and iron (Fe) concentrations. Mg (a), Cd (b), Fe (c) and Mn (d) concentrations are expressed per unit dry weight of tissue in immature leaves and rosette buds (IL), expanding and young mature leaves (YL), old mature leaves (OL) and roots (R). = 3 individual plants at d7 and = 4 at (d7)+7 ± SE. Note the different scales on the y-axes for the Fe and Mn concentrations. For a description of the treatments, see the legend to Fig. 1. The different letters denote values significantly different at < 0.05, by an LSD test (uppercase letters for comparisons between organs for the same treatment; lowercase letters for comparisons between treatments for the same organ).

Low Mg status alleviates Cd toxicity symptoms in leaves

Cd exposure resulted in chlorotic symptoms primarily in IL and YL (Fig. 1a). Interestingly, the Mg starvation pretreatment alleviated these chloroses upon Cd exposure (Fig. 1a). After 7 d of Cd exposure, a decrease was observed in the total chlorophyll content in YL of −58% between the +Mg−Cd and +Mg+Cd treatments and of −20% between the −Mg−Cd and −Mg+Cd treatments (Fig. 1b). Also, the photosynthetic performance index (PI), characterizing overall photochemical processes in photosystem II (Hermans et al., 2003), decreased, respectively, by −94% between +Mg−Cd and +Mg+Cd and by −46% between −Mg−Cd and −Mg+Cd in YL (Fig. 1b). Note that, because IL were too small, no photosynthetic analysis could be performed on these leaves. After (7 d)+7 of toxic metal exposure, Cd accumulation was greater in roots than in leaves, in accordance with previous studies with the same species (Gong et al., 2003; Smeets et al., 2008; Li et al., 2010), and Cd concentrations were equal among the three leaf groups (Fig. 2b). Cd concentrations in roots and leaves (except for OL) did not differ significantly (< 0.05) between Mg-deficient and Mg-fully supplied backgrounds (Fig. 2b). The impact of Cd exposure was also assessed on the whole ionic profile. The addition of Cd decreased manganese (Mn) concentrations in roots (−61%) and leaves, with the highest difference observed in IL (−55%), compared with the +Mg–Cd treatment (Figs 2d, S2). The decrease in Mn concentration was similar in roots of +Mg and –Mg backgrounds but smaller in leaves of the −Mg background (Figs 2d, S1). Upon Cd exposure, the sulphur (S) concentration was increased in leaves of both backgrounds but to a greater extent in the −Mg background (up to +79% in OL between the −Mg+Cd and +Mg−Cd treatments) and Zn concentration was increased (+92%) in −Mg roots only (Fig. S2). Furthermore, ratios of mineral elements are reported to be crucial for chloroplast stability and chlorophyll contents and often considered as diagnostic criteria for chlorosis (Nenova & Stoyanov, 1999; Fernández-Falcón et al., 2006). We found that Cd treatment markedly increased the Mg : Mn ratio in YL and IL, and the Fe : Mn ratio in the same organs and in roots (Fig. S3). Mg deficiency pretreatment led to a more modest increase in the Fe : Mn ratio in leaves (Fig. S3).

Low Mg status improves Fe deficiency symptoms in leaves

Fe starvation applied for 1 wk also resulted in chlorotic symptoms but affected the very young leaves less and the mature leaves more, compared with Cd exposure (Fig. 1a). The Fe starvation treatment depleted the Fe content by −55% in roots and in leaves (up to −28% in IL) after (7 d)+7 in +Mg conditions (Fig. 2c). The Mg deficiency pretreatment was also able to prevent the chlorotic symptoms produced by Fe deficiency and to improve photosynthetic performance (Fig. 1b). Under those conditions, there was less of a reduction in the Fe concentrations of Mg-deficient leaves (up to −12% in IL between −Mg−Fe and +Mg−Fe) compared with Mg-sufficient conditions (Fig. 2c). The Fe depletion treatment had much more pronounced effects on the ionic profile than Cd exposure: copper (Cu), Mn and Zn (+553%) concentrations increased in roots and Cu (up to +197% in IL), Mn, S and Zn concentrations increased in leaves (Fig. S1). These differences were similar in the +Mg and −Mg backgrounds, with the exception of Cu, for which lower concentrations were found in −Mg leaves (up to +110% in IL between −Mg−Fe and +Mg−Fe).

It has been reported that total Fe content does not necessarily reflect the nutritional status of the plant (Köseoğlu & Açıkgöz, 1995; Pestana & Varennes, 2003). Therefore, we investigated Fe deficiency markers at the molecular level in Arabidopsis, which uses Fe(III) reduction/Fe(II) uptake strategy I (Robinson et al., 1999). A comparative examination of the following genes was carried out: FERRIC REDUCTION OXIDASE 2 (FRO2) (Robinson et al., 1999), IRON-REGULATED TRANSPORTER 1 (IRT1) (Eide et al., 1996) and some transcription factors involved in the activation of Fe acquisition genes: Fe-DEFICIENCY INDUCED TRANSCRIPTION FACTOR 1 (FIT1/BHLH29) (Colangelo & Guerinot, 2004), BHLH101 (Wang et al., 2007) and POPEYE (PYE) (Long et al., 2010). Upon Fe depletion, we observed the up-regulation of FRO2, IRT1, FIT1, BLHL101 and PYE in roots (Fig. 3a), in accordance with previous published reports (Eide et al., 1996; Robinson et al., 1999; Wintz et al., 2003; Lucena et al., 2006; Cassin et al., 2009; García et al., 2010; Long et al., 2010). Upon Cd exposure, some transcription factors were also induced but the main observation was that expression of downstream targets FRO2 and IRT1 was not triggered (Fig. 3a). Overall, the Mg starvation pretreatment triggered the expression of FIT1, BLHL101 and PYE to a lesser extent and resulted in equal or lower FRO2 and IRT1 expression levels in roots upon application of the +Cd and −Fe treatments (Fig. 3a).

Figure 3.

Expression patterns of several marker genes related to iron (Fe) deficiency. (a) Genes monitored in Arabidopsis roots were Fe-DEFICIENCY INDUCED TRANSCRIPTION FACTOR 1 (FIT1; At2g28160), BHLH101 (At5g04150), POPEYE (PYE; At3g47640), FERRIC REDUCTION OXIDASE 2 (FRO2; At1g01580) and IRON-REGULATED TRANSPORTER 1 (IRT1; At4g19690). (b) Genes monitored in young mature and expanding leaves (YL) and old mature leaves (OL) were ACONITASE 2 (ACO2; At4g26970) and FERRIC REDUCTION OXIDASE 1 (FER1; At5g01600). Expression levels were monitored by quantitative RT-PCR. Values shown are the average of two biological replicates ± SE. Three technical replicates were performed for each sample. For a description of the treatments, see the legend to Fig. 1.

Chloroses of Cd-treated plants are partially recovered by application of Fe foliar spray

Cd exposure and Fe depletion both induced chlorotic symptoms but on somewhat different leaves (Fig. 1a). The two treatments resulted in similar decreases in Fe concentrations in leaves but only Fe starvation resulted in a decrease in roots (Fig. 2c). To determine whether chloroses on Cd-treated leaves are related to an induced Fe deficiency or are attributable to other types of damage, we monitored the expression of markers related to Fe homeostasis in leaves such as FERRETIN 1 (FER1) (Thimm et al., 2001; Murgia et al., 2007) and three isoforms of ACONITASE (ACO1, ACO2 and ACO3) (Arnaud et al., 2007). Upon Fe depletion, we observed the down-regulation of FER1 and ACO2 in leaves at (7d)+3 and (7d)+7. ACO1 and ACO3 turned out not to be reliable markers for Fe deficiency at most of the time-points (data not shown). In Cd-treated leaves, FER1 and ACO2 were also repressed (Fig. 3b). Those results tend to indicate that leaves treated with Cd were subjected to an induced Fe deficiency. However, application of an Fe spray to rosette leaves exposed to Cd only partially alleviated chlorotic symptoms, while it almost completely reversed these symptoms in Fe-deficient plants (Fig. S4).


In this study, as demonstrated in Fig. 1, we found a clear protective effect of a low Mg concentration in the plant against the bleaching of young leaves typically caused by Cd exposure (Collin et al., 2008; Morel et al., 2009). No impact of Mg starvation on Cd translocation from the root to the shoot was found (Fig. 2b). This observation is in contrast to the finding that Cd toxicity is alleviated by a high supply of calcium in Arabidopsis (Suzuki, 2005) or of Mg in Brassica (Kashem & Kawai, 2007), where Cd uptake and accumulation have been reported to be reduced. However, in our case, the cellular localization of Cd could have been different and/or protection mechanisms against this heavy metal may have been enhanced by the low-Mg conditions. In recent genome-wide transcriptomic studies of the response to Mg starvation in Arabidopsis (Hermans et al., 2010a,b), we identified several Mg-responsive genes involved in the response to oxidative stress, the photoprotection of the photosynthetic apparatus, the detoxification process for drugs and heavy metals and the homeostasis of other essential elements (Table S1). We discuss here the possibilities that Mg deficiency (1) increases the antioxidative capacity, (2) protects chloroplasts against Cd-induced injuries, (3) alters Fe homeostasis to prevent Cd toxicity and (4) favours cytosolic efflux transport or vacuolar storage of Cd.

Low Mg status may enhance defence mechanisms against oxidative stress and protection of the photosynthetic apparatus against Cd

Previously, we identified Mg-responsive genes that potentially detoxify ROS (Table S1, Fig. 4) and higher oxidation states of the leaf ascorbate and glutathione pools (Hermans et al., 2010b). Mg deficiency is known to increase antioxidative capacity (Cakmak & Marschner, 1992; Tewari et al., 2006; Cakmak & Kirkby, 2008) and Cd exposure to induce oxidative stress (Cho & Seo, 2005; Smeets et al., 2008; Bi et al., 2009; Rodríguez-Serrano et al., 2009; Sandalio et al., 2009). Therefore, defence mechanisms against ROS induced upon Cd exposure seem to be activated by an Mg starvation pretreatment. Transcriptomic data also indicate an early protection of the photosynthetic apparatus and the enhancement of mechanisms preventing the accumulation of free chlorophyll (Hermans et al., 2010b). The up-regulation was detected of REDOX RESPONSIVE TRANSCRIPTION FACTOR 1 (RRTF1), which encodes an AP2 transcription factor involved in the control of redox homeostasis upon photosystem II inhibition (Khandelwal et al., 2008), and EARLY LIGHT-INDUCED PROTEINS 1 and 2 (ELIP1;2), which encode proteins fulfilling a photoprotective function (Hutin et al., 2003) or working as sensors to prevent accumulation of free chlorophyll (Tzvetkova-Chevolleau et al., 2007) (Fig. 4, Table S1).

Figure 4.

Potential targets of magnesium (Mg) deficiency that can alleviate the toxic effects of cadmium (Cd) in the leaf cell. (i) Higher expression of genes in the ‘oxygen and radical detoxification’ MIPS category such as GLUTAREDOXIN (GRX) and GLUTATHIONE S-TRANSFERASE TAU (GSTU) and a higher ratio of oxidized ascorbate and gluthatione forms are indicative of increased antioxidative capacity. (ii) Photoprotection of the pigment–protein complexes is supported by higher expression of EARLY LIGHT-INDUCED PROTEINS 1 and 2 (ELIP1;2) and REDOX RESPONSIVE TRANSCRIPTION FACTOR 1 (RRTF1). (iii) Higher iron (Fe) concentrations in shoots could overcome the Fe shortage caused by Cd and the induction of FERRIC REDUCTION OXIDASE 1 (FER1) could result in additional complexation of Cd by ferritin. (iv) Higher vacuolar sequestration and efflux of Cd out of the cytosol are supported by higher expression of MULTIDRUG RESISTANCE PROTEIN 3 (MRP3) and DETOXIFICATION 1 (DTX1), 2, 3 and 4, respectively.

Low Mg status may indirectly alter mineral homeostasis to prevent Cd toxicity

An antagonistic interaction between Mg and Fe is widely recognized to occur in plants (Agarwala & Mehrotra, 1984; Ward et al., 2008). Particularly in Arabidopsis, Mg deficiency promotes a slight increase in the Fe content in leaves (Fig. 2c; Hermans et al., 2010b) and elevated Fe concentrations have been reported to prevent the harmful effects of Cd (Solti et al., 2008). Before Cd exposure at (d7)+0, we already observed a higher Fe concentration in Mg-deficient leaves and higher expression of FER1 (Fig. 4, Table S1), which encodes a well-characterized chloroplastic Fe-storage protein that accumulates under conditions of Fe excess (Murgia et al., 2007). The major physiological role of ferritins in plants is to fine-tune the quantity of metal available to meet metabolic requirements by sequestering Fe (Briat et al., 2009). In addition to Fe detoxification, another function of ferritins may be to prevent oxidative stress (Ravet et al., 2009) and to detoxify heavy metals, as these proteins are also capable of binding Cd2+ (Sczekan & Joshi, 1989; Rama Kumar & Prasad, 1999). Therefore, ferritins could also play an important role in controlling the interaction between Fe homeostasis, Cd excess and oxidative stress generated by Cd.

An influence of modified Fe and Cd supply on the uptake and distribution of both metals and photosynthesis has been reported (Siedlecka & Krupa, 1999; Sávári et al., 2008). Upon mild Cd concentration exposure, we did not find differences in Fe concentration (Fig. 2c) and in the expression of the Fe transporter IRT1 (Fig. 3a) in roots of the +Mg+Cd treatment in comparison with the control. This is in contrast to the findings of other studies which have reported a transient induction of IRT1 expression at higher Cd concentrations in different species (Hodoshima et al., 2007; Besson-Bard et al., 2009). Also, we found here that chloroses on Cd-treated leaves were not directly related to Fe deficiency but may be in part attributable to interactions with other elements such as Mn (Figs 2d, S2) or other types of Cd-induced damage. Overall, an Fe foliar spray had a positive influence on the chlorophyll content and photosynthetic activity of Cd-treated plants (Fig. S4). Incomplete restoration of Cd-induced chloroses by application of the Fe foliar spray may be a result of the direct inhibitory effect of Cd on the transcription of genes encoding apoLHC proteins (causing incomplete recovery of the light-harvesting complex of photosystem II), as previously suggested by Solti et al. (2008).

Finally, we note that the Mg deficiency pretreatment resulted in other ionic profile variations (Figs 2, S2); however, it is difficult to ascertain whether such variations account for the protective effect against Cd toxicity. Reference mineral concentrations in tissues, below which there is deficiency and above which there is toxicity, are available for some land and crop species (Bennett, 1993; Marschner, 1995; Krämer, 2010) but are not necessarily relevant for Arabidopsis. In addition, these thresholds are dependent on the stage of development, the mode of culture and other environmental factors.

Low Mg status may enhance Cd detoxification processes

We observed the induction of four DETOXIFICATION family members (DTX1, 2, 3 and 4) (Table S1). In Arabidopsis, the DTX family consists of 56 members and belongs to the multidrug and toxic efflux (MATE) superfamily. To date, these transporters have been poorly characterized; nevertheless, some members have been shown to transport ethidium bromide and secondary metabolites (Li et al., 2002). Interestingly, one DTX member could be involved in Cd efflux out of the cell (Li et al., 2002). On the basis of these observations, it is possible that DTXs mediate at least in part the beneficial effect of low Mg status against Cd exposure by lowering the cytosolic Cd concentration (Fig. 4). In order to evaluate the impact of DTX activity on the Cd protection phenotype, we challenged dtx2 and dtx3 knockout mutants with Mg starvation followed by Cd treatment. However, no change in the phenotype compared with the wild type was observed (data not shown). In addition, we also observed in Mg-deficient leaves a strong induction of MULTIDRUG RESISTANCE PROTEIN 3 (MRP3) (Table S1, Fig. 4), which encodes a vacuolar ABC transporter that transports glutathione conjugates and chlorophyll catabolites (Tommasini et al., 1998) and has a possible role in toxic metal detoxification (Zientara et al., 2009).

Several knockout mutants of Mg deficiency-responsive genes (dtx and rrtf1) were challenged with a Cd excess, but the protective effect of Mg deficiency persisted in these mutant backgrounds. At this stage we cannot exclude the possibility that the protective effect of low Mg status is polygenic. To dissect the Cd–Mg interaction, further studies are underway. In conclusion, the low Mg status protective effect against Cd is pleiotropic and could be caused by modification of Fe, Mn, Zn and S leaf tissue concentrations, enhancement of antioxidative capacity, detoxification and/or protection of the photosynthetic apparatus.


This work was supported by a grant from the Interuniversity Attraction Poles Programme (IUAP VI/33), initiated by the Belgian State, Science Policy Office, FRFC (no. 2.4.583.08F), Crédit aux Chercheurs (no. from the Fonds National de la Recherche Scientifique (FNRS-FRS) and Fondation Jaumotte-Demoulin at ULB. C.H. is a postdoctoral fellow of the FNRS. We thank C.L. Meyer for statistical advice and P. Salis for technical assistance.


While our paper was in press, Chou et al. (2011) published that Mg deficiency also protected rice seedlings from Cd stress. However these authors reported higher Cd concentrations in Mg-deficient plants compared with Mg-sufficient plants upon Cd exposure.