is brassica juncea a suitable plant for phytoremediation of cadmium in soils with moderately low...

Soil Science and Plant Nutrition (2006) 52, 32–42 doi: 10.1111/j.1747-0765.2006.00008.x

© 2006 Japanese Society of Soil Science and Plant Nutrition

Blackwell Publishing LtdORIGINAL ARTICLEPhytoremediation of Cd-contaminated SoilsS. ISHIKAWA et al.

Is Brassica juncea a suitable plant for phytoremediation of cadmium in soils with moderately low cadmium contamination? – Possibility of using other plant species for Cd-phytoextraction

Satoru ISHIKAWA1, Noriharu AE2, Masaharu MURAKAMI1 and Tadao WAGATSUMA3

1Department of Environmental Chemistry, Heavy Metal Group, Soil Biochemistry Unit, National Institute for Agro-Environmental Sciences, Ibaraki 305-8604, 2Department of Biological and Environmental Sciences, Faculty of Agriculture, Kobe University, Hyogo 657-8501 and 3Laboratory of Plant Nutrition and Soil Science, Faculty of Agriculture, Yamagata University, Yamagata 997-8555, Japan

Abstract

We evaluated the ability of Brassica juncea (L.), which has already been recognized as a plant suitable formetal phytoremediation, and of several other plant species (maize, rice and sugar beet) to extract cadmium(Cd) from soils with moderately low levels of Cd contamination. Two of the 56 cultivars of B. juncea werepreliminarily screened as high-Cd accumulators using a hydroponic culture solution containing a high levelof external Cd (1 mg L−1). Thereafter, 7 cultivars within 4 plant species (maize, B. juncea [2 cultivars], rice[3 cultivars with different subspecies] and sugar beet) were grown in a hydroponic culture solution containinga low Cd level (0.05 mg Cd L−1) or in pots filled with 2 types of contaminated soils containing moderatelylow Cd levels under upland conditions. The 2 soils consisted of a Fluvisol and an Andosol and contained1.82 and 4.01 mg Cd kg−1 on a dry soil weight basis, respectively, determined using 0.1 mol L−1 HCl-extraction. The results indicated that B. juncea was less able to accumulate Cd in shoots compared withhydroponically cultured rice and sugar beet, and was even less effective when grown in soil culture. Riceand sugar beet displayed a higher accumulation not only of Cd but also of other heavy metals (Cu, Fe, Mnand Zn) in their shoots than B. juncea when they were grown in the two Cd-contaminated soils. Maizedisplayed the lowest metal accumulation among the plant species tested. Growing the rice cultivars in bothsoil types led to the most significant decrease in soil Cd concentration determined using extraction with0.1 mol L−1 HCl. In contrast, we did not observe any significant decrease in soil Cd concentration inB. juncea. Sequential Cd extraction of soil revealed that rice was more effective than B. juncea in phytoex-tracting Cd from less-soluble fractions in soils. Based on the plant and soil analyses, it was suggested thatB. juncea does not offer much promise for phytoextraction of Cd from soils with relatively low contamina-tion, and that rice may be an eligible plant for metal phytoremediation of such soils.

Key words: cadmium, phytoremediation, rice, sequential Cd extraction, upland conditions.

INTRODUCTION

Japanese arable lands, particularly paddy fields, arecontaminated to some extent with moderately lowlevels of cadmium (Cd) because of irrigation with river

water passing through mines or because of emissionsfrom smelters. Several techniques for addressing thisproblem have been developed. Soil dressing, achievedby removing the Cd-polluted soil layer and replacing itwith non-polluted soil, has been the most reliable tech-nique to reduce Cd contamination in rice. However,remediation of large areas of soil with low Cd contami-nation using soil dressing is not practical because of thehigh cost ($300,000–$500,000 per ha) and the shortageof unpolluted replacement soil.

Phytoremediation, the use of plant systems to restorecontaminated environments, has attracted a great deal

Correspondence: Dr S. ISHIKAWA, Department ofEnvironmental Chemistry, Heavy Metal Group, SoilBiochemistry Unit, National Institute for Agro-EnvironmentalSciences, 3-1-3 Kannondai, Tsukuba, Ibaraki 305-8604,Japan. Email: [email protected] 23 June 2005.Accepted for publication 19 October 2005.

Phytoremediation of Cd-contaminated soils 33


of attention recently as an alternative to soil dressing.Phytoremediation is a highly valuable method becauseit is: (1) cost-effective, (2) applicable to large areas, (3)environment-friendly because it does not dramaticallydisturb the landscape. The most important factor in thedevelopment of successful phytoremediation technologyfor Cd-polluted soils is the selection of promisingplants. In many reports, emphasis has been placed onthe use of hyperaccumulator plants such as Thlaspicaerulescens (alpine pennycress). This plant hasattracted much attention as a promising plant for phy-toremediation because it displays favorable characteris-tics, namely the ability to take up and translocateextremely large amounts of Cd to shoots, and a hyper-tolerance to Cd (Baker et al. 1994; Brown et al. 1994).However, this species has come under criticism becauseof its low shoot biomass production and slow growth(Ebbs et al. 1997). It is essential that cleanup plantsaccumulate a high Cd concentration in their shoots andare characterized by high shoot biomass production.

Brassica juncea has also been reported to be a prom-ising plant for metal phytoremediation. It comparesfavorably with T. caerulescens in Zn and Cd removalfrom contaminated soil because of its moderate to highZn and Cd accumulation and high shoot biomass (Ebbset al. 1997). Researchers who have investigatedB. juncea genotypes have found wide variations inshoot Cd uptake ranging from 200 to 1200 mg Cd kg−1

on a shoot dry weight basis, when the plants werehydroponically exposed to 25 µmol L−1 Cd (Li et al.2004). Brassica juncea displays a high tolerance to Cdbecause of sequestration with heavy metal-binding pep-tides referred to as phytochelates (Zhu et al. 1999).Therefore, high Cd accumulation in B. juncea mayresult in relatively normal shoot growth under Cd expo-sure. Most studies reporting heavy metal accumulationof B. juncea have been carried out using water or soilculture with a relatively high Cd level (Ebbs et al. 1997;Li et al. 2004). For example, Ebbs et al. (1997) evalu-ated the phytoextraction potential of B. juncea bygrowing plants in Cd-contaminated soil that containedtotal Cd (digestion in nitric acid/hydrogen peroxide) andavailable Cd (1 mol L−1 ammonium acetate-extraction)at concentrations of 40 mg kg−1 and 16 mg kg−1 dry soil,respectively. Our targeted areas for phytoremediationwere arable lands that might potentially produce paddyrice or upland crops contaminated with Cd. Speculatingfrom soil Cd concentrations in several fields in whichpaddy rice contaminated with Cd have been produced,the soil Cd concentration for phytoremediation rangefrom at most 1 to 5 mg kg–1 (extraction with 0.01 m HCl),depending on the soil properties, such as soil type, soilpH, carbon content and so on. Under such soil condi-tions, it remains to be determined whether B. juncea

displays a high ability for Cd phytoextraction. In addi-tion, there are very few studies on the selection of plantspecies that could be used for phytoremediation of soilswith relatively low levels of Cd, such as those found inJapanese paddy fields (Kurihara et al. 2005).

Our study consisted of three components. In the firstpart, we identified 2 cultivars of B. juncea that werehigh accumulators of Cd after screening them hydro-ponically at a relatively high Cd level (1 mg Cd L−1).We determined whether they would display a high Cdaccumulation in shoots when exposed to a lower Cdconcentration (0.05 mg Cd L−1): in which case theplants would not exhibit any phytotoxic effects. Wecharacterized the Cd accumulation of the 2 cultivars ofB. juncea by comparing them with several other plantspecies. We used maize, rice (3 cultivars with differentsubspecies) and sugar beet, all of which also exhibithigh shoot biomass production and are cultivated inJapanese arable lands. We grew these plant species inpots to investigate their potential to extract Cd from 2different types of soils containing relatively low levels ofCd. We applied a sequential extraction method to theCd-contaminated soils in each pot to identify therespective bioavailable Cd fractions in the soils for eachplant species tested. Based on the data obtained, weevaluated the potential of B. juncea and other plant spe-cies for phytoremediation of Japanese arable lands con-taminated with relatively low levels of Cd.

MATERIALS AND METHODS

Plant materialsIn the preliminary experiment, we grew 56 cultivars ofB. juncea in a hydroponic culture solution containing1 mg Cd L−1 for 20 days. From these, we screened 2 cul-tivars (Daulal and 6-2825) that displayed a high Cdconcentration in their shoots and roots (approximately200 mg kg−1 and 1600 mg kg−1, respectively). We con-firmed that the selected 2 cultivars showed a constantlyhigher shoot Cd concentration than several other culti-vars when treated with 4 levels of lower Cd concentra-tions (0.1, 0.2, 0.3 and 0.5 mg L−1). Sugar beet (Betavulgaris L. var. saccharifera Alef., cv. Sugarman gold)was also selected as a high-Cd-accumulating plant usingthe same preliminary screening in hydroponic culture.In our previous studies, we found that several paddy rice(Oryza sativa L.) cultivars accumulated high Cd levelsin grains and shoots when cultivated in Cd-pollutedsoils under upland conditions (Arao and Ae 2003). Threerice cultivars, Koshihikari (Japonica), Kasalath (Indica)and Milyang 23 (Japonica-Indica hybrid) with differentsubspecies, were used in this study. Maize (Zea mays L.,cv. Gold dent) was also tested because of its fast-growing habit and high shoot biomass production.

34 S. Ishkawa et al.


To investigate the extraction potential of Cd andother heavy metals (Cu, Fe, Mn and Zn) by 7 cultivarswithin 4 plant species, we grew the cultivars either in hydro-ponic or soil culture as described below. The plantswere grown in a greenhouse at ambient temperature(18–30°C) under sunlight in both types of cultures.

Hydroponic culture experimentSeeds of B. juncea, sugar beet, rice and maize were sownin artificially prepared fertilized peaty soil Supermix-A(Sakata Seed Co. Ltd., Yokohama, Japan). The Cdconcentration in Supermix-A was negligible. For fourseedlings of each cultivar that had survived for 10–15 days after sowing, the artificial soil was washed andthe seedlings were carefully transferred to 40-L vesselshalf-full of the following nutrient solution (mg L−1): N,20 (NH4NO3); K, 30 (K2SO4); Ca, 40 (CaCl2·2H2O);Mg, 20 (MgSO4·7H2O); Fe, 2 (Fe[III]EDTA); Mn, 1(MnSO4·7H2O); B, 0.2 (H3BO3); Zn, 0.1 (ZnCl2); Cu,0.01 (CuSO4·5H2O); and Mo, 0.005 (NaMoO4·2H2O).The solution was continually aerated, replaced with anew solution as often as once per week, and the pH wasadjusted to 5.5 every day. Ten days after the transfer,we exposed the seedlings to a full nutrient solution con-taining 0.05 mg Cd L−1 in the form of CdSO4. The Cdconcentration of the treatment solution was selected toensure that it would not cause phytotoxic effects in anyplants. After 1 week, 4 seedlings of each cultivar weretaken from the vessels, washed thoroughly in distilledwater, separated into roots and shoots, and dried in anoven at 60°C.

Soil culture experimentTwo types of Cd-polluted soils were used in the presentstudy: a Fluvisol and an Andosol. The Fluvisol was col-lected from the upper layer (0–15 cm depth) of soil in apaddy field, which was contaminated with Cd from irri-gation with river water passing through mines. TheAndosol was collected from the upper 0–15 cm layer ofsoil in a mulberry field that was polluted by Cd and alsoZn through emission from zinc smelters. Both soil typeswere air-dried and sieved (2-mM mesh). Plastic pots werefilled with 500 g of each kind of Cd-polluted soil.We applied fertilizers in the form of ammonium sulfate,single superphosphate and potassium sulfate, at rates of0.1 g N, 0.1 g P2O5 and 0.1 g K2O per pot, respectively.The seeds of B. juncea, sugar beet, rice and maize weredirectly sown in each pot filled with the Cd-polluted soils.

After emergence, the seedlings were thinned to 3 per potfor B. juncea and sugar beet, and 2 per pot for maize.The rice seedlings were thinned to 2 rice stubbles (3seedlings per stubble). The soil was watered daily withtap water at 60% of field capacity. Thirty days aftersowing, the plants were harvested by cutting the stems

approximately 2 cm above the soil level. The B. junceacultivars flowered because they were early-ripeningcultivars. Shoot samples were rinsed with distilled waterto remove soil, heated in an oven at 60°C until theybecame dry, and then weighed. Soil samples were col-lected from each pot, air-dried and sieved (2-mm mesh).The roots in the soils were removed for metal analysisof soils. Each treatment of the soils was applied using arandomized block design in three replicates.

Heavy metal analyses of plants and soilsPlant samples were ground to a fine powder using astainless steel grinder (P-14, Fritsch, Kastl, Germany).The powder (0.5 g) was digested with 10 mL of aHNO3–HClO4–H2SO4 (5:1:1, v/v) mixed solution usingthe digest system (K-437, BÜCHI, Flawil, Switzerland).

We measured soil pH with a glass electrode (ModelpH81, Yokogawa Co, Ltd., Tokyo, Japan) after extrac-tion with deionized water (1:5, w/v). The concentra-tions of heavy metals in soils were evaluated based onconcentrations of 0.1 mol L−1 HCl-extractable Cd, Cu,Fe, Mn and Zn. Fractionation of Cd in soils was carriedout following the sequential method of Sadamoto et al.(1994). First, exchangeable Cd in soil was extracted byshaking with 0.05 mol L−1 Ca(NO3)2 (1:10, w/v). Next,inorganically bound Cd was extracted with 2.5% aceticacid (1:10, w/v), and then organically bound Cd wasextracted with 2.5% acetic acid (1:10, w/v) afterdecomposition of organic matter with 6% H2O2. Finally,the oxide-occluded Cd fraction in soil was extracted byheating with a mixed solution of acid ammonium oxalateand ascorbic acid (1:30, w/v).

The concentrations of Cd, Cu, Fe, Mn and Zn in theplant tissues and soils were determined using induc-tively coupled plasma optical emission spectroscopy(ICP-OES, Vista-Pro, Varian, Mulgrave, Australia).

RESULTS

Uptake and distribution of cadmium in plants grown in hydroponic cultureNo plants showed any visible symptoms of Cd excesswhen treated with 0.05 mg Cd L−1 in hydroponic culturefor 1 week (data not shown). The shoot Cd concentra-tion was lowest in maize (2.8 mg kg−1 of shoot dryweight), followed by B. juncea (13.7 mg kg−1 and17.2 mg kg−1 for Daulal and 6-2825, respectively), rice(19.6 mg kg−1, 25.2 mg kg−1 and 18.6 mg kg−1 forKoshihikari, Kasalath and Milyang 23, respectively),and sugar beet (36.3 mg kg−1) (Fig. 1a).

Both cultivars of B. juncea showed the lowest rootCd concentration: less than half the root Cd concentra-tion of maize (Fig. 1b). The roots of the rice cultivars and



of sugar beet accumulated Cd at approximately 150–250 mg kg−1 (root dry weight basis). The Cd concentrationof the B. juncea cultivar, 6-2825, was slightly higher inboth shoots and roots than that of Daulal. Among therice cultivars, Kasalath accumulated the highest Cd con-centration in shoots compared with the other cultivars.The order of Cd concentration in the roots was as follows:Koshihikari > Kasalath > Milyang 23.

The Cd-distribution ratio in each tissue was calcu-lated based on the proportion of Cd in each type of tissueto the total amount of Cd uptake (Fig. 1c). Maize rootsaccounted for more than 90% of the total Cd uptake.

The highest Cd ratio in shoots was observed in B. juncea;the shoot ratio accounted for approximately 70% of thetotal Cd uptake. Cadmium uptake in the shoots of riceand sugar beet was 30% and 40% of the total Cduptake, respectively. There were no conspicuous differ-ences in the Cd distribution ratio between the 2 culti-vars of B. juncea or among the 3 rice cultivars.

Dry weight, cadmium concentration and content of shoots grown in the 2 types of cadmium-polluted soilsNo plants displayed phytotoxic effects when grown inthe two Cd-contaminated soils for 30 days (data notshown). The shoot dry weight of plants grown inFluvisol was highest for maize and for the rice cultivarKasalath (Fig. 2a). Among the Fluvisol-grown plants,the shoot dry weight of sugar beet was the lowest, prob-ably because shoot growth was inhibited by a relativelylow soil pH (5.1). Except for the rice cultivar, Kasalath,all the plants showed a similar shoot dry weight whengrown in Andosol (Fig. 2d). Shoot Cd concentrations inmaize were less than 1 mg kg−1 dry weight in both soiltypes (Fig. 2b,e). The average value of the shoot Cdconcentration for the 2 cultivars of B. juncea wasapproximately half that for the 3 rice cultivars thatwere grown in Fluvisol (Fig. 2b). These differences weremore significant in the Andosol-grown plants (Fig. 2e):the average value of the shoot Cd concentration wasapproximately fourfold higher for the rice cultivarsthan for the B. juncea cultivars. Among the plantspecies tested, sugar beet showed the highest shoot Cdconcentrations in both types of soils (Fig. 2b,e).

A similar pattern was observed for the shoot Cd content(Fig. 2c,f). Maize showed the lowest Cd content amongthe shoots grown in the 2 soil types. The average valueof the shoot Cd content for the 2 cultivars of B. junceawas much lower than that for the 3 rice cultivars and forsugar beet, irrespective of soil type. Except for Milyang23, the Cd contents of the shoots of the rice cultivars werealmost equal to those of sugar beet. The differences in theconcentrations and contents of Cd in the shoots weremuch smaller between the 2 cultivars of B. juncea andamong the 3 rice cultivars than among the 4 plant species.

Uptake of other heavy metals (Cu, Fe, Mn and Zn) by plants grown in the 2 types of cadmium-polluted soilsFluvisol and Andosol showed concentrations (mg kg−1

dry soil weight) of 10.9 and 6.1 for Cu, 248 and 6.0 forFe, 46.1 and 65.6 for Mn, and 16.4 and 146 for Zn,respectively, when extracted with 0.1 mol L−1 HCl.

As in the case of the shoot Cd concentrations, theconcentrations of other heavy metals (Cu, Fe, Mn and Zn)

Figure 1 (a) Shoot cadmium (Cd) concentration, (b) root Cdconcentration and (c) Cd distribution of shoots and roots for 7cultivars of 4 plant species grown hydroponically in a nutrientsolution containing 0.05 mg Cd L−1 for 1 week. Bars denotemeans and standard errors of four replicates.



in shoots varied widely among the plant species (Fig. 3).The lowest concentrations for all the heavy metals werefound in the shoots of maize grown in both soil types,whereas sugar beet showed the highest concentration.For plants grown in Fluvisol, shoots of the 3 rice culti-

vars exhibited higher concentrations of all the metals,except for Zn, than shoots of the 2 cultivars of B. juncea.

As for the heavy-metal content per pot among theshoots of the plants grown in Fluvisol, the 3 rice cultivarsregistered the highest contents of all metals, except for

Figure 2 (a,d) Shoot dry weight, (b,e) shoot cadmium (Cd) concentration and (c,f) shoot Cd content for 7 cultivars of 4 plantspecies grown in a Fluvisol (a,b,c) and an Andosol (d,e,f) for 30 days after sowing. The concentrations of 0.1 mol L−1 HCl-extractable Cd for plants grown in Fluvisol and Andosol are 1.82 and 4.01 mg kg−1 (dry soil weight basis), respectively.Bars denote means and standard errors of three replicates.



Figure 3 (a,e) Concentrations of Cu, (b,f) Fe, (c,g) Mn and (d,h) Zn in shoots of 7 cultivars of 4 plant species grown in a Fluvisol(a,b,c,d) and an Andosol (e,f,g,h) for 30 days after sowing. Bars denote means and standard errors of three replicates. Theconcentrations of 0.1 mol L−1 HCl-extractable Cu, Fe, Mn and Zn for plants grown in Fluvisol and Andosol, respectively, are Cu,10.9 and 6.1; Fe, 248 and 6.0; Mn, 46.1 and 65.6; and Zn, 16.4 and 146 (mg kg−1 dry soil weight).



Zn (Fig. 4a–d). The 2 cultivars of B. juncea were com-parable to the 3 rice cultivars in shoot Zn content (Fig. 4d).The shoots of sugar beet accumulated the highest amountof Zn among the plant species tested. As for the shoots ofthe plants grown in Andosol, sugar beet showed the highest

amounts of Mn and Zn (Fig. 4g,h). The rice cultivar Kasalathaccumulated the highest amounts of Cu and Fe in shoots(Fig. 4e,f), presumably because of its high shoot dry weight.

These results revealed that the accumulation of all theheavy metals tested, including Cd, was higher in the

Figure 4 (a,e) Amounts of Cu, (b,f) Fe, (c,g) Mn and (d,h) Zn in shoots of 7 cultivars of 4 plant species grown in a Fluvisol(a,b,c,d) and an Andosol (e,f,g,h) for 30 days after sowing. Bars denote means and standard errors of three replicates.



shoots of rice and sugar beet than in those of maize andB. juncea. Differences in the concentrations and amountsof heavy metals were much smaller among the shoots ofthe rice and B. juncea cultivars than among those of theother plant species. During the 30-day-growing stage,rice and sugar beet were found to exhibit a superiorability for the uptake of Cd and the other heavy metalscompared with B. juncea.

Changes in pH and cadmium concentration of each fraction in the 2 soil types throughout plant growthCultivation of maize in both Fluvisol and Andosolresulted in a higher soil pH compared with unplantedsoil (control) (Tables 1,2). In contrast, the pH of Fluvi-sol decreased after cultivation with rice and sugar beet.No changes in soil pH occurred after the growth of the2 cultivars of B. juncea, regardless of soil type.

The Cd concentrations in the control soils extractedwith 0.1 mol L−1 HCl were 1.82 and 4.01 mg kg−1 on adry soil weight basis for Fluvisol and Andosol, respec-tively. There were differences in the Cd concentration ineach of the fractions between the two soil types.Exchangeable Cd and inorganically bound Cd were thepredominant fractions in Fluvisol. In contrast, Andosolcontained a much higher level of Cd in the form of inor-ganically and organically bound Cd fractions than ofexchangeable Cd.

After cultivation of rice and sugar beet in Fluvisol,there was a significant decrease in 0.1 mol L−1 HCl-extractable Cd concentration (Table 1). A similar pat-tern was also observed among the Andosol-grown ricecultivars, although it was not statistically significant(Table 2). No significant decrease in soil Cd concentra-tion was found in either soil after the growth of maizeand B. juncea.

Table 1 Post-growth pH and cadmium (Cd) concentration in each soil fraction for plants grown in a Fluvisol

Plant species CultivarsSoil pH (H2O)

0.1 mol L−1 HCl-extractable Cd

(mg kg−1 dry soil wt)

Fractionation of Cd in soils (mg kg−1 dry soil weight)

Exchangeable Inorganically bound

Organically bound

Occluded intosesquioxides

None (Control) 5.10ab 1.82a 0.92a 0.95a 0.66a 0.84a

Maize Gold dent 5.27a 1.81a 0.74b 0.77b 0.51b 0.81ab

B. juncea Daulal 5.11ab 1.83a 0.80b 0.76b 0.51b 0.84a

6-2825 5.11ab 1.60a 0.74b 0.64b 0.43bc 0.62ab

Rice Koshihikari 4.88b 1.17b 0.56c 0.44c 0.33d 0.66ab

Kasalath 4.94b 1.15b 0.56c 0.43c 0.30d 0.57b

Milyang 23 4.84b 1.22b 0.60c 0.46c 0.32d 0.71ab

Sugar beet Sugarman gold 4.85b 1.20b 0.60c 0.44c 0.35cd 0.67ab

Differences among cultivars were compared usng Bonferroni’s multiple comparison test (P < 0.05). Values with different letters are significantly different.

Table 2 Post-growth pH and cadmium (Cd) concentration in each soil fraction for plants grown in an Andosol

0.1 mol L−1 HCl-extractable Cd

(mg kg−1 dry soil wt)

Fractionation of Cd in soils (mg kg−1 dry soil weight)

Plant species CultivarsSoil pH (H2O)

Exchangeable Inorganically bound

Organically bound

Occluded into sesquioxides

None (control) 5.72bc 4.01ab 0.48ab 2.74a 2.06a 2.44a

Maize Gold dent 5.86a 4.01ab 0.40bc 2.56ab 2.04a 2.59a

B. juncea Daulal 5.78ab 3.97ab 0.44abc 2.60ab 2.10a 2.28a

6-2825 5.72bc 4.17a 0.50a 2.66ab 2.12a 2.02a

Rice Koshihikari 5.72bc 3.90ab 0.39c 2.52b 2.08a 2.49a

Kasalath 5.73bc 3.69b 0.41bc 2.47b 1.97a 2.09a

Milyang 23 5.75bc 3.70b 0.39c 2.50b 2.01a 2.45a

Sugar beet Sugarman gold 5.68c 3.99ab 0.46abc 2.55ab 2.10a 2.45a

Differences among cultivars were compared with Bonferroni’s multiple comparison test (P < 0.05). Values with different letters are significantly different.



Among the Fluvisol-grown plants (Table 1), the Cdconcentrations of all the fractions, except for thoseoccluded into sesquioxides, decreased significantly fol-lowing the growth of all plant species used. In particu-lar, rice and sugar beet caused a significant decrease inCd concentrations of the 3 soil fractions: exchangeableCd, inorganically bound Cd and organically bound Cd.In Andosol, the Cd concentrations of 2 soil fractions,exchangeable Cd and inorganically bound Cd,decreased significantly after the cultivation of rice(Table 2). Cadmium concentrations of the organicallybound fractions appeared to decrease after the growthof the 2 rice cultivars, Kasalath and Milyang 23, inAndosol, although this decrease was not statisticallysignificant. There was no significant decrease in the Cdconcentrations of the fractions occluded into sesquiox-ides in either soil type, except for the Kasalath cultivarin Fluvisol. For both soil types, no significant differ-ences in Cd concentrations in any soil fractions wereobserved between the 2 cultivars of B. juncea or amongthe 3 rice cultivars.

DISCUSSION

Brassica juncea is effective in phytoextracting Cd andZn from soils with relatively high Cd contamination(Ebbs et al. 1997). In our preliminary hydroponic cul-ture experiment with high levels of Cd (1 mg L−1), weconfirmed that the B. juncea cultivars used in thepresent study could accumulate a high level of Cd(approximately 200 mg kg−1 on a shoot dry weightbasis) (data not shown). However, in the case of minorCd contamination, the data presented here suggest thatB. juncea is less effective than rice or sugar beet in phy-toextracting Cd in hydroponic culture (Fig. 1), and evenmore so in soil culture (Fig. 2). Similar results werereported by Kurihara et al. (2005), who observed thatthe Cd uptake was much lower in B. juncea than inkenaf (Hibiscus cannabinus) when the plants weregrown in soil with a relatively low Cd level (1 mg kg−1

dry soil, 0.1 mol L−1 HCl-extraction) in pot-scale experi-ments. Plant Cd accumulation is controlled by severalfactors, including root Cd uptake, shoot-to-root trans-location, Cd tolerance and utilization of Cd in soil.Cadmium translocated to the shoots in B. junceaaccounted for 70% of its total uptake in hydroponicculture (Fig. 1c). A similar high Cd ratio in shoot con-centrations was found in B. juncea grown in soils artifi-cially contaminated with Cd (Yanai et al. 2004).Although rice and sugar beet primarily accumulated Cdin their roots (Fig. 1b,c), a higher shoot Cd accumula-tion was found in rice and sugar beet than in B. junceawhen the plants were grown in both soil types (Fig. 2).In addition, no plants showed phytotoxic effects (data

not shown), suggesting that Cd tolerance may not berequired for plants to ameliorate soils that are slightlycontaminated with Cd. These results suggested thataccess to the Cd pool in the soil and subsequent Cduptake by roots could be a more important factor forphytoremediation of soils with low Cd contamination.

When the concentrations of 0.1 mol L−1 HCl-extractable Cd and each soil fraction were comparedbetween unplanted soils (controls), only the exchangeableCd fractions were higher in Fluvisol than in Andosol(Tables 1,2). The shoot-Cd accumulation of all plantspecies was also higher in Fluvisol than Andosol(Fig. 2). These results suggest that bioavailable Cdin soil is mainly present in exchangeable fractions.A significant decrease in the Cd concentrations of theexchangeable fractions was observed following thegrowth of all plant species in Fluvisol, and to a lesserextent in Andosol. Among the Fluvisol fractions, riceand sugar beet were more effective than B. juncea ormaize in phytoextracting Cd, not only from theexchangeable fractions but also from the less-solublefractions (inorganically and organically bound frac-tions). Rice was most effective in phytoextracting Cdfrom these 3 fractions in Andosol as well. These resultssuggest that rice exhibits a superior ability for mobiliz-ing and absorbing Cd from less-soluble fractions ofsoils under upland conditions, compared with B. junceaor any other plant species tested in the present study.Researchers have proposed several hypotheses relatedto metal mobilization and absorption by plants toaccount for this ability, including root exudates (Cak-mak et al. 1996), rhizosphere acidification (Römheldand Marschner 1986) and enhancement of root growthin contaminated soil (Whiting et al. 2000). Among theplants tested, only Fluvisol-grown rice and sugar beetdecreased the soil pH compared with the control,although the decrease was not statistically significant(Table 1). A decrease in soil pH may lead to the releaseof free ionic Cd2+ into the rhizosphere (Salam andHelmke 1998). Using the rhizobox method, Goto et al.(2003) and Yanai et al. (2004) observed a significantpH decrease in the rhizosphere of B. juncea (within1 mm or 4 mm from the root-growing zone compart-ment, respectively) when the plants were grown in Cd-contaminated Andosols. In the present study, the pH ofthe bulk soils and not merely of the rhizosphere wasmeasured. Further studies should be carried out todetermine whether rhizosphere acidification or someother mechanisms are involved in Cd mobilization.

We did not examine the accumulation of Cd in theroots of all the plants grown in the soils because of thedifficulty in collecting whole roots from soil and remov-ing Cd in the soil particles adhering to the root surfacefor Cd analysis. Therefore, it remains to be determined



whether whole plant Cd uptake corresponds to thedecrease in the amount of Cd in the soil. We estimatedthe ratio of Cd accumulation in the roots based on thevalues of shoot Cd accumulation and the soil Cd con-centration extracted with 0.1 mol L−1 HCl after plant-ing. In the case of Kasalath, which showed the highestCd content in shoots, the root ratio accounted forapproximately 80% of the total Cd uptake in both soiltypes. This was similar to the root Cd ratio in hydro-ponic culture, in which approximately 70% of the totalCd uptake accumulated in the roots. Thus, the signifi-cant decrease in the soil Cd content after rice plantingmay result from the effective access of roots to Cd poolsin soil and subsequent root accumulation.

Among the plant species used, rice accumulated ahigh amount of Cd, Cu, Fe, Mn and Zn (Figs 3,4). Highaccumulation of multiple metals is particularly advanta-geous in phytoremediation because soils are often con-taminated with multiple metals (McGrath et al. 2001).Brassica juncea compared favorably to rice for Znremoval from Fluvisol (Figs 3d,4d) and may display ahigher ability to accumulate Zn than Cd (Ebbs et al.1997). However, shoot Zn accumulation was lower inB. juncea than in rice when the plants were grown inthe Andosol (Figs 3h,4h). Brassica juncea is probablyunable to gain easy access to inorganically and organi-cally bound Zn, which are the predominant Zn frac-tions, when grown in an Andosol (data not shown).

Thus, B. juncea was not a promising plant for theextraction of Cd and other heavy metals from soils. Inthe present experiment, we harvested all the plants30 days after sowing because the 2 cultivars of B. juncea,which were early-ripening cultivars, started flowering.Using water culture, we observed that the Cd uptake ofB. juncea markedly decreased after the flowering stage,indicating that no additional Cd might be extractedfrom the soil (data not shown). Throughout this study,metal accumulation by maize was the lowest among theplant species tested, even though maize is known to be afast-growing high-biomass plant, which is one of thekey features needed for successful phytoremediation.Sugar beet displayed an equal or superior ability to thatof rice to accumulate heavy metals. However, sugarbeet may not be a suitable plant for phytoremediationof Japanese paddy fields because cultivation of thisspecies in Japan is usually restricted to cold regions. Inaddition, the soil pH would have to be adjusted to nearneutral values before sugar beet cultivation (a neutralpH would induce a low Cd availability in soil). It is wellknown that Cd uptake by rice is suppressed under sub-merged conditions (reductive) because of the formationof less-soluble cadmium sulfide (CdS). Conversely, ricewas able to accumulate high levels of Cd in shootswhen grown under upland conditions, that is, oxidative

conditions. As most of our targeted areas for phytore-mediation are paddy fields, rice is considered to be themost suitable plant in fields after drainage for Cdextraction. We are currently carrying out field experi-ments to select high Cd-accumulating rice cultivars andto develop agronomic practices to maximize Cd removalby these cultivars.

ACKNOWLEDGMENT

We thank Dr Hiroyuki Kurihara, Plantech ResearchInstitute, Japan, for supplying seeds of B. juncea andsugar beet.

REFERENCES

Arao T, Ae N 2003: Genotypic variations in cadmium levelsof rice grain. Soil Sci. Plant Nutr., 49, 473–479.

Baker AJM, Reeves RD, Hajar ASM 1994: Heavy metal accu-mulation and tolerance in British populations of the met-allophyte Thlaspi caerulescens J. & C. Presl (Brassicaceae).New Phytol., 127, 61–68.

Brown SL, Chaney RL, Angle JS, Baker AJM 1994: Phytore-mediation potential of Thlaspi caerulescens and bladdercampion for zinc- and cadmium-contaminated soils.J. Environ. Qual., 23, 1151–1157.

Cakmak I, Sari N, Marschner H et al. 1996: Phytosiderophorerelease in bread and durum wheat genotypes differing inzinc efficiency. Plant Soil, 180, 183–189.

Ebbs SD, Lasat MM, Brady DJ, Cornish J, Gordon R, KochianLV 1997: Phytoextraction of cadmium and zinc from acontaminated soil. J. Environ. Qual., 26, 1424–1430.

Goto S, Hayashi H, Yoneyama T, Chino M 2003: Behavior ofCd and Zn in rhizosphere of Brassica plants grown in anAndosol contaminated with Cd and Zn. Soil Sci. PlantNutr., 49, 735–739.

Kurihara H, Watanabe M, Hayakawa T 2005: Phytoremedia-tion with kenaf (Hibiscus cannabinus) for cadmium-contaminated paddy field in southwest area of Japan.Jpn J. Soil Sci. Plant Nutr., 76, 27–34 (in Japanese).

Li W, Higashi T, Fujimura T 2004: Uptake of cadmium andinorganic nutrients by Brassica plant under water cultureconditions. Jpn J. Soil Sci. Plant Nutr., 75, 329–337 (inJapanese).

McGrath SP, Zhao FJ, Lombi E 2001: Plant and rhizosphere pro-cesses involved in phytoremediation of metal-contami-nated soils. Plant Soil, 232, 207–214.

Römheld V, Marschner H 1986: Mobilization of iron in therhizosphere of different plant species. Adv. Plant Nutr.,2, 155–204.

Sadamoto H, Iimura K, Honna T, Yamamoto S 1994:Examination of fractionation of heavy metals in soils.Jpn. J. Soil Sci. Plant Nutr., 65, 645–653 (in Japanese).

Salam AK, Helmke PA 1998: The pH dependence of freeionic activities and total dissolved concentrations ofcopper and cadmium in soil solution. Geoderma, 83,281–291.



Whiting SN, Leake JR, McGrath SP, Baker AJM 2000:Positive responses to Zn and Cd by roots of the Zn andCd hyperaccumulator Thlaspi caerulescens. New Phytol.,145, 199–210.

Yanai J, Mabuchi N, Moritsuka N, Kosaki T 2004: Distributionand forms of cadmium in the rhizosphere of Brassica juncea

in Cd-contaminated soils and implications for phytore-mediation. Soil Sci. Plant Nutr., 50, 423–430.

Zhu YL, Pilon-Smits EAH, Jouanin L, Terry N 1999: Overex-pression of glutathione synthetase in Indian mustardenhances cadmium accumulation and tolerance. PlantPhysiol., 119, 73–79.

is brassica juncea a suitable plant for phytoremediation of cadmium in soils with moderately low...

Documents