Arsenic and heavy metal accumulation by Athyrium yokoscense from contaminated soils

Abstract

Athyrium yokoscense, a type of fern that grows vigorously in mining areas in Japan, is well known as a Cd hyperaccumulator as well as a Cu, Pb and Zn tolerant plant. However, no information is available on As accumulation of A. yokoscense, although it often grows on soils containing high levels of both heavy metals and As. In this study, young ferns collected from a mine area were grown in media containing As-spiked soils or mine soil in a greenhouse for 21 weeks. Athyrium yokosense was highly tolerant to arsenate and survived in soils containing up to 500 mg As (V) kg⁻¹. The addition of 100 mg As (V) kg⁻¹ resulted in the highest fern biomass (1.95 g plant⁻¹) among As-spiked soils. Although the As concentration of the fern was lower than other As hyperaccumulators, such as Pteris vittata, A. yokoscense could hyperaccumulate As in mature and old fronds. Arsenic was accumulated most efficiently in old fronds (922 mg kg⁻¹) in the media containing 5 mg As (III) kg⁻¹. Moreover, higher As accumulation was found in the roots of the ferns, with a range from 506 to 2,192 mg kg⁻¹. In addition, in the mine soil with elevated concentrations of As and heavy metals, A. yokoscense not only hyperaccumulated As (242 mg As kg⁻¹ in old fronds), but also accumulated Cd, Pb, Cu and Zn at concentrations much higher than those reported for other terrestrial plants. Athyrium yokoscense accumulated Cd mostly in fronds in high concentrations, up to 1095 mg kg⁻¹, while it accumulated Cu, Zn and Pb mainly in the roots and the concentrations were 375, 2040 and 1165 mg kg⁻¹, respectively.

Key words:

• accumulation
• arsenic
• Athyrium yokoscense
• heavy metal
• hyperaccumulator

INTRODUCTION

Arsenic and heavy metal contamination in soils has become an environmental concern worldwide. In the environment, metal contaminated soils are difficult to remediate and can adversely impact human health. Even in low doses, some metals still pose a major threat. Therefore, cost-efficient and environmentally friendly remediation techniques are urgently required. Phytoremediation can be defined as the combined use of plants, soil amendments and agronomic practices to remove pollutants from the environment or to decrease their toxicity (Chaney et al. 1999; Cunningham et al. 1995; Ward and Singh 2004). This technique has many advantages compared with other remediation procedures: low economic costs, the generation of recyclable residues and the possibility of being applied to soils and water with minimum environmental impact (Cunningham et al. 1995; Tu and Ma 2002; Ward and Singh 2004). Phytoextraction techniques involve the elimination of the pollutants from the soil by accumulating the toxic elements in the harvestable parts of the plants (Caille et al. 2004; Francesconi et al. 2002; Lombi et al. 2001; Tu et al. 2002). Finding the optimum plant species for remediation of a given soil will be the most important point controlling the success of the process, as well as the selection of adequate soil amendments that allow plant survival and good growth.

A plant that takes up abnormally large quantities of contaminants from soils and sequesters them in its aboveground biomass is known as a hyperaccumulator (Adriano 2001; Tu et al. 2002). More precisely, a Cd and As hyperaccumulator is defined as a plant that accumulates more than 100 mg kg⁻¹ dry weight (DW), whereas a Co, Cu, Cr and Pb hyperaccumulator is defined as a plant that contains more than 1,000 mg kg⁻¹ DW (Baker and Brooks 1989). In general, the concentrations of metals in hyperaccumulator plants are approximately 100 to 1,000-fold higher than those in normal plants growing on soils with background metal concentrations, and approximately 10 to 100-fold higher than most other plants growing on metal-contaminated soils (McGrath et al. 2002). There are more than 400 plant species known to hyperaccumulate heavy metals, including Arabis gemmifera (Kubota and Takenaka 2003), Thlaspi caerulescens (Lombi et al. 2000), Berkheya coddii (Robison et al. 2003), Sonchus asper L., and Corydalis pterygopetala F (Zu et al. 2005). Only recently, some types of ferns such as Pteris vittata (Ma et al. 2001), Pityrogramma calomelanos L (Francesconi et al. 2002), Pteris cretica, Pteris longifolia and Pteris umbrosa (Zhao et al., 2002) have been reported to be hyperaccumulators of As. However, the applicability of these plants to sites with multiple mixed contaminants is limited because these hyperaccumulators often accumulate a specific element only (McIntyre 2003). At many hazardous waste sites, where cleanup is required, the contaminated soil, groundwater and wastewater contain a mixture of several contaminants at widely varying concentrations. Especially in Japan, heavy metal contaminated soils often coincide with high levels of As in the soil (Tsutsumi 1981). The simultaneous cleanup of multiple contaminants using conventional methods such as landfills, soil washing, acid leaching or vapor stripping are all technically difficult and expensive; and these methods also destroy the biotic component of the soils (Francesconi et al., 2002; Tu and Ma, 2002). Therefore, it is desirable to find plant species that can accumulate multiple contaminated elements concurrently.

Athyrium yokoscense (Aspideaceae), a type of ferns, is distributed throughout Japan, particularly in the mine areas with metalliferous soil and is also distributed in Korea, Sakhanlin, northeastern China and eastern Siberia (Hiroi 1981). It is an indicator plant of heavy metals and grows vigorously in mining areas. It grows thickly on unstable slag heaps, flat damp grounds and floodplains. Through a preliminary survey of soil and vegetation around a closed old silver mine in Hyogo Prefecture, Japan, we have found that this fern propagates on soils containing high levels of As together with heavy metals (Kang et al. 2005). This fern has been reported to be a Cd and Pb hyperaccumulator. Morishita and Boratynski (1992); indicated that the leaflets (pinnule) of A. yokoscense could accumulate 996 mg kg⁻¹ of Cd. Yoshihara et al. (2005) also reported that the tissues of this fern accumulated high levels of Cd (maximum 3.3 mg g⁻¹ dry weight in roots). In addition, results of the studies on Pb accumulation showed that Pb concentration of this fern was more than 10 000 mg kg⁻¹ in gametophytes (Kamachi et al. 2005) and more than 2000 mg kg⁻¹ in the leaf blade (Sakai et al. 1991) as well as a Cu and Zn tolerant plant. Not only shows hyperaccumulation of Cd and Pb, A. yokoscense also reported as Cu and Zn tolerant plant. The root of A. yokoscense growing in areas contaminated with metals (Cu 3303 mg kg⁻¹; Zn 9195 mg kg⁻¹) contained maximum amount of Cu (5989 mg kg⁻¹ dry weight) (Nishizono et al. 1987b). Usui et al. (1975) also indicated that the root of A. yokoscense contained 2463.6 ppm Cu in average. However, no information is available on As accumulation by A. yokoscense. Therefore, we examined the uptake of As and its interaction with heavy metals in A. yokoscense grown in mine soil and in media containing As-spiked soils.

MATERIALS AND METHODS

Field survey and sampling

The survey was conducted in the Silver Ikuno Park, which contains a closed silver mine (latitude 35^ο10′0 N, longitude 134^ο48′0 E), in Hyogo prefecture, Japan. More than 90% of the lkuno area is mountainous and it is approximately 300 m a.s.l. The main ores distributed in this area are chalcopyrite, sphalerite, galena, arsenopyrite, argentite, cassiterite and wolframite. These ores contain Cu, Cd, Zn, Pb, As, Ag, Sn and W (Asami 1981). Intensive exploiting and smelting began in 1542 and its operation was ceased in 1973. At that time, the waste gases and the untreated wastewater from mining and smelting activities had caused long-term and severe contamination to local air, water, soil, vegetation and crops. At present, however, the contaminated areas are completely remediated and the mine area has been converted to a natural park.

In June 2004, more than 200 fern plants (with 4–5 fronds) with roots, and surface soils (0–10 cm) adjacent to the growing site of the fern were collected from the Ikuno mine area for greenhouse experiments. The size of the A. yokoscense plants was approximately 10 cm in height, and this was considered to be a young plant because the plants grow to a height of more than 20 cm in summer.

Greenhouse experiments

The mine soil (SIK) and an uncontaminated soil from a paddy field at Kochi University (FKU) were used for the greenhouse experiments. The SIK soil is classified as Dystrochrepts and the FKU soil is classified as Endoaquepts based on the USDA Soil Taxonomy classification system (Classification Committee of Cultivated Soils 1996).

Experiment 1

Potassium arsenate (KH₂AsO₄) or sodium arsenite (NaAsO₂) were added to the FKU soil at rates of 50, 100, 500 mg As (V) kg⁻¹ or 5, 10, 50 mg As (III) kg⁻¹, respectively. Both the FKU soil and the As-spiked FKU soil were thoroughly mixed with bark compost at a ratio of 2:1 (v:v), and kept for 2 weeks before use as the fern cultivation medium. FKU soil mixed with bark compost was used as a control. Each soil was then placed in a 1.5 L plastic pot, and one small 10 cm high fern plant with 4–5 fronds was planted in each pot. Each treatment was replicated four times. In addition, one set of pots containing media only (i.e. without a fern) was used to determine the influence of both watering and aging on the As concentration in the soils. After 21 weeks of cultivation, plants were harvested and washed with tap water to remove the soil and then quickly rinsed with 0.1 mol L⁻¹ HCl solution to remove the metals adsorbed onto the root surface, followed by several rinses with deionized water. Then the plants were separated into aboveground and belowground parts (roots including rhizomes). The aboveground part of the plants was separated into young, mature and old fronds based on age. Biomass was measured on a DW basis (after 2 days at 55^oC). Dried plant samples were ground to fine powder before analysis. Soil samples were taken from the pots both before transplanting and after harvest, air-dried and sieved with a 2-mm mesh for analytical use.

Experiment 2

Athyrium yokoscense was cultivated in the mine soil (SIK). One small live fern with 4–5 fronds and 10 cm in height was replanted in the SIK soils in a 1.5 L plastic pot and cultivated in a greenhouse at the Faculty of Agriculture, Kochi University, Kochi, Japan. This experiment was carried out with four replications. Neither chemical treatment nor fertilization was conducted to keep their original growing condition in nature. The plants were watered daily with tap water when necessary. The duration of the experiment was 21 weeks. Plant and soil samples were collected in the same way as experiment 1.

Analytical methods

General properties of the soils

The general soil properties were determined by following standard methods. Soil pH was measured using a pH meter with a combined electrode (Horiba, F-21, Kyoto, Japan). Cation exchange capacity (CEC) and the amounts of exchangeable bases (Ca, Mg, K and Na) were measured after successive extraction using 1 mol L⁻¹ NH₄-OAc

Table 1 Selected properties of the soils used in this study

Properties	FKU soil	SIK soil
pH	5.98	5.18
Total C (g kg^−1)	22.0	18.8
Total N (g kg^−1)	0.96	1.39
Cation exchange capacity (cmol (+) kg^−1)	10.8	7.4
Exchangeable Ca (cmol (+) kg^−1)	0.38	0.35
Exchangeable Mg (cmol (+) kg^−1)	0.21	0.69
Exchangeable K (cmol (+) kg^−1)	0.23	0.12
Sand (%)	74	72.3
Silt (%)	12.2	13.7
Clay (%)	13.8	13.9
Total As (mg kg^−1)	7.57	814
Total Cd (mg kg^−1)	0.84	10.5
Total Pb (mg kg^−1)	28.6	3464
Total Zn (mg kg^−1)	158	2422
Total Cu (mg kg^−1)	28.8	343
FKU soil, uncontaminated soil from a paddy field at Kochi University; SIK soil, mine soil.

(pH 7.0) and 100 g L⁻¹ NaCl. The amount of NH₄ ⁺ replaced by Na was determined using the stream distillation and titration method. The amounts of exchangeable bases were determined by atomic absorption spectrophotometry (Ca, Mg, and K) and by flame photometry (Na) (Shimadzu AA-6,800, Kyoto, Japan). Total carbon (T-C) and nitrogen (T-N) contents were determined using the dry combustion method with an NC Analyzer (Sumigraph NC-80 Tokyo, Japan). Particle size distribution was determined using the pipette method. Selected physical and chemical properties of FKU and SIK soils are summarized in Table 1.

Determination of As in soil and plant samples

A 0.1 g portion of the soil samples was weighed into a 120 mL Teflon digestion vessel, mixed with a mixture of concentrated HClO₄, HNO₃, HF and 20 g L⁻¹ KMnO₄ (2:3:5:2), and then heated on a hot plate at 100^oC for 20 min. If the purple color of KMnO₄ disappeared, 1 mL of 20 g L⁻¹ KMnO₄ was added and the vessel was again heated for 20 min. This procedure was repeated until the purple color did not disappear after 20 min of heating. Next, the temperature of the hot plate was set at 180–220^oC until the solution in the vessel became almost dry. The hot plate was turned off and 20 mL of HCl (1 + 1) was added to dissolve the ingredients using the remaining heat of the hot plate. After cooling, the sample solution was transferred into a 50 mL polypropylene volumetric flask and filled up to 50 mL with deionized water and then filtered (Advantec No. 5C) (Terashima 1984).

A 0.1 g portion of the plant samples was digested in a mixture of concentrated H₂SO₄ and 300 g L⁻¹ H₂O₂ (3:2) in a Kieldal flask. After vigorous reaction, the solution became clear and the Kieldal flask was heated on a micro-Keildal digester at 300^oC. If the solution became brown in color, the flask was removed from the digester; after several minutes of cooling, 2 mL of 300 g L⁻¹ H₂O₂ was added and the flask as placed on the digester again; this procedure was repeated until the color of the solution remained colorless after 60 min consecutive heating at 300^oC and white fumes of SO₂ could be observed. After cooling, the solution was transferred into a 50 mL polypropylene volumetric flask filled up to 50 mL with deionized water and then filtered (Advantec No. 5C).

Available As in soils was extracted from a 5.0 g soil sample using 25 mL of 1 mol L⁻¹ HCl by reciprocal shaking for 30 min at 30^oC and then filtered (Advantec No. 5C). Water-soluble As in the soil was extracted from approximately 2.0 g of soil with 20 mL deionized water by reciprocal shaking for 30 min at 30^oC. The suspension was then filtered (Advantec No. 5C).

For the measurement of As in the soil and plant digests and in the extracts of soils, 5 mL of the solution was mixed with 5 mL of concentrated HCl, 4 mL of 250 g L⁻¹ ascorbic acid, 2.0 mL of 300 g L⁻¹ KI and 2.0 mL of 20 g L⁻¹ AlCl₃, and then filled up with water to 20 mL immediately before measurement. Determination of As concentration was carried out using an inductively coupled plasma atomic emission spectrometer (ICP-AES, Shimadzu, ICPS-1,000 IV) combined with a hydride vapor generator (Shimadzu, HVG-1). The accuracy of the method was assessed using the certificated reference material soils (JSO-1, JSO-2) and plants (NIES CRM No. 1, NIES CRM No. 9) provided by the Geological Survey of Japan (GSJ) and the National Institute for Environmental Studies (NIES), Japan. Recoveries of the element ranged from 82 to 113%.

Distribution of As in soils

Arsenic distribution in the soil was estimated according to the sequential extraction method proposed by Keon et al. (2001) with some modifications. A 1.5 g soil sample that was collected after fern cultivation was extracted sequentially with water, magnesium chlorite (1 mol L⁻¹ at pH 7.0), sodium dihydrogenphosphate (1 mol L⁻¹ at pH 5.0) and hydrochloric acid (1 mol L⁻¹) and finally with a mixture of concentrated HClO₄, HNO₃, HF and 2% KMnO₄ (2:3:5:2). The five sequential extraction steps were assumed to produce operationally defined fractions corresponding to water-soluble As (III) and As (V), ionically bound forms of As (III) and As (V), strongly bound forms of As (V) by monodentate or bidentate ligand exchange, specifically adsorbed As

Figure 1  Dry biomass of Athyrium yokoscense after 21 weeks of cultivation in a greenhouse. SIK soil, mine soil.

(V) and occluded As by Mn oxides and amorphous Fe oxides, and As strongly bound to the residual phases, respectively.

Determination of total and available heavy metals in soil and plant samples

Soil samples for measuring total heavy metals were microwave-digested in a mixture of HNO₃ and HF at a ratio of 3:1 (Perkin Elmer, Multi-wave, Anton Paar, A-8054 Graz, Austria). Available heavy metals in soils were extracted by 0.1 mol L⁻¹ HCl. Concentrations of heavy metals (Cd, Cu, Zn and Pb) in the sample solutions were determined by Atomic Absorption Spectrometer (AAS; Shimadzu, AA-6,800). The digested solutions of the plant samples previously obtained were used for heavy metal determination by AAS.

Statistical analyses

anovas were carried out using SPSS Version 11.0. Differences among treatment means were separated using Turkey's multiple range tests. Significance was tested at the 0.05% probability level. Pearson's correlations were also conducted.

RESULTS AND DISCUSSION

Both the SIK and FKU soils used in this study were acidic and had a sandy texture. As shown in Table 1, the FKU soil is an uncontaminated soil both in terms of As and heavy metals (Cd, Cu, Pb and Zn). The total As content of FKU soil was 7.6 mg kg⁻¹, which is in the range of natural abundance of Japanese soils, 0.4–70 mg kg⁻¹ with a mean of 11 mg kg⁻¹, Iimura (1981), while SIK soil had an extremely high content of As and heavy metals.

Biomass production of A. yokoscense cultivated in a greenhouse

Growth of A. yokoscense was significantly affected by the chemical forms of As, and varied with the As levels in the soils (Fig. 1) because of the phytotoxic effect of As (Kabata and Pendias 1991). The biomass of A. yokoscense with the addition of 500 mg As (V) kg⁻¹ and 50 mg As (III) kg⁻¹ FKU soils reduced both the aboveground and belowground biomass by 33% to 57% compared to the control (Fig. 1). In contrast, the biomasses obtained on the media containing 50 mg As (V) kg⁻¹, 5 and 10 mg As (III) kg⁻¹ FKU soils were not significantly different compared with that of the control, and the addition of 100 mg As (V) kg⁻¹ FKU soil significantly increased the aboveground biomass. Among all treatments, the biomass (both aboveground and belowground) of the fern grown in mine soil was highest and differed significantly from the other treatments. Thus, in this experiment, the mine soil was the most favorable medium for fern growth. After 21 weeks of cultivation in mine soil, fern biomass increased markedly from 0.26 ± 0.08 to 4.3 ± 0.40 g plant⁻¹ DW.

In the media treated at 500 mg kg⁻¹ of As (V), and at 10, 50 mg kg⁻¹ of As (III) FKU soils, symptoms of As toxicity appeared in the old fronds after 2 weeks, but the plants survived throughout the experiment. Thus, it is concluded that moderate As levels might promote the growth of the ferns. At much higher concentrations, As inhibited the growth of the plants and led to plant death.

Distribution of As in A. yokoscense cultivated in media containing As-spiked soils and mine soil

Arsenic concentration in fronds increased from young to old fronds grown in the media containing As-spiked soil. Arsenic concentrations in the mature and old fronds increased with As levels in the medium when As was supplied as As (V) (Fig. 2). In old fronds, the As concentration increased from 162 mg kg⁻¹ (control) to 562 mg kg⁻¹ at the 500 mg As (V) kg⁻¹ level. However, As concentrations in young fronds were not significantly different between the control and the other treatments. These results suggested that detoxification or sequestration mechanisms may be present in mature and old fronds of A. yokoscence as described by Zhang et al. (2002) for P. vittata.

In the case of As (III) addition, the accumulation of As in fronds decreased with the addition of 10 and 50 mg As (III) kg⁻¹ FKU soils. Among all treatments, the addition of 5 mg As (III) kg⁻¹ FKU soil resulted in the highest As accumulation, with up to 922 ± 68 mg kg⁻¹ in the old fronds. When levels of As in soils were not toxic for plant growth, a greater part of the absorbed As was probably translocated to the fronds, leading to greater concentrations in older fronds because they have been receiving As for a longer time. Most plants have relatively low As concentrations, typically < 1.5 mg kg⁻¹ (Kabata and Pendias 1991) and 0.02–7 mg kg⁻¹ DW (Bowen 1979). As the As concentration of A. yokoscense was much more than 100 mg kg⁻¹ in fronds, it could be said that A. yokoscense is an As hyperaccumulator.

Figure 2  Arsenic concentration in different parts of Athyrium yokoscense cultivated in As-spiked soils and mine soils. Solid line represents the concentration of As in hyperaccumulation plants, 100-fold of central value of normal range, AS: 3.5 mg kg−1 (Bowen, 1979). SIK soil, mine soil.

The concentration of As in the ferns planted in the mine soil was a little lower than that in the media containing As-spiked soils. This may be because the fern grown in the mine soil had a greater biomass production or there was less soluble As in the mine soil than the media containing As-spiked soils. Before transplanting, As concentrations of young and old fronds were 7.8 ± 0.3 and 57.7 ± 2.2 mg kg⁻¹, respectively. After 21 weeks of cultivation in the mine soil, As concentration had increased to 167.6 ± 5, 242.4 ± 25 and 2,175.2 ± 97 mg kg⁻¹, respectively (Fig. 2).

Contrary to the fronds, concentrations of As in the roots slightly decreased with increasing As levels in soils (Fig. 2). Arsenic concentration in roots with the addition of 50 mg As (V) mg⁻¹ FKU soil was highest, followed by that in the mine soil. The addition of As (III) to the soils reduced the As concentrations in roots. As both the biomass and the As concentrations in the roots of the ferns grown in the media containing As (III) were lower than those in the control, it demonstrated that the addition of As (III) at levels greater than 10 mg kg⁻¹ to the FKU soil inhibited normal root growth, thus, resulting in As toxicity. Compared to other As hyperaccumulators, such as P. vittata or P. calomelanos, A. yokoscense has a lower capacity to uptake As to its biomass. Athyrium yokoscense accumulated As mostly in the roots, while As is predominantly concentrated in the fronds of P. vittata and P. calomelanos. The concentrations of As in fronds of P. vittata and P. calomelanos were 7,230 and 8,000 mg kg⁻¹, respectively (Francesconi et al. 2002; Tu et al. 2002). Thus, the As tolerant mechanisms in A. yokoscense might be different from those in P. vittata and P. calomelanos.

Arsenic tolerance and detoxification mechanisms in P. vittata have been reported by Zhang et al. (2002). Although a strong and positive correlation between As concentration and acid-soluble thiols (phytochelatin) in plant leaflets was found, they suggested that phytochelatin could play a minor detoxification role in this hyperaccumulator because only a very small part of As is complexed by phytochelatin. Although metal tolerant mechanisms have been reported in A. yokoscense, to date, As detoxification and tolerant mechanisms are unclear. Further study should be conducted to elucidate As species in plants to understand its hyperaccumulation as well as the detoxification mechanisms, and the influence of nutrients (e.g. P, Fe) on the growth and the simultaneous accumulation of As and heavy metals in A. yokoscense.

Distribution of As in media containing As-spiked soils after cultivation

Mobility, bioavailability and toxicity of trace elements depend strongly on their chemical form and type of binding. Consequently, determination of total As concentrations is insufficient to estimate potential risks of remobilization and uptake of liberated As by biota.

The sum of each fraction obtained by sequential extraction procedures agreed well with total As in soils obtained by digestion (recovery varied from 81.2 to 99.6%). The forms of As in the soil media containing As-spiked FKU soils with and without ferns (designed as soil + f and soil − f, respectively) are presented in Fig. 3. Despite the difference in total As concentration, the proportion of As among the five sequentially extracted fractions was similar for all soil media, except for the mine soil. Almost 10% of the total As content was extracted during the first step (water-soluble As) and less than 7.5% of As was found in the second fraction, which corresponds to ionically bound forms of As. The majority of the As was extracted in the third fraction, which included strongly adsorbed As via the ligand exchange mechanism. This result agrees with another study that examined soils in Cornwall in southwest England (Caille et al. 2004). They found that more than 60% of As was extracted in the fraction that corresponds to As adsorbed to amorphous and poorly crystalline hydrous oxides of Fe and Al. A considerable amount of As was also associated with the amorphous Fe hydroxides fraction. The residual fraction accounted for approximately 10% of the total As in the media containing As (V) spiked soils, and 10–14% of the total As in the media spiked with As (III). The fractional distribution of As in the mine soil (SIK) indicated that only a small proportion (less than 3%) of total As was present in water-soluble and MgCl₂ fractions. The data also indicated that mine soil contained higher proportions of As in the residual fraction (accounting for 62%) compared with the media containing As-spiked FKU soils.

Heavy metal accumulation in A. yokoscense cultivated in mine soil

Athyrium yokoscense could accumulate not only As but also heavy metals (Cd, Cu, Pb and Zn) when cultivated in the mine soil. The concentrations of heavy metals in A. yokoscense were significantly higher than the values reported for terrestrial plants (Cd 0.1–2.4 mg kg⁻¹; Pb 1–13 mg kg⁻¹; Cu 1–30 mg kg⁻¹; Zn 20–400 mg kg⁻¹; Bowen 1979). The fern accumulated Cd mainly in the fronds, with concentrations of 335 ± 84 (young), 959 ± 96 (mature) and 1,095 ± 120 (old) mg kg⁻¹, respectively (Fig. 4). The concentration of Cd in the tissue of A. yokoscense was in the following order: roots < young frond < mature frond < old frond, which is typical of an accumulator species. Morishita and Boratynski (1992) detected that at polluted areas in Bandai, Fukushima, A. yokoscense exhibited high levels of Cd up to 996 mg kg⁻¹ in leaflets, while

Figure 3  Arsenic distribution in the As-spiked soils and mine soils after cultivation. f, fern; SIK soil, mine soil; W-s: water soluble-As; Res: residual.

Figure 4  Heavy metal accumulation in Athyrium yokoscense cultivated in mine soil. Solid lines represents the concentration of As in hyperaccumulation plants, 100-fold of central value of normal range, Cd: 1.2 mg kg−1 (Bowen, 1979).

Yoshihara et al. (2005) demonstrated that the callus and regenerated tissues of A. yokoscense tolerated at least 1 mmol L⁻¹ Cd for more than 1 month, and accumulated high levels of Cd (maximum 3,300 mg kg⁻¹ DW in root). Athyrium yokoscense accumulated Pb in the roots up to 2,040 mg kg⁻¹. In addition, the concentrations of Cu and Zn in the roots of the ferns were 375 and 1,165 mg kg⁻¹, and those in fronds were 88 and 973 mg kg⁻¹, respectively. These values were much higher than values recorded in other terrestrial plants, although this accumulation did not reach hyperaccumulation in this experiment. It is well known that A. yokoscense accumulates large amounts of heavy metals in the tissues, particularly in the roots. It is reported that the roots of this fern accumulated a large amount o Pb (∼22400 ppm) than that of leaf blade (∼2080 ppm) (Sakai et al. 1991) Greater amount of Cu and Zn were accumulated in the root than in the leaf of A. yokoscense was also reported by Nishizono et al. (1987b) They found that Cu content of the root of this fern was 22 times of that in the leaf. Similarly, Usui et al. (1955) also found that the root of A. yokoscense showed far greater accumulation of Cu than on the leaves. Thus, a tolerant mechanism appears to have evolved to allow A. yokoscense to survive in soils containing high As and metal concentrations. Nishizono et al. (1987a, 1988) suggested that a 9.5-kDa cysteine-rich peptide that was induced by Cu exposure may be involved in the mechanism of Cu tolerance in ferns growing in Cu-contaminated soil. In addition, the compartmentalization of excess Cu in the cell wall appears to contribute to Cu tolerance and accumulation in A. yokoscense. Similarly, Xian (1996) also indicated that in the roots of A. yokoscense, root cell walls appear to act as an outstanding heavy metal accumulation part, possessing more than 80% of the total Cd and Zn in the roots. Moreover, Sakai et al. (1991) postulated that the main chemical species of Pb in the tissues of A. yokoscense might be lead sulfate (cortex and epidermis parts) and metal complexes with pectic acid and some amino acids (stele part). Furthermore, proanthocyanidins (condensed tannins) have been proposed to have a latent ability to complex with Pb ions and play an important role in Pb tolerance and accumulation in A. yokoscense (Kamachi et al. 2005).

Caille et al. (2004) reported that P. vittata could hyperaccumulate As from naturally contaminated soil, but grew poorly on soil that had extremely high contents of Zn and Cu (1221 and 5,500 mg kg⁻¹, respectively). These authors reported that the As and Cu concentration in P. vittata was 3,510 and 7.0 mg kg⁻¹ when the fern was grown on soil containing 475 mg kg⁻¹ As and 370 mg kg⁻¹ Cu (Roscroggan soil), respectively. In the present study, A. yokoscense accumulated 242 mg kg⁻¹ As and 88 mg kg⁻¹ Cu from soil containing 814 mg kg⁻¹ As and 343 mg kg⁻¹ Cu. Thus, at a similar Cu concentration in soil, P. vittata can accumulate much more As than A. yokoscense, but Cu accumulation was lower than that by A. yokoscense. In contrast, when P. vittata was cultivated in a soil containing 4,550 mg kg⁻¹ As, 3,903 mg kg⁻¹ Pb and 1,242 mg kg⁻¹ Zn, the accumulation of As in P. vittata (362 mg kg⁻¹ in fronds) was almost the same as A. yokoscense that was grown in the mine soil (814 mg kg⁻¹ As, 3,464 mg kg⁻¹ Pb and 2,422 mg kg⁻¹ Zn) (Caille et al. 2004). Fayiga et al. (2004) also indicated that, although effective in taking up As (up to 4,100 mg kg⁻¹), P. vittata had a limited capacity to take up other metals. Compared to P. vittata, it is proposed that the potential for heavy metal accumulation from As and heavy metal contaminated soils is higher for A. yokoscense, particularly when the As level in the soil is not extremely high.

Potential of A. yokoscense for As and heavy metal phytoremediation

For the practical application of phytoremediation, the total amount of metal accumulation by A. yokoscense must be considered to measure the potential effectiveness of a plant for phytoextraction (Francesconi et al. 2002; Tu et al. 2002). The original soil metal removed by one crop is expressed as a percentage of metal accumulation by the plant (plant metal concentration × biomass) to the metal in the soil (soil metal concentration × soil mass) (Tu et al. 2002; Zhao et al. 2003). Approximately 0.2–8.8% of the initial soil As was removed by A. yokoscense after 21 weeks of cultivation. Among the six As levels tested, the highest removal of As was in the media containing 50 mg As (V) kg⁻¹ FKU soil, accounting for 8.84%. The removals of the other treatments were 0.89, 6.16, 2.27, 0.26, 5.15, 2.37 and 0.82% for the mine soil, control-FKU soil, the media containing 100 mg As (V) kg⁻¹, 500 mg As (V) kg⁻¹, 5 mg As (III) kg⁻¹, 10 mg As (III) kg⁻¹ and 50 mg As (III) kg⁻¹, respectively. Assuming a constant rate of As extraction as above, it would take 12–15 harvests (250–300 weeks) to remediate soil contaminated with 50 mg As kg⁻¹. For P. vittata, the accumulation of As in its biomass accounted for up to approximately 26% of the initial As in the soil spiked with As at a level of 100 mg kg⁻¹ (Tu et al. 2002). Thus, P. vittata would take less time than A. yokoscense, only 4 or 5 harvests (80–100 weeks), to phytoremediate a soil contaminated with 100 mg As kg⁻¹. In addition, 13.7, 0.37, 0.24 and 0.20% of Cd, Cu, Pb and Zn, respectively, in the mine soil were removed by A. yokoscense in the greenhouse cultivation.

Although the direct application of A. yokoscense to As and heavy metal phytoremediation is not a promising approach practically, soil amendments or the application of fertilizers might assist A. yokoscense to increase its As and heavy metal removal from contaminated soils.

ACKNOWLEDGMENTS

We thank Mitsubishi Material for providing soils and fern samples. This research was supported by a Sasakawa Scientific Research Grant (No. 16–317) from The Japan Science Society to T. K. Van and by the Grant-in-Aid for Scientific Research (B, No. 15380223) from Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan to K. Iwasaki.