Arbuscular mycorrhizal fungus (Glomus aggregatum) influences biotransformation of arsenic in the rhizosphere of sunflower (Helianthus annuus L.)

Abstract

The role of arbuscular mycorrhizal fungus (AMF) on biotransformation of arsenic (As) in the rhizosphere of sunflower (Helianthus annuus L.) was studied in a rhizobag system with As-contaminated soil. The treatments consisted of a combination of two levels of AMF (Glomus aggregatum) inoculation (–AM and +AM) under sterile and non-sterile soil conditions (S and NS). Each treatment was designated as –AM/S, –AM/NS, +AM/S and +AM/NS. Sunflower seedlings were cultured in the rhizobag for 6 weeks. Rates of root AMF colonization in +AM treatments were approximately 40% and 0% in –AM/S treatment. The AMF inoculation reduced As toxicity symptoms and improved plant growth. Shoot As, but not the root As, concentrations were reduced by AMF inoculation. Shoot and root P concentrations increased in +AM treatments regardless of soil sterilization. Analyses of water soluble (WS) As in the soils at the end of the cultivation indicated that the amount of WS-arsenite (As^III) was higher in the central compartment compared with the outer compartment, while WS-arsenate (As^V) was dominant in the outer compartment. Dimethylarsinic acid (DMAA) was detected only in +AM treatments and the concentration was higher in the +AM/NS treatment compared to the +AM/S. Trimethylarsine oxide (TMAO) was also detected in measurable amounts in the NS condition and AMF inoculation increased the amount of TMAO, especially in the central compartment. No methylated As species was detected in the –AM/S treatment. From these results, it is clear that the mycorrhizal roots colonized by Glomus aggregatum are primarily involved in DMAA formation. Aside from these mycorrhizal roots, the indigenous soil microorganisms in the mycorrhizosphere are responsible for promoting the transformation of DMAA into TMAO. These As biomethylation processes would favor the detoxification of As, especially at the interface of plant roots and rhizosphere soil.

Key words:

• arbuscular mycorrhiza fungi
• dimethylarsinic acid
• rhizosphere
• sunflower
• trimethylarsine oxide

INTRODUCTION

Arsenic (As) pollution is a serious environmental problem, especially in Bangladesh, Vietnam, Taiwan and the West Bengal District of India (Abedin et al. 2002; Meharg 2004). To understand the movement of As in the soil–water environment, several researchers have examined the vertical distribution of As in sediment columns taken by boring operations (Yamazaki et al. 2003a), As extractability (Yamazaki et al. 2003b), and background As contents in soils (Ali et al. 2003). In contrast, findings of As-accumulating ferns have created an opportunity for developing strategies that would remediate As-contaminated soils by using plant uptake ability (Ma et al. 2001; Van et al. 2006). For the development of effective phytoremediation technology and for alleviating As toxicity in crop plants, an understanding of the dynamics of As in the plant–soil system is indispensable.

Soil microorganisms affect metal speciation and availability through their ability to modify its solubility (Bently and Chasteen 2002; Turpeinen et al. 1999). Mobilization occurs through processes such as protonation, chelation and chemical transformation, while immobilization occurs through crystallization, sorption and assimilation into biomass. Such abilities of soil microorganisms to transform metals and metalloids in soils can be utilized as a means for removal, recovery or detoxification of inorganic and organic metal pollutants. Soil microorganisms also mediate As transformation processes, including methylation, demethylation, reduction of arsenate (As^V) and oxidation of arsenite (As^III) (Gadd 2004). Such biotransformation influences the fate, mobility and bioavailability of As in soils, controlling its effects on soil organisms and its uptake by higher plants (Fitz and Wenzel 2002).

Mycorrhiza is an integral and functional part of plant roots that plays an important role in alleviating metal toxicity in plants. One of the dominant forms of mycorrhizas for symbiosis with plants is arbuscular mycorrhizal fungi (AMF), which belongs to the groups of endomycorrhiza (Smith and Read 1997). The role of AMF in plant uptake of heavy metals has been extensively studied. Several studies have shown that AMF often protect plants against high concentrations of non-essential metals in their shoots by enhancing metal retention in the roots (Christie et al. 2004; Joner et al. 2000; Leyval et al. 1997), in addition to the improvement of phosphorus (P) nutrition status (Li et al. 1991) and uptake of some essential elements, resulting in greater plant biomass (Marschner 1995). Because As has great similarity in chemical behavior to P, it is thought that AMF would also influence As dynamics in soils. The results of several studies on the role of mycorrhiza on As uptake by plants have been recently reported. The general consensus is that plants colonized by mycorrhizal fungi display enhanced resistance to As contamination in soils (Meharg and Hartley-Whitaker 2002). Several studies suggest that mycorrhiza reduces the As concentration in shoots (Asher and Reay 1979; Gonzalez-Chaves et al. 2002; Liu et al. 2005; Meharg et al. 1994; Sharples et al. 2000; Ultra et al. 2007) and acts as an As filter to maintain low As concentrations in the shoots, while improving P nutrition of the host plant.

As a rule, inorganic As species are more toxic than organic As species (Adriano 2001). The AMF are known to change the chemical composition of root exudates and to affect the activities of soil enzymes, and to introduce physical modifications to the environment surrounding the roots (Smith and Read 1997). These AMF-induced changes in the rhizosphere and the development of AMF mycelia, which can act as a nutrient source for other microorganisms, are likely to influence the structure and functioning of soil microbial communities in the mycorrhizosphere. Similarly, there could be specific modification in the environment between the plant roots and the mycorrhizosphere, such as changes in pH and redox potential, that could possibly promote the biotransformation process of As (Jeffies et al. 2002). Possible mechanisms to reduce As toxicity may include the enhancement of indigenous microflora that are capable of metabolizing As, the AMF association triggering the evolution of As co-metabolism between plant and AMF, and the direct involvement of AMF on As biomethylation (Barea et al. 2005). As there are large numbers of bacteria (including actinomycetes) and fungi that are capable of As biomethylation (Frankenberger and Arshad 2002), increased growth and activity of these indigenous microorganisms can be expected to promote the production of organic As species.

Only a few studies have directly investigated the fates of As in the rhizosphere with respect to the potential interactions in the soil–plant–microbe system. Recently, we examined the effects of AMF inoculation and P application on As dynamics in the rhizosphere of sunflower and detected the presence of dimethylarsinic acid (DMAA) in the soil inoculated with AMF (Ultra et al. 2007). Based on this result, we hypothesized that the chemical behavior and toxicity of As were altered in the mycorrhizosphere, which could be contributing to better survival of mycorrhizal plants in As-contaminated soil. However, it has not yet been clarified whether AMF and mycorrhizal roots are primarily involved in the transformation of inorganic As into the organic forms or whether the indigenous microorganisms are responsible for the process. Therefore, in this study, we attempted to elucidate the role of AMF on As biomethylation in the rhizosphere of sunflower through the evaluation of different As species and enzymatic activities in the rhizosphere soil.

MATERIALS AND METHODS

Soil sampling and rhizobag assembly

The As-contaminated soil used in these experiments was collected from the sub-surface layer (20–30 cm) of a cypress forest near a closed silver mine in Isotake, Ohda city, Shimane prefecture, Japan. The soil properties and the heavy metal contents were reported in a previous paper (Ultra et al. 2007). After air-drying, soil samples for cultivation and chemical analysis were passed through 5 mm and 2 mm mesh sieves, respectively. Soil for the sterile treatments was autoclaved at 121°C, 1 h, for three consecutive days prior to use. The comparison of some selected chemical properties before and after sterilization is given in Table 1. Methylated As species were not detected for the sterile and non-sterile soils prior to the experiment.

A rhizobag system composed of two compartments was used for the experiment. The system had one central compartment in which the plants were grown, and an outer compartment serving as the mycorrhizosphere (Fig. 1). The central compartment was separated using a 25 µm nylon mesh mounted on a cylindrical plastic

Table 1 Chemical properties of the soil collected from Shimane Prefecture, Japan, as affected by sterilization

Parameter		Non-sterile	Sterile
pH (1:5, water)		5.69	5.82
Electrical conductivity (1:5)	dS cm^−1	4.11	4.13
Total carbon	g kg^−1	5.36	5.22
Total nitrogen	g kg^−1	0.483	0.457
Available P (Bray 1)	mg kg^−1	6.03	8.09
Water soluble P (1:100)	mg kg^−1	14.3	18.2
Total As	mg kg^−1	620	619
1.0 mol L^−1 HCl-extractable As	mg kg^−1	6.50	8.30
Water Soluble (WS) As
Total WS-As	mg kg^−1	0.190	0.260
WS-As^III	mg kg^−1	0.114	0.137
WS-As^V	mg kg^−1	0.076	0.123
MMAA	mg kg^−1	nd	nd
DMAA	mg kg^−1	nd	nd
TMAO	mg kg^−1	nd	nd
“Free” Fe-oxide	g kg^−1	2.30	2.26
nd, not detected.

frame (60 mm diameter, 120 mm height). There was 230 g of soil in the central compartment and 320 g of soil in the outer compartment, summing up to 550 g soil per pot. Although root growth was confined to the central compartment, nutrients and As could be delivered to the roots by diffusion and through water movement caused by transpirational flow from the soil in the outer compartment. Therefore, in this rhizobag system, we assumed that a total of 550 g soil was available both for –AM and +AM plants, as a source of nutrients and As. After assembling the experimental pots, urea and K₂HPO₄ were supplied to all pots as basal fertilizers at a rate of 60 mg pot⁻¹ (50 mg N kg⁻¹) and 90 mg pot⁻¹ (30 mg P kg⁻¹), respectively.

Description of the treatments and plant culture

The treatments consisted of two levels of AMF (Glomus aggregatum) inoculation (–AM and +AM) under sterile and non-sterile soil conditions (S and NS), comprising four-treatment combinations. Each treatment was designated as –AM/S, –AM/NS, +AM/S and +AM/NS. Air-dried soil was used for non-sterile soil and autoclaved soil was used as sterile soil. Five grams of sterile (121°C, 30 min for three consecutive days) or non-sterile AMF inoculants (Kinkon-King, Hokko Chemical, Tokyo, Japan) were mixed with the soil in the central compartment of the –AM and +AM treatments, respectively. The only beneficial microorganism in the AMF inoculant used in these experiments was AMF (Glomus aggregatum) (Iwasaki K, 1998, pers. comm.). Sunflower (Helianthus annuus L. cv. “Sweet”) seeds were surface sterilized using a 70% ethanol solution for 5 min, then washed

Figure 1  Design of the rhizobag system.

three times with sterile distilled water and germinated on sterile petriplates lined with moistened filter paper. The pre-germinated seed was then planted in the central compartment. Pots were maintained at a moisture content of 40% of the field capacity under glasshouse conditions. Three replicates were used for each treatment.

Plant sampling and chemical analysis

During the 6-week cultivation period, plants were periodically examined for toxicity symptoms (brown spots appearing on the leaves). At harvest, the toxicity score was evaluated as: (4) more than two-thirds of the leaf area turned brown; (3) two-thirds to one-third of the leaf area turned brown; (2) less than one-third of the leaf area turned brown; (1) only the edges of the leaf blade turned brown; (0) no symptoms. The scores for individual leaves were totaled for each plant and divided by the total number of leaves in each plant (referred to as the normalized toxicity score). After cultivation, the shoots were excised and the roots were carefully separated from the soil, and their fresh and dry weights were determined.

A 0.5 g portion of the fresh roots from each plant was subjected to trypan blue staining for microscopic examination of mycorrhizal infection (Brundrett 1994). The root length colonized by AMF was quantified using the gridline intersection method (Giovennetti and Mosse 1980). The remaining roots and all the shoots were oven-dried at 90°C for 2 days and ground for chemical analysis. The tissues were digested using a sulfuric acid and hydrogen peroxide mixture in a Kjeldahl flask at 300°C and analyzed for As using hydride generation-inductively coupled plasma-atomic emission spectrometry (HG-ICP-AES; ICPS-1000IV equipped with HVG-1, Shimadzu, Kyoto, Japan). The detection limit for As was 1.0 µg L⁻¹ in the digested sample solution. The P content was determined for the same digest using ICP-AES. To verify the accuracy of As and P determination, standard reference materials of pepperbush (CRM No.1) and sargasso (CRM No. 9) provided by the National Institute for Environmental Studies, Tsukuba, Japan, were analyzed. The recovery rates of As and P with respect to their certified values were 90–95%.

Soil sampling and chemical analysis

After the sunflower roots were sampled from the central compartment, the soils in the central and outer compartments were passed through a 2 mm sieve. Portions of 20 g were stored at 4°C for the analysis of acid phosphatase activity and dehydrogenase activity; the remaining soil was freeze-dried and stored at −30°C for chemical analysis.

Soil samples were shaken with distilled water (1:100) for 2 h at room temperature to obtain a water-soluble (WS) fraction. The WS fraction was analyzed for different forms of As using cold-trap vapor-reduction atomic absorption spectrometry (AA-6800 equipped with ASA-2sp, Shimadzu), and the concentrations of As^III, As^V, monomethylarsonic acid (MMAA), dimethylarsinic acid (DMAA) and trimethylarsine oxide (TMAO) were determined (Masscheleyn et al. 1991; Ultra et al. 2007). Calibration curves for each As species were prepared using standard solutions (Tri Chemical Laboratories, Yamanashi, Japan). The detection limit for each As species was 0.1 µg L⁻¹ in the sample solution. Water-soluble P content was determined using ICP-AES in the same soil extract.

Acid phosphatase activity in the soil was estimated based on p-nitrophenol release after incubation with p-nitrophenyl phosphate (Alef et al. 1995). In brief, 1 g of the soil sample was treated with 0.25 mL of toluene, 4 mL of modified universal buffer (MUB; pH 6) and 1 mL of 15 mmol L⁻¹ p-nitrophenyl phosphate, and incubated for 1 h at 37°C. The MUB was prepared from a stock solution (fivefold strength) containing 12.1 g of Tris, 11.6 g of maleic acid, 14 g of citric acid and 6.3 g of boric acid dissolved in 500 mL NaOH (1 mol L⁻¹); diluted with sterile distilled water. To terminate the reaction, 1 mL of 0.5 mol L⁻¹ CaCl₂ and 4 mL of 0.5 mol L⁻¹ NaOH were added to the mixture prior to filtration. The absorbance at 400 nm resulting from p-nitrophenol was measured using a spectrophotometer (UV-1200, Shimadzu) and the acid phosphatase activity was expressed in mg (p-nitrophenol) g⁻¹ h⁻¹. The dehydrogenase activity was estimated based on the incubation of soil with an electron acceptor 2(p-iodophenyl)-3-(p-nitrophenyl)-5-phenyl tetrazolium chloride (INT) at 40°C for 2 h (Von Mersi and Schinner 1991). The reaction product, iodonitrotetrazolium formazan (INF) was extracted using N,N-dimethylformamide and an ethanol mixture (1:1, v/v) and quantified using the spectrophotometer at 464 nm against a blank. Dehydrogenase activity was expressed in µg INF g⁻¹ h⁻¹.

Statistical analysis

For the data on toxicity symptoms, plant growth, and As and P concentrations, Tukey's tests were used to determine any significant differences (P ≤ 0.05) between the treatments. Main effects and interaction effects of treatments and compartments on As distribution and the proportion of different forms of As in the total WS-As were subjected to a two-way anova.

RESULTS

Plant growth, As toxicity and AM infection in roots

Arsenic toxicity symptoms began to develop in the sunflower plants 3 weeks after transplanting in the –AM/S treatment and approximately 4 weeks after transplanting in the other treatments. Initial symptoms were a browning at the edges of the leaf blades and red-brown necrotic spots on older leaves, followed by the senescence of the leaves. At harvest, the average normalized toxicity scores in the +AM treatments were lower compared with the –AM, and the symptoms were most clearly alleviated in the +AM/S treatment (Table 2). Shoot dry weight tended to increase with AMF inoculation and the highest value was obtained in the +AM/S treatment. Root dry weight in the S condition was relatively higher compared with the NS condition, regardless of AMF treatment. A significant increase in root dry weight as a result of AMF inoculation was only observed in the NS condition, while the value of the +AM/NS treatment was also higher than that of –AM/NS. The root length colonized in the +AM treatments was approximately 42% and 38% under S and NS conditions, respectively, while plants in the –AM/NS treatments showed less than 10% and no roots were colonized in the –AM/S treatment. The colonization observed in the –AM/NS treatment probably resulted from mycorrhizal fungi initially existing in the soil.

Concentration and contents of As and P in plants

The concentration of As in the shoots was significantly reduced by AMF inoculation under both S and NS conditions (Table 2). Shoot As content did not differ

Table 2 Normalized As toxicity scores, AM root length colonized, dry matter yield, and the As and P concentrations in shoots and roots of sunflower grown in As contaminated soil†

Treatment	Normalized toxicity scores	Root length colonized (%)	Dry matter		As concentration		P concentration
			Shoots (g plant^−1)	Roots (g plant^−1)	Shoots (mg kg^−1)	Roots (mg kg^−1)	Shoots (mg kg^−1)	Roots (mg kg^−1)
+AM/NS	1.06 b	35 b	1.86 b	0.39 a	1.56 c	85.6 a	2210 a	1150 a
–AM/NS	1.53 a	7 c	1.46 b	0.20 b	2.16 a	67.7 ab	1330 b	814 b
+AM/S	0.66 b	42 a	2.74 a	0.55 a	1.36 c	78.6 ab	2210 a	1210 a
–AM/S	1.38 a	0 d	2.04 b	0.42 a	1.87 b	61.7 b	1420 b	856 b
^†Means within the same column followed by the same letter(s) are not significantly different from each other based on Tukey's test at P < 0.05.

Table 3 Contents of As and P, ratios of shoot content to root content, and ratios of P content to As content in the shoots and roots of sunflower plants grown on As-contaminated soil†

Treatment	As			P			P/As		Shoot uptake^‡
	Shoots (µg plant^−1)	Roots (µg plant^−1)	Shoots/ roots	Shoots (µg plant^−1)	Roots (µg plant^−1)	Shoots/ roots	Shoots	Roots	As (µg g^−1rdw)	P (mg g^−1rdw)
+AM/NS	2.90 a	33.5 a	0.09 b	4.10 b	0.45 b	9.35 a	1420 a	14 a	7.40 b	12.0 a
–AM/NS	3.16 a	13.5 b	0.24 a	1.95 c	0.16 c	12.20 a	624 b	12 a	15.90 a	10.7 a
+AM/S	3.72 a	43.0 a	0.09 b	6.06 a	0.66 a	9.24 a	1640 a	16 a	6.80 b	12.3 a
–AM/S	3.83 a	26.1 ab	0.15 ab	2.89 bc	0.36 b	8.54 a	753 b	14 a	9.07 ab	8.1 a
^†Means within the same column followed by the same letter(s) are not significantly different from each other based on Tukey's test at P < 0.05.
^‡Calculated as shoot uptake divided by root dry weight (rdw).

significantly between the treatments, although there was a tendency for +AM treatments to be lower than –AM treatments regardless of the sterilization conditions (Table 3). The P concentration and content in the shoots and roots were significantly increased by +AM inoculation and tended to be higher under the S condition. The ratio of P/As content in the shoots was significantly higher in the +AM treatments than in the –AM treatments, while the P/As ratio in roots was not significantly different, although +AM treatments tended to be higher than –AM treatments. Under the NS condition, the shoot As uptake per root dry weight in the +AM treatment was significantly lower than in the –AM treatment, and the same tendency was observed under the S condition, although the difference was not significant (Table 3).

Concentration of water soluble P in soil

The concentration of water soluble P (WS-P) in the soils after sunflower cultivation was lower in the central compartment than in the outer compartment, especially in the +AM treatments (Fig. 2). Regardless of the AMF inoculation, the concentrations under S conditions were slightly higher than NS conditions.

Concentration and forms of WS-As in soils

In general, the dominant As species in the WS fraction was As^V in the outer compartment and As^III in the central compartment (Fig. 3). The lowest concentration of

Figure 2  Amounts of water soluble P (WS-P) in the soil collected from the rhizosphere and bulk compartments of the rhizobag system. Data are mean ± standard deviation for three replications. Bars on the top right indicate the least significant differences (LSD) for comparison between treatments (left), between compartments (middle) and between any pair of data (right) at P < 0.05. +AM, plus arbuscular mycorrhizal fungus inoculation; –AM, without arbuscular mycorrhizal fungus inoculation.

WS-As^V was observed in the central compartment in the –AM/NS treatment. In contrast, WS-As^III concentrations were higher in the central than in the outer compartment for all treatments.

Figure 3  Amounts of (A) water soluble (WS)-AsIII and (B) WS-AsV in the soil collected from different compartments of the rhizobag. Bars on the top right indicate the least significant differences (LSD) for comparison between treatments (left), between compartments (middle) and between any pair of data (right) at P < 0.05. +AM, plus arbuscular mycorrhizal fungus inoculation; –AM, without arbuscular mycorrhizal fungus inoculation.

Under S conditions, WS-DMAA was detected only in the central compartment of the +AM treatment. The concentration of WS-DMAA and WS-TMAO in the +AM/NS treatment was significantly higher in the central compartment than in the outer compartment (Fig. 4). While WS-DMAA was only detected in the +AM treatments and the values in the +AM/NS treatment were higher than those in the +AM/S treatment, WS-TMAO was only detected under NS conditions with or without AM inoculation and the amounts in the central compartment were relatively higher than those in the outer compartment. WS-MMAA was not detected in any soil samples (data not shown).

Acid phosphatase and dehydrogenase activity in soils

Acid phosphatase activity in the soil samples from the central compartment was significantly higher than that

Figure 4  Amounts of (A) water soluble dimethylarsinic acid (WS-DMAA) and (B) water soluble trimethylarsine oxide (WS-TMAO) in the soil collected from different compartments of the rhizobag. Bars on the top right indicate the least significant differences (LSD) for comparison between treatments (left), between compartments (middle) and between any pair of data (right) at P < 0.05. +AM, plus arbuscular mycorrhizal fungus inoculation; –AM, without arbuscular mycorrhizal fungus inoculation.

in the outer compartment (Fig. 5). Inoculation with AMF increased the activities, and the activity was higher in the NS condition than in the S condition. Dehydrogenase activity also followed similar trends to the phosphatase activity: that is, higher in the central compartment, and increased by AMF inoculation and decreased by soil sterilization (Fig. 5).

DISCUSSION

Sterilization of soils prior to experimental use resulted in the elimination of native mycorrhiza, while inoculated G. aggregatum successfully colonized sunflower roots even at high levels of As in the soils (Table 2). Root

Figure 5  (A) Acid phosphatase and (B) dehydrogenase activity in the soil collected from different compartments of the rhizobag planted with sunflower. Bars on the top right indicate the least significant differences (LSD) for comparison between treatments (left), between compartments (middle) and between any pair of data (right) at P < 0.05. +AM, plus arbuscular mycorrhizal fungus inoculation; –AM, without arbuscular mycorrhizal fungus inoculation; INF, iodonitrotetrazolium formazan.

colonization in the –AM/NS treatment indicated that indigenous mycorrhizal fungi were viable in soils with high levels of As. Significantly higher root colonization in the +AM treatments than in the –AM showed that G. aggregatum formed a stable association with sunflower in As-contaminated soil.

The AMF inoculation promoted plant growth and alleviated As toxicity symptoms in the sunflower plants, especially in the sterile soil. Higher growth in the sterile soil compared with the non-sterile soil could be attributed, in part, to an increase in nutrient availability in soils after sterilization (Table 1), especially P and NH₄-N (Sakamoto et al. 1992; Yamamoto et al. 2006). However, regardless of soil sterilization, AMF inoculation increased plant P uptake, but lowered the As concentration in shoots as well as the shoot As uptake per root dry mass. In contrast, the As concentration in roots was relatively higher in the +AM treatments than in the –AM treatment, and a significant increase in root As content was observed as a result of AMF inoculation under NS conditions (Tables 2,3). As evident from the high P/As ratio in AMF inoculated plants (Table 3), AMF inoculation reduced the translocation of As to the shoots and promoted P uptake, and increased the tolerance against As toxicity through an As dilution effect by growth enhancement (Meharg et al. 1994; Ultra et al. 2007).

These observations are consistent with our previous results and similar to the reports by Leung et al. (2005) and Liu et al. (2005), which showed that AMF restricted the transfer of As from the soil to the above-ground plant organs, resulting in reduced As concentration in shoots, although the overall As uptake was increased because of an AMF-induced increase in biomass. Several mechanisms have been proposed to explain how AMF affects As transport and translocation in tolerant and resistant plants. Gonzalez-Chaves et al. (2002) reported a reduction of As^V influx in AMF-colonized Holcus lanatus by suppression of a high-affinity phosphate/As^V transport system. In contrast, Sharples et al. (2000) showed that ericoid mycorrhizal fungus Hymenocyphus ericae decreased the net As uptake in plants by inducing As^III efflux from roots. It is also plausible that AMF will be able to sequester As by binding on hypae and other fungal tissues in the roots, thus restricting the transport and translocation of As.

The amount of WS-P in the soil after sunflower cultivation indicates the residual effects of AMF inoculation, plant uptake, soil sterilization and their interaction. As we have already discussed the As–P dynamics in soil as affected by AMF (Ultra et al. 2007), the following discussion focuses on the role of AMF in As transformation.

Distribution of WS-As between the compartments was different for As^V and As^III (Fig. 3). The increase in WS-As^III in the central compartment could be attributed to several factors/processes, such as higher mobility of As^III in the soil and transformation of As mediated by microorganisms in the rhizosphere (Cervantes et al. 1994; Fitz and Wenzel 2002). As^III, which is more mobile than As^V (Hering and Dixit 2005; Marschner 1995), could be carried by mass flow and accumulate with time in the rhizosphere, especially when transport exceeds uptake (Vetterlein and Jahn 2004).

In the case of As^V, a depletion zone could be formed near the root surfaces after plant uptake as with the case of P (Marschner 1995). Therefore, a lower amount of WS-As^V in the central compared with the outer compartment could be explained in part by plant uptake. However, it should be noted that a significant difference between the central and the outside compartment occurred in the –AM and not in the +AM treatment. This observation suggests an AMF-induced modification of microbial activities and community structures in the mycorrhizosphere (Marschner and Baumann 2003), which might have affected the speciation and dynamics of As.

In our previous experiment, we detected DMAA in the mycorrhizosphere of sunflower plants. Because methylated As species have been reported to be less toxic than inorganic As (Bentley and Chasteen 2002; Mukhopadhyay et al. 2002; Schmidt et al. 2004), we hypothesized that the chemical behavior and toxicity of As are altered in the mycorrhizosphere, which contributes to better survival of mycorrhizal plants in As-contaminated soil (Ultra et al. in press). However, it has not yet been clarified whether mycorrhizal roots alone can methylate As or whether AMF can indirectly induce the biomethylation process by indigenous microorganisms in the rhizosphere. In this experiment, either DMAA or TMAO was detected in the +AM/S or +AM/NS treatments, and these organic As species were not detected in the –AM/S treatment. These facts clearly indicated that mycorrhizal roots were primarily involved in As biomethylation in the rhizosphere of sunflowers.

It is noteworthy that in the sterile soils, DMAA was detected only in the central compartment of +AM treatments, which would serve as evidence that mycorrhizal roots were capable of DMAA formation in the rhizosphere. However, our experimental system could not determine whether or not AMF in association with sunflower roots is primarily responsible for As biomethylation. According to a report by Raab et al. (2005), sunflower plants can methylate As in the form of MMAA and DMAA. Therefore, we cannot deny the possibility that the DMAA detected in our experiment is formed by sunflowers or by some possible solubilization from other fractions in the soil by AMF or some other reason. However, taking into consideration the fact that DMAA was not detected in the central compartment of the –AM/S treatment, it is more likely that AM and mycorrhizal roots could synergistically enhance As biomethylation.

Furthermore, a higher concentration of WS-DMAA was detected in the central compartment compared to the outer compartment, especially in the +AM/NS treatment (Fig. 4). This suggests that the presence of AMF and mycorrhizal roots enhanced the capability of the indigenous microorganisms for biomethylation by means of stimulating their activity, especially in the rhizosphere (Turpeinen et al. 1999). The number of microorganisms in the rhizosphere is typically one order of magnitude larger than in the non-rhizosphere soil because of the continuous input of root-derived organic substrate, resulting in a more diverse, active and synergistic community (Jeffries et al. 2002; Marschner 1995). Concurrent evidence of higher microbial activity in the rhizosphere of +AM treatments is shown by high phosphatase and dehydrogenase enzyme activities (Fig. 5).

Trimethylarsine oxide was detected in measurable amounts only under NS conditions. This result would indicate that its formation is primarily mediated by indigenous microorganisms and less by AMF. However, in the central compartment of the +AM treatment, the concentration of WS-TMAO was higher than in the outer compartment of the +AM treatment and in the central compartment of the –AM treatment. These results imply that the presence of mycorrhizal roots has the ability to enhance TMAO formation. It has been generally accepted that DMAA is a precursor of TMAO (Thomas et al. 2004) and a substrate for fungal conversion to arsenobetaine and arsenolipids found in marine animals (Cervantes et al. 1994). Therefore, it would also be true in this experiment that TMAO was transformed from DMAA, which was converted from As^V. These As transformation processes could be possible mechanisms for As detoxification by AMF and sunflower roots. Thus, there is a possibility that AMF and mycorrhizal roots contribute to the formation of TMAO, indirectly through the supply of DMAA as a precursor of TMAO.

In conclusion, consistent with our previous report, AMF inoculation of sunflower grown on As-contaminated soil alleviated As toxicity by improving P nutrition and reducing As concentration in the shoots. It was clarified that mycorrhizal roots are primarily involved in DMAA formation, which would affect the fate of As in the rhizosphere soil. Aside from the mycorrhizal roots, indigenous soil microorganisms in the mycorrhizosphere may possibly be capable of promoting the transformation of DMAA into TMAO.

ACKNOWLEDGMENTS

This research was supported by a Sasakawa Scientific Research Grant (No. 17-207) from The Japan Science Society and by a Grant-in-Aid for Scientific Research (B, No. 15380223) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan. The authors gratefully acknowledge Dr Shingo Matsumoto, Shimane University, for his assistance in sampling the As-contaminated soils.

Notes

Present address: University of Eastern Philippines, Catarman, Northern Samar 6400, Philippines.