Introduction

Arsenic (As) contamination of groundwater has been frequently reported (e.g., Stone 2008; Zhu et al. 2008a, b; Meharg and McGrath 2009). In the As-affected areas of Bangladesh, groundwater contains up to 2 mg As L−1 as compared to the WHO recommended provisional limit of 0.01 mg As L−1. In areas where the arsenic-contaminated water is used for irrigation, soil As concentration can be up to 83 mg kg−1 (Abedin et al. 2002a, b). Three billion people, predominantly in Asia, eat rice as a staple. However, rice grown on As-contaminated soils usually contain high As levels in shoots (including grains) (Meharg 2004). Rice grains with As levels of 1.8 mg kg−1 have been recorded in the arsenic-affected tube well areas of Bangladesh (Zhu et al. 2008b; Williams et al. 2009). Therefore, the food that sustains half of the world’s population also increases the risk of cancer (Stone 2008). It is crucial that the physiology and genetics of rice uptake of As is understood to counteract this widespread contamination of the food chain (Meharg 2004).

Waterlogging is a typical characteristic growth condition for wetland plants including paddy rice, resulting in deficiency of oxygen and essential nutrients, low redox potential, and accumulation of Fe2+, Mn2+, H2S, S2−, HS−, and organic acids (McDonald et al. 2001). On the other hand, wetland plants have developed some special features, such as root anatomy (i.e., aerenchyma), radial oxygen loss (ROL), and the formation of iron (Fe) plaque on root surface, to cope with the adverse environmental conditions (Armstrong 1979; Colmer 2003a, b). The formation of Fe plaque could be an adaptation to stressed environments, as Fe hydroxides forming on roots of wetland plants are capable of reacting with metals and may therefore immobilize phytotoxic metals (Kuo 1986; Deng et al. 2010; Pi et al. 2010). Transport of oxygen (O2) from shoot to root via aerenchyma can occur in wetland plants and O2 can diffuse from root to rhizosphere, which is termed as radial oxygen loss (Armstrong and Armstrong 1994; Comer et al. 2006). ROL from root to the rhizosphere is essential for the detoxification of phytotoxins by direct oxidation or indirectly by oxidizing aerobic microorganisms maintained in the rhizosphere regions (Revsbech et al. 1999), which presumably contribute to the waterlogging tolerance of wetland plants (Colmer 2003a, b).

It has been found that Fe plaque and ROL are related with As tolerance and uptake in rice (Liu et al. 2004a, b; Chen et al. 2005; Mei et al. 2009). However, most of the previous studies have focused on As uptake, root anatomy, and Fe plaque conducted in solution culture (Liu et al. 2004a, b; Deng et al. 2010), which is substantially different from the real rhizosphere condition in the environment (Fitz and Wenzel 2002; Liu et al. 2006). Moreover, As toxicity depends not only on its total contents, but also on its chemical speciation, with inorganic As considered much more toxic than organic forms (Meharg et al. 2008; Norton et al. 2009a, b). It is therefore crucial to investigate As speciation in rice plants and grains to assess the associated human health risks.

The objectives of the present study were (1) to investigate the effects of As levels (low, medium, and high) on growth, Fe plaque formation, and the amount of ROL in rice; (2) to determine As accumulation, distribution, and speciation in rice plants grown in soils contaminated with low, medium, and high levels of As.

Materials and methods

Plant culture

Rice seeds of three genotypes (Nanyangzhan, Yuxiangyouzhan, CNT 87059-3) were germinated on moist filter papers and grown in Yoshida Nutrient solution (Yoshida et al. 1976) for 2 weeks. The genotypes were chosen based on their differences on radial oxygen loss and As accumulation levels according to the previous study (Wu et al. 2011). The plants were then transplanted into bags (30 μm nylon mesh, 6 cm diameter, 9 cm height, with one plant per bag) filled with acid-washed quartz sand. The sand–soil culture system was employed to mitigate the damage to rice roots for determination of ROL and Fe plaque, and to simulate the soil condition (Chen et al. 2005). The nylon bags were then transferred to a PVC pot (10 cm diameter, 14 cm height, with one plant per pot) and the gap between nylon bag and PVC pot was filled with 1.5 kg soil, previously air dried and sieved through a 2-mm sieve. Figure 1 shows the rhizobag design of the experiment. Soils were collected from a paddy field in the campus of South China Agricultural University (sandy clay with pH of 6.43 and average As concentration of 8.6 mg kg−1). Arsenic was added as arsenate (Na2HASO4·7H2O) with concentrations of 0, 50, and 100 mg As kg−1 dry weight (treatments designated as control, As 50, and As 100). Arsenate was added as a solution and mixed thoroughly with soils. Soils without arsenate amendment received the same volume of distilled water. All the soils were equilibrated for 2 weeks. There were six replicates in each treatment and each genotype.

Fig. 1
figure 1

Design of the rhizobag system

Full size image

Plants were allowed to grow under submerged conditions until maturity in a greenhouse (with a temperature of 25 °C during the day and 20 °C during the night, relative humidity of 70 %, and the natural light supplemented with sodium light [1,200 lx], with a photoperiod of 12:12 h [day/light]).

Iron plaque extraction

Upon harvest and before drying, half of root samples were extracted using dithionite–citrate–bicarbonate (DCB) (Otte et al. 1989) to measure As and Fe on the root surface, while the other half were used for determination of ROL. After washing in deionized water, 0.5 g root tissue was incubated for 60 min at room temperature (20–25 °C) in 40 ml solution consisting of 0.03 M sodium citrate and 0.125 M sodium bicarbonate, with addition of 0.6 g sodium dithionite. After incubation, the roots were rinsed three times with deionized water which was added to the DCB extract. The resulting solution was made up to 100 ml with deionized water for analysis.

Measurement of ROL

ROL measurements were determined 60 days (stem elongation stage) and 90 days (grain filling stage) after transplanting. Each nylon bag was removed carefully from the PVC pot, and roots were washed under tap water to remove any quartz particles adhering to the root surface. Care was taken during this process to ensure the roots were intact and exposed to open air for a minimal period for the determination of ROL.

The ROL rates of the entire root system of rice plants were measured using the titanium (III) (Ti3+) citrate buffer method (Kludze et al. 1994). The method was described in more detail in previous studies (Mei et al. 2009; Wu et al. 2011). The released O2 was calculated using the following formula (Kludze et al. 1994):

$$ \mathrm{ROL}=\mathrm{c}\left(\mathrm{y}-\mathrm{z}\right) $$

where ROL = radial oxygen loss, μmol O2 plant−1 day−1, c = initial volume of Ti3+ added to each test tube, L; y = concentration of Ti3+ in solution of control (without plant), μmol Ti3+ L−1; and z = concentration of Ti3+ in solution after 6 h treatment with plants, μmol Ti3+ in solution plant−1 L−1.

$$ \mathrm{Rate}\ \mathrm{of}\ \mathrm{ROL}=\mathrm{c}\left(\mathrm{y}-\mathrm{z}\right)/\mathrm{G} $$

where rate of ROL = rate of radial oxygen loss, μmol O2 g−1 dry weight day−1; G = root dry weight, g.

Plant analysis

Plant analysis for total As

Sampling procedures followed those described by Abedin et al. (2002a) and Liu et al. (2006). At maturity, after measuring plant height, half of the plants were harvested, carefully washed, and separated into grains, straws, and roots, then oven-dried at 50 °C. Another half of plants were freeze-dried, and stored at −20 °C for the analysis of As speciation. After recording the dry weight, all plant samples were ground to fine powder, and digested with 5 ml HNO3 until the digestion solution became clear. The certified reference material [(CRM) 1568a rice flour from National Institute of Standards and Technology, USA (NIST)] was used to verify the accuracy of metal determination. The acid digests of plant material (grains, straws, and roots) and DCB extracts were analyzed for total As and Fe [determined by inductively coupled plasma spectrometer (ICP), PerkinElmer, Elan 9000] (Allen 1989). The recoveries of As in 1568a ranged from 105.1 to 107.3 %.

Plant analysis for As speciation

Two genotypes (Nanyangzhan and Yuxiangyouzhan) were used in this investigation. The speciation method is described in more detail in Wu et al. (2011). Trifluoroacetic acid (TFA) extraction method was used (Heitkemper et al. 2001; Williams et al. 2005). Milled subsamples (0.2 g) were weighed into quartz digestion tubes and 2 mL of 2 M TFA was added. The mixture was allowed to stand overnight. The tubes were then placed on a heating block at 100 °C for 6 h. The digest was evaporated to dryness at 120 °C. The residues were resuspended in distilled water and filtered through a 0.45 μm filter (cellulose nitrate, Micro Filtration Systems, California, USA), then made up to 10 mL with ultrapure (>18 MΩ) deionized water before analysis.

A Hamilton PRP-X100 10-μm anion-exchange column (4.1 × 150 mm) with an appropriate precolumn (containing the same material) and an Agilent 1100 series HPLC system (Agilent Technologies) were used for all analyses. The mobile phase employed for anion-exchange chromatography, consisted of ultrapure (>18 MΩ) deionized water and 50 mM ammonium bicarbonate (from Aldrich Chemical Co.).

Each analysis was performed within 24 h of sample extraction to minimize any changes in speciation during prolonged storage. Post-column element-specific detection of arsenic was achieved using an ICP-mass spectrometer (PerkinElmer, Elan 9000). NIST CRM 1568a rice flour was used to validate the method, which was also used to characterize its speciation (Williams et al. 2005; Liu et al. 2006). The mean total recovery [(sum of species recovered from the TFA extraction/total As from acid digestion) × 100 %] ranged from 83–111 %, which was consistent with other studies (Heitkemper et al. 2001; Williams et al. 2005).

Statistical analyses

Analysis of variance (ANOVA) on plant biomass, concentrations of As and Fe, and rates of ROL was performed using the statistical package SPSS 13.0 for Windows (SPSS Inc., USA).

Results

Effects of As on growth, iron plaque formation, and ROL of entire roots in three genotypes of rice

Plant height varied from 84 to 98 cm among the three genotypes. The As treatments did not exert a significant effect on the height of rice plants (P > 0.05); and there was no significant genotypic effect on the height of rice plants (P > 0.05). Root biomass varied from 2.8 to 5.9 g pot−1, with the highest root biomass in As 100 treatment of genotype Nanyangzhan (Table 1). Straw biomass varied from 6.1 to 8.2 g pot−1 (Table 1). Application of As did not significantly (P > 0.05) change the biomass of root and straw, while there was a significant genotypic effects (P < 0.05) on the biomass of root and straw. There were significant genotypic effects on grain yield (P < 0.01), but there were no significant As treatment effects on grain yield (P > 0.05, Table 1).

Table 1 Biomass (g pot−1, DW) and As concentrations (mg kg−1) in roots, straws, and grains of three genotypes of rice grown in soils amended with different concentrations of arsenic (0, 50, and 100 mg kg−1)
Full size table

Iron plaque was clearly visible as reddish coatings on the root surface when harvested. There were no significant differences (P > 0.05) in the amounts of Fe plaque formed for different genotypes and for different treatments (Table 2). However, the increase of As led to a slight increase (P > 0.05) of Fe plaque formed on root surface. Arsenic concentration in Fe plaque ranged between 29 mg kg−1 in control treatment and 665 mg kg−1 in As 100 treatment (Table 2). There were significant differences in As concentration in Fe plaque between As treatments (P < 0.01), while no significant difference between genotypes (P > 0.05).

Table 2 Concentrations of Fe (g kg−1) and As (mg kg−1) in DCB extracts of three genotypes of rice grown in soils amended with different concentrations of arsenic (0, 50, and 100 mg kg−1)
Full size table

Rice plants showed relatively higher rate of ROL on day 60 compared with day 90, showing the decrease of ROL in mature stage of rice compared with stem elongation stage. Yuxiangyouzhan showed the highest total amount of ROL in all treatment groups, while Nanyangzhan and CNT 87059-3 had relatively lower amounts in all treatments. On days 60 and 90, increasing As concentration caused a decrease in total ROL rates (P < 0.01; Fig. 2). This suggested that total ROL was inhibited by arsenic stress. Total ROL on entire roots varied significantly referred to genotypes × treatments combination (P < 0.05) according to the two-way ANOVA (Fig. 2). There were significant differences between genotypes in ROL rates only in As 100 treatment (P < 0.05).

Fig. 2
figure 2

Total amounts of radial oxygen loss (ROL) from entire root systems of three genotypes of rice growing in treatments amended with different concentrations of arsenic (0, 50, and 100 mg kg−1) for 60 days (a) and 90 days (b). All data are shown as means ± standard deviation; different letters in each genotype indicated that they were significantly different in three treatments at P < 0.05 determined by Tukey’s HSD test

Full size image

Arsenic uptake and speciation in rice plants

Arsenic concentration in root increased significantly (P < 0.001) with the increase of soil As concentration, and As concentrations in rice straws or grains showed significant (P < 0.01) differences between As treatments. Arsenic concentration in grains remained statistically similar in As 50 and As 100 treatments, which implicated a lower translocation from straws to grains for rice as a protection process. There were significant genotypic effects on As concentrations in the straw and grain but not root (P < 0.05; Table 1).

According to the World Health Organization’s 10 μg l−1 limit for As in drinking water, 0.05 mg kg−1 As in rice contributes about 60 % of dietary As exposure (WHO 2001). The maximum contaminant levels for inorganic As in rice grains was set at 0.2 mg kg−1 in China (Chinese Food Standards Agency 2012). In this investigation, As concentrations in grains of rice grown in soils spiked with different As levels exceeded 0.2 mg kg−1. It reflects a great risk when growing rice with As-contaminated soils in the field. However, the situation in the field may be different, due to the combination of different environmental factors, soil As speciation and behavior, water management, cultural practices, and genetic differences.

Arsenic species, As(III), As(V), DMA, and MMA were analyzed in different parts of rice for two genotypes (Yuxiangyouzhan and Nanyangzhan). There were genotypic differences (P < 0.05) in levels of DMA and inorganic As (Asi) in the grain, with Yuxiangyouzhan having higher inorganic As in grains than Nanyangzhan. The percentages of inorganic As decreased with increasing As concentrations, while percentages of DMA increased (Fig. 3). Inorganic As was the predominant As species in the root and straw, accounting for 94–99 % of the total As, whereas grains contained substantially higher DMA, accounting for 53–70 % of the total As in As 100 treatment (Fig. 3).

Fig. 3
figure 3

As speciation a in grains of rice plants grown in soils amended with different concentrations of arsenic (0, 50, and 100 mg kg−1); b in different parts of rice plants grown in soils with 100 mg As kg−1. White bar inorganic arsenic, hatched bar DMA, gray bar MMA. All data are shown as means − standard deviation

Full size image

Discussion

Lower shoot biomass of rice subjected to high arsenate treatment has been observed by Marin et al. (1993) and Abedin et al. (2002a). However, stimulation of growth by arsenate addition has also been reported for rice (Marin et al. 1992; Carbonell et al. 1998). The present study showed no significant reduction in plant biomass when subjected to different As levels, possibly because of the different growth conditions. Arsenate addition may displace phosphate from the soil in certain situations, increasing plant P availability (Jacobs et al. 1970).

The waterlogged anoxic conditions of wetland plants would cause the accumulation of potentially toxic, reduced solutes, such as Fe2+, Mn2+, and Pb2+ in the pore water. On the other hand, wetland plants can develop some special features, such as radial oxygen loss, root anatomy, and the formation of Fe plaque on root surface, to cope with the adverse environments (Armstrong 1979; Visser et al. 2000). Iron plaque could serve as a barrier to prevent excess pollutants entering roots of wetland plants (Pi et al. 2010; Liu et al. 2004a, b). A slight increase of Fe plaque formed in As added treatments was observed in the present study. This might be due to the tolerance strategy demonstrated by wetland plants and rice to cope with pollutants (Pi et al. 2010). Pi et al. (2009) observed significantly higher amounts of Fe plaque formed on the root surface in two mangrove plants treated with wastewater discharge. This was due to aerobic degradation of nutrients and organic matter by microorganisms, leading to a more anaerobic soil environment and induced more ROL around the rhizosphere to oxidize ferrous ions, thus formed more Fe plaque. Hu et al. (2007) reported the sulfur-induced enhancement of plaque formation, probably due to an increase in concentrations of Fe2+ and Mn2+ resulting from S transformation in soil. In the present study, the stress of As caused a slight increase of iron plaque formation, but there was no significant difference between As treatments and control. It may be due to the fact that As has influenced the rhizosphere microbial activities. However, more investigations are needed to further our understanding of the mechanism. Garnier et al. (2010) have shown that the As contents of roots and Fe plaque raise to 1,000–1,500 mg kg−1 towards the middle of the growth season, then decline to ∼300 mg kg−1 in the field. The As contents in roots and Fe plaque in the present study were within this range in the field (Table 2).

ROL could be affected by environmental factors, such as the oxygen content, light condition, redox potential, and microbial oxygen demand (Laskov et al. 2006). Rahman et al. (2007) revealed that the content of photosynthetic pigments were reduced, and subsequently affect photosynthesis, under As treatments. Furthermore, Connell et al. (1999) indicated that photosynthesis can affect ROL of Halophila ovalis roots. The decrease of ROL might be due to the decrease of the photosynthetic pigments and photosynthesis, which further affected rates of ROL in rice. However, this should be clarified further in future studies. Cheng et al. (2010) suggest that ROL from root tip is a potential biomarker of environmental pollution (such as heavy metals). The decreases of ROL observed in rice are mainly related to the alteration of root anatomical structure and decreased root porosity induced by heavy metals (Liu et al. 2009; Mei et al. 2009; Cheng et al. 2010). The reduced permeability of the roots which are induced by the pollutants seems to be a defense response to prevent excessive toxins entering the root and causing possible fungal infection (Hose et al. 2001; Armstrong and Armstrong 2005).

The concentration of Fe plaque is profoundly influenced by the amount of the ROL of plant roots, the more ROL present around the rhizosphere would induce more Fe plaque formation on the root surface of wetland plants (Armstrong 1979; Otte et al. 1991). However, Møller and Sand-Jensen (2008) found that if excessive Fe plaque formed on the root surface of Lobelia dortmanna, it may act as a ‘barrier’ and prevent oxygen from being released from the root, leading to lower ROL around rhizosphere. Another study also showed that the concentration of Fe plaque formed in two mangrove plants was negatively correlated with the rate of ROL along the lateral root when treated with wastewater (Pi et al. 2010). The present results of the significantly decreased ROL and slightly increased Fe plaque formation in As treatments indicated that the increased Fe plaque might inhibit oxygen release from the roots.

Zavala and Duxbury (2008) speculated that As speciation in rice grain is under genetic control. However, it has been demonstrated that As speciation in rice grain can be strongly influenced by the environmental conditions such as watering regime and As bioavailability in soils (Xu et al. 2008; Arao et al. 2009; Li et al. 2009). Zhao et al. (2013) reviewed past literature and indicated that grain As speciation are primarily attributed to environmental factors, and methylated As species in rice are derived from the soil, while rice plants lack the As methylation ability. The present study demonstrated that As speciation varied between different genotypes and As treatments. The genotypic variation of As speciation may be due to the variation in the root uptake or the internal translocation efficiency of methylated As of different genotypes (Zhao et al. 2013).

Studies showed that the majority of As present in rice grain were DMA or inorganic As, while in roots and straws the majority of As were inorganic As (Smith et al. 2008; Zavala et al. 2008; Zheng et al. 2011). Zheng et al. (2011) reported that the unloading of inorganic As and DMA into rice grain are different, with the latter accumulating mainly in the caryosis before flowering and inorganic As mainly transported into the caryopsis during grain filling. Moreover, inorganic As is considered more toxic than methylated As (Abedin et al. 2002a, b; Zhu et al. 2008a, b; Zhao et al. 2009), indicating As speciation exerts important implications for human health.

Conclusions

The stress of As caused a slight increase of iron plaque formation (P > 0.05), a decrease in the rates of ROL (P < 0.01). The results of As speciation showed that the percentages of DMA increased from 19–28 to 53–58 %, while the percentages of inorganic As decreased from 53–58 % to 36–42 % with the increasing soil As concentrations, indicating a strong environmental influence on As species in rice grain. The present study provided useful information on As tolerance and accumulation in rice to reduce the health risk posed by As contamination in rice.