Methylated Arsenic Species in Rice: Geographical Variation, Origin

Mar 22, 2013 - Rice is a major source of inorganic arsenic (iAs) in the human diet because paddy rice is efficient at accumulating As. Rice As speciat...
0 downloads 13 Views 376KB Size
Download PDF
Critical Review pubs.acs.org/est

Methylated Arsenic Species in Rice: Geographical Variation, Origin, and Uptake Mechanisms Fang-Jie Zhao,*,†,‡ Yong-Guan Zhu,§ and Andrew A. Meharg∥ †

College of Resources and Environmental Sciences, Nanjing Agricultural University, Nanjing 210095, China Rothamsted Research, Harpenden, Hertfordshire AL5 2JQ, U.K. § Key Lab of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen, China ∥ Institute for Global Food Security, Queen’s University Belfast, David Keir Building, Stranmillis Road, Belfast BT9 5AG, U.K. ‡

ABSTRACT: Rice is a major source of inorganic arsenic (iAs) in the human diet because paddy rice is efficient at accumulating As. Rice As speciation is dominated by iAs and dimethylarsinic acid (DMA). Here we review the global pattern in rice As speciation and the factors causing the variation. Rice produced in Asia shows a strong linear relationship between iAs and total As concentration with a slope of 0.78. Rice produced in Europe and the United States shows a more variable, but generally hyperbolic relationship with DMA being predominant in U.S. rice. Although there is significant genotypic variation in grain As speciation, the regional variations are primarily attributed to environmental factors. Emerging evidence also indicates that methylated As species in rice are derived from the soil, while rice plants lack the As methylation ability. Soil flooding and additions of organic matter increase microbial methylation of As, although the microbial community responsible for methylation is poorly understood. Compared with iAs, methylated As species are taken up by rice roots less efficiently but are transported to the grain much more efficiently, which may be an important factor responsible for the spikelet sterility disorder (straight-head disease) in rice. DMA is a weak carcinogen, but the level of ingestion from rice consumption is much lower than that of concern. Questions that require further investigations are identified.

1. INTRODUCTION

2. ARSENIC SPECIATION IN RICE FROM MARKET-BASKET SURVEYS

It has now become clear that rice is a major source of inorganic As (iAs), a class one carcinogen, in the human diet.1,2 Rice is normally cultivated in flooded paddy soil, an environment that leads to a mobilization and, hence, a much enhanced bioavailability of As to rice plants.3−6 Rice is also a strong accumulator of the macro-nutrient silicon, an element that plays an important role in the defense against a range of biotic and abiotic stresses.7 Arsenite, which can be the dominant form of As in flooded soils, is a silicic acid analogue and is assimilated by roots via silicic acid transport systems.8 However, iAs species are not the only As species found in rice; several methylated As species, predominantly dimethylarsinic acid (DMA), are also present with variable proportions.1 Although several recent reviews have addressed As uptake and metabolism in plants focusing on the iAs species (e.g., 9, 10) less is known about the origin and the transport mechanisms of methylated As, and the implications of these As compounds on rice production or potential health risk to humans. These are the aspects reviewed here, with the knowledge gaps also being identified. © 2013 American Chemical Society

Studies have shown that As speciation in rice grain is dominated by iAs and DMA.1,11−13 MMA (monomethylarsonic acid) is occasionally detected in some samples but it is present only as a minor component. Tetramethylarsonium was found in rice grain samples collected from a contaminated paddy field in China, accounting on average for 5.8% of the total As concentration.14 Inorganic As in rice comprises mainly arsenite (As(III)), although conversion between As(III) and arsenate (As(V)) may occur during some extraction procedures and, for this reason, the sum of the two species is usually reported as total iAs. In-situ speciation analysis using synchrotron-based Xray absorption near edge structure (XANES) suggests that some As(III) in rice grain may be complexed by thiol groups of Received: Revised: Accepted: Published: 3957

October 21, 2012 February 27, 2013 March 22, 2013 March 22, 2013 dx.doi.org/10.1021/es304295n | Environ. Sci. Technol. 2013, 47, 3957−3966

Environmental Science & Technology

Critical Review

Figure 1. Regional differences in arsenic speciation in rice from market-basket surveys. (A) and (B) are plots of inorganic As concentration or inorganic As percentage versus total As (sum of As species) concentration, respectively; (C) and (D) are plots of DMA concentration or DMA percentage versus total As (sum of As species) concentration, respectively. Data are from 11−13, 17, 18. MMA was present as a minor component in a few samples and was included in the sum of methylated As species. For the data of 17, the difference between total and iAs was assumed to represent DMA.

sulfur-rich amino acid oligomers.15 However, these complexes dissociate during the extraction and analysis unless specific care is taken to preserve them.16 Both the concentrations and the relative percentages of iAs and DMA vary widely among samples of rice collected from market basket surveys.1 Williams et al.11 compared As speciation in commercial polished rice produced in Bangladesh, India, Europe, and the United States. They found high percentages of iAs (∼80%) in Bangladeshi and Indian rice. In comparison, European and U.S. rice had lower percentages of iAs with a mean of 64% and 42%, respectively, with corresponding high percentages of DMA. Similarly, Zavala et al.12 analyzed 24 U.S. rice samples and found on average 46% iAs with DMA making up the rest. Zhu et al.13 reported high percentages of iAs in Chinese rice, with a mean of 78%, which is similar to the Indian and Bangladeshi rice. Torres-Escribano et al.17 reported a mean iAs percentage of 62% for Spanish rice. Adomako et al.18 reported low concentrations of total As in Ghanaian rice with high percentages of iAs. When the data from different market-basket surveys are combined (Figure 1), some broad regional differences in the relationships between iAs or DMA with total As concentration (calculated from the sum of the As species) become apparent. For the rice produced in Asia (including Bangladesh, India, China, and Thailand), there is a strong linear relationship between iAs and total As (R2 = 0.97, n = 77); the slope of the regression is 0.78, meaning that on average 78% of the total As in Asian rice is inorganic (Figure 1A). A few Ghanaian rice samples also fall on this regression line, but in the low

concentration range. The U.S. rice, in contrast, shows a hyperbolic pattern in the relationship, approaching a maximum of approximately 0.15 μg g−1 iAs. European (Italy, Spain, and France) rice samples appear to be more variable, and the iAs/ total As relationship exhibits a pattern that is intermediate between those of Asian and U.S. rice. The relationships between DMA and total As are linear for the rice produced in the three regions, but the slope decreases in the order of U.S. > European > Asian (Figure 1B). In general, the iAs percentage tends to decrease, while the DMA percentage increases, with the total As concentration in the U.S. and European rice (Figure 1C and 1D). These trends are not apparent in the Asian rice. Within the combined data set, the percentages of iAs and DMA vary from 10 to 100% and from 0 to 90%, respectively.

3. ENVIRONMENTAL VERSUS GENETIC VARIATION IN THE ARSENIC SPECIATION IN RICE GRAIN The apparent regional differences in rice As speciation prompt Zavala et al.12 to classify rice into the iAs or DMA types and to suggest that As speciation is under genetic, rather than environmental, control. However, emerging evidence shows that U.S. rice cultivars are unlikely to be genetically different from others with regard to As speciation, but rather the regional pattern is primarily controlled by environmental factors. When different rice cultivars were grown under the same conditions in pot or field experiments, significant differences in As speciation were indeed found.11,19−24 For example, Liu et al.19 showed that the percentage of iAs varied from 40 to 70% and that of 3958

dx.doi.org/10.1021/es304295n | Environ. Sci. Technol. 2013, 47, 3957−3966

Environmental Science & Technology

Critical Review

with total As concentration, surpassing the concentration of iAs at the total As concentration of around 0.6 mg kg−1. When the data were expressed as percentages of the total As concentration, it can be seen that the iAs% decreased, while DMA% increased, with increasing total As concentration in a hyperbolic pattern (Figure 2). Therefore, for a given rice cultivar, environmental variation (different soils or different water management regimes) can cause iAs percentage to vary from 100% to 20% and, concurrently, DMA% to vary from 0% to 80%. These ranges are nearly wide enough to cover the entire ranges reported for market rice samples analyzed to date (Figure 1), although this statement does not mean that genotype has no influence on As speciation.

DMA ranged from 30 to 55% among six rice cultivars grown in a pot experiment. In another pot experiment, twenty rice cultivars varied in the percentage of iAs from 19 to 95% and in the percentage of DMA from 2 to 81%.21 Norton et al.22 grew a common panel of rice cultivars at six field sites in Bangladesh, India, and China and analyzed As speciation in the grain samples of seven cultivars. They found significant effects of site, genotype, and genotype × site interactions on As speciation, but the effect of site outweighed that of genotype substantially according to the results from a two-way analysis of variance (for iAs percentage the F ratios were 32.5 and 6.6 for site and genotype, respectively, and for DMA percentage the corresponding values were 45.5 and 6.8, respectively). Pillai et al.24 grew ten rice cultivars belonging to the indica or japonica subtype and originating from the U.S. or Asia in a paddy field in Arkansas, U.S. They found high percentages of DMA in the grain from all cultivars (48−77%, mean 63%), similar to the data from the market basket surveys. There was no evidence that Asian cultivars belong to the “iAs type” and the U.S. cultivars belong to the “DMA type” as suggested by Zavala et al.12 Environmental influence on As speciation in rice grain is clearly demonstrated in experiments that used the same genotype but varied soil or environmental conditions.3,20,25 Maintaining anaerobic soil conditions during rice growth greatly increased As bioavailability and As accumulation in rice grain of a single rice cultivar; this also resulted in a large effect on grain As speciation by increasing the percentage of DMA.3,20 When rice was grown under aerobic conditions, very little DMA was present in rice grain. In another pot experiment testing a range of soils from Bangladesh, India, China, and the UK with a single rice cultivar, the percentage of iAs decreased, while the percentage of DMA increased, with increasing total As concentration.26 Although the two types of experiments described above used two different rice cultivars, the results regarding the relationships between As speciation and total As concentration in rice grain were similar; hence the data are combined in Figure 2. It is clear that the concentration of iAs increased with increasing total As concentration in a hyperbolic pattern, reaching a maximum of approximately 0.5 mg kg−1, whereas DMA started from undetectable levels at low total As concentrations and then rose in a slightly exponential pattern

4. ORIGIN OF METHYLATED ARSENIC SPECIES IN PLANTS Because DMA can represent a substantial proportion of the total As in rice grain, it is pertinent to ask where the methylated As species originate. It was previously thought that higher plants were able to methylate As, thus converting iAs into organic forms.12,27,28 There are many reports on the presence of methylated As species in terrestrial plants either collected from field surveys or grown in controlled environmental conditions, usually in small percentages of the total As in the plant tissues.29−35 However, the presence of methylated As species in plants grown in soil or other nonsterile media does not necessarily indicate an ability of plants to methylate As, because methylated As species may originate from the media, either from the residues of the past uses of methylated As compounds or from the products of microbial methylation. This issue has been addressed recently by Lomax et al.,36 who grew rice, tomato, and clover in axenic cultures with additions of either arsenate or arsenite. They did not detect any methylated As species in the plants under these conditions. When rice plants were exposed to MMA(V) in axenic culture, this As species was taken up by roots and partly reduced to MMA(III) but was not further methylated to DMA.36 In contrast, in the same experiment rice grown in nonsterile nutrient solution or nonsterile soil contained small amounts of methylated As species. Further evidence for a lack of As methylation ability in rice comes from a recent study by Jia et al.37 They grew rice plants in a sterile soil using a two-chamber system that allowed separate collections of volatile As from the above-ground plant tissues and from the soil. Volatilization of trimethylarsine (TMA) from the plants was detected only when trimethylarsine oxide (TMAO) was added to the sterile soil, not in the treatments with the additions of iAs, MMA, or DMA. These results indicate that rice plants were able to reduce TMAO to volatile TMA, similar to previous observations on the ability of plants to reduce MMA(V) to MMA(III),36,38,39 but were not able to transfer methyl groups to As. In a hydroponic experiment under nonsterile conditions, DMA was detected in the nutrient solution after the addition of arsenite or MMA, whereas an addition of the antibacterium agent chloramphenicol dramatically suppressed the production of DMA.40 Taken together, it can be concluded that methylated As species such as DMA in rice are derived from microbial methylation in the medium. This may also be true for other plant species, at least for tomato and clover that have been investigated,36 although more plant species should be tested under axenic conditions and, if possible, to plant maturity. There is a single report in the literature on the measurement of the in vitro activity of As methylation in the cell extracts of

Figure 2. Arsenic speciation in rice grain as influenced by soil and environmental conditions in pot experiments. (A) and (B) are plots of As species concentration or As specie percentage versus total As concentration, respectively. The soil experiment used the indica cultivar Shatabdi26 while the water management experiments used the japonica cultivar Oochikara.3,20. 3959

dx.doi.org/10.1021/es304295n | Environ. Sci. Technol. 2013, 47, 3957−3966

Environmental Science & Technology

Critical Review

bentgrass (Agrostis capillaris).27 These authors used S-[3Hmethyl]adenosyl-L-methionine (3H-SAM) with either arsenite or arsenate as substrate and measured the 3H radioactivity in the supposedly methylated As fractions. It remains unproven whether 3H-methyl group had been transferred to As in the assay because no direct evidence was provided. In a study using cell suspension cultures of the Madagascar periwinkle (Catharanthus roseus), Cullen et al.41 briefly mentioned that a small proportion of MMA added to the growth medium was further methylated to DMA. Whether the experiment was maintained in axenic conditions was not stated. Given that rice appears to lack an in planta ability to methylate As, the genotypic variation in the As speciation in rice grain observed in numerous studies is likely to be caused by the variation in the root uptake or the internal translocation efficiency of methylated As, or other indirect factors discussed below.

uptake efficiency of methylated As versus iAs. For example, Marin et al.47 reported an efficient uptake of MMA by rice roots compared with the uptake of arsenate, probably because the nutrient solution was maintained at a low pH (4.0), whereas at higher pHs (∼6), MMA was taken up less efficiently than iAs.42,48 In contrast to the decreasing uptake efficiency, the translocation from roots to shoots generally increases with the increasing number of methyl group in the As compounds.37−39,42,44,47 DMA is highly mobile during both xylem and phloem transport in rice and castor bean.39,49,50 Carey et al.50 showed that DMA was taken up from the feeding solution by cut panicles of rice and transported to the grain at a much greater efficiency (>10 fold) than arsenite. Moreover, DMA was delivered to the grain via both the xylem and the phloem, whereas for arsenite the main route of transport was via the phloem. When As species were fed to cut flag leaves, DMA and MMA were also efficiently translocated to the grain, whereas arsenate was largely reduced to arsenite and retained in the flag leaves.49 Lomax et al.36 exposed rice plants to 10 μM arsenite or DMA at different growth stages in a hydroponic experiment. At maturity, the ratio of grain to straw As concentration was about 2 orders of magnitude larger in the DMA treatments than in the arsenite treatments, indicating that DMA was preferentially accumulated in the rice grain, in contrast to iAs which was accumulated mainly in the vegetative tissues. It is now clear that, although methylated As species are not absorbed by roots as efficiently as iAs species, they are much more readily transported to the reproductive organs such as rice grain. In contrast, iAs, mostly in the oxidation state of As(III), accumulated mainly in the vegetative tissues, especially roots and to a lesser extent stems and leaves, with only a very small proportion being translocated to the grain.50−52 Due to this marked difference in the mobility within plant tissues, the relative abundance of DMA in rice grain can be magnified considerably compared with that in the soil solution. For example, it has been shown that the rate of DMA uptake by rice roots is about one twentieth of that of arsenite45 but the translocation to grain is approximately 100 times greater;36 thus, a concentration ratio of DMA to inorganic As in the soil solution of 1:5 could lead to equal proportions of the two As species in rice grain. However, the reasons behind these different mobilities are still unclear. One possible explanation is that arsenite, either absorbed by roots or produced from the reduction of arsenate, is readily complexed by thiol compounds such as phytochelatins (PCs) 9,53,54 and the complexes are subsequently sequestered in the vacuoles,55,56 resulting in decreased mobility within plant tissues. This hypothesis is supported by the evidence that the Arabidopsis thaliana mutants lacking PC synthesis have a higher root to shoot translocation ratio of As than the wild-type plants.54 Suppressing PC synthesis by a chemical inhibitor also led to an increased As mobility.54,57 MMA(III) (reduced from MMA) can also be complexed by PCs, but perhaps not as extensively as arsenite, whereas no DMA-PC complexes have been detected.53 The lack of DMA− PC complexation may allow this methylated As species to move more freely between plant cells and tissues. However, why DMA is preferentially accumulated in rice grain is not known. Membrane transporters that mediate the loading and unloading of DMA into rice grain remain to be identified. The high mobility of DMA compared with iAs occurs not only in the translocation from roots to shoots and from leaves

5. UPTAKE AND TRANSLOCATION OF METHYLATED ARSENIC SPECIES BY PLANTS Plant roots are able to take up MMA, DMA, or TMAO, but the rates of uptake are lower than those of iAs species and also decrease with increasing numbers of the methyl groups.37,38,42−45 For example, Raab et al.42 compared the uptake of arsenate, MMA, and DMA by 46 plant species in hydroponic experiments and found the mean root absorption factor for arsenate was 5 and 2.5 times higher than that for DMA and MMA, respectively. Increasing hydrophobicity may be a reason for the decreasing uptake rate of methylated As species. Li et al.38 demonstrated that undissociated molecules of MMA and DMA can permeate through the aquaporin channel OsNIP2;1 (also called Lsi1 as it was initially identified as a silicic acid transporter46) in rice roots. This transporter has previously been shown to be a major entry route for arsenite into rice roots.8 Compared with the normal wild-type plant, a rice mutant with a single amino acid mutation in the OsNIP2;1 protein lost about 80% and 50% of the uptake capacity for MMA and DMA, respectively.38 When the OsNIP2;1 gene was expressed in the Xenopus laevis oocytes, MMA uptake increased significantly compared with the control. In the same study, DMA uptake into oocytes was very limited and was not significantly enhanced by the heterologous expression of OsNIP2;1; however, more recent studies in our laboratory have shown a significant increase in DMA uptake by OsNIP2;1 expression (Lomax et al., unpublished). It is thus clear that the OsNIP2;1 aquaporin channel, which is highly expressed in rice roots,7 is permeable to a range of neutral molecules including silicic acid, arsenite, MMA, and DMA. While silicic acid and arsenite have a high pKa (9.2) and, therefore, are present mostly as neutral molecules under neutral or acidic conditions, MMA and DMA have relatively low pKas (4.2 and 6.1, respectively). As a result, MMA and DMA will dissociate significantly in the acidic to neutral pH range. Because aquaporin channels are permeable only to uncharged molecules, pH can influence the uptake of MMA and DMA by shifting the equilibrium between protonation and dissociation, thus affecting the availability of the substrate for the membrane transporters. Indeed, Li et al.38 showed that increasing pH from 4.5 to 6.5 decreased MMA uptake by rice greatly, while increasing pH from 5.5 to 6.5 also decreased DMA uptake considerably. This pH dependency also explains the inconsistency in the literature with regard to the ranking of the 3960

dx.doi.org/10.1021/es304295n | Environ. Sci. Technol. 2013, 47, 3957−3966

Environmental Science & Technology

Critical Review

containing the three strictly conserved cysteine residues,65 supporting the notion that rice lacks the As methylation ability. Using a microarray-based Geochip that contains ArsM probes for 66 microbial species, Lomax et al.36 detected the presence of 35 sequences in a Bangladashi paddy soil contaminated with As from irrigation with high As groundwater, including several sulfate-reducing bacteria and methanogens which are likely to be particularly active under anaerobic conditions of flooded paddy soils. Because the Geochip probes are designed based on the known sequences in the Genebank, it is highly likely that many more soil microbes are able to methylate As. The dominant As methylation microbes may be different among different soils or change with the soil condition (e.g., the redox potential). It is also possible that different rice genotypes could modify the rhizosphere differently, thus affecting the As methylation microbes in the rhizosphere and indirectly influencing the accumulation of DMA by rice. To date, there is little knowledge of the key microbial species responsible for As methylation in different soils, and how their activities are regulated. The evidence that ArsM mediates microbial As methylation has been provided by several recent studies. Qin et al.66 investigated the function of ArsM from the soil bacterium Rhodopseudomonas palustris by heterologous expression in an As-sensitive strain of Escherichia coli. They showed that ArsM catalyzed the formation of a number of methylated intermediates from As(III), with trimethylarsine as the end product that was volatilized. The ArsM gene from R. palustris has been transferred to rice.67 The transgenic plants produced 10-fold more volatile As than the untransformed control, although the amount of As volatilized accounted for a very small percentage (0.06%) of the total As in the plants. More recent studies have demonstrated the As methylation ability of ArsM from a thermoacidophilic eukaryotic alga and three cyanobacteria.68,69 Methylation of iAs leading to the production of volatile TMA(III) is considered a detoxification mechanism in microorganisms because volatilization lessens the cellular burden of As. This also explains why ArsM is often located downstream of the As-responsive transcriptional repressor ArsR in prokaryotic and archaeal microoganisms.63 Challenger proposed a pathway for As methylation in fungi that involves sequential reduction of As(V) and oxidative methylation in which a methyl group is added to As(III), with the reactions being repeated leading to the production of volatile TMA(III).70 This scheme has been widely documented and considered plausible.71,72 Recently, Hayakawa et al.73 proposed a new scheme of As methylation that is somewhat different from the Challenger pathway, based on in vitro biochemical studies with the recombinant human AS3MT. In the new scheme, As(III)−glutathione (GSH) complexes are the substrate for the AS3MT enzyme and methylation is not an oxidative step. Conjugates of MMA(III) or DMA(III) with GSH are the intermediates, which can dissociate from the enzyme, be hydrolyzed and oxidized to MMA(V) or DMA(V) when GSH concentration is lower than 2 mM. This pathway is supported by a recent study on the structure of the alga Cyanidioschyzon merolae ArsM.74

to grain, but also in the distribution inside the grain. Inorganic As, especially arsenite, shows a strong accumulation in the ovular vascular traces, which are located on the surface of the grain and are the conducting tissues that transport water and minerals into the grain.15,49,50 In contrast, DMA permeates readily into the endosperm49,50,58 (Figure 3). This difference explains why DMA is present at a greater proportion in polished rice (i.e., endosperm) than in the rice bran; the reverse is true for iAs.15,59

Figure 3. Synchrotron X-ray fluorescence reveals contrasting distribution patterns of inorganic As and DMA in immature rice grain. Reprinted with permission from ref 58. Copyright 2013 Springer.

As mentioned in the previous section, genotypic variation in the As speciation in rice grain may be attributed to the differences in DMA uptake or translocation. Differences in DMA or MMA uptake and translocation have been observed among different plant species,42 but there are no reports on genotypic variation in these measurements among rice genotypes. Susceptibility to “straight-head” disease (spikelet sterility), which is suspected, among other potential triggers, to be caused by the toxicity of methylated As species,58,60 differs greatly among rice cultivars.61,62 Whether the susceptibility is related to the accumulation of methylated As species or to different levels of tolerance remains to be investigated. In a hydroponic experiment DMA treatments caused deformation of the reproductive organs in rice similar to the symptoms of “straight-head” disease.58 The extremely high translocation efficiency of DMA toward the grain may be an important factor in the induction of this physiological disorder. Straight-head disease is most prevalent in the south-central states of the U.S., where rice grain also contains high percentages of DMA.

6. ARSENIC METHYLATION BY SOIL MICROORGANISMS Many archaea, bacteria, fungi, and eukaryotic algae are able to methylate As, producing mono-, di-, tri-, or even tetra-methyl As species. Methylation is catalyzed by the arsenic methyltransferase enzymes, designated as AS3MT for the human enzyme and ArsM for the microbial enzymes.63 The amino acid sequences of arsenic methyltransferases from divergent organisms are highly conserved, with three strictly conserved cysteine residues being involved in the catalytic function.63−65 A search of the rice genome reveals no ArsM-like genes

7. ENVIRONMENTAL FACTORS INFLUENCING ARSENIC METHYLATION IN SOIL Several studies have shown that As methylation, as measured by the amounts of methylated As species in the soil solution or volatilized, was greatly enhanced by the anaerobic conditions 3961

dx.doi.org/10.1021/es304295n | Environ. Sci. Technol. 2013, 47, 3957−3966

Environmental Science & Technology

Critical Review

which developed after soil flooding, especially when decomposable organic matter was added to the soil.37,75−78 This observation explains why rice grown under flooded conditions contained proportionally more DMA than that grown under aerobic conditions,3,20,25 as well as why other crops grown on aerobic soil generally contain little or no methylated As species.60 There are several possible explanations for the enhanced As methylation under anaerobic conditions. First, the abundance or activities of anaerobic As methylation microorganisms in soil are likely to be enhanced. Flooding has a major impact on soil microbial communities, enhancing microorganisms such as iron-reducing bacteria.79,80 Second, more arsenite is mobilized into the soil solution under anaerobic conditions,3,6,20,81 thus increasing the availability of the substrate for methylation. In a pot study investigating the effect of phytoextraction with the As hyperaccumulator Ptteris vittata, Ye et al.82 showed that the phytoextraction treatment reduced DMA in rice grain much more than iAs, suggesting that a decrease in soil available As led to a decreased As methylation in the soil. Third, the reductive dissolution of iron oxides/hydroxides under anaerobic conditions could release previously sorbed MMA and DMA into the soil solution. The effect of organic manure additions can be explained by enhanced microbial activities, leading to decreased redox potential, enhanced As mobilization, and subsequently enhanced microbial As methylation. Increasing dissolved organic C was found to correlate with increased As concentration in the soil pore water.83 There are only a few studies comparing the As methylation capacity of different soils. Mestrot et al.76 reported a variation of about 2 orders of magnitude in the volatilization of methylated As among 9 soils in a microcosm study. This variation was not related to the total concentration of As in the soils, but correlated significantly with the level of As in the soil pore water, again suggesting that the availability of iAs in the soil solution is a factor influencing the extent of As methylation. However, the abundance and activities of As methylation microorganisms may be more important, as our recent studies (Zhao et al. unpublished) show a lack of correlation between As concentration in the pore water and As methylation in a range of soils collected from different countries. In light of the emerging evidence on the variation in the As methylation capacity among soils and the lack of in planta methylation ability in rice, we propose that the regional variation in the As speciation in rice grain reflect the extent of As methylation in the soil, either due to the difference in the microbial community engaged in As methylation or to the environmental conditions that either favor or inhibit methylation. Another factor could be the rate of the degradation of methylated As species in soil,84,85 which may also vary among different soils. Historically, MMA and DMA were commonly used as pesticides and defoliants in cotton production in the south central states of the U.S.,86 where rice is now widely grown. Residues of methylated As pesticides in soil could have contributed to elevated concentrations of DMA in the rice grain produced from previous cotton fields; however, it is difficult to determine the extent of this influence because MMA and DMA are biodegradable with the half-life of about 20 days according to one study.87

8. IMPLICATIONS FOR HUMAN HEALTH It has become clear in recent years that consumption of rice is an important route of As exposure in the populations that are not exposed to elevated levels of As in drinking water.1,88−90 Recent discussions on the potential health risk of As in rice are based on the content of iAs only, because there is scarce information with regard to the toxicology of methylated As species. Excess cancer risk from the exposure to iAs in rice has been estimated, and varies from