The Rice Aquaporin Lsi1 Mediates Uptake of Methylated Arsenic Species

Pentavalent methylated arsenic (As) species such as monomethylarsonic acid [MMA(V)] and dimethylarsinic acid [DMA(V)] are used as herbicides or pesticides, and can also be synthesized by soil microorganisms or algae through As methylation. The mechanism of MMA(V) and DMA(V) uptake remains unknown. Recent studies have shown that arsenite is taken up by rice (Oryza sativa) roots through two silicon transporters, Lsi1 (the aquaporin NIP2;1) and Lsi2 (an efflux carrier). Here we investigated whether these two transporters also mediate the uptake of MMA(V) and DMA(V). MMA(V) was partly reduced to trivalent MMA(III) in rice roots, but only MMA(V) was translocated to shoots. DMA(V) was stable in plants. The rice lsi1 mutant lost about 80% and 50% of the uptake capacity for MMA(V) and DMA(V), respectively, compared with the wild-type rice, whereas Lsi2 mutation had little effect. The short-term uptake kinetics of MMA(V) can be described by a Michaelis-Menten plus linear model, with the wild type having 3.5-fold higher V_max than the lsi1 mutant. The uptake kinetics of DMA(V) were linear with the slope being 2.8-fold higher in the wild type than the lsi1 mutant. Heterologous expression of Lsi1 in Xenopus laevis oocytes significantly increased the uptake of MMA(V) but not DMA(V), possibly because of a very limited uptake of the latter. Uptake of MMA(V) and DMA(V) by wild-type rice was increased as the pH of the medium decreased, consistent with an increasing proportion of the undissociated species. The results demonstrate that Lsi1 mediates the uptake of undissociated methylated As in rice roots.Arsenic (As) contamination affects millions of people worldwide, particularly in South Asia where As-contaminated groundwater has been extracted for drinking (Chakraborti et al., 2002; Nordstrom, 2002). Recent studies have shown that foods, especially rice (Oryza sativa), are an important source of inorganic As to populations dependent on a rice diet (Kile et al., 2007; Ohno et al., 2007; Mondal and Polya, 2008). Paddy rice is more efficient than other cereal crops in accumulating As (Williams et al., 2007). This is because anaerobic conditions in submerged paddy soils lead to mobilization of arsenite [As(III); Takahashi et al., 2004; Xu et al., 2008], which is then taken up by rice roots mainly through the highly efficient transport pathway for silicon (Si; Ma et al., 2008). The relatively high accumulation of As in rice is of concern, as it may pose a significant health risk (Zhu et al., 2008; Meharg et al., 2009).A number of As species may be present in soil depending on soil conditions and the history of As contamination. These include arsenate [As(V)], As(III), and methylated As species such as monomethylarsonic acid [MMA(V): CH₃AsO(OH)₂] and dimethylarsinic acid [DMA(V): (CH₃)₂AsO(OH)]. As(V) is the main species in aerobic soils, while As(III) dominates in anaerobic environments such as flooded paddy soils. Both MMA(V) and DMA(V) have been found in paddy soils (Takamatsu et al., 1982), which may have been derived from microbial and algal biomethylation and/or past uses of methylated As compounds. MMA(V), as sodium or calcium salt, and DMA(V), as sodium salt or free acid (also called cacodylic acid), are herbicides widely used for weed control on cotton (Gossypium hirsutum), orchards, and lawns, or as a defoliant of cotton (U.S. Environmental Protection Agency, 2006). Conversion of cotton fields for the production of paddy rice in the United States may be a reason for the high levels of methylated As reported for the U.S. rice (Meharg et al., 2009).The mechanism of As(V) uptake by plants through the phosphate transport system has been well established (for review, see Zhao et al., 2009). In contrast, As(III) is taken up into the cells by aquaglyceroporins in Escherichia coli, yeast (Saccharomyces cerevisiae), and mammalian tissues (for review, see Bhattacharjee and Rosen, 2007). Recent studies have shown that several plant aquaporin channels belonging to the Nodulin 26-like Intrinsic Protein (NIP) subfamily are permeable to As(III) when expressed heterologously in yeast (Bienert et al., 2008; Isayenkov and Maathuis, 2008; Ma et al., 2008). The rice Si transporter Lsi1 (OsNIP2;1; Ma et al., 2006) is also permeable to As(III) when expressed in yeast or Xenopus laevis oocytes (Ma et al., 2008). Furthermore, the lsi1 rice mutant lost 57% of the influx capacity for As(III) compared to the wild type in short-term assays, suggesting that Lsi1 is an important entry route for As(III) (Ma et al., 2008). In rice roots, a second Si transporter, Lsi2, functions as an efflux carrier to transport Si efflux from the exodermis and endodermis cells toward the stele for xylem loading (Ma et al., 2007). This transporter also mediates As(III) efflux; two independent lsi2 mutants had 73% to 91% lower concentrations of As(III) in the xylem sap than their wild types (Ma et al., 2008). The shared uptake pathway between Si (silicic acid) and As(III) (arsenous acid) is consistent with their physiochemical properties; both are present predominantly as undissociated neutral molecules at the normal environmental and physiological pH range (pK_a = 9.2, >99% undissociated at pH ≤ 7.0), and the two molecules have similar sizes.The uptake mechanisms of methylated As species by plant roots are not known. Previous studies showed that both MMA(V) and DMA(V) can be taken up by roots and translocated to shoots in a number of plant species (Marin et al., 1992; Carbonell-Barrachina et al., 1998, 1999; Burló et al., 1999). Marin et al. (1992) found that uptake by rice followed the order of As(III) > MMA(V) > As(V) > DMA (V), although DMA(V) was more efficiently translocated from roots to shoots than other As species. Raab et al. (2007) reported large variations in the absorption and translocation efficiencies for As(V), MMA(V), and DMA(V) among the 46 plant species tested. On average, root absorption of As(V) was 2.5- and 5-times higher than MMA(V) and DMA(V), respectively. The translocation efficiency, defined as the shoot-to-root concentration ratio after 24-h exposure, was highest for DMA(V) (0.8), followed by MMA(V) (0.3) and As(V) (0.09). The concentration-dependent uptake kinetics of MMA(V) in rice roots could be described by the Michaelis-Menten equation, whereas the limited uptake of DMA(V) appeared to be linear in relation to the increasing concentration in the uptake medium (Abedin et al., 2002). Abbas and Meharg (2008) showed that DMA(V) uptake by maize (Zea mays) seedlings was enhanced by more than 10-fold by a pretreatment of phosphorus starvation; this compared with only 2-fold increase in As(V) uptake. They thought that DMA(V) might be taken up by the phosphate transporters, or that phosphorus starvation altered expression of a range of membrane transporters or even membrane permeability itself.In addition to the root uptake of methylated As species, some plants appear to be able to biomethylate As, but the pathway and enzymology remains unclear (Wu et al., 2002; Zhao et al., 2009). In microbes, As methylation follows the Challenger pathway involving repeated steps of As reduction and oxidative methylation (Bentley and Chasteen, 2002). As(V) is first reduced to As(III), which is methylated by S-adenosylmethyltransferase using S-adenosyl-l-Met as the methyl donor. The product of this reaction is pentavalent MMA(V), which is reduced by a reductase to trivalent MMA(III) with thiols (e.g. glutathione). Methylation and reduction steps continue to produce di- and trimethyl As compounds. MMA(III) and DMA(III) are intermediates in the As methylation pathway, which is not very stable (Gong et al., 2001). In rice grain, DMA(V) is the main form of methylated As, and can account for up to 80% of the total As (Zavala et al., 2008; Meharg et al., 2009). In light of the significant presence of methylated As in rice, it is important to elucidate the transport and assimilation pathways of these As species in plants.In this study, we present evidence that MMA(V) and DMA(V) are taken up by rice roots, at least partly, through the NIP aquaporin channel Lsi1, and that this process is strongly pH dependent. We also show that MMA(V) can be reduced to MMA(III) in planta.