NH4H2PO4-extractable arsenic provides a reliable predictor for arsenic accumulation and speciation in pepper fruits (Capsicum annum L.)

Dietary arsenic (As) intake from food is of great concern, and developing a reliable model capable of predicting As concentrations in plant edible parts is desirable. In this study, pot experiments were performed with 16 Chinese upland soils spiked with arsenate [As(V)] to develop a predictive model for As concentrations in pepper fruits (Capsicum annum L). Our results showed that after three months' aging, concentrations of bioavailable As (extracted by 0.05 M NH4H2PO4) in various soils varied widely, depending on soil total As concentrations and soil properties such as soil pH and amorphous iron (Fe) contents. Furthermore, both the bioconcentration factor (BCF, denoted as the ratio of fruit As to soil As) and total As concentrations in pepper fruits were largely determined by concentrations of bioavailable As, which explained 27% and 69% variations in the BCF and fruit As concentrations, respectively. Apart from bioavailable As, soil pH and Fe contents were another two important factors influencing As accumulation in pepper fruits. Taking the three factors into account, concentrations of fruit As can be well predicted using a stepwise multiple linear regression (SMLR) analysis (R-2= 0.80, RMSE = 0.17). Arsenic species in soils and edible parts were also analyzed. Although As(V) predominated in soils (>96%), As in pepper fruits presented as As(V) (46%) and arsenite [As(III)] (39%) with small amount of methylated As (<15%). Aggregated boosted tree (ABT) analysis revealed that inorganic As concentrations in pepper fruits were determined by concentrations of bioavailable As, phosphorus (P) and Fe in soils. In contrast to inorganic As, methylated As concentrations were not correlated with those factors in soils. Taken together, this study established an empirical model for predicting As concentrations in pepper fruits. The predictive model can be used for establishing the As threshold in fruit vegetable farming soils. (C) 2019 Elsevier Ltd. All rights reserved.