Abstract
Despite its high toxicity and widespread occurrence in many parts of the world, arsenic (As) concentrations in decentralized water supplies such as domestic wells remain often unquantified. One limitation to effective monitoring is the high cost and lack of portability of current arsenic speciation techniques. Here, we present an arsenic biosensor assay capable of quantifying and determining the bioavailable fraction of arsenic species at environmentally relevant concentrations. First, we found that inorganic phosphate, a buffering agent and nutrient commonly found in most bioassay exposure media, was in fact limiting As(V) uptake, possibly explaining the variability in As(V) detection reported so far. Second, we show that the nature of the carbon source used in the bioassay differentially affects the response of the biosensor to As(III). Finally, our data support the existence of non-specific reduction pathways (non-ars encoded) that are responsible for the reduction of As(V) to As(III), allowing its detection by the biosensor. To validate our laboratory approach using field samples, we performed As(III) and As(V) standard additions on natural water samples collected from 17 lakes surrounding Giant Mine in Yellowknife (NWT), Canada. We found that legacy arsenic contamination in these lake water samples was accurately quantified by the biosensor. Interestingly, bioavailability of freshly added standards showed signs of matrix interference, indicative of dynamic interactions between As(III), As(V) and environmental constituents that have yet to be identified. Our results point toward dissolved organic carbon as possibly controlling these interactions, thus altering As bioavailability.