Bioconcentration of munitions compounds in plants and worms: experiments and modeling
Date
2016
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
Elevated concentrations of munitions compounds (MCs) – which include explosives and propellants – have been found in soils at military ranges and adjacent areas exposed to off–site migration of contaminants. Organisms such as plants and worms inhabiting these soils are exposed to and may take up the MCs, posing a risk to higher trophic levels. Experimental measurements and modeling tools are required to estimate the degree of bioconcentration to be expected.
Plant uptake assays, plant–water partitioning experiments, and two partition–based models for the estimation of MCs bioconcentration in plants and worms are presented. An experimental protocol for the plant uptake assays to obtain bioconcentration factors (BCFs), defined as the steady state ratio of the concentration in the organism to that available in the growth medium, was tested using barley (Hordeum vulgare L.). Unlike conventional methods, this protocol separated the effects of soil characteristics on the MCs bioavailability by using coarse quartz sand (99%, 0.85–1.27 mm effective diameter particles) rather than more complex field or synthetic soils. Applying the proposed protocol, steady state concentrations in both plant and exposure medium were achieved within a one–month period that produced BCFs. Standard partitioning experiments with plant biomass and water were also performed. The resulting plant–water partition coefficients effectually predicted the upper–bound of the experimental BCFs.
The models developed for the prediction of concentrations in plants and worms from soil exposures use polyparameter linear free energy relationships (pp–LFERs) to estimate the partition coefficients of MCs between soil solids and soil interstitial water, and between organism biomass and water. The pp–LFERs were applied with a set of numerical descriptors computed from chemical structure only. These computations used quantum chemical methods that quantitatively characterize the molecular properties by which a MC interacts with soil solids, water, and organism biomass. Specifically soil organic carbon, plant cuticle, worm lipid, and worm protein were the phases considered in the soil–water–organism system. Concentrations of MCs in plants observed in independent validation uptake assays were predicted using pp–LFERs for the partitioning between soil organic carbon and interstitial water, and, subsequently, between water and plant cuticle. The resulting RMSE, root mean square error (log predicted - log observed concentration) of prediction, was 0.433. Similarly, concentrations in worms observed in independent validation uptake assays were predicted with the estimated concentrations in the soil interstitial water and pp–LFERs for the partitioning between water and worm lipid, and water and worm protein. The resulting RMSE was 0.396.
These results highlight the major role played by partitioning in the uptake of MCs by plants and worms from soil. Furthermore, these partition–based models yield estimates without the need for experimental measurements. They require only parameters computed from a compound’s molecular structure using quantum chemical methods. These models are particularly useful when: (i) data for a specific organism are scarce, (ii) predictions need to be made for large libraries of compounds, and/or (iii) environmental risk needs to be assessed for compounds in the development stage.