Article pubs.acs.org/est
Dissolved and Particulate Copper Exposure Induces Differing Gene Expression Profiles and Mechanisms of Toxicity in the Deposit Feeding Amphipod Melita plumulosa Sharon E. Hook,* Hannah L. Osborn, Lisa A. Golding, David A. Spadaro, and Stuart L. Simpson Centre for Environmental Contaminants Research, CSIRO Land and Water, Locked Bag 2007, Kirrawee, New South Wales 2232, Australia S Supporting Information *
ABSTRACT: Uptake of metals via ingestion is an important route of exposure for many invertebrates, and it has been suggested that the toxic response to metals accumulated via food differs from that of metals accumulated via the dissolved phase. To test this hypothesis, the deposit-feeding epibenthic amphipod Melita plumulosa was exposed to nontoxic or reproductively toxic concentrations of copper via the overlying water, via ingestion of sediment, or via a combination of the two. Rates of copper uptake from the two exposure routes were predicted using a biokinetic model. Gene expression profiles were measured via microarray analysis and confirmed via quantitative polymerase chain reaction. Differences in expression profiles were related to the exposure route more than to individual or combined rates of copper uptake. Chitinase and digestive protease transcript expression levels correlated to the copper uptake rate from sediment, rather than from the dissolved phase or combined total uptake rate. Overall, this study supports the hypothesis that metals accumulated via ingestion have a different mode of toxic action than metals taken up from water. Consequently, guidelines that only consider dissolved metal exposure, including equilibrium-partitioning-based guidelines, may underestimate the potential effects from deposited or resuspended metal-contaminated sediments.
■
INTRODUCTION For many benthic invertebrates, the contribution of dietary exposure to contaminants (e.g., via ingestion of sediment particles) to the overall body burden can be greater than the contribution of the dissolved exposure.1−5 Despite this, many guidelines only consider the dissolved contaminant exposure when assessing potential for tissue accumulation or toxic effects.6−8 It is now well recognized that accumulated metals bind to a range of biological ligands at sites within the organism that have different functions, potentially resulting in toxicity arising from a number of different mechanisms.1,9,10 As a consequence, these complex metal handling and sequestration strategies make metal toxicity difficult to predict.11 Toxicity occurs if the metal influx rate exceeds the combined rates of detoxification and metal excretion, resulting in the metabolically available metal concentration exceeding the toxic threshold concentration for that metal at specific internal sites.12 The bioavailability of metal in sediments is strongly influenced by the partitioning between the dissolved and particulate phases,3,13,14 as well as the relative contribution of the exposure pathways applicable to the organism, e.g., pore water and overlying water and sediment particles or food via ingestion.1,2,4,5,15 Consequently, understanding the processes that affect the rate of uptake and the mechanism of toxicity from each exposure route may be necessary to accurately predict toxic effects of metals in sediments.11,15−17 Previous studies Published 2014 by the American Chemical Society
have shown that metal toxicity to invertebrates is manifested differently depending on the exposure route.18−20 Initially, Hook and Fisher19 were able to show that reproductive toxicity occurred when copepods ingested algal cells with elevated metal concentrations and did not occur when the copepods were exposed to similar concentrations of dissolved metals. Other studies were able to demonstrate both a decrease in reproduction and growth in daphnids that ingest copper or nickel in their diets20,21 and a decrease in reproduction without a change in growth in daphnids that ingest zinc in their diets.22 By contrast, daphnids exposed to waterborne zinc show declines in weight and energy reserves, without much change in reproduction,23 suggesting that toxicity occurs via different modes of action depending on the exposure route. Additional work has also shown that the tissue distribution for waterborne metals differs from that taken up with food.24 The benefits of the field of toxicogenomics, which has emerged over the past decade, include the ability to compare modes of toxic action.25−28 Toxic responses that occur via the same biological pathways would be expected to cause changes in the abundance of either the same transcripts or transcripts Received: Revised: Accepted: Published: 3504
November 28, 2013 February 16, 2014 February 19, 2014 February 19, 2014 dx.doi.org/10.1021/es405322s | Environ. Sci. Technol. 2014, 48, 3504−3512
Environmental Science & Technology
Article
μm), and brought to a room temperature of 21 ± 1 °C. M. plumulosa does not reproduce successfully in water alone; therefore, dissolved exposures were undertaken in clean sand that had been equilibrated with copper-spiked seawater. The nominal dissolved copper concentrations were 0, 10, 25, 50, and 60 μg/L. The clean sand had a low copper-binding affinity and was 0.2−1 mm in size to prevent ingestion. The test water was equilibrated with the sand for 24 h before renewal to allow for saturation of copper-binding sites present on the sand. Preliminary experiments indicated that measured dissolved copper concentrations would be approximately 30% lower than nominal. The silty sediment comprised 100% Bonnet Bay sediment, and the silty sand comprised a 1:1 mixture (based on dry weight) of Bonnet Bay sediment and Sydney sand. The sediments were spiked with copper using the superspike approach described previously39 and following the general considerations outlined previously.40 The spiked sediments were thoroughly mixed, and where necessary, a small volume of 1 M NaOH was used to adjust the pH to achieve approximately pH 8 following the 4 week equilibration period. The copper concentrations were 0, 100, 200, 350, and 500 mg/kg for the silty sediments and 0, 75, 150, and 250 mg/kg for the silty sand sediments (all dry weight). Renewal of the overlying test water was made every day, and measurement of dissolved copper concentrations was made before and after each water change to allow time-averaged copper concentrations in the overlying waters to be determined. Analyses of the sediments involved total recoverable metals (TRMs; by microwave-assisted aqua regia digestion) and dilute acid-extractable metals (AEMs; 1 M HCl). Additional details of sediment characterization, including sediment particle size, particulate organic carbon, and acidvolatile sulfide (AVS), are provided in Table S1 of the Supporting Information. Exposure Effect Model. A biokinetic model was used to describe the uptake of dissolved and particulate copper by M. plumulosa, as described previously.16,41 This model was developed using Cu radiotracer uptake experiments with exposure times similar to those used in this study.16 Copper bioaccumulation is described by the steady-state equation
that have similar molecular functions.29−33 Previous studies have compared transcriptomic profiles of different chemical compounds to determine whether compounds exert toxicity via the same mechanism.28,34−36 In the present study we investigate the links between the dissolved and particulate exposure route for copper and the gene expression (transcriptomic) responses in the deposit feeding epibenthic amphipod Melita plumulosa. This amphipod is used in whole sediment toxicity tests in Australia because it is sensitive to environmental contaminants and amenable to laboratory culture.37 Amphipods were exposed to coppercontaminated silty sediment, copper-contaminated silty sand, or copper in the overlying water with sand provided as a very weak copper-binding substrate that could not be ingested (>180 μm). A biokinetic model was used to predict the proportion of copper exposure from the particulate and the dissolved phase resulting from these different exposure scenarios. Gene expression profiles, as determined via microarray analysis, were measured for two doses of the dissolved exposure (sand), two doses of the particulate exposure (silty sediment), and one dose of the mixed exposure (silty sand). Selected transcripts were also profiled at additional doses via qPCR (quantitative polymerase chain reaction). The resultant transcriptomic profiles were compared to determine whether toxicity occurs via the same mode of toxic action under the different exposure scenarios.
■
MATERIALS AND METHODS Test Organisms and Bioassays. M. plumulosa were obtained from continuous laboratory-stock cultures, and descriptions of the species life cycle and procedures for culturing, handling, and 10 d chronic (reproduction) tests were described previously.37 Full details of the procedures are provided in the Supporting Information. All exposures were undertaken in an environmental chamber on a 12 h light/12 h dark cycle (21 ± 1 °C; 3.5 μmol of photons/s/m2) for the test duration. Physicochemical parameters, including dissolved oxygen, pH (8.0 ± 0.2), salinity (30 ± 2), temperature (21 ± 1 °C), and ammonia (0.5 mm, 2−5 months old) and were performed in 250 mL glass beakers containing 5 mm (20 g) of substrate and 220 mL of overlying water. All exposures were undertaken with five replicates. After 48 h, replicate exposures were terminated and the organisms collected. Animals were collected by gently sieving the sediment through 180 μm mesh and washing in a sorting tray to remove any sediment. The exposed animals were pipetted into a 2 mL cryogenic tube, excess seawater was removed via syringe, and then the tubes were flash frozen in liquid nitrogen, which both instantly euthanized the organisms and preserved the tissues for RNA extraction. Test Substrates and Analyses. Three copper exposures were created by using substrates that had differing abilities to bind copper: sand, silty sand, and silt. The methods for preparing these exposures are described in detail in the Supporting Information. In brief, clean seawater was collected from Port Hacking, Sydney, Australia, membrane filtered (0.45
CO = [(k u‐WC W )/(ke‐W + g )] + [((AE)(IR)CS) /(ke‐S + g )]
(1)
where CO is the amount of copper taken up by the organisms (μg g−1 dry wt) for an exposure time t (days), ku‑W is the uptake rate constant from the dissolved phase (L g−1 day−1), CW is the copper concentration in the dissolved phase (μg L−1), AE is the copper assimilation efficiency from the ingested particles (%), IR is the ingestion rate of the organism (g g−1 day−1), CS is the copper concentration in the ingested particle (μg g−1), ke‑W and ke‑S are efflux rate constants for copper taken up from the dissolved and particulate phases, respectively, and g is the growth rate constant (day−1). The model assumes that uptake from dissolved and sediment sources is additive. In the current application, this model was used to predict the copper exposure rate (ER) from individual and combined copper exposure routes (dCu and AE-Cu), and the ER has been demonstrated to be useful in predicting the toxicity of copper to this amphipod for dissolved and sediment copper exposures.41 This differs from the biodynamic metal bioaccumulation model,12 which was used to calculate the steady-state body concen3505
dx.doi.org/10.1021/es405322s | Environ. Sci. Technol. 2014, 48, 3504−3512
Environmental Science & Technology
Article
Figure 1. (a) Partitioning of copper between the dissolved phase (dCu, overlying water) and sediments (AE-Cu) for dissolved copper with sand, and copper-spiked silty-sand and silty sediment. dCu is shown in black triangles, mixed dCu-pCu is shown in blue circles, and pCu is shown in red squares. (b) Copper exposure rate (ER, where Mp refers to the amphipod, Melita plumulosa), showing the relative portion contributed from dissolved (dCu) and particulate (pCu) copper sources for all sediments with treatments used in gene expression/qPCR experiments indicated by arrows. The dashed lines represent the threshold (EC10) between rare and frequent adverse effects to reproduction.46
Microarray Data Analysis. Microarray data were analyzed using the approach described previously, and more details regarding our approach are provided in the Supporting Information..32,33,43 Briefly, data files generated using Agilent’s feature extraction software were LOWESS normalized,44 and then GeneSpring 12.7 (Agilent) was used for subsequent analysis. Genes were filtered such that the features had to be considered “present” (with signal levels above mean background levels) in four of five arrays and the features had to be 2-fold different relative to those of controls cultured in the same sediment and substrate type. If the features on the array met these criteria, expression ratios were tested for significance (p < 0.05) relative to controls using a t test using GeneSpring’s option for tests with replicates, employing the Benjamini− Hochberg multiple testing correction. Relationships among treatments were depicted using Euclidean hierarchical clustering with centroid linkages, performed on both treatments and features. Relationships among treatments were interrogated further using the principal components analysis (PCA). qPCR Validation. Quantitative reverse-transcriptase PCR (qRT-PCR) validation of the microarray results was performed as described previously.33 All exposures were analyzed via twostep qRT-PCR using 5′-nuclease assays. Briefly, 200 ng from each aliquot was reverse transcribed into cDNA using Qiagen’s QuantiTect cDNA synthesis kit. Primers with dual-labeled 6′FAM−ZEN−Iowa Black probes were designed for the transcripts in Table S3 of the Supporting Information using IDT DNA’s RealTime PCR software. Transcript levels were measured using a relative quantification protocol on Applied Biosystems Fast 7500 system using the manufacturer’s protocols and reagents, with the exception of an added primer anneal step, which was performed at the temperature given in Table S3 of the Supporting Information. All PCR efficiencies were comparable (approximately 90%), and no amplification was measured in the NTC. Fold change was calculated by the ΔΔCt method with elongation factor 1 alpha used as a housekeeping gene, as described by Livak and Schmittgen.45 Our previous work indicates that this transcript’s abundance is not altered following exposure to metals.33,42 Significance was determined at p < 0.05 using Student’s t test.
tration. This parameter was considered independent of organism size, growth, or efflux and is described by ER (μg of Cu/g of organism/day) = (k u‐WC W ) + ((AE)(IR)CS)
(2)
This epibenthic amphipod forages for food to shallow sediment depths, and this feeding behavior results in exposure to copper from the diet (ingestion of particles) and also dissolved copper in the water near the sediment−water interface (overlying water and surficial pore water).3,5 The biokinetic model calculations used the dissolved copper concentration in the overlying water and the dilute acidextractable copper (AE-Cu) concentration in the
