Solid phase microextraction fibers are useful for investigating bioavailable organic contaminants in complex environmental matrixes such as aquatic sediments. Solid phase microextraction fibers are polymer-coated silica fibers that sorb dissolved organic compounds from water and sediment. We investigated their concentrations as measures of exposure for 2,4,6-trinitrotoluene and its degradation products in acute sediment and water-only toxicity tests with Tubifex tubifex, Chironomus tentans, and Ceriodaphnia dubia. Results from these exposures allowed us to compare solid phase microextraction fibers concentrations to two conventional measures of toxicant dose: external matrix (water, sediment) and internal (organism) concentrations. Because trinitrotoluene degrades within sediment and organisms, doses based on sediment, water, and organism concentrations were calculated using the molar sum of trinitrotoluene and its nitroaromatic degradation products. Among species and matrixes, median lethal doses based on solid phase microextraction fiber and organism concentrations ranged from 12.6 to 55.3 (μmol nitroaromatic per ml polyacrylate and 83.4 to 172.3 nmol nitroaromatic per gram tissue, wet weight, respectively. In contrast to matrix concentrations, which are specific to sediment or water, both organism and solid phase microextraction fiber concentrations appeared to provide measures of dose independent of exposure scenario (sediment or water). Median lethal doses based on fiber concentrations in whole-sediment and water-only Tubifex tubifex toxicity tests were within a factor of 1.1 (18.7 and 21.3 (μmol nitroaromatic per ml polyacrylate, respectively). Median lethal doses based on organism concentrations were within a factor 1.4 for Chironomus tentans exposed in water-only or whole-sediment scenarios (118.0 and 83.4 nmol nitroaromatic per gram tissue, wet weight, respectively). Solid phase microextraction fibers may provide a powerful chemical estimate of exposure with which to understand bioavailability and toxicity of organic compounds to benthic organisms.
Vigorous strong-solvent extractions, which attempt to sample 100% of sediment-associated organic compounds, usually overestimate contaminant bioavailability to sediment-dwelling or benthic biota (Landrum et al., 1994, 1996). As an alternative to traditional chemical measures of water-, soil-, or sediment-associated organics, several techniques have been investigated to provide an estimate of the bioavailable fraction of compound. These techniques are often referred to as biomimetic sampling (Hermens et al., 2001). Biomimetic techniques include using C18 disks (Verhaar et al., 1995; van Loon et al., 1996, 1997), weak solvent extractions (Cuypers et al., 2002; Liste and Alexander, 2002), gut fluid extractions (Mayer et al., 1996), Tenax beads (Kraaij et al., 2002), semi-permeable membrane devices (SPMDs; Lu and Huckins, 2002), and solid phase microextraction fibers (SPMEs; Leslie et al., 2002a, b; Conder et al., 2003).
The goal of the biomimetic approach is to provide a better chemical estimate of exposure than that provided by strong-solvent extractions. Biomimetic devices are not substitutes for organisms, but can provide complementary or preliminary information allowing a more efficient use of biota and additional evidence regarding toxicant availability in complex matrixes. Biomimetic devices are usually less expensive than organisms, in both material and labor costs. In some cases, exposure times are less than that required for organisms (Leslie et al., 2002b). Most biomimetic devices are able to measure available compounds in situations in which the use of organisms may be complicated or impossible (e.g., low pH, anoxia, turbidity; Leslie et al., 2002b). However, biomimetic devices are subject to a number of general disadvantages for measuring bioavailable organic compounds. The primary criticism of biomimetic devices is that they are overly simplistic (Salazar and Salazar, 2000). In aquatic scenarios, biomimetic devices primarily mimic dermal or respiratory (gills) routes of exposure; they cannot mimic dietary uptake. Biomimetic devices cannot simulate the complex compartmentalization of organics among different tissues, organs, and/or biomolecules, nor can they mimic biotransformation due to metabolic reactions with biomolecules, active elimination, or detoxification systems. They cannot account for biomagnification of persistent compounds across trophic levels or behavior which may affect exposure. Solid phase microextraction fibers cannot mimic organism movement or other behaviors which alter exposure scenarios.
These fibers are a promising biomimetic approach for estimating the availability of sediment-associated organic compounds (Lu and Huckins, 2002; Conder et al., 2003). They are thin silica fibers (∼110 μm diameter, 1 cm long) coated with a microlayer (5–100 μm thick) of organic polymer. During exposure to water, soil, sediment, or air, the polymer sorbs organic compounds to concentrations several orders of magnitude higher than the surrounding matrix (Pawliszyn, 1997). Disposable 1-cm long SPMEs, contained within steel mesh envelopes, can be deployed directly into sediment, retrieved, and extracted with solvent for analysis (Conder et al., 2003) or inserted directly into a GC (Mayer et al., 2000). In contrast to extracts associated with analyses of organisms or other biomimetic devices, SPME extracts do not require purification or preparation to aid chromatographic analysis. The SPME sampling procedure is rapid, inexpensive, and simple (Conder et al., 2003). Relevancy to most physicochemical conditions surrounding the organism can be achieved, as the fibers can be deployed in sediments during toxicity tests or in situ (Conder et al., 2003).
Solid phase microextraction fiber-available organic compounds closely predict body burdens for a variety of compounds in water (Leslie et al., 2002a, b) and soil (Wells and Lanno, 2000). Compound partitioning to polyacrylate-coated SPMEs may represent partitioning to biological membranes (Vaes et al., 1997, 1998; Verbruggen et al., 2000; Poerschmann and Kopinke, 2001). Because the site of toxic action for many compounds is the cell membrane, the amount of organic compound absorbed by polyacrylate-coated SPMEs may more accurately estimate dose at the site of toxic action (Leslie et al., 2002a). The objective of this study was to investigate the use of polyacrylate-SPMEs to estimate lethal doses of the explosive organic compound 2,4,6-trinitrotoluene (TNT) and its nitroaromatic (NA) degradation products to benthic invertebrates in TNT-spiked sediment and/or water. Trinitrotoluene is a contaminant of concern at military sites worldwide and is toxic to benthic invertebrates (Dave et al., 2000; Lotufo et al., 2001; Nipper et al., 2002; Steevens et al., 2002). Our goal was to compare SPME concentrations to conventional measures of toxicological dose: matrix (water or sediment) concentrations and organism concentrations.
Materials and methods
Tubifex tubifex toxicity tests
A sediment toxicity test using the aquatic oligochaete Tubifex tubifex was conducted using uncontaminated sediment collected from the top 15 cm of ponds at the University of North Texas Water Research Field Station, Denton, TX, USA spiked with crystalline TNT (98% pure, Chem Service Inc., West Chester, PA, USA). The sediment was prepared as described in Conder et al. (2003) and summarized here. Sediment was dried for 24 h (105°C), ground and sieved (<250 (m particle size). Crystalline TNT (molecular mass = 227.1 g mole−1) was dissolved in pure high performance liquid chromatography (HPLC)-grade acetonitrile and added in 4 ml aliquots to each replicate of 100 g dry sediment, such that initial sediment concentrations were 440, 1409, and 4403 nmol TNT g−1, dry weight (dw), with nine replicates per concentration (27 replicates total). Solvent control replicates received 4 ml acetonitrile containing no TNT. After mixing the sediment and acetonitrile thoroughly by hand (30 sec), the solvent was allowed to evaporate for 24 h. Sediment was then wet with 500 ml ultrapure water, mixed vigorously with a magnetic stirrer for 1 min, and placed in an environmental chamber (23 ± 1°C, 16:8 gold light:dark photoperiod). To maintain dissolved oxygen, the overlying water of each beaker was aerated gently and continuously using a pipette immersed below the water's surface.
Nine replicates of each TNT spiking concentration were divided into three subsets which were allowed to age 1, 8, or 29 d before chemical and toxicological evaluation (Conder et al., 2004). The experimental design consisted of 28-d toxicity experiments using 1-d aged sediments (1–29 d after sediments were spiked and wetted), 8-d aged sediments (8–36 d after sediments were spiked and wetted), and 29-d aged sediments (29–57 d after sediments were spiked and wetted). Twenty-seven replicate beakers were used in this experiment. Because TNT degrades rapidly in contact with sediment in laboratory experiments (Conder et al., 2004), this resulted in nine TNT-spiked sediments with different levels of NAs.
At the beginning of the toxicity test, sediment samples (∼5 g, wet wt (ww)) were removed from replicates for determination of acetonitrile-extractable NAs (see below). Measurement of NAs by SPMEs was conducted directly in sediment during organism exposures (Conder et al., 2003). Each experimental replicate received one stainless steel mesh envelope containing a 1-cm long SPME (85 μm polyacrylate coating). The SPME envelope was inserted within the sediment such that the SPME fiber was exposed 2 (1 cm below the surface. Immediately after SPME deployment, each replicate received four adult 9 to 10-week old Tubifex. The envelope was retrieved 24 h later, and the fiber extracted (see below). This exposure time period is only sufficient to reach ∼74% of steady state SPME concentrations (Conder et al., 2003); all reported SPME concentrations have been multiplied by 1.35 to correct for this. At the termination of the 28-d toxicity test, sediments were sequentially sieved through 250- and 500-μm sieves to recover adult Tubifex, young Tubifex, and cocoons.
A 4-d water-only acute toxicity test was conducted using TNT-spiked reconstituted hard water and 1-mm diameter glass beads. To spike test water, an aliquot of 50:50 (v/v) acetonitrile:water solution containing dissolved TNT was added to a volumetric flask. Acetonitrile was evaporated from the aliquot under a gentle stream of He, leaving crystalline TNT in water. This was re-dissolved in and diluted to volume with reconstituted hard water to 72 nmol TNT ml−1 before serial dilution to generate the remaining concentrations (2, 6, 12, 36 nmol TNT ml−1). Solvent control water was generated via the same procedure as the 72 nmol TNT ml−1 without the addition of TNT. Exposure solutions were not renewed during the test. Each replicate (three replicates per concentration) received 100 ml water, 10 cm3 1-mm diameter glass beads, 10 adult 8 to 9 week old Tubifex, and one 1-cm long SPME (85-μm polyacrylate coating) inserted through a small thin teflon disc to facilitate handling of the thin, transparent fiber. Fiber uptake is independent of the holder (steel mesh envelope or teflon disc); this was verified by measurement before these experiments. The SPMEs were exposed during the first 48 h of the experiment to achieve steady state uptake in the fiber (Conder et al., 2003). This toxicity test was conducted in an environmental chamber under the same conditions as described above for the sediment test. Tubifex were observed frequently during exposure. Individuals not responding after a gentle prod with a probe were considered dead. Dead Tubifex were removed from solution, blotted on absorbent paper, and stored frozen until tissue analysis (see below). At the end of the test, live Tubifex were prepared and stored in the same manner. Each experimental replicate of ten Tubifex was pooled for tissue analysis. Where mortality was greater than 0 but less than 100%, the tissue analysis replicate was comprised of both live and dead worms.
Chironomus tentans toxicity tests
A 4-d sediment toxicity test using 3rd-instar midge larvae Chironomus tentans was conducted using uncontaminated sediment collected from Brown's Lake, Waterways Experiment Station, Vicksburg, MS, USA spiked with TNT as described in Conder et al. (2003). The Chironomus tentans sediment toxicity test was performed in a different laboratory with slightly different spiking procedures and storage conditions. After collection, sediment was wet-sieved (<250 μm) and stored at 4°C. Trinitrotoluene was added to the wet sediment via TNT-coated quartz sand in order to achieve initial sediment concentrations of 220, 440, 881, 1321, and 1761 nmol TNT g−1, dw. To coat the sand, TNT dissolved in acetone was added to 30 g of dry sand under a fume hood. After the solvent evaporated, the sand was added to sediment and vigorously mixed with a mechanical stirrer. After a 3-h mixing period, sediments were sampled for chemical analysis and added to exposure beakers (100 g sediment/beaker, three replicates per concentration). Each beaker received 200 ml dechlorinated water. Overlying water was not aerated in this experiment, but aerobic dissolved oxygen levels (>40% saturation) were verified. Test conditions were as described above for the Tubifex sediment toxicity test, except for temperature (25 ± 1°C). Each replicate received ten midge larvae and one SPME, deployed as described above. Midge larvae exposed to the three highest sediment TNT concentrations did not completely burrow into the sediments. After 7.5 h of exposure, these midge larvae were presumed dead after failing to respond to gentle prodding by a blunt probe and were immediately collected for tissue analysis. Midge larvae collected at this time and those collected at the end of the toxicity test (alive) were blotted on absorbent paper and stored frozen until tissue analysis (see below). Because of the amount of tissue needed for tissue analysis, midge larvae from all replicates of each concentration were pooled.
A 4-d water-only toxicity test in TNT-spiked dechlorinated tap water and sand was conducted using 4th-instar midge larvae. To spike test water, aliquots of acetone containing dissolved TNT were added directly to test water (0.5 ml acetone l−1). There were seven treatment levels: six TNT concentrations (1.1, 2.2, 4.4, 8.8, 17.6, and 26.4 nmol TNT ml−1) and one solvent control which contained acetone at the concentration used to prepare the TNT-spiked treatments. Exposure solutions were changed daily in all treatments. Each replicate (four replicates per concentration) received 250 ml water, 5 g quartz sand, and ten 4th-instar Chironomus. The toxicity test was conducted under the same conditions as described above for the Chironomus sediment toxicity test, except that exposure temperature was 23 ± 1°C and SPMEs were not deployed. At the end of the test, surviving midge larvae were removed from solution, blotted on absorbent paper, and stored frozen until tissue analysis. Replicates at each concentration were pooled for tissue analysis; where mortality was <100%, tissue analysis included only live midge larvae. Dead midge larvae were not included because the experiment was not checked frequently enough to collect organisms that had not decayed significantly. Water samples were taken at the end of the exposure period.
Ceriodaphnia dubia toxicity test
A 2-d Ceriodaphnia dubia acute toxicity test was conducted using reconstituted hard water spiked with TNT in the same manner and at the same concentrations as in the water-only Tubifex toxicity test. Waters were not renewed during the test. Each replicate (four replicates per concentration) received 15 ml water, five neonate (<24 h old) Ceriodaphnia dubia, and one 1-cm long SPME (85-μm polyacrylate coating) inserted through a small, thin teflon disc. The SPME exposure period and environmental conditions were identical to those of the water-only Tubifex test.
Because TNT degrades rapidly in TNT-spiked sediments (Green et al., 1999; Lotufo et al., 2001; Steevens et al., 2002; Conder et al., 2004), sediment concentrations were determined for TNT and its primary degradation products 2-amino-4,6-dinitrotoluene (2ADNT), 4-amino-2,6-dinitrotoluene (4ADNT), 2,4-diamino-6-nitrotoluene (2,4DANT), and 2,6-diamino-4-nitrotoluene (2,6DANT) following USEPA method 8330A (USEPA, 1998). Sediment samples (1–5 g, ww) were extracted with 10 ml HPLC-grade acetonitrile, mixed for 16 h using an end an end-over-end rotating mixer (60 rpm), and centrifuged. Supernatants (extracts) were filtered with 0.45-μm glass microfiber filters and diluted with an equal volume of ultrapure water. To minimize possible compound degradation and/or disappearance, samples were not air-dried before extraction. Sediment sample dry weights were determined via gravimetric moisture determination (dried at 105°C for 24 h) of the post-extraction sediment.
Solid phase microsextraction fibers exposed to TNT-spiked sediment and water were removed from their holders (steel mesh envelopes or teflon discs) and placed into HPLC autosampler vials containing 400 μl of 50:50 HPLC-grade acetonitrile:ultrapure water to desorb the compounds from the SPMEs. The one-cm lengths of SPME fiber used in this study contain 0.521 μl polyacrylate; we present SPME data as concentrations (number of molecules absorbed by the polyacrylate divided volume of the polyacrylate) as in previous studies (Mayer et al., 2000; Verbruggen et al., 2000; Leslie et al. 2002a, b).
Tubifex and Chironomus tissue samples (0.02–0.1 g) were weighed and homogenized for 1 min in 750 l HPLC-grade acetonitrile using 1.0-mm glass beads in a Mini-beadbeater (Biospec Products, Inc., http://www.biospec.com). The homogenate then received 750 μl 1% CaCl2 and was sonicated for 1 h in a cool water bath (16 ± 2°C). After sonication, the homogenate was centrifuged for 10 min at 10,000 g. The supernatant was then filtered through a 0.45-μm glass microfiber filter before HPLC analysis.
Sediment extracts, SPME extracts, organism extracts, and water were analyzed via HPLC using a Waters Nova Pak C-18 column with an 84:16 isopropanol:water isocratic mobile phase (Fixed UV detector, 254 nm) or a Supelco Discovery RP-Amide C-16 column with a 55:45 methanol:water isocratic mobile phase (Photodiode array UV, variable wavelengths). Quality assurance/quality control measures included spike and method blank analyses for sediment, organism, and SPMEs; average percent recoveries for the analytes measured ranged from 99–109%, 90–100%, and 99–103%, respectively. For NA compounds, approximate method detection limits were 0.5–2.5 nmol NA g−1, dw for sediment, 1.5–6.9 nmol NA g−1, ww for organisms, 0.06–0.23 μmol NA ml−1 for 1-cm SPME fibers, and 0.07–0.30 nmol NA ml−1 for water.
Median lethal dose estimates were calculated via the Trimmed Spearman-Karber method (Hamilton et al., 1977) using up to four different measures of NA dose: nominal or measured water concentrations, measured sediment concentrations at the beginning of exposure, SPME NA concentrations, and organism NA concentrations. Control mortality was accounted for during calculation of the median lethal dose estimate. For each dose metric, a simple molar sum of the five analytes was used: ∑(TNT + 2ADNT + 4ADNT + 2,4DANT + 2,6DANT). Although this approach to mixture toxicity is routinely followed for polycyclic aromatic hydrocarbons (Swartz et al., 1995), it assumes that the compounds are relatively equitoxic and that toxicity of the mixture is additive. Although this may be an oversimplification, data exist which both support (Lotufo et al., 2001; Steevens et al., 2002) and refute (Dodard et al., 1999; Sunahara et al., 1999; Siciliano et al., 2000) these assumptions for NAs.
Results and discussion
Among the five toxicity tests, we used three types of ‘dose’ to describe the exposure of TNT and NAs to invertebrates: matrix concentration (sediment or water), organism concentration, and SPME concentration. All dose measures increased with respect to mean mortality (Figures 1–5), enabling the calculation of median lethal doses (Table 1).
Median lethal doses based on SPME concentrations in the whole-sediment and water-only T. tubifex toxicity tests measured at 24 h were within a factor of 1.1, suggesting that SPMEs may be a matrix-independent measure of dose for NAs. Among all species and matrixes, SPME NA concentrations ranged from 12.6 to 55.3 μmol NA ml−1. Our results were very similar to median lethal doses based on SPME concentrations (polyacrylate SPME) for the polar narcotic 2,4,5-trichloroaniline (32.2 μmol ml−1) and the narcotic 1,2,3,4-tetrachlorobenzene (30.5 μmol ml−1) tested with Chironomus riparius in a water-only exposure (Leslie et al., 2002a). However, the SPME data represent only the initial 24 h of exposure, and are not indicative of dose during the full duration of the 4- or 28-d toxicity tests. In the Tubifex sediment test, mortality likely occurred within 48 h. During the Tubifex experiment, it was possible to observe sediment burrowing (galleries) and tail waving activity in each replicate. In replicates with 100% Tubifex mortality (verified after 28-d), these activities were not observed or ceased approximately 24 to 48 h after addition of the organisms. Thus, the 28-d duration was not necessary to generate lethal toxicity data. We assume that lethality also occurred on a similar time scale in the 4-d Chironomus tentans sediment test. Because mortality probably occurs quickly in TNT-spiked sediment, initial SPME data during the first 24 h is probably sufficient to estimate lethal dose.
Median lethal dose estimates based on organism concentrations ranged from 83.4 to 172.3 nmol NA g−1, ww. Because NAs degrade rapidly in matrixes during toxicity tests (especially in sediment) and NA depuration kinetics are rapid in organisms (Lotufo, unpubl. data; Conder unpubl. data), concentrations in organisms at the time of retrieval from sediment may be lower than concentrations at time of death. Thus, our estimates of organism concentrations associated with lethality may be lower than at the time of toxic action and median lethal doses based on these data may overestimate toxicity. Organism concentrations appear to be independent of matrix, for median lethal doses based on organism concentrations were similar for Chironomus tentans exposed to water or sediment (factor of 1.4 difference). Our range of organism concentrations associated with lethality lie within the range of critical body residues for marine amphipods (Leptocheirus plumulosus) and polychaetes (Neanthes arenaceodentata) exposed to radiolabeled TNT-spiked sediments, 37.9 and 220 nmol NA g−1, ww, respectively (Green et al., 1999). However, these body residues are likely an overestimate of the concentrations of HPLC-identifiable, solvent-extractable NAs in the organisms, since concentrations in the Green et al. (1999) study were not measured by HPLC, but estimated via activity of the radiolabeled NAs within the tissue. Significant proportions of radiolabeled NA body burden are unidentifiable by the HPLC method for nitroaromatics or unextractable by the acetonitrile tissue extraction (Renoux et al., 2000; Lotufo, unpubl. data; Conder unpubl. data). It is unknown whether unidentifiable or unextractable compounds contribute to lethal toxicity.
Median lethal dose, based on organismal NA concentrations normalized by literature-derived lipid levels (Chironomus tentans = 1.08%; West et al., 1997; T. tubifex = 2.74%, Kraaij et al., 2002) were an order of magnitude lower (range 6288–10926 nmol NA g−1 lipid, ww) than the 40000–160000 nmol g−1 lipid, ww range for acute lethal critical body residues for compounds with nonspecific modes of toxic action (McCarty et al., 1992; Rand et al., 1995). The mode of toxic action for NAs is thought to be via the oxidation of biomolecules by oxyradicals (Di Giulio et al., 1995; Stahl and Aust, 1995). The range of acute lethal critical body residues for NAs is above that for respiratory uncouplers (30-6000 nmol g−1 lipid, ww; McCarty et al., 1992; Rand et al., 1995) and below that for polar narcotics (12000–38000 nmol g−1 lipid, ww; McCarty et al., 1992; Rand et al., 1995).
Median lethal doses based on measured NA concentrations in water for Chironomus tentans (8.54 nmol NA ml−1) were similar to those based on nominal TNT concentrations for T. tubifex and Ceriodaphnia (17 and 34 nmol TNT ml−1, respectively). Since NA concentration decrease to as little as 50% of nominal TNT levels in water containing biota after a few days, the latter two values likely underestimate toxicity based on a time-weighted average exposure. However, dose metrics derived from aqueous exposures are not very relevant for assessing potential biological effects of contaminants to sediment-dwelling organisms such as Tubifex and Chironomus tentans. Toxic doses based on water concentrations derived from water-only tests are difficult to convert to sediment quality criteria without using models which often require measuring many other sediment-specific physicochemical parameters (Bierman, 1990). Empirical approaches using the measurement of porewater toxicant concentrations may also be used to enhance relevance to water-only exposures (Nipper et al., 2002; Kraaij et al., 2003).
Whereas whole-sediment toxicity tests may provide a more realistic exposure scenario, results are much more difficult to interpret than results of water-only tests. Trinitrotoluene degrades exponentially following its addition to sediments in the laboratory, making exposure dose difficult to quantify in freshly spiked sediments (Conder et al., 2004). It is best to base measures of dose in sediments on measured NA concentrations (molar sum of TNT and its degradation products; Conder et al., 2004). Because mortality probably occurred quickly (within approximately 48 h), median lethal doses were based on initial measured sediment NA concentrations. Initial sediment concentrations provide a reasonable estimate of dose during the time period sufficient to induce lethality, for there is only an approximate 10% difference in initial sediment NA concentrations and those measured 48 h later (Conder et al., in press). Estimates based on initial sediment concentrations were very similar between the organisms studied (141–185 nmol NA g−1, dw). Normalization of sediment concentrations by total organic carbon (OC) content is often used to account for bioavailability differences due to sorption of non-polar contaminants by OC in sediment (Bierman, 1990; Thomann and Kolmos, 1999). Normalization by sediment OC (0.964% and 1.31% for Denton and Vicksburg, respectively (Conder et al., 2003)) only increased variability in the median lethal dose estimates for these species (range 10773–19091 nmol NA g−1 OC, dw) for these more polar compounds.
Our results suggest SPME and organism concentrations may be useful as measures of dose for several different organisms exposed to TNT and NAs in sediment and/or water. SPMEs are inexpensive, rapidly analyzed, and simple to use and can be deployed directly into contaminated sediment during toxicity tests.
Although median lethal doses based on water concentrations and sediment concentrations varied little under the laboratory exposure conditions among species tested with TNT, concentrations of organic compounds in soils, sediments, and waters are usually poor measures of exposure dose. Median lethal concentrations can range several orders of magnitude for a single toxicant tested with one species in a particular matrix. Partitioning of a compound to a solid phase (such as an organism or SPME) may provide similar estimates of toxicological dose in water-only or whole-sediment exposure scenarios, as observed for Tubifex toxicity in TNT-spiked water and sediment.
The use of SPMEs, like all chemical techniques, has inherent disadvantages compared to biological organisms. The use of SPMEs and other biomimetic approaches are not intended to replace the use of organisms in risk assessment. Instead, they may be able to augment or replace traditional chemical measures of dose (matrix concentrations based on vigorous solvent extractions). In this way, they may provide an additional line of evidence in weight-of-evidence approaches (Burton et al., 2002; Chapman et al., 2002) for assessing the effects of sediment-associated organic compounds.
This research was supported by the Department of Army, Environmental Quality Technology (EQT) Program (Project BT-25 BR204, Program Manager, M. John Cullinane). Permission to publish this material was granted by the Chief of Engineers. The authors thank Rod Millward for a helpful review of the manuscript. J. Conder received support in the form of a Science to Achieve Results (STAR) graduate research fellowship from the US Environmental Protection Agency during this study.