Use of Passive Samplers To Mimic Uptake of Polycyclic Aromatic

First, organisms are often difficult to maintain in a laboratory exposure system. ... Chelsea Creek, Tenean Beach, Savin Hill Cove, Deer Island, Marin...
0 downloads 0 Views 100KB Size
Download PDF
Environ. Sci. Technol. 2004, 38, 1154-1160

Use of Passive Samplers To Mimic Uptake of Polycyclic Aromatic Hydrocarbons by Benthic Polychaetes A M Y E . V I N T U R E L L A , * ,† ROBERT M. BURGESS,‡ BRENT A. COULL,§ KIMBERLY M. THOMPSON,| AND JAMES P. SHINE† Departments of Environmental Health, of Biostatistics, of Health Policy and Management, and of Maternal and Child Health, Harvard School of Public Health, Boston, Massachusetts 02115, and U.S. EPA ORD/NHEERL Atlantic Ecology Division, Narragansett, Rhode Island 02882

Experiments were conducted to test whether passive samplers made of low-density polyethylene (polyethylene devices, or PEDs) can estimate the extent of uptake of polycyclic aromatic hydrocarbons (PAHs) by benthic polychaetes (Nereis virens) in contaminated marine sediments. For a variety of PAHs, PEDs reached 90% equilibrium with sediment PAHs in 60 days or less. Using 60-day sediment bioaccumulation tests, we have demonstrated a significant relationship between PAH concentrations in the polychaetes and the PEDs (R2 ) 0.67, p ) 0.002), with the PEDs taking up less PAHs than the polychaetes. Because of this relationship, PEDs can potentially be used in a regulatory context to simulate uptake of bioavailable PAHs in contaminated marine sediments. The PED PAH concentrations were also used to calculate porewater PAH concentrations that allowed for the estimation of a linear free-energy relationship between the lipid-water distribution coefficient (Klip) and the octanol-water distribution coefficient (KOW) for PAH uptake in marine polychaetes (R2 ) 0.94, p < 0.0001).

Introduction Assessing the bioavailability of organic contaminants in sediments is critical to understanding the ecological risks of contamination in aquatic ecosystems. Standardized sediment bioaccumulation tests using benthic organisms (1) are often performed to determine the relative bioavailability of sediment contamination and to guide options for future testing or disposal (2). These tests operate under the premise that they reflect the bioavailability of polycyclic aromatic hydrocarbons (PAHs) within a given sediment and thus illuminate the potential for adverse effects on aquatic ecosystems and human health (2). * Corresponding author address: Department of Environmental Health Sciences, Tulane School of Public Health and Tropical Medicine, 1440 Canal St., Suite 2100, New Orleans, LA 70112; phone: (504) 988-8897; fax: (504) 584-1726; e-mail: [email protected]. † Department of Environmental Health, Harvard School of Public Health. ‡ U.S. EPA ORD/NHEELR Atlantic Ecology Division. § Department of Biostatistics, Harvard School of Public Health. | Departments of Health Policy and Management and of Maternal and Child Health, Harvard School of Public Health. 1154

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 4, 2004

While the theory is sound, and the tests are standardized, they do have their drawbacks. First, organisms are often difficult to maintain in a laboratory exposure system. They are subject to variables such as mortality, metabolic shifts, reproductive development, growth, feeding needs, chemical stressors, and others. All of these factors can directly affect the outcome of the exposure. Live organisms also require specific living conditions. Temperature, salinity, light, oxygen, ammonium, and anthropogenic contaminants all must be kept at levels that are agreeable to the test organisms (1). In addition, direct measurement of tissue residues in organisms is a complicated task. Microextraction procedures or compositing of sample tissues must be adopted when the organisms are small and cannot offer sufficient tissue samples for traditional tissue extractions (1, 3). Lipid fractions must be dealt with during the extraction process, as well as accounted for in the data analysis. Cleanup procedures are almost always required on the tissue extracts to remove the inherent confounding compounds associated with organism tissues (4, 5). With the need to overcome the limitations of using live organisms in standard bioaccumulation tests, much research has focused on developing alternative, biomimetic models and passive sampling devices to simulate the bioaccumulation of contaminants by organisms. One such device is the semipermeable membrane device (SPMD), which is a tube of low-density polyethylene filled with the synthetic lipid triolein. This synthetic lipid equilibrates with organic contaminants in the water column in the same way as the lipid fraction of an organism (6). Since the genesis of the SPMD, it has been shown that SPMDs without the triolein compartment work almost as well as the SPMDs themselves (7). In this configuration, the polyethylene sheets themselves equilibrate with organic contaminants in the water. Researchers in the food chemistry field have also reported the migration of PAHs from water into polyethylene sheets (8, 9). Consequently, two teams of researchers have adapted more parsimonious sampling devices consisting only of a sheet of low-density polyethylene (10, 11). The contaminant diffuses into the polyethylene where it establishes an equilibrium with the surrounding water, with the length of the equilibration period depending on the physicochemical properties of the contaminants and environmental variables. Without the triolein, these devices equilibrate faster, in hours to days (7, 10). In contrast, SPMDs can take up to 60 days to equilibrate with organic contaminants in water (7), which is why SPMDs were designed to be integrative samplers, not equilibrium samplers, that are able to determine timeweighted averages of bioavailable aqueous concentrations (6). This difference in equilibration times becomes extremely important in sediment deployment of these sampling devices. Since the equilibration times between porewater and sediments are always much longer than those in the water column, due to tortuosity and the slow desorption of contaminants from sediments to porewater (12), an efficient passive sampler of the sediment-porewater system should be able to equilibrate quickly with the porewater. Therefore, PEDs may be preferred over SPMDs for use in sediment equilibration experiments, since they seem to be able to equilibrate faster with organic contaminants in water than SPMDs. The goal of this study was to examine the utility of polyethylene-only samplers as biomimics of contaminant uptake from sediments. Using the familiar construct of the standard sediment bioaccumulation assay, dual exposure studies were con10.1021/es034706f CCC: $27.50

 2004 American Chemical Society Published on Web 01/15/2004

ducted with marine polychaetes (Nereis virens) and polyethylene devices (PEDs) in sediments from 14 sampling sites in and around Boston Harbor. The worms and the PEDs were analyzed for accumulation of a suite of PAHs after their equilibration with the sediment. By comparing the levels of PAHs in the worms and in the PEDs, the hypothesis that PEDs are good simulators or biomimics of organic contaminant uptake by benthic organisms could be assessed.

Materials and Methods Construction of PEDs. The polyethylene device in this study consists of a circle of low-density polyethylene (Brentwood Plastics, Inc., Brentwood, MO) within an aluminum foil frame (approximate diameter 4 cm). The low-density polyethylene (LDPE) has a density of 0.92 g/cm3 and a thickness of 74-84 µm and contains no additives. The PED was constructed by stacking two aluminum foil weighing pans (VWRbrand, VWR Scientific Products) inside one another and forming a 2.4 cm hole in the center of the pans using a hole punch (15/16 in.). A polyethylene circle (roughly 3 cm in diameter) was then placed between the two pans. The edges of the pans were folded down to seal in the polyethylene. The aluminum pans were used to give the PED sufficient rigidity to be pressed down into the sediments without bending or folding. The exposed circle of polyethylene in the aluminum frame measures roughly 2.4 cm in diameter; on the basis of the limit of detection (LOD) and depletion calculations (see PED Proof of Concept), a slightly smaller circle (2.2 cm) was cut out at the end of the exposure period for extraction. Before use in the experiments, PEDs were pre-extracted in amber glass jars with methylene chloride (DCM). The jars were set in a class 100 clean fume hood for 4 days, with the DCM being changed once during the 4-day period. Once these PEDs were pre-extracted, they were allowed to dry in a class 100 clean fume hood for a minimum of 6 h. The PEDs were stored in amber glass jars with Teflon-lined lids until ready for deployment. Study Sediments. Sediments were collected from various sites in Boston Harbor (Fort Point Channel, Chelsea Creek, Tenean Beach, Savin Hill Cove, Deer Island, Marina Bay, Admiral’s Hill Cove) and the vicinity (Fox Creek, Ipswich, MA; New Bedford Harbor, MA). Sediments were also obtained from the well-studied Long Island Sound site (13), and used as reference sediments for worm toxicity. Sediments were collected from the top 5 cm of the sediment layer with an Eckman dredge or shovels. Before use in the bioaccumulation tests, the sediments were sieved to 0.96 and p < 0.0001 for all PAH curves). Having solved for a and b, the equation was then solved for t, given that CPED,t was at 90% of its maximum value. These times to 90% equilibration (t90) for a few representative PAHs in the Savin Hill Cove sediments are shown in Table 2. For these and all other PAHs analyzed, the PEDs seem to approach equilibrium within 60 days or less. No significant relationship was observed between t90 and KOW. We recognize that a device made solely of LDPE might face criticism since the material itself is not available in a standardized and controlled formulation, whereas triolein is always uniform and may offer a more reproduceable result for researchers using SPMDs. However, the tradeoff in equilibration times was too great to warrant using triolein in our sediment sampler, given that we have empirically determined equilibration times nearing 60 days for our sediment samplers, and SPMDs would require more time. To minimize uniformity issues, LDPE was ordered from the same distributor used by most SPMD researchers (e.g., Huckins et al. (6)) to ensure consistency. It is therefore recommended that, should this PED be developed for use in future sediment biomimetic studies, the same cares be taken to ensure reproducibility in the manufacture of the PEDs. Relationship between PEDs and Polychaete Worms. Parts a and b of Figure 2 show the relationship between PAH concentrations in the worms and their associated PEDs from the exposure studies. Figure 2a shows this relationship for all PAHs individually, and Figure 2b shows the relationship for ∑PAH (the sum of all PAHs). Both figures show a positive linear relationship between PED and worm PAH concentrations, as described by the regression

log Cworm ) a log CPED + b

(3)

where Cworm is the concentration of a single PAH or the ∑PAH in the worm (ng/g of lipid), CPED is the concentration of a single PAH or the ∑PAH in a PED (ng/g of polyethylene), a is the slope of the linear relationship, and b is the intercept of that relationship. Since the regression of individual PAHs in this manner may introduce correlations within sampling sites, the regression with individual PAHs was performed as a mixed VOL. 38, NO. 4, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1157

FIGURE 1. Equilibration curves for the uptake of selected PAHs into the PEDs over the 63-day equilibration period with the sediments. Error bars are (1 standard deviation from the mean. N ) 3 for each of 12 time intervals.

TABLE 2. Time to 90% Equilibrium for PAHs in PEDs time to 90% equilibrium (days) fluoranthene pyrene benz[a]anthracene chrysene benzo[a]pyrene

time to 90% equilibrium (days)

30.5 27.4 55.8 52.6 52.7

perylene benzo[ghi]perylene benzo[k]fluoranthene benzo[e]pyrene

model with sampling site taken as a random effect:

(log Cworm)ij ) a(log CPED)ij + b + βi

(4)

where i and j denote the site and a sample within the site, respectively, b is the average fixed intercept for all sites, and βi is the random effect correction factor for each intercept by site. Note that the slope, a, is a fixed effect and thus is the same for all sites. Once the site effect was controlled for in calculating the slope and average intercept parameters of the mixed model, the linear relationship for individual PAHs was found to be

log Cworm ) 0.6 log CPED + 1.8

(5)

This regression explains about 65% of the variability in worm concentrations on the basis of what is accumulated by the PED (R2 ) 0.65), according to the new methods for determining R2 in mixed models (26). The slope was also statistically significant (p < 0.0001). Likewise, the fixed effects linear regression for ∑PAH was determined to be

log Cworm ) 0.6 log CPED + 2.4

(6)

Similar to the regression with individual PAHs, the slope is 0.6 and the relationship explains about 67% of the variability in worm concentrations on the basis of what is taken up by the PED (R2 ) 0.67, p ) 0.002). Given that these relationships 1158

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 4, 2004

54.0 35.5 33.0 51.1

rely on highly variable biological data, they show a promising correlation between PED and worm PAH uptake. It should be noted that, for all of the data points, the worms take up more PAHs per gram of lipid than the PEDs do per gram of polyethylene. Similar results were found in comparison studies with SPMDs and mussels (27, 28). Since both the PEDs and worms have achieved equilibrium at the end of the 60-day exposure, it seems that the worms may be actively taking up the PAHs by some mechanism other than the passive diffusion that the PEDs use for PAH uptake (29, 30). However, regardless of the differences in uptake mechanism, the relationship between the worm and PED PAH burdens remains. This result highlights the strength of equilibrium samplers as biomimics. Even if the route of uptake is different between the sampler and the organism, the sampler is in equilibrium with every compartment in the sediment system, including the compartment that the organism is subjected to as its main route of exposure (31). Therefore, while there may never be a 1:1 linearity between a passive sampler and an organism, there will still be a useful relationship, if the sampler is at equilibrium with the sediment system. The effectiveness of the PED as a biomimic is further highlighted in a comparison of the PAH signatures of the PEDs and the worms. This relationship was evaluated by comparing individual PAH concentrations, expressed as a log-transformed fraction of the total PAH concentration, in PEDs and worms for each site (Figure 3). This comparison shows a strong linear relationship (R2 ) 0.56, p ) 0.012),

FIGURE 4. Relationship between log Klip and log KOW, where each Klip (L/g of lipid) is estimated from the lipid-normalized polychaete concentration of a given PAH and the paired freely dissolved porewater concentration of that PAH, via the equation Klip ) Corganism,lipid/Cdiss. Error bars are (1 standard deviation from the mean.

FIGURE 2. (a) Relationship between individual PAH residues in PEDs (ng/g of polyethylene) and N. virens (ng/g of lipid), when the PED and the polychaete were simultaneously exposed to sampled sediments in the bioaccumulation studies. Note that each point on the graph represents the average paired polychaete and PED concentration of a given individual PAH for a given sampling site. These points represent such pairs for all of the analyzed parent PAHs and all of the sampling sites. (b) Relationship between ∑PAH residues in PEDs (ng/g of polyethylene) and N. virens (ng/g of lipid), when the PED and the polychaete were simultaneously exposed to sampled sediments in the bioaccumulation studies. Note that each point on the graph represents the average sum of all analyzed parent PAHs in the polychaete and PED pairs for a given sampling site. These points represent such pairs for all of the sampling sites.

though the magnitude of total PAH uptake is different in PEDs and in worms. Both the PED and the worm seem to be “seeing” the PAHs in the same way, and thus accumulating PAHs in the same relative amounts. PED-Worm Effects. An ANOVA procedure was performed for each group of six beakers (four with worms, two without) pertaining to a site. It was found that, within each site, there were no significant differences in PED concentrations among the PEDs in beakers with worms and those in beakers without worms (p ) 0.003). Therefore, it can reasonably be assumed that the PED data were unaffected by the presence of worms in the beakers. Derivation of the Klip Linear Free-Energy Relationship for Marine Polychaetes. Klip is defined here as the lipidwater distribution coefficient for a specific PAH (L/g of lipid) and is often estimated in the literature according to the general equation (reviewed in ref 32)

log Klip ) a log KOW + b

(7)

where KOW is the octanol-water distribution coefficient of the given PAH. Since there were no available constants for a and b in the literature for marine polychaetes, Klip values for each PAH were empirically determined from the exposure data presented above, and then related to the respective PAH KOW values to derive constants for this relationship. The Klip term describes the relationship between the porewater PAH concentration and the lipid-normalized tissue PAH concentration as follows:

Klip ) Corg,lip/Cdiss

FIGURE 3. Relationship between individual PAH concentrations in PEDs and worms, where the concentration of an individual PAH is expressed as a log-transformed fraction of the total PAHs in the PED or worm. Each point represents an individual PAH as a fraction of the total PAHs in the worm or PED for a particular site. which means that, for each PAH, the relative amounts of that PAH taken up by the PED and the worm are similar, even

(8)

where Corg,lip is the lipid-normalized concentration of a given PAH in the organism (mol/g of lipid) and Cdiss is the freely dissolved concentration of that PAH in the porewater (mol/L of porewater). Empirical Klip values for each PAH were calculated according to this equation on the basis of the porewater PAH concentrations (as measured by the PEDs) and the worm tissue PAH concentrations from the exposure experiments. For each PAH, one Klip was calculated for every experimental unit (i.e., for every pair of PEDs and worms), and an average of those Klip values was taken. These average Klip values were then regressed against the KOW values for the respective PAHs to get the following relationship for PAH uptake in marine polychaetes, as shown in Figure 4:

log Klip ) 0.76 log KOW - 0.96

(9)

This relationship is statistically significant (R2 ) 0.94, p < 0.0001, N ) 12) and correlates well with the results found by VOL. 38, NO. 4, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1159

Veith et al. (33) for fish (log Klip ) 0.85 log KOW - 0.7, R2 ) 0.897, N ) 55). General Features and Future Developments. PED PAH concentrations correlate well with worm tissue PAH concentrations when the PED and worm are subjected to the same sediment exposure. Therefore, PEDs seem to be reliable and effective simulators of PAH uptake from contaminated sediments by benthic polychaetes, with the potential to simulate other benthic organisms as well. As such, they may eliminate the necessity of using live organisms in traditional sediment bioaccumulation tests to determine the bioavailability of organic contaminants. Further research should explore possibilities of (1) decreasing the equilibration times, such as exposing the PEDs to agitated sediment slurries, (2) decreasing the time allowed for preextraction and drying of the PED upon its construction and extraction of the PED upon its retrieval from the sediment, as well as decreasing the amount of solvent used in preextraction and extraction of the PED, (3) creating PED-organism uptake relationships for other benthic organisms and organic compounds, (4) adapting the PEDs for field deployment, and (5) understanding the mechanism controlling the systematic underprediction of worm PAH accumulation by the PEDs and other similar passive samplers in the literature.

Acknowledgments We are grateful to Larisa Altshul, Brian LaBrecque, Raya Stolyar, and Scott Forsberg for their assistance with the organics analysis, and to Mark G. Cantwell and Marguerite C. Pelletier for their collection of sediments. This paper was made possible by Grant Number 5 P42 ES05947 from the National Institute of Environmental Health Sciences (NIEHS), NIH, and Kresge Center for Environmental Health Grant Number ES00002, from the NIEHS. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIEHS, NIH.

Literature Cited (1) Lee, H., II; Boese, B. L.; Pelletier, J.; Winsor, M.; Specht, D. T.; Randall, R. C. Guidance Manual: Bedded Sediment Bioaccumulation Tests; EPA-600/R-93-183; Office of Research and Development, U.S. Environmental Protection Agency: Washington, DC, 1993. (2) Bioaccumulation Testing and Interpretation for the Purpose of Sediment Quality Assessment: Status and Needs; EPA-823/R00-001; Bioaccumulation Analysis Workgroup, U.S. Environmental Protection Agency: Washington, DC, 2000. (3) Fay, A. A.; Brownawell, B. J.; Elskus, A. A.; McElroy, A. E. Environ. Toxicol. Chem. 2000, 19 (4), 1028. (4) Ozretich, R. J.; Schroeder, W. P. Anal. Chem. 1986, 58, 2041. (5) Pelletier, M. C.; Burgess, R. M.; Cantwell, M. G.; Serbst, J. R.; Ho, K. T.; Ryba, S. A. Environ. Toxicol. Chem. 2000, 19 (11), 2691. (6) Huckins, J. N.; Manuweera, G. K.; Petty, J. D.; Mackay, D.; Lebo, J. A. Environ. Sci. Technol. 1993, 27, 2489.

1160

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 4, 2004

(7) Booij, K.; Sleiderlink, H. M.; Smedes, F. Environ. Toxicol. Chem. 1998, 17 (7), 1236. (8) Sˇ imko, P.; Simon, P.; Khunova´, V. Food Chem. 1999, 64, 157. (9) Pascall, M. A.; Zabik, M. E.; Zabik, M. J.; Hernandez, R. J. Food Chem. 1997, 60 (4), 563. (10) Adams, R. Masters Thesis, Massachusetts Institute of Technology, 2000. (11) Booij, K.; Hofmans, H. E.; Fischer, C. V.; Weerlee, E. M. V. Environ. Sci. Technol. 2003, 37 (2), 361. (12) Berner, R. A. Early diagenesis: A theoretical approach; Princeton University Press: Princeton, NJ, 1980; p 36. (13) Ho, K. T.; Burgess, R. M.; Pelletier, M. C.; Serbst, J. R.; Ryba, S. A.; Cantwell, M. G.; Kuhn, A.; Raczelowski, P. Mar. Pollut. Bull. 2002, 44, 286. (14) Solorzano, L. Limnol. Oceanogr. 1969, 14, 799. (15) Ferretti, J. A.; Calesso, D. F.; Hermon, T. R. Environ. Toxicol. Chem. 2000, 19 (8), 1935. (16) Voparil, I. M.; Mayer L. M. Environ. Sci. Technol. 2000, 34 (7), 1221. (17) McElroy, A. E. Aquat. Toxicol. 1990, 18, 35. (18) McElroy, A.; Leitch, K.; Fay, A. Mar. Environ. Res. 2000, 50, 33. (19) Bock, M. J.; Miller, D. C. Limnol. Oceanogr. 1999, 44 (4), 1132. (20) Comprehensive Description of Trace Organic Analytical Methods; from the Technical Memorandum NOS ORCA 71; National Status and Trends Program for Marine Environmental Quality, National Oceanographic and Atmospheric Administation: Silver Springs, MD, 1993; Vol. IV. (21) Pruell, R. J.; Norwood, C. B.; Bowen, R. D.; Boothman, W. S.; Rogerson, P. F.; Hackett, M.; Butterworth: B. C. Mar. Environ. Res. 1990, 29, 77. (22) Lake, J. L.; McKinney, R.; Lake, C. A.; Osterman, F. A.; Heltshe, J. Arch. Environ. Contam. Toxicol. 1995, 29, 207. (23) Ryba, S. A.; Burgess, R. M. Chemosphere 2002, 48, 139. (24) Shine, J. P.; Ika, R.; Ford, T. E. Environ. Sci. Technol. 1995, 29, 1781. (25) Leslie, H. A.; Laak, T. L. T.; Busser, F. J. M.; Kraak, M. H. S.; Hermens, J. L. M. Environ. Sci. Technol. 2002, 36 (24), 5399. (26) Xu, R. Measuring explained variation in linear mixed effects models; Technical Report; Department of Biostatistics, Harvard School of Public Health, Department of Biostatistics: Boston, Massachusetts; submitted for publication. (27) Axelman, J.; Næs, K.; Na¨f, C.; Broman, D. Environ. Toxicol. Chem. 1999, 18 (11), 2454. (28) Hofelt, C. S.; Shea, D. Environ. Sci. Technol. 1997, 31 (1), 154. (29) Forbes, V. E.; Forbes, T. L Environ. Toxicol. Chem. 1997, 16 (5), 1002. (30) Krauss, M.; Wilcke, W. Environ. Sci. Technol. 2001, 35 (19), 3931. (31) West, C. W.; Kosian, P. A.; Mount, D. R.; Makynen, E. A.; Pasha, M. S.; Sibley, P. K.; Ankley, G. T. Environ. Toxicol. Chem. 2001, 20 (5), 1104. (32) Spacie, A.; Hamelink, J. L. Environ. Toxicol. Chem. 1982, 1, 309. (33) Veith, G.; Defoe, D.; Bergstedt, B. J. Fish. Res. Board Can. 1979, 36 (9), 1040. (34) Luellen, D. R.; Shea, D. Environ. Sci. Technol. 2002, 36 (8), 1791.

Received for review July 3, 2003. Revised manuscript received November 19, 2003. Accepted December 8, 2003. ES034706F