Screening Bioavailable Hydrophobic Toxicants in Surface Waters with Semipermeable Membrane Devices: Role of Inherent Oleic Acid in Toxicity Evaluations

Ecotoxicology and Environmental Safety 44, 160}167 (1999) Environmental Research, Section B Article ID eesa.1999.1802, available online at http://www.idealibrary.com on

Screening Bioavailable Hydrophobic Toxicants in Surface Waters with Semipermeable Membrane Devices: Role of Inherent Oleic Acid in Toxicity Evaluations D. Sabaliu nas,*? J. Ellington,- and I. Sabaliu niene ? *Department of Ecology, Lund University, S 223 62, Sweden; -National Exposure Research Laboratory, United States Environmental Protection Agency, 960 College Station Road, Athens, Georgia 30605-2700; and ?Faculty of Nature Sciences, Vilnius University, C[ iurlionio 21, 2009 Vilnius, Lithuania Received September 4, 1998

Semipermeable membrane devices (SPMDs) were deployed for 4 weeks in two rivers in Lithuania. The SPMD dialysates were tested in the Microtox assay and, surprisingly, the sample from the relatively clean U la River exhibited three times more toxicity than the sample from the Vilnia River receiving discharge from several industrial enterprises and municipal wastewater. The whole dialysates were subjected to bioassay-directed fractionation on silica gel columns. Toxicity testing of each fraction revealed that most of the toxicity was contained in fraction 10, which was eluted with 100% acetone. GC/FID, GC/ECD, and GC/FTIR/MS analysis of the fractions indicated that the major component of this fraction was oleic acid. The oleic acid was most likely the hydrolysis product of methyl oleate, the major impurity of the SPMD triolein. It can be inferred that oleic acid was responsible for the toxicity of this fraction in Microtox, as a threefold di4erence in the toxicity between the two samples was also marked by a threefold di4erence in their oleic acid content. Toxicity of unsaturated fatty acids in various tests, including Microtox, has been demonstrated elsewhere. Vilnia fraction 2, which was eluted with 100% hexane, exhibited the most toxicity of the remainder of silica gel fractions. The spectral analysis demonstrated that other toxic fractions contained a number of halogenated compounds and PAHs. In general, SPMDs have proved to be a useful way to screen for hydrophobic toxicants in water. However, sample clean-up procedures to remove oleic acid may be required prior to toxicity testing for the estimation of the true toxic potential of the accumulated pollutants.  1999 Academic Press Key Words: semipermeable membrane devices (SPMDs); toxicity testing; environmental monitoring; screening of aquatic pollutants; bioassay-directed fractionation. INTRODUCTION

Rapid, e!ective, and low-cost integrated environmental monitoring methods are of great value, especially in East  To whom correspondence should be addressed at Swedish address. E-mail: [email protected].

European countries, which are facing acute environmental problems while they are also undergoing rapid economic and infrastructure changes. Although resources for the protection of the environment are limited, the implementation of low-tech and cost-e!ective monitoring methods is often feasible. Ideally, such methods should allow one not only to directly monitor the fate and concentrations of chemicals in the environment but also to evaluate their e!ects and assess the potential hazard these chemicals pose for both the ecosystem and human health. Semipermeable membrane devices developed by Huckins et al. (1990, 1993) have proved to be a useful tool in the time-integrated monitoring of hydrophobic pollutants in aquatic ecosystems. In these devices, the uptake of chemicals is based on the process of passive partitioning of a compound between water and the synthetic lipid triolein that is enclosed in thin-walled lay-#at polyethylene tubing. Since triolein is a major component of the lipid pool of the aquatic organisms (Chiou, 1985; Ewald and Larsson, 1995) and the molecular size-exclusion limit of the polyethylene membrane is similar to that of biological membranes (Opperhuizen et al., 1985), SPMDs can be used as indicators of bioavailability of chemical pollutants. SPMDs are timeintegrated samplers, and thus re#ect the true extent of exposure of living organisms to contaminant chemicals in the aquatic environment. Since their development in the late 1980s, SPMDs have been successfully applied under a variety of experimental and environmental settings (Lebo et al., 1995; Petty et al., 1995, 1998; Prest and Jacobson, 1995; Prest et al., 1995; Ellis et al., 1995; Gale et al., 1997; Strandberg et al., 1997; Ockenden et al., 1998). In many cases, they proved superior to aquatic organisms in terms of the concentrations and the number of accumulated pollutants as well as in terms of total amounts available for chemical analysis. Although most of these studies have focused on the identi"cation and quanti"cation of chemicals accumulated by SPMDs, it has been demonstrated that the SPMD

160 0147-6513/99 $30.00 Copyright  1999 by Academic Press All rights of reproduction in any form reserved.

OLEIC ACID IN TOXICITY EVALUATIONS

technique can be readily integrated with standard bioassays to measure toxic and mutagenic e!ects of accumulated pollutants (Cleveland et al., 1997; Sabaliu nas and SoK dergren, 1997; Sabaliu nas et al., 1998). Such an approach allows rapid and low-cost screening of bioavailable hydrophobic chemicals in the aquatic environment and provides information about the extent of hazard these chemicals pose to the living components of the ecosystem. Depending on the results of bioassays, the SPMD dialysates may be analyzed to identify the compounds responsible for the observed e!ects applying, for example, bioassay-directed fractionation and toxicity identi"cation evaluation (TIE) techniques. The main goal of the present study was to further investigate the use of SPMDs in screening the e!ects of chemical pollutants in the aquatic environment. The SPMDs were deployed in two rivers in Lithuania, and the bioluminescence inhibition in the marine bacteria
Semipermeable Membrane Devices SPMDs were prepared at the Department of Ecology, Lund University, Sweden, in a manner similar to that described earlier (Sabaliu nas and SoK dergren, 1997). The preextracted polyethylene tubing (membrane thickness 75}80 lm, diameter 2.5 cm, from Brentwood Plastics, MO) was cut into 50-cm-long segments which were "lled with 0.5 ml (0.455 g) of triolein (95% purity, from Sigma Chemical, St. Louis, MO) con"gured to form a thin "lm. The e!ective length of the SPMDs (distance between the two thermosealed ends of the segment) was 45 cm, resulting in a membrane surface area-to-lipid volume ratio of 520 cm/ml. Freshly prepared SPMDs were stored in clean metal containers (gas tight paint cans) at !203C, and they were transported from Sweden to the study locations in Lithuania at about 43C. Study Locations and SPMD Deployment Two sites were selected for the study: the Lower River Vilnia in the capital city of Vilnius and the River U la in the Dzu kija National Park in the south of the country. They are both relatively small rivers with a similar #ow rate. The River Vilnia receives pretreated wastewater from several industrial enterprises as well as municipal wastewater from one district of the city, while the River U la is regarded as

161

a clean natural river with a very limited in#ow of industrial and municipal discharge. In May 1996, six SPMDs were deployed at each site. They were placed in metal protective shields and submerged in the running river water. Field blank consisted of six SPMDs that accompanied the other SPMDs to the study locations. The exposure lasted for 4 weeks. The mean (n"3) water temperature during the exposure was 14.53C in the Vilnia River and 13.23C in the U la River. At the end of the exposure, the SPMDs were transported back to Sweden where they were processed at the laboratory of the Department of Ecology, Lund University, as follows. Each SPMD was rinsed with running tap water, and debris was removed manually. The SPMDs were examined for cuts and leaks and air-dried after a quick rinse with acetone. All six SPMDs from each site and the "eld blank were cut into "ve to six pieces each and extracted in glass bottles with 500 ml of hexane for 48 h (with solvent replacement after 24 h). The volume of the extracts was reduced to about 5 ml using rotary evaporation and nitrogen blow-down. The extracts were then passed through a kilned (4003C) and solvent prewashed glass column packed with 2 g of anhydrous Na SO , which was supported by a glass "ber "lter (70 lm).   Several solvent rinses were used to ensure recovery of the applied chemicals. The volume of the combined rinses was reduced to 2.5 ml using nitrogen blow-down. The concentrated samples were transferred into preextracted 2.5-cmwide polyethylene tubes with one end sealed, which were placed upright into 250-ml grad cylinders. The length of the tubes matched the height of the cylinders. Hexane (250 ml) was added carefully to each cylinder, and the extracts in the tubes were dialyzed for 48 h with solvent replacement after 24 h. The dialysates were evaporated to a volume of 2 ml and again "ltered through Na SO and glass "ber "lter. The   "ltrates were divided into two equal parts. One part was solvent-exchanged to 1 ml of acetone for testing in the Microtox. The other part was sealed in glass ampules and shipped to the National Exposure Research Laboratory, U.S. EPA, Athens, Georgia, for fractionation and chemical analysis. Silica Gel Fractionation of SPMD Extracts The dialysates were concentrated to 200 ll by nitrogen blow-down. Silica gel (SG) columns (6 ml column, 1 g silica gel, from J.T. Baker, Phillipsburg, NJ) were prerinsed with 10 ml hexane; elution continued until the last portion of hexane reached the surface of the silica gel. The Vilnia or U la concentrates were then added to the SG column. The stopcock was opened to lower the hexane level to the top of the column. The fractionation on the silica gel column was performed with 3-ml volumes of hexane}methylene chloride}acetone}methanol in the proportions 3 : 0 : 0 : 0, 2.85 : 0.15 : 0 : 0, 1.2 : 2.8 : 0 : 0, 0 : 3 : 0 : 0, 0 : 0 : 3 : 0, and

SABALIU NAS, ELLINGTON, AND SABALIU NIENE

162

TABLE 1 Percentage Distribution of the Components of the Standard Mixture in the Silica Gel Fractionation 100% Hexane

Fraction No.: C-19 IC-16 2,6-DMN p,p DDT Chry 4-CIP PA

1

5% Methylene chloride/ 95% hexane

2

3

98 70

2 30

60% Methylene chloride/ 40% hexane

4

5

6

95 43

5 57 23

77

100% Methylene chloride

7

8

100% Acetone

9

10

100% Methanol

11

12

100 100

Note. C-19, nonadecane; IC-16, iodohexadecane; 2,6-DMN, 2,6-dimethylnaphthalene; Chry, chrysene; 4-CIP, 4-chlorophenol; PA, palmitic acid.

0 : 0 : 0 : 3, ml : ml : ml : ml. Twelve 1.5-ml fractions were collected. The "rst 1.5 ml was the void volume. Each fraction (1.5 ml) was concentrated to 500 ll by nitrogen blow-down and subsequently analyzed on capillary gas chromatographs equipped with an ECD, FID, or FT-IR/MS. After the GC analyses were complete, the remainder of the SG fractions was quantitatively solvent-exchanged into acetone, sealed in glass ampules, and shipped to the Department of Ecology, Lund University, to be tested in the Microtox assay. To determine functional group elution sequence from the silica gel columns, a mixture that contained nonadecane (C-19), iodohexadecane (IC-16), 2,6-dimethylnaphthalene (2,6-DMN), p,p-DDT, chrysene (Chry), 4-chlorophenol (4-CIP), and palmitic acid (PA) was eluted from an SG column using the solvent elution sequence described above. Fractions (1.5 ml) were collected, concentrated to 0.5 ml, and analyzed by capillary GC with detection by either FID or ECD. The sequence of elution and distribution between fractions is given in Table 1.

Gas Chromatography/Fourier Transform}Infrared Spectroscopy/Mass Spectrometry (GC/FT-IR/MS)

Gas Chromatography/Electron Capture and Flame Ionization Analysis (GC/ECD and GC/FID)

Microtox Assay of the SPMD Extracts

The SPMD samples were injected on a Hewlett Packard (Palo Alto, CA) 5890 Series II gas chromatograph equipped with a DB-5 capillary column (30-m length, 0.32-mm ID, and 0.25-lm "lm, J&W Scienti"c, Folsom, CA), a Nickel 63 electron capture and #ame ionization detectors. Helium was used as the carrier gas at 25 cm/s and nitrogen was the make-up gas. The GC oven was held at 353C for 4 min, programmed at 93C per min to 2403C, and held there for 10 min. Splitless injections (1 ll) were made at an inlet temperature of 2503C.

GC/FT-IR/MS analyses were performed on a Hewlett Packard 5890 Series II gas chromatograph interfaced with a 5965B infrared detector and a 5971 series mass selective detector. For injections, the capillary columns and oven temperature program were the same as for ECD and FID analysis. The FT-IR lightpipe and transfer line were held at 2903C and the MS transfer line was maintained at 2853C. Before any samples were analyzed, acceptable system performance was veri"ed by injection of a mixture of more than 70 organics covering a broad range of functional groups, volatilities, polarities, and molecular weights; all at 20 mg/liter in methylene chloride. Analysis of this standard prior to sample injection also allowed the estimation of sample component concentrations based on detector response relative to selected components of the standard. Under acceptable system performance, an injection of 3 ll of this standard a!orded easy identi"cation of all test mix compounds with a signal-to-noise ratio of about 20 for both IR and MS.

Both nonfractionated SPMD extracts and their fractions were tested in the Microtox bioassay, which is a commercial toxicity test from Azur Environmental, Carlsbad, CA. It is based on the inhibition of luminescence in the marine bacteria <. ,scheri as a result of toxic action of chemicals. The experiments were carried out following the standard procedures of the manufacturer. The luminescence intensity was measured with the Microbics Model 500 Analyzer at 5and 15-min exposure times. Each test was run in duplicate using four sample concentrations and a negative control. Concentration of the carrier solvent (acetone) in the test

163

n.t. (134.0) 160.5 142.1}181.3

medium, including controls, was 1%. The calculated EC  values denote sample concentration that reduces the light output by 50%. RESULTS

Toxicity and Chemical Composition of Whole Samples

? The concentration values are based on the amount of SPMD triolein. @ Not toxic (highest concentration tested).

n.t. (134.0) 34.7 21.9}55.0 Vilnia Estimate 95% CI

0.182 1.148}0.224

11.95 11.15}12.80

122.9 103.9}145.2

50.64 48.39

170.1 116.1}249.1

n.t. (134.0)

n.t. (134.0)

n.t. (134.0)

192.9 170.5}218.3

0.136 0.115}0.162

79.8 77.6}82.1 52.1 45.4}59.8 0.066 0.064}0.068 n.t. (134.0) 212.9 149.8}302.5 n.t. (134.0) n.t. (134.0) 197.1 142.9}271.8 110.2 101.1}120.2 72.7 40.7}152.4 n.t. (134.0) n.t. (134.0)@ 0.062 0.057}0.067 34.7 21.9}55.0 U la Estimate 95% CI

100% methanol 100% acetone 100% methylene chloride 60% methylene chloride: 40% hexane 5% methylene chloride: 95% hexane 100% hexane

Fraction 9 Fraction 8 Fraction 7 Fraction 6 Fraction 5 Fraction 4 Fraction 3 Fraction 2 Fraction 1 Whole sample Field blank EC  (5 min), mg/liter?

TABLE 2 Toxicity of SPMD Dialysates and Their Silica Gel Fractions in the Microtox Assay

Fraction 10

Fraction 11 Fraction 12

OLEIC ACID IN TOXICITY EVALUATIONS

Both U la and Vilnia samples exhibited high toxicity in the Microtox assay with the calculated EC values in the  range of micrograms of the SPMD triolein per milliliter of the bacterial suspension (Table 2). The toxicity of the U la sample was about three times higher than that of the Vilnia sample. A slight toxicity was also observed in the "eld blank, and it can probably be attributed to minor impurities in the solvents which were used in large volumes for the SPMD dialysis with the subsequent radical reduction of the dialysate volumes and possible concentration of some of these impurities with toxic properties. The Vilnia and U la whole-sample FID chromatograms were essentially identical (Fig. 1). The notable di!erence was that the U la extract contained three times as much oleic acid as the Vilnia extract. The other major peaks in the chromatograms identi"ed as a homologous series of straight chain alkyl hydrocarbons by IR/MS and coinjection of a mixture, which contained the C-10 through C-19 straight-chain hydrocarbons. The chromatograms contained a continuum of unresolved peaks in the C-13 to C-19 retention time window, which appeared as a &&hump'' in the chromatogram. A notable di!erence was observed in the Vilnia wholesample ECD chromatogram when compared to the U la ECD chromatogram. The Vilnia chromatogram contained both a much greater number and more intense peaks than the U la chromatogram, especially in the retention times bracketed by the C-13 to C-19 retention time window. One peak, with the same retention time and almost the same area in all three chromatograms, was used as a relative reference peak. The U la extract contained only three additional peaks of larger area than the reference peak, while the Vilnia extract contained 15 peaks of equal or greater area. The response to the ECD was an indication of the possible presence of compounds, which contained halogens and heteroatoms such as sulfur and oxygen. The complexity of the whole-sample chromatograms precluded identi"cation of minor components by FT-IR/MS. Diwerences between the Fractions The results of the toxicity testing (Table 2) clearly indicate that almost all of the toxicity of both samples was contained in fraction 10 (eluted with 100% acetone). This fraction was about two times more toxic in the U la sample than in the Vilnia sample (Fig. 2). There is no doubt that the latter fraction predetermined the overall higher toxicity of the U la

164

SABALIU NAS, ELLINGTON, AND SABALIU NIENE

FIG. 1. Mirrored FID chromatograms of the whole SPMD dialysates.

sample compared to the Vilnia sample. Interestingly, the sum of the relative toxicities (de"ned as 1/EC ) of the rest  of the fractions is higher for the Vilnia sample (Fig. 3). This is largely due to the pronounced di!erence between the two samples in the toxicity of fraction 2 (100% hexane). The ECD chromatograms indicated that in both the Vilnia and U la samples, fractions 2, 3, and 4 contained the majority of the ECD-sensitive compounds. Compared to the SG fractions of the U la sample, the ECD response of the

Vilnia fractions was more intense, as indicated by the number of peaks and peak areas (Fig. 4). The MS and IR spectra for peaks in fraction 2 of both samples were typical for alkyl and branched-chain alkyl hydrocarbons; the &&hump'' in the C-13 to C-19 region as seen in the whole-sample chromatograms was contained in this fraction. Fraction 3 in both samples contained the remainder of the alkylhydrocarbons and C-1, C-2, and C-3 naphthalenes. Several mass spectra in the Vilnia fraction 3

FIG. 2. Relative toxicity (1/EC ) of the whole SPMD dialysates and their silica gel fractions in the Microtox test. 

OLEIC ACID IN TOXICITY EVALUATIONS

165

FIG. 3. Relative toxicity (1/EC ) of the SPMD silica gel fractions 1}9 and 11}12 in the Microtox test. 

exhibited &&isotopic clusters'' indicative of halogen containing compounds. Interestingly, Vilnia fractions 4 and 5 contained pyrene and #uoranthene but neither compound was detected in the U la fractions. However, phenanthrene, anthracene, 9H-#uorene, methyl 9H-#uorene, biphenyl, alkyl substituted biphenyls, and alkyl substituted naphthalenes and phenanthrenes were detected in fractions 4 and 5 in both the Vilnia and the U la samples. A compound was identi"ed by MS and IR as benzene, 1,11-sulfonylbis in fractions 8 and 9 for both extracts.

The results of the chemical analysis of fractions 2 to 5 that indicated a higher content of halogenated organic compounds and PAHs in the Vilnia sample compared to the U la sample are in line with the Microtox toxicity data, as the summaric relative toxicity of these fractions is about three times higher for Vilnia than for U la. The most toxic U la and Vilnia fraction 10 contained the same suite of compounds based on retention times and their MS and IR spectra, except the major component identi"ed as oleic acid was at a three times higher level in the U la sample.

FIG. 4. Mirrored ECD chromatograms of the SPMD silica gel fraction 2.

SABALIU NAS, ELLINGTON, AND SABALIU NIENE

166 DISCUSSION

The overall higher toxicity of the U la sample compared to the Vilnia sample was rather unexpected since, as mentioned above, the River U la was regarded as one of the cleaner rivers in the country and it is often used as a reference site in various environmental monitoring studies. However, testing of separate fractions has revealed that the di!erence was due to the highly toxic fraction 10 which contained oleic acid. Most of the other fractions were more toxic and contained more anthropogenic compounds in the Vilnia sample than in the U la sample, which is in line with expectations based on the pollution history and geographic location of these two sites. The toxicity testing and chemical analysis results of fraction 10 suggest that oleic acid may have been responsible for the high toxicity of this fraction and di!erences in the toxicity between the two samples. Toxicity of fatty acids has been demonstrated in a number of tests, including Microtox (Has kanson et al., 1991). The toxicity of unsaturated fatty acids is attributable to their membrane-disturbing properties: their incorporation in the lipophilic phase of the double layer of the cell membrane may result in increased membrane #uidity followed by increased membrane permeability and leakage of cell components (Ewald and Sundin, 1993). Apart from the increased membrane #uidity, the oleic acid may cause a disruption of the calcium pump and other ion transport mechanisms due to the formation of metal salts. Finally, by disrupting the membrane structure, oleic acid may also increase the susceptibility of the cell to some minor, not-identi"ed components of the toxic fraction. The oleic acid present in the samples was either the hydrolysis product of the SPMD triolein or, even more likely, methyl oleate, which constitutes most of the 5% impurities of commercial triolein used in the SPMDs (James Huckins, personal communication). Methyl oleate readily di!uses to the exterior surface of the SPMD membrane. Once a periphytic community is established on the membrane surface, methyl oleate is degraded by the microorganisms with hydrolysis to oleic acid being a probable pathway. As concentration of oleic acid becomes higher on the exterior membrane surface compared to the inside, a concentration gradient is established and a net increase in oleic acid in the SPMD will occur over time. Also, methyl oleate may be hydrolyzed abiotically. The rate of this process during the "eld exposure would depend on a number of factors, such as the water temperature during the SPMD exposure, abundance, diversity and properties of periphytic organisms, oxygen content, light intensity, and others. These factors apparently caused the di!erence in the oleic acid content of the SPMDs deployed in the two rivers. Oleic acid was not detected in the "eld blank, as the blank SPMDs were stored under conditions not amenable to hydrolysis or biodegradation.

If oleic acid was indeed a major contributor to the toxicity of the samples, it may have important implications for the use of SPMDs in screening the toxicity of hydrophobic pollutants in the aquatic environment. It may be necessary to employ a set of alternative bioassays insensitive to unsaturated fatty acids to verify the toxicity of the sample or to re"ne the sample clean-up procedures to eliminate oleic acid prior to toxicity testing in <. ,scheri. Similarly, we have demonstrated (Environmental Pollution, in press) that in the anaerobic environment such as some sediments, SPMDs tend to accumulate elemental sulfur, which is mainly of natural origin. Since sulfur is highly toxic to <. ,scheri (Jacobs et al., 1992; Salizzato et al., 1998), it is also necessary to pretreat such samples for sulfur removal before testing. One must note, however, that the possibility of at least some of the oleic acid being sequestered by SPMDs from the water cannot be completely ruled out. It has been reported that unsaturated fatty acids, including oleic acid, were responsible for 87 to 97% of the toxicity of e%uent water from forest product and pharmaceutical industries in <. ,scheri (Svenson et al., 1996). Unsaturated fatty acids were also major components of bleached kraft mill e%uent toxicity (Cherr et al., 1987; Crooks and Sikes, 1990). CONCLUSIONS

SPMDs were highly e!ective in concentrating bioavailable hydrophobic chemicals from surface waters. The sequestered chemicals subsequently induced a toxic response in the Microtox assay. Only a small amount of the sample was needed to perform a full-scale test and obtain accurate EC values. In this respect, testing SPMD dialysates is  clearly advantageous over testing extracts from water where large volumes usually need to be processed to obtain su$cient amount of material for bioassays. Moreover, contrary to the extracts from water samples obtained by &&grab sampling,'' SPMDs accumulate pollutants over a period of time, and they re#ect a truer exposure situation of aquatic organisms. The bioassay-directed fractionation of the SPMD dialysates indicated that oleic acid was likely to be one of the major contributors to the toxicity of the samples. Further research is needed to establish the role of oleic acid in determining the toxicity of SPMD concentrates and to re"ne the method to separate it from the sample prior to toxicity testing. Toxic fractions, other than the oleic acid fraction, contained a number of halogenated compounds and PAHs which were present at higher concentrations in the Vilnia sample compared to the U la sample. This was not surprising since the Vilnia River receives waste-water from several industrial enterprises, whereas the U la River is virtually free from point source pollution. Even though the applied fractionation scheme has helped to tentatively identify major

OLEIC ACID IN TOXICITY EVALUATIONS

toxicants of the SPMD dialysates, the complexity of GC/ECD, GC/FID, and GC/FT-IR/MS chromatograms of the individual fractions suggests that a more re"ned fractionation is needed to accurately identify and quantify the speci"c chemicals, and especially the minor components of the pollutant mixture. ACKNOWLEDGMENTS The authors thank Prof. A. SoK dergren at Lund University, Sweden, and Prof. J. Lazutka at Vilnius University, Lithuania, for comments and critical review of the manuscript. Special thanks go to Dr. James Huckins at the U.S. Geological Survey, Columbia, Missouri, for valuable comments and suggestion of the methyl oleate hydrolysis hypothesis. The authors further acknowledge Dr. T. Collette and Mr. George Yager at the National Exposure Research Laboratory, U.S. EPA, Athens, Georgia, for providing the IR/MS spectra. This study was supported by grants from the Swedish Environmental Protection Agency and the Swedish Institute. This manuscript has been reviewed in accordance with the o$cial U.S. EPA policy. Mention of trade names does not indicate endorsement by the U.S. EPA or the U.S. federal government.

REFERENCES Cherr, G. N., Shenker, J. M., Lundmark, C., and Turner, K. O. (1987). Toxic e!ects of selected bleached kraft mill e%uent constituents on the sea urchin sperm cell. Environ. ¹oxicol. Chem. 6, 561}569. Chiou, C. T. (1985). Partition coe$cients of organic compounds in lipidwater systems and correlations with "sh bioconcentration factors. Environ. Sci. ¹echnol. 19, 57}62. Cleveland, L., Little, E. E., Petty, J. D., Johnson, B. T., Lebo, J. A., Ora, E., Dionne, J., and Crockett, A. (1997). Toxicological and chemical screening of Antarctica sediments: Whole sediment toxicity tests, Microtox, Mutatox, and semipermeable membrane devices (SPMDs). Mar. Pollut. Bull. 34, 194}202. Crooks, R., and Sikes, J. (1990). Environmental e!ects of bleached kraft mill e%uents. Appita 43, 67}76. Ellis, G. S., Huckins, J. N., Rostad, C. E., Schmidt, C. J., Petty, J. D., and MacCarthy, P. (1995). Evaluation of lipid-containing semipermeable membrane devices for monitoring organochlorine contaminants in the upper Mississippi River. Environ. ¹oxicol. Chem. 14, 1875}1884. Ewald, G., and Larsson, P. (1995). Partitioning of 14C-labelled 2,2,4,4tetrachlorobiphenyl between water and "sh lipids. Environ. ¹oxicol. Chem. 13, 1577}1580. Ewald, G., and Sundin, P. (1993). ATP leakage from ELD cells after exposure to stearic, monochlorostearic, dischlorostearic, and oleic acids. Pharmacol. ¹oxicol. 73, 159}162. Gale, R. W., Huckins, J. N., Petty, J. D., Peterman, P. H., Williams, L. L., Morse, D., Schwartz, T. R., and Tillitt, D. E. (1997). Comparison of the uptake of dioxin-like compounds by caged channel "sh and semipermeable membrane devices in the Saginaw River, Michigan. Environ. Sci. ¹echnol. 31, 178}187. Has kanson, H., Sundin, P., BrunstroK m, B., Dencker, L., Engwall, M., Ewald, G., Gilek, M., Holm, H., Honkasalo, S., Idestam-Almquist, J., Jonsson, P., Kautsky, P., Lundberg, G., Lund-Kvernheim, A., Martinsen, K., Norrgren, L., Personen, M., Rundgren, M., Stas lberg, M., Tarkpea, M., and WeH sen, C. (1991). In vivo and in vitro toxicity of fractionated "sh

167

lipids, with particular regard to their content of chlorinated organic compounds. Pharmacol. ¹oxicol. 69, 459}471. Huckins, J. N., Manuweera, G. K., Petty, J. D., Mackay, D., and Lebo, J. A. (1993). Lipid-containing semipermeable membrane devices for monitoring organic contaminants in water. Environ. Sci. ¹echnol. 27, 2489}2496. Huckins, J. N., Tubergen M. W., and Manuweera, G. K. (1990). Semipermeable membrane devices containing model lipid: A new approach to monitoring the bioavailability of lipophilic contaminants and estimating their bioconcentration potential. Chemosphere 20, 533}552. Jacobs, M. W., Del"no, J. J., and Bitton, G. (1992). The toxicity of sulfur to Microtox2+ from acetonitrile extracts of contaminated sediments. Environ. ¹oxicol. Chem. 11, 1137}1143. Lebo, J. A., Gale, R. W., Petty, J. D., Tillitt, D. E., Huckins, J. N., Meadows, J. C., Orazio, C. E., Echols, K. R., and Schroeder, D. J. (1995). Use of the semipermeable membrane device as in Situ sampler of waterborne bioavailable PCDD and PCDF residues at sub-parts-per-quadrillion concentrations. Environ. Sci. ¹echnol. 29, 2886}2892. Ockenden, W. A., Prest, H. F., Thomas, G. O., Sweetman, A., and Jones, K. C. (1998). Passive air sampling of PCBs: Field calculation of atmospheric sampling rates by triolein-containing semipermeable membrane devices. Environ. Sci. ¹echnol. 32, 1538}1543. Opperhuizen, A., van der Velde, E. W., Gobas, F. A., Liem, A. K., van der Steen, J., and Hutzinger, O. (1985). Relationship between bioconcentration in "sh and steric factors of hydrophobic chemicals. Chemosphere 14, 1871}1896. Petty, J. D., Huckins, J. N., Orazio, C. E., Lebo, J. A., Poulton, B. C., Gale, R. W., Charbonneau, C. S., and Kaiser, E. M. (1995). Determination of waterborne bioavailable organochlorine pesticide residues in the lower Missouri River. Environ. Sci. ¹echnol. 29, 2561}2566. Petty, J. D., Poulton, B. C., Charbonneau, C. S., Huckins, J. N., Jones, S. B., Cameron, J. T., and Prest, H. F. (1998). Determination of bioavailable contaminants in the Lower Missouri River following the #ood of 1993. Environ. Sci. ¹echnol. 32, 837}842. Prest, H. F., and Jacobson, L. A. (1995). Passive sampling of water and coastal air via semipermeable membrane devices. Chemosphere 30, 1351}1361. Prest, H. F., Richardson, B. J., Jacobson, L. A., Vedder, J., and Martin, M. (1995). Monitoring organochlorines with semipermeable membrane devices and mussels (Mytilus edulis) in Corio Bay, Victoria, Australia. Mar. Pollut. Bull. 30, 543}554. Sabaliu nas, D., Lazutka, J., Sabaliu niene , I., and SoK dergren, A. (1998). Use of semipermeable membrane devices (SPMDs) for studies of e!ects of organic pollutants: Comparison of pesticide uptake by SPMDs and mussels. Environ. ¹oxicol. Chem. 17. Sabaliu nas, D., and SoK dergren, A. (1997). Use of semipermeable membrane devices to monitor pollutants in water and assess their e!ects: A laboratory test and "eld veri"cation. Environ. Pollut. 96, 195}205. Salizzato, M., Pavoni, B., Ghirardini, A. V., and Ghetti, P. F. (1998). Sediment toxicity measured using