Introduction

For over 40 years, Dechlorane Plus (DP, C₁₈H₁₂Cl₁₂) has been used as an additive flame retardant in electrical wires and cables, computer connectors, and plastic roofing materials. Commercial grade DP is synthesized by the Diels-Alder addition of 2 mol fully chlorinated cyclopentadiene to 1 mol of cyclooctadiene. The main products of the reaction are two stereoisomers, syn and anti, and small amounts of byproducts arising from impurities within the cyclooctadiene starting material.

Despite its commercial longevity, DP has only recently been detected in the environment. Hoh et al. (2006) were the first to report on the isomers in air, fish, and sediment samples from the Great Lakes region (1). The chemical structure of the DP-isomers would suggest that their physical-chemical properties should be similar to those of the higher brominated flame retardants (BFRs) like decabromodiphenylether (BDE-209); based on what is known about the environmental behavior of BDE-209, the detection of the DP-isomers in the environment comes as no surprise. What is surprising is that the compound has remained unreported for so long.

Dechlorane Plus is unregulated for use and is considered a high production volume (HPV) chemical, so it is subject to the United States Environmental Protection Agency's HPV challenge. It is also listed on Canada's Domestic Substances List (DSL) which has been the subject of review and classification for persistent, bioaccumulative, and toxic compounds. With the move to phase out production of some BFR-based formulations, it is possible that DP might be used as their replacement in some applications.

The recent detection of the DP-isomers in Great Lakes biota spurred our efforts to examine the bioaccumulation potential and trophic transfer of the isomers in food webs from Lake Winnipeg and Lake Ontario (Canada). The current study builds on our earlier work reporting on select BFRs in both food webs (2, 3). The syn- and anti-isomers of DP that are predominant in the technical mixture are now com mercially available and were used in this study to produce more accurate environmental occurrence data and to assess whether there were differences in the accumulation potential between them.

Materials and Methods

Chemicals. Individual solutions of the syn- and anti-isomers (50 g/mL, in toluene, purity >95%), brominated diphenyl ether congeners (BDE) 71, 126, 156, 197, and 207, chlorinated diphenyl congener 99 (CDE-99) and the mass labeled HBCD isomers (¹³C and d₁₈-) were supplied by Wellington Laboratories (Guelph, ON, Canada).

Sample Information. Detailed sample information can be found in Law et al. and Tomy et al. (2, 3). Samples for both studies were collected between 2000 and 2003. Additional sediment samples from the central basin of Lake Ontario collected by Environment Canada (EC) in 1998 were also analyzed. All samples were processed at the Freshwater Institute (FWI) laboratory. Biological samples were homogenized with dry ice, spiked with a suite of recovery internal standards, and extracted using accelerated solvent extraction (ASE), lipid removal by gel permeation chromatography (GPC), and further cleanup using Florisil. Sediment samples were first freeze-dried and extracted in a manner identical to that of biota with the exception of GPC. Stable isotope analysis of nitrogen was previously determined on biota to define trophic levels (TL).

Gas Chromatography/Mass Spectrometric Analysis. Lake Winnipeg and Lake Ontario extracts were analyzed at FWI and EC, respectively. Both laboratories employed identical methodologies: analyses were performed on Agilent 5973 GC-mass selective detectors (Mississauga, ON, Canada) fitted with a 10 m DB-5 capillary column (0.25 m film thickness × 0.25 mm i.d; J&W Scientific, Folsom, CA). Splitless injections of 2 L were made onto an injector set isothermally at 280 C. The initial oven temperature was set at 90 C with no hold time, ramped at 20 C/min to 310 C, and held for 5 min. The MS analysis was performed in the electron capture negative ion mode using methane as the buffer gas. Source and quadrupole temperatures were both set to 150 C. The dominant peak in the molecular ion cluster of the two isomers (m/z 651.8; spectra were identical) was used for quantitation while the second most abundant peak (m/z 653.8) was used for confirmation. The retention time for the syn- and anti-isomers was 11.17 and 11.42 min, respectively. Extraction efficiencies of the BDE congeners 71, 126, 197, and 207 were measured using the [Br]^- ions (m/z 79 and 81) while that of the CDE-99 congener was done using the [Cl]^- ions (m/z 35 and 37).

Quality Control. The extracts employed in this study were those previously used for our work on BFRs in both lakes. In those studies, chlorinated diphenyl ethers (CDEs) and mass labeled hexabromocyclododecane (HBCD) isomers were used as recovery internal standards. To test if BFRs are extracted in a manner similar to that of the DP isomers, a controlled experiment was conducted whereby a suite of BDE isomers, including 71, 126, 197, and 207, CDE-99, and ¹³C--HBCD, along with the syn- and anti-isomers (using the technical product) were spiked into an ASE cell containing hydromatrix (baked for 6 h at 600 C) at an elevated (200 pg, n = 6) and lower (50 pg, n = 6) dose. The spiked ASE cells were immediately extracted under conditions identical to those reported previously (2, 3). Extracts were then further processed through the GPC and adsorption chromatography cleanup steps followed by the addition of an instrument performance internal standard (IPIS), BDE-156, and injected onto the GC/MS; for HBCD analysis the d₁₈--HBCD was added as an IPIS and analysis was based on liquid chro matography tandem mass spectrometry as described in Tomy et al. (4). Mean recoveries of the four BDE recovery standards in the low dose range were 93.5 ± 6.8% (mean ± 1 × standard error), CDE-99 was 95.1 ± 8.5%, ¹³C--HBCD was 89.6 ± 7.9%, and DP-isomers were 76.4 ± 3.4%, while in the elevated dose respective mean recoveries of the four BDEs, CDE-99, ¹³C--HBCD, and DP were 90.3 ± 1.9, 98.2 ± 6.5, 96.5 ± 6.4, and 102.1 ± 3.8%. One-way ANOVA testing indicated that there was no difference between the recoveries of the BFRs or DP isomers at either treatment level. Therefore, these results suggest that it would be suitable to use the recoveries of the internal standards analyzed previously to correct our DP concentrations. However, because recoveries of CDEs in the Lake Winnipeg and ¹³C₂-labeled HBCD in the Lake Ontario samples were consistent in the sample extracts and greater than 82% no recovery correction was applied to the data. Duplicates of sculpin from Lake Ontario, extracted separately and analyzed to check for repeatability, were within 94% of each other for both isomers suggesting good repeatability.

An interlaboratory comparison (EC and FWI) on the technical mixture gave excellent agreement: the respective contributions of the syn- and anti-isomers in the mixture were estimated to be 36 and 64% by EC laboratory and 34 and 66% by FWI laboratory (Table 1). Four sediment extracts were also analyzed in both laboratories and showed good agreement (63 to 86%) between the measured DP isomer values and suggested good precision between laboratories.

Method detection limits (MDLs) were estimated from the procedural blanks which consisted of Ottawa sand. Trace amounts of both isomers (syn, 0.6 pg; anti, 2.6 pg) were present in the blanks. Using an average sample mass of 15 g, MDLs of 0.3 and 1.5 pg/g for syn- and anti-isomers, respectively, were determined. The linear dynamic range of the instruments was 10-2500 pg on column (r ² > 0.995) for both isomers. The ratio of the quantitation and confirmation ions in samples was within 15% of measured standard values in all cases.

Statistical Analysis. Statistical treatment of the data was done using SigmaStat (Version 9.01, Systat Software Inc.).

Results and Discussion

Lake Winnipeg. Concentrations of DP isomers in the Lake Winnipeg food web are given in Table 2. The syn-isomer was consistently detected in all samples while the anti-isomer was less frequently detected (~ 45% of samples). In biota, concentrations of the syn-isomer were greatest in burbot (range, 67-773 pg/g, lw; median, 415 pg/g), zooplankton (range, 469-647 pg/g lw; median, 542 pg/g), and mussels (range, 76-823 pg/g lw; median, 504 pg/g), while the anti-isomer was greatest in walleye (range, 608-883 pg/g lw; median, 714 pg/g) and goldeye (range, 594-932 pg/g lw; median, 763 pg/g). Concentrations of both isomers were similar in whitefish but varied considerably in walleye and goldeye; respective concentrations of the anti-isomer were 25 and 14 times greater than that of the syn-isomer in these species. Sediments contained small pg/g (dry weight) concentrations (syn, 11.7 pg/g; anti, 18.3 pg/g dry wt) of both isomers.

Lake Ontario. Concentrations of the isomers in the Lake Ontario food web are given in Table 3. Both isomers were detected in all samples with anti-DP consistently greater than that of the syn isomer. Similar concentrations of both isomers were observed in trout (median, syn = 44.3, anti = 47.2 pg/g lw), smelt (median, syn = 5.5, anti = 6.5 pg/g lw), alewife (median, syn = 48.3, anti = 54.2 pg/g lw) and sculpin (median, syn = 626, anti = 777 pg/g lw). Concentrations of the anti-isomer were approximately 2.5 times greater than the syn-isomer in Diporeia, 3 times greater in Mysis, and 2 times greater in plankton. In sediments, the anti-isomer comprised ~85% of the total DP concentrations (mean DP = 206 ng/g dry wt), which were orders of magnitude greater than those from Lake Winnipeg.

Concentrations of BFRs determined in our other studies on the same samples are presented along with the DP concentration data in Tables 2 and 3. In general, DP concentrations were 2-3 orders of magnitude smaller than those of PBDE and HBCD in biota from Lake Winnipeg. These discrepancies in concentration are likely due to differences in production and usage volume patterns among the compounds, but differences in bioavailability, bioaccumulation, and biotransformation potentials cannot be ruled out. In sediments, DP concentrations were only slightly less than HBCD concentrations supporting the hypothesis that reduced bioavailability of the DP-isomers may be a contributing factor to the small concentrations observed in biota.

Interestingly, the syn-isomer was detected more frequently than the anti-isomer in biota from Lake Winnipeg. Of the five species exhibiting no detectable anti-DP, the syn-isomer levels in burbot, zooplankton, and mussels were an order of magnitude higher than those in the other species. An enrichment of the syn-isomer may be expected since the structural conformation of the anti-isomer should be more susceptible to biological attack relative to the less sterically hindered syn-isomer, as noted by Hoh et al. (1), although significant concentrations of the syn-isomer with no detection of the anti-isomer was surprising. However, factors such as differences in bioavailability, and bioaccumulation, and/or biotransformation efficiencies are likely to affect the isomeric distribution as well.

In Lake Ontario, differences among contaminant concentrations are even more pronounced. For example, compared to PCBs, DP concentrations were 5 orders of magnitude smaller in biota. Slightly smaller concentration differences were seen for the other contaminants relative to DP. In sediments, however, mean DP concentrations (206 ng/g) were approximately 1.5 and 4 times greater than whole-lake mean PCBs and DDT, respectively (5); this is not too surprising considering that Lake Ontario is downstream to the major DP manufacturer. Taken together these results suggest that DP-isomers are less bioavailable than PCB congeners.

Age, sex, lipid content, length, mass, and proximity to exposure of contaminants are some of the factors that can affect contaminant burdens in biota. Some of these variables along with body burdens of other contaminants were regressed against individual DP isomer concentrations (Spearman Rank Order test). Results for both lakes are presented in separate correlation matrices (Tables 4 and 5). For Lake Winnipeg, strong positive statistically significant (p < 0.05) correlations were observed for concentrations of the syn-isomer and concentrations reported by Law et al. (2) of HBCD, 1,2-bis(2,4,6-tribromophenoxy)ethane (BTBPE), and lipid content. Concentrations of the anti-isomer were not significantly correlated with any of the other contaminants but were negatively correlated with overall animal length. For Lake Ontario, neither DP isomer was correlated to HBCD, polychlorinated naphthalene (PCN) levels, or mass and lipid content (6).

It is puzzling that the DP-isomer concentrations are correlated in biota from Lake Ontario but not so in Lake Winnipeg. This lack of correlation between the isomers in the Lake Winnipeg biota samples may be influenced by undetectable concentrations of the anti-isomer in many of the biotic species from the lake. The considerable differences in syn- and anti-DP concentrations and variation in the predominance of the isomers between species also contribute to the poor correlation. If selective biotransformation of the isomer occurs, species-specific biotransformation rates may explain these findings.

Biomagnification. Two measures of trophic transfer were calculated for the DP-isomers. The first method calculated trophic level (TL) adjusted biomagnification factors (BMF_TL) and was based on the ratio of the lipid-corrected concentra tions in the predator-prey relationship given by ref 7:

Trophic magnification factors (TMFs) can also be used as a descriptor for biomagnification and are derived from the slope of the regression between an organism's lipid-normal ized contaminant concentrations and trophic position, as determined by stable isotopes of nitrogen. TMFs represent the average increase in contaminant concentration in food webs rather than the variability shown between species in BMF_TL calculations, which represent only specific predator-prey relationships.

There are a few interesting features of the BMF_TL values for the isomers in both food webs (Table 6). In the walleye/whitefish feeding relationship, only the anti-isomer had a BMF_TL >1; the large BMF_TL for the anti-isomer suggests that either there is a stereoselective elimination of the syn-isomer in preference to the anti-isomer by walleye or that walleye can metabolize the syn-isomer more readily. In Lake Ontario, only the trout/smelt feeding relationship showed BMF_TL values >1. Similar BMF_TL values between DP-isomers suggest that lake trout, unlike walleye from Lake Winnipeg, are not stereoselectively accumulating or metabolizing the isomers. Taken together, these results support the hypothesis that interspecies differences in bioaccumulation and biotransformation are likely.

Regressions of the TL against concentrations were statistically significant for both isomers only in the Lake Winnipeg food web (Figure 1). For Lake Ontario, plots were constructed using wet weight (not presented) and lipid weight concentrations but neither plot yielded significant relation ships for either isomer. Regression analysis for the Lake Winnipeg food web suggested that with a lipid weight TMF value of 2.5 (r ² = 0.12, p = 0.04), the anti-isomer was biomagnifying throughout the entire food web while the syn-isomer with a lipid weight TMF value of 0.45 (r ² = 0.17, p = 0.01) was being diluted with increasing TL. The lack of biomagnification potential of the syn-isomer in the Lake Winnipeg food web is consistent with the other data presented in this study. TMF values (wet weight) for the same Lake Winnipeg food web for HBCD of 6.3 (2) and for perfluorooctane sulfonate (PFOS) in the Lake Ontario food web of 6.1 (8) were greater than that of the anti-isomer.

Figure 1 Relationship between concentrations (ng/g, lipid weight) of syn-isomer (top left panel) and anti-isomer (top right panel) in Lake Winnipeg food web, and syn-isomer (bottom left panel) and anti-isomer (bottom right panel) in Lake Ontario food web. Error bars represent standard errors, and results of the regression analysis are given in each of the plots.

Isomeric Profiles. Tables 1 and 2 clearly show that there are differences in the relative concentrations of the isomers in environmental samples from both lakes. This can be expressed numerically as the mean fractional abundance (f_anti) of the anti-isomer which is the mean concentration of the anti-isomer divided by the sum of the mean concentra tions of syn- and anti-isomers. Because the response factors of the isomers determined from repeated injections of the individual isomers are not statistically different (n = 6, p > 0.05), the f_anti values in samples derived from peak areas or from concentrations determined using the individual isomer standard are comparable.

There are three industrial formulations of the technical mixture which differ only in the particle size of the final product (9). This suggests that the isomeric compositions of the technical mixtures are likely to be similar, although the variation between batches is unknown. The mean f_anti of the technical mixture obtained in this study was calculated to be 0.650.

The mean f_anti values in sediments from both lakes are remarkably dissimilar. While the isomeric profile in Lake Winnipeg sediments (mean f_anti = 0.610) closely resembles that of the technical mixture, sediments from the central basin of Lake Ontario were clearly depleted in the syn-isomer (mean f_anti = 0.860). These observations are in contrast to those of Hoh et al. where air and sediment samples furthest from a known production source area exhibited smaller f_anti than sites closer to the Niagara source region which had values approaching that of the technical mixture (1). Although the isomeric composition of the different formulations of technical DP should be the same, it cannot be assumed that both lakes are receiving DP inputs with the same isomeric composition. The input of environmentally "weathered" or "aged" DP would confound any interpretation of interlake variability in f_anti values in sediments. Another consideration is that if heat is required to infuse the technical DP-mixture into the end-product, changes in the isomeric ratios of the technical mixture may occur. A similar chemical change occurs during the thermal infusion of technical HBCD into its intended product. As well, DP sources may not be limited to manufacturing location(s) of the raw product, but may also extend to areas where DP is applied as a flame retardant as part of a commercial formulation (i.e., cable coatings industry).

There are also clear interlake differences in the ac cumulation pattern of the isomers in lower TL organisms. In plankton, for example, mean f_anti values in Lake Ontario (0.65) are very similar to that of the technical mixture, while the syn-isomer is dominant in plankton from Lake Winnipeg, although the syn-isomer is enriched in plankton relative to sediments in both Lakes. It is not clear why the same species would show such a different accumulation pattern. If accumulation of DP-isomers by zooplankton is solely by diffusion from water, then differences in concentration profiles of the isomers in water between the lakes might explain these differences. Clearly, a better understanding of limnology of the two lakes might be needed to tease out these differences.

Another noteworthy difference in mean f_anti values are for Diporeia and sediment from Lake Ontario. Diporeia are benthic organisms and accumulation of contaminants from sediments is thought to be their primary exposure route. The mean f_anti value of 0.704 for Diporeia from Lake Ontario compared to the 0.860 for sediment indicates that stereoselective accumulation may be taking place. Biotransformation is less likely as Diporeia are thought to have low metabolic capabilities.

In biota, there were no clear trends between mean f_anti values and TL. However, the range of f_anti values in biota with increasing trophic level from the lakes (Lake Winnipeg, 0.962 to 0.003; and Lake Ontario, 0.766 to 0.512) strongly suggests that the overall extent of depletion and resulting enrichment of the anti- and syn-isomers, respectively, is much greater in the Lake Winnipeg ecosystem. These observations lend further support to the argument that interspecies differences in bioaccumulation and biotransformation are likely driving factors in the distribution of the isomers in the environment.

While the concentrations of the DP-isomers were small relative to those of other chemicals like PBDEs and HBCD, their inclusion in monitoring programs should be relatively straightforward because they are not discriminated against in the extraction and cleanup procedures. Further monitoring for this compound should be considered as production and usage patterns for flame retardants are shifting as a result of regulation of PBDEs. However, assessments of the chemical fate and bioaccumulation/biotransformation processes which may impact environmental levels of DP would lead to improved understanding of monitoring observations.

Acknowledgment

Gilles Arsenault and Brock Chittim (both of Wellington Laboratories, Canada) are thanked for providing the individual Dechlorane Plus isomers. We thank Sheryl Tittlemier (Health Canada) for her helpful comments on an earlier version of the manuscript.