Experimental exposure of eggs to polybrominated diphenyl ethers BDE-47 and BDE-99 in red-eared sliders (Trachemys scripta elegans) and snapping turtles (Chelydra serpentina) and possible species-specific differences in debromination
Karen M. Eisenreich,
University of Maryland Center for Environmental Science, Chesapeake Biological Laboratory, Solomons, Maryland, USA
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Polybrominated diphenyl ethers (PBDEs) are compounds that have persistent, bioaccumulative, and toxic properties, and because of their extensive use, have been banned from production or are being phased out of production 1–3. Although production of PBDEs has declined, they are still widely detected in aquatic and wildlife samples. The lower brominated congeners, including BDE-47, -99, -100, and -153, are most widely detected and usually in the greatest quantity, which has been attributed to metabolism of the higher brominated congeners through debromination 4, 5 as well as exposures to the penta-BDE commercial mixture present in older products that remain in use. There is evidence that in fish, rats, and birds, reductive debromination is a potential pathway for metabolism of PBDEs 6–11. The environmental persistence of PBDEs as well as metabolism to the more bioaccumulative and toxic lower brominated congeners suggests that they are likely to remain contaminants of concern despite recently imposed restrictions on use. Therefore, there is a need for additional animal models that can provide information on potential effects resulting from major pathways of exposure to bioaccumulative and persistent compounds such as PBDEs.
The common snapping turtle (Chelydra serpentina) and the red-eared slider (Trachemys scripta elegans) possess life history and ecological traits that make them well suited to studies of persistent, bioaccumulative compounds. The snapping turtle does not reach sexual maturity until 11 to 16 years of age, and the red-eared slider can require up to 10 years to reach maturity, providing both species the potential to accumulate persistent lipophilic compounds for long periods prior to reproduction. This chronic bioaccumulation can result in the maternal transfer of those compounds to their offspring during embryogenesis 12. Several studies have documented maternal transfer of persistent contaminants to eggs in turtles as well as resultant effects on embryonic, hatchling, and juvenile health and development 13–16. Kelly et al. 15 demonstrated maternal transfer of polychlorinated biphenyls (PCBs) in snapping turtles collected from the upper Hudson River, NY, USA. Hatchlings from eggs collected from the same area showed increased mortality beginning eight months after hatching, and mortality was correlated with total PCB concentrations in eggs 16. High incidences of deformities have also been documented in snapping turtles and painted turtles (Chrysemys picta) in areas where high concentrations of polycyclic aromatic hydrocarbons (PAHs) have been documented in turtle fat 13.
Quantifying impacts of specific contaminants on embryos is difficult in environmental settings, because natural habitats often contain complex chemical mixtures. To overcome such complexities, controlled laboratory exposure studies are required. In some cases, exposures of captive adults to contaminants and subsequent assessment of their offspring have been used to provide a more controlled but natural maternal transfer. Typically, these types of studies use species with short reproductive cycles or are conducted where there is access to captive populations or colonies of the organisms 17–20. However, individual variations in physiological traits among adults can bring about considerable variability in transfer of contaminants to offspring, making establishment of exposure-response relationships tenuous.
Alternatives to natural maternal transfer studies that can eliminate potential parental affects have been to deliver controlled doses to the eggs directly via injection or to employ topical egg dosing techniques in which the compound is applied directly to the egg surface and allowed to passively diffuse through the shell. Egg injections have been successfully employed in avian studies with various contaminants 8, 19–22. However, injection of reptilian eggs has frequently resulted in very high embryonic mortality rates ranging from approximately 44 to 79% 23–26, suggesting that the reptilian egg is more sensitive to the physical damage associated with injection relative to the avian egg. Schnars and colleagues 27 reported the highest hatching success (61%) after injection in snapping turtle eggs harvested directly from the oviducts of adult females prior to laying, which is a much lower success rate than in studies employing topical dosing techniques. Compared with reptile egg injection techniques, topical dosing methods do not have such a negative impact on hatching success, as shown in several studies reporting hatching success greater than 85% in the control groups 26, 28, 29.
Although topical dosing has been shown to be less detrimental to reptile embryos than egg injections, only a few studies have verified the percentage of the topically applied dose (transfer efficiency) that ultimately is incorporated into the egg contents, an issue that must be addressed to determine exposure and effect relationships 29. The present study sought to quantify transfer of two PBDEs of different molecular weights and log KOWs (BDE-47 and BDE-99, molecular weights 485.8 and 564.7 g/mol, log KOWs 6.81 and 7.32, respectively) from the eggshell into the egg contents of red-eared sliders and snapping turtles to validate the method for embryonic exposure and effect studies of PBDEs. In addition, the concentrations of the two congeners on the eggshells and in the egg contents analyzed with and without the chorioallantoic membrane were quantified in an attempt to determine the location of the topically applied dose and to determine species and congener differences.
MATERIALS AND METHODS
Eggs were collected from Concordia Turtle Farm in Wildsville, Louisiana, USA. All eggs for both species were laid on May 23rd, with a total of 70 eggs from 15 red-eared slider clutches and 54 eggs from seven snapping turtle clutches used in the present study (see Fig. 1). Thirty red-eared slider and 14 snapping turtle eggs were used for background PBDE concentration determination, and 40 eggs from each species were used for dosing. On collection, eggs were weighed and measured for diameter (snapping turtle eggs) and length and width (red-eared slider eggs), after which they were placed in bins containing damp vermiculite mixed with water in a 1:1 ratio. Eggs were stored at ambient temperature at approximately 18 to 20°C until June 12 to slow development but maintain egg viability during collection, transportation, and study preparation 30. Prior to incubation beginning on June 12 at 26°C, a temperature known to produce only males, the eggs were candled to determine egg viability (see Fig. 1) 30, 31. Only viable eggs were used for dosing. Moisture and humidity were maintained by misting the eggs and nest substrate with water at 2- or 3-d intervals. Previous research has found that within-clutch variation in contaminant concentrations in the egg contents is very low 32. Thus, prior to incubation, two eggs from each clutch (15 red-eared slider clutches and seven snapping turtle clutches) were randomly selected and egg contents homogenized together for contaminant analysis of background concentrations of PBDEs. All procedures were approved by the University of Maryland Center for Environmental Science Institutional Animal Care and Use Committee (S-CBL-07-04).
Dimethyl sulfoxide (DMSO; Fisher Scientific) was used as a vehicle for the topical application of BDE-47 and -99 to the upper surface of the eggshells for both species, because it has been employed in topical dosing studies and shown to have very little toxicity to reptilian embryos 33–35. Working stock solutions for both BDE congeners were prepared by adding neat BDE-47 or -99 (Accustandard) to DMSO, followed by further dilution of the working stock in DMSO to prepare the topical dosing solutions for both congeners to achieve a target embryonic exposure of 40 ng/g for each egg. The target concentration was selected because it represents the concentration of BDE-47 and -99 found in snapping turtle eggs collected from areas of known environmental contamination 36. The concentrations of the doses for each species were calculated based on a 20% transfer rate of BDE-47 and -99 across the eggshell and chorioallantoic membrane into the embryo as determined in a previous pilot study (K.M. Eisenreich and C.L. Rowe, unpublished data). Doses for each species were further adjusted for the average egg mass (red-eared sliders 11.6 g and snapping turtles 14.5 g). All dosing solutions were analytically verified (see below). Prior to the start of incubation, eggs from each clutch were randomly assigned to either the BDE-47 or the BDE-99 treatments for both species and placed into incubation bins separately by species and treatments to limit cross-contamination of BDE-47 and -99. Initial sample sizes for dosed eggs were 20 and 21 red-eared slider eggs and 27 and 26 snapping turtle eggs topically dosed with BDE-47 and -99, respectively.
Solutions were topically applied to the vascularized upper surface of the eggshell over a period of 8 d to avoid an acutely toxic embryonic exposure starting on the first day of incubation, for a total applied solution volume of 40 µl (see Fig. 1). The range in egg masses for the red-eared sliders and snapping turtles was 7.9 to 16.3 g and 10.7 to 18.6 g, respectively, so the dose applied to each egg was adjusted for individual egg mass to prevent drastically under- or over-exposing the embryos of eggs in a given treatment. To account for the variation in mass, the eggs were categorized into 1-g mass classes with a total of eight classes for each species, and dosing solution volumes were adjusted for each of the eight classes. The dosing solution was then administered each day in 5-µl volumes until the entire dose had been applied. For the largest egg mass class, 8 d of dosing was required to administer the entire dose, resulting in a total applied volume of 40 µl. All smaller mass classes received an equal volume, but of a more dilute BDE solution (e.g., 5 d of doing solution and 3 d of DMSO for smallest mass class). Dosing was based on nominal target concentrations of 40 ng/g.
After 8 d of dosing, eggs were incubated for 14 d, after which they were frozen (approximately 55 d from hatching) for analysis of BDE-47 and -99 on the eggshell and in the contents (yolk and albumin; see Fig. 1). Prior to extraction, eggs were thawed and contents removed from the shell. For 10 eggs from each treatment and species, the chorioallantoic membrane was extracted along with the shell; and for an additional 10 eggs, the chorioallantoic membrane was extracted along with the egg contents. This was done to estimate the concentration of the two BDE congeners in the chorioallantoic membrane, because the membrane alone did not contain enough tissue mass to extract on its own.
To assess background concentrations, total PBDEs (34 congeners representing the tri- through deca-homolog groups) were analyzed in two eggs from each clutch of each species that had been homogenized and combined for a single composite sample. In addition, eggshells and contents were analyzed for BDE-47 and -99 for estimation of transfer efficiency of the two congeners across the eggshell and chorioallantoic membrane for both species. Individual egg contents (n = 20 for both species) were homogenized and the eggshells (n = 20 for both species) cut into small pieces, after which water from the samples was removed by adding sodium sulfate and grinding to further homogenize the samples. All samples were then extracted using accelerated solvent extraction (ASE 300; Dionex) with dichloromethane. Samples were packed atop deactivated alumina in stainless-steel extraction cells to remove potential interfering lipids and other polar compounds. Prior to extraction, 13C-BDE-15 and 13C-BDE-118, surrogate standards were added to each extraction cell for calculation of analyte recoveries. Extracts were then concentrated and subjected to purification using deactivated Florisil column chromatography for removal of nonpolar interferences. A sodium sulfate blank sample was run concurrently with each set of extractions as a means of quality assurance for measurement of any laboratory contamination during the extraction and purification procedures. All samples were blank-corrected to account for any laboratory background contamination. Dosing solutions were diluted in hexane, with BDE-47 and -99 concentrations directly determined, along with all extracted egg contents and shell sample BDE concentrations, using a gas chromatograph (Agilent 6890N) coupled to a mass-selective detector (Agilent 5973N) operated in negative chemical ionization mode. Prior to analysis, 13C-CDE-86 (2,2´,3,4,5-pentachlordiphenyl ether) and 13C-BDE-209 were added as internal standards to all samples and calibration standards. All BDE standards were purchased from Cambridge Isotope Laboratories, Wellington Labs, and Accustandard, or were received from the U.S. National Institute of Standards and Technology (NIST). The programmable temperature vaporization (PTV) injector was used in pulsed splitless mode with 5-µl injections and a 15-m DB-5MS column (J&W Scientific) having an inner diameter of 0.25 mm and 0.1 µm film thickness. Instrument program specifications follow those methods routinely used in our laboratory 37. The mass fragments m/z −79 and −81 were monitored for di- to octa-BDEs, −487 and −409 for the nona-BDEs and BDE-209, −318 and −316 for 13C-BDE-209, and −495 and −415 for 13C-BDE-209 for quantitative and qualitative ions, respectively.
Three times the analyte mass in the laboratory blanks divided by the mass of the sample extracted was determined to be the method detection limit (MDL) for all analytes. The MDLs averaged 0.123 and 0.164 ng/g wet weight for BDE-47 and BDE-99, respectively. Mean recoveries (± 1 standard error) for BDE surrogate standards 13C-BDE-15 and 13C-BDE-118 in red-eared slider (n = 15) and snapping turtle (n = 7) eggs used for background concentrations were 121 ± 4.5% and 118 ± 5.3% as well as 117 ± 6.7% and 102 ± 6.3%, respectively. Mean recoveries for eggshells were 87.0 ± 2.5% and 95.6 ± 2.6% for the red-eared slider and 86.8 ± 2.3% and 92.8 ± 2.1% for the snapping turtle. Finally, recoveries for the red-eared slider contents of dosed eggs were 89.2 ± 1.9% and 99.4 ± 2.2% as well as 86.1 ± 1.9% and 95.3 ± 2.2% for the snapping turtle dosed egg contents. The recoveries of the surrogate standards were abnormally high in the eggs used for background concentrations. Sample values were not corrected for the surrogate recoveries.
All data were analyzed using Minitab Statistical Software (Version 15; Minitab). Concentrations of BDE-47 and -99 on the eggshell and in the contents of eggs for both species as well as the transfer efficiencies defined as the percentage nanograms in the egg contents were analyzed by analysis of variance (ANOVA), followed by Tukey's pairwise comparisons. Analysis of the chorioallantoic membrane BDE-47 and -99 content was completed using two sample t tests to comparing differences in membrane contributions to egg content and shell concentrations for each species and congener. The relationship between BDE-99 concentrations and BDE-47 concentrations (ng/g wet wt) in all red-eared slider egg contents of the eggs topically dosed with BDE-99 was assessed by linear regression. Statistical significance was judged based on a type I error rate of α = 0.05. Prior to statistical analyses, data were tested, and we verified that they met the assumptions of the statistical model using Levene's test for homogeneity of variance and Shapiro–Wilk (W) statistic for normality in distribution, employing log10 transformations as necessary with the exception of BDE-47 and -99 content in the shell, for which the data were rank transformed.
Background concentrations and dosing solutions
Background concentrations of total PBDEs detected in the eggs were extremely low, with the snapping turtle egg composites having a greater number of detectable congeners and higher concentrations than the red-eared slider egg composites. Red-eared slider egg composites had background total PBDE concentrations (mean ± SE) of 0.324 ± 0.073 ng/g wet weight. Only BDE congeners -100, -153, and -154 were detected in all red-eared slider egg composites. Additional congeners detected were -183 in 12 egg composites, -99 in three egg composites, and -47, -196, -197, -198/203, and -204 in one egg composite. In the snapping turtle eggs analyzed for background PBDEs, total PBDEs was 4.614 ± 1.124 ng/g wet weight, with congeners -47, -99, -100, -153, -154, and -155/85 detected in all egg composites. Congeners -28/33 and -183 were detected in two egg composites, and -138 was detected in one egg composite. The only red-eared slider egg composite that contained BDE-47 had a concentration of 0.137 ng/g wet weight, whereas the mean background concentration of BDE-47 in snapping turtle egg composites was 10 times higher (1.35 ± 0.338 ng/g wet wt). Similarly, mean BDE-99 in the three red-eared slider egg composites in which it was detected was 0.0690 ± 0.002 ng/g wet weight compared with 0.769 ± 0.386 ng/g wet weight in the snapping turtle egg composite. Overall, the background concentrations of both BDE-47 and -99 in egg composites for both species were much lower than the target concentration in the contents of dosed eggs (40 ng/g wet wt).
Red-eared slider BDE-47 and -99 dosing concentrations were analytically verified to be 63.7 and 70.4 ng/µl, respectively, lower than the calculated concentrations (77.1 ng/µl) needed to achieve the target 40 ng/g in the egg contents over the 8-d dosing period. Dosing concentrations of BDE-47 and -99 for the snapping turtle were verified to be 104 and 111 ng/µl, respectively, higher than the calculated concentrations (96.3 ng/µl). No other PBDE congeners were detected in the dosing solutions.
BDE-47 and BDE-99 in egg contents and eggshells
Overall, regardless of dose, the egg contents contained higher concentrations of BDE-47 than BDE-99 in both species. Concentrations of BDE-47 in egg contents of snapping turtle eggs dosed with BDE-47 were significantly greater than BDE-47 in red-eared slider egg contents (p < 0.0001; Fig. 2). In addition, BDE-47 concentrations in eggs of both species dosed with BDE-47 were significantly greater than the BDE-99 in eggs dosed with BDE-99 (all comparisons p < 0.0001; Fig. 2). Although not statistically significant, BDE-99 concentrations in red-eared slider egg contents appeared lower than BDE-99 in snapping turtle egg contents in eggs dosed with BDE-99 (p = 0.0546; Fig. 2).
Concentrations of BDE-47 and -99 on the eggshells of both species were reversed from differences found in the egg contents. Across species and congeners, BDE-47 concentrations were lowest on red-eared slider eggshells and significantly lower than BDE-47-dosed snapping turtle eggs (p = 0.0412; Fig. 2) and BDE-99-dosed eggs of both species (both comparisons p < 0.0001; Fig. 2). Likewise, snapping turtle eggs dosed with BDE-47 had significantly lower concentrations than BDE-99 on eggshells of both species dosed with BDE-99 (both comparisons p < 0.0001; Fig. 2). Finally, concentrations of BDE-99 on red-eared slider eggshells were significantly lower than BDE-99 on eggshells of snapping turtles dosed with the same congener (p = 0.0004; Fig. 2).
Transfer efficiency comparisons
Transfer efficiency, calculated as the percentage of administered compound detected in the egg contents, was compared across treatments and species. To estimate the total egg content tissue mass, the shell tissue mass was subtracted from the total egg mass, and this was used to convert the BDE concentrations in the 3-g aliquot of egg contents to total BDE mass (in nanograms) in the total egg contents tissue. Percentage transfer of BDE-47 across the eggshell was similar for both the red-eared sliders and the snapping turtles (25.8 ± 1.9% vs 31.3 ± 1.6%, respectively), as was the case for the transfer of BDE-99 (9.9 ± 1.1% vs 12.5 ± 1.4%, respectively). There was significantly greater transfer of BDE-47 across the eggshell relative to BDE-99 in both species (p < 0.0001 for both species; Fig. 3).
Impacts of the chorioallantoic membrane
The chorioallantoic membrane did not appear to act as an accumulator of BDE-47 or BDE-99 or a barrier to transfer from the shell to the egg contents of either congener. Only red-eared slider eggs dosed with BDE-47 showed a significant difference in concentrations when comparing samples of egg contents with the chorioallantoic membrane included to the samples without the chorioallantoic membrane (40.3 ± 3.51 and 57.7 ± 5.88 ng/g wet wt, respectively). The contents analyzed with the chorioallantoic membrane included in eggs dosed with BDE-47 had significantly (p = 0.042) lower concentrations of BDE-47 than contents without the chorioallantoic membrane. Similar comparisons were made with the eggshells, and it was found that snapping turtle eggshells dosed with BDE-47 and -99 had significantly lower concentrations of both congeners when the chorioallantoic membrane was analyzed with the shell, compared with the shell alone (p = 0.002 and p = 0.007, respectively). For snapping turtle eggs dosed with BDE-47, the concentration of BDE-47 on the shell without the chorioallantoic membrane was 106 ± 8.91 ng/g wet weight, compared with 61.8 ± 7.95 ng/g wet weight when the membrane was analyzed with the shell. For snapping turtle eggs dosed with BDE-99, the concentration of BDE-99 on the shell without the chorioallantoic membrane was 710 ± 76.5 ng/g wet weight, compared with 472 ± 28.8 ng/g wet weight when the membrane was analyzed with the shell.
Total dose accounting
The entire eggshell (10 with and 10 without the chorioallantoic membrane) was analyzed for BDE concentrations, and 3-g aliquots of egg contents were analyzed (10 with and 10 without the chorioallantoic membrane). To gain an estimate of the total egg content tissue mass, the shell mass was subtracted from the total egg mass, and this was used to convert the BDE concentrations in the egg contents to total mass (in nanograms) of BDE in the contents. With this estimation of total BDE within or on the egg, it was determined that, for both compounds and species, we could not account for more than 50% of the total dose topically applied to the eggs (Fig. 4). Specifically, 25.8 ± 1.9% of the total dose of BDE-47 was accounted for in the red-eared slider egg contents. This was significantly different from the 9.9 ± 1.1% in the egg contents accounted for of the total BDE-99 dose applied to red-eared slider eggs (p > 0.0001). However, the percentage of BDE-47 accounted for in the red-eared slider egg contents was not different from the amount accounted for in the egg contents of the BDE-47-dosed snapping turtle eggs (31.3 ± 1.6%, p = 0.0536) but was significantly greater than the amount of BDE-99 in the egg contents of the BDE-99-dosed snapping turtle eggs (12.5 ± 1.4%; p < 0.0001). The percentage of BDE-99 accounted for in red-eared slider egg contents was significantly less than the BDE-47 in snapping turtle eggs (p < 0.0001) but was not significantly different from the BDE-99 accounted for in snapping turtle egg contents (p = 0.5891). In snapping turtle egg contents, the percentage of BDE-47 accounted for was significantly greater than the percentage BDE-99 accounted for (p < 0.0001).
Substantial concentrations of BDE-47 were detected in the egg contents of red-eared slider eggs dosed with BDE-99 (2.70 ± 0.650 ng/g wet wt). The BDE-47 detected cannot be accounted for by the background concentrations, because BDE-47 was nondetectable in the contents of all but one slider egg composite analyzed for background concentrations. In addition, the BDE-99 concentrations in all red-eared slider egg contents showed a positive and significant relationship with the concentrations of BDE-47, although only 25.7% of variation was explained by the linear regression model (p = 0.022, r2 = 0.257; Fig. 5). In contrast, the BDE-47 concentrations in the egg contents of snapping turtles dosed with BDE-99 were lower than the concentrations recorded in the background snapping turtle eggs (0.722 ± 0.093 ng/g wet wt vs 1.35 ± 0.338 ng/g wet wt), and there was no relationship between BDE-99 and BDE-47 concentrations in the egg contents (p = 0.366).
Studies using topical egg-dosing techniques for exposing reptile embryos to chemicals have often either not quantified transfer across the eggshell into the egg contents or done so in too few samples to achieve adequate characterization of variability 29. However, topical dosing techniques can possibly serve as a viable option for controlled embryonic exposure studies of dose-response relationships as long as exposure is verified. Although it is known that a low percentage of the total compound (i.e., dichlorodiphenyltrichloroethane, 2,3,7,8-tetrachlorodibenzo-P-dioxin, and 3,3´,4,4´,5-pentachlorobiphenyl) 38–40 placed on top of the egg is transferred to the contents, the total dosing concentration applied can be calibrated to achieve an amount in the egg contents at concentrations found in environmental samples and/or those suspected to cause adverse effects 38–40.
The eggs used in the present study were collected from a turtle farm, so the egg contents did not reflect environmental concentrations of PBDEs as may occur in contaminated sites 36. Background BDE-47 and -99 concentrations were very low to negligible, so any BDE-47 or BDE-99 found in the egg contents of those eggs used in the dosing study were the result of dosing and transfer across the eggshell. Use of eggs from captive turtles reduces the potential influence or interference of background concentrations of bioaccumulative contaminants such as PBDEs.
Findings from the present study indicate that PBDE congeners -47 and -99 can be transferred across the eggshell and chorioallantoic membrane into the egg contents following our protocol. Transfer efficiency was higher for BDE-47 than for BDE-99 in both species but was generally low (BDE-47: red-eared slider 25.8%, snapping turtle 31.3%; BDE-99: red-eared slider 9.9%, snapping turtle 12.5%). These transfer efficiencies are within the range reported from the few other reptilian topical dosing studies in which transfer of other compounds was quantified. Gale et al.  reported that 4% of total 2,3,7,8-tetrachlorodibenzodioxin (TCDD) and 10% of total PCB-126 topically applied to the egg crossed the eggshell into the egg contents over a 16-d period in red-eared slider eggs. A shorter period of dosing (72 h) resulted in 1.6 to 20% transfer of p,p′-dichlorodiphenyldichloroethylene (p,p-DDE) across the eggshell of snapping turtles 39, whereas a longer period (49 d) resulted in approximately 33% transfer of p,p′-DDE across the eggshell of the green sea turtle (Chelonia mydas) 38. Similarly, less than 2% transfer occurred in alligator (Alligator mississippiensis) eggs topically exposed to DDE, dieldrin, and chlordane for 14 d 26. In several species of birds, eggs injected with a penta-BDE mixture (DE-71) that included both BDE-47 and -99 resulted in only 18.8 to 29.6% of the total injected amount being absorbed into the egg contents by date of pipping (26 d of incubation) 8.
Transfer efficiency was higher for BDE-47 than for BDE-99 in both species, but there also appeared to be a difference in transfer between the two species. Although not statistically different, the observed pattern of greater transfer efficiency of BDE-47 compared with BDE-99 likely is due to the greater molecular size of BDE-99 (564.7 vs 485.8 g/mol) or larger log KOW (7.32 vs 6.81), preventing it from being as readily transferred through the eggshell pores. Lower absorption of the higher brominated congeners was also described by McKernan et al. 8, who injected chicken, mallard, and American kestrel eggs with either a penta-BDE mixture (DE-71) or an octa-BDE mixture (DE-79) and found significantly less DE-79 in the egg contents. In addition, they observed preferential uptake of the lower brominated congeners across the incubation period in the eggs injected with DE-71 8.
Differences in transfer of the same congener between species are less easily explained. The eggshells and chorioallantoic membranes of both species have similar compositions, with eggshells having similar distributions and sizes of pores 41, so it is unlikely that differences in shells or chorioallantoic membranes between the two species would explain transfer efficiency differences. The red-eared slider egg, because of its oval shape, has a larger surface area to spread out the multiple topical doses, compared with the round snapping turtle egg, suggesting potential for higher transfer across the eggshell of the red-eared slider. However, it is unclear why transfer was higher across the snapping turtle egg, especially for BDE-47.
The chorioallantoic membrane did not seem to retain either BDE-47 or BDE-99. In this study, chorioallantoic membranes were analyzed either with the eggshell or with egg contents to obtain an estimate of how much BDE may be retained in the membrane. Egg contents analyzed without the chorioallantoic membrane had higher concentrations of the two BDE congeners than egg contents analyzed with the membrane for both species. This suggests that inclusion of the chorioallantoic membrane might have diluted the concentrations of the congeners in the egg contents. Among the prior studies that measured transfer across the eggshell after topical application, none reported concentrations in the membrane or indicated whether the membrane was analyzed with shell or egg contents. However, McKernan et al. 8 analyzed pooled air cell membranes collected from chicken eggs injected with DE-71 and determined that 4.3% of the injected dose was associated with the membrane. They also estimated that an additional 32% of the administered dose could potentially be associated with the inner membrane of the chicken egg 8. These observations suggest that a large concentration of the injected dose remained associated with the membranes.
For both congeners in both species, less than 45% of the dose applied to the egg was accounted for. There are several possible explanations for the loss of the remaining dose. We observed that during the 8 d of dosing, the DMSO absorbed moisture from the air in the incubation chamber. This absorption of water caused the total volume of liquid on the egg to increase and occasionally roll off the egg. In these instances, a portion of the dose was potentially lost from the eggshell. In addition, it is possible that a portion of the PBDE remaining on the eggshell volatilized over the 22 d of incubation. A third possibility for PBDE loss from the eggshell is handling of the egg and eggshell for storage and analysis, potentially removing a portion of the dried dose from the eggshell.
An additional mode of loss, metabolism of the parent compound, cannot be ruled out as a potential mechanism of loss in red-eared slider eggs. It was not within the scope of this study to characterize metabolites of BDE-47 and -99 such as methoxylated or hydroxylated compounds, but we observed that egg contents of red-eared sliders (but not snapping turtles) that had been dosed only with BDE-99 contained a significant concentration of BDE-47 above the background concentration, suggestive of reductive debromination of BDE-99 to BDE-47 in ovo. McKernan et al. 8 reported the presence of several PBDE congeners in the egg contents of several species of birds that had been injected with DE-71 but were not detected in the dosing solution or control eggs. Although early embryonic development of red-eared sliders and snapping turtles is similar 42, 43, species-specific differences are likely in enzymatic systems potentially responsible for differential capacities for debromination, as has been shown in fish exposed to BDE-99 7, 10. Several studies of fish have shown debromination of BDE-99 to BDE-47 in carp, but other species such as Chinook salmon and rainbow trout demonstrate different reductive metabolite products and slower formation rates 6, 7, 9, 10. Deiodinase enzymes (DIs)-which are responsible for the removal of iodine atoms from thyroid hormones, resulting in activation or deactivation of the hormones-have been suggested as the primary enzymatic system for debromination of BDE-99 to BDE-47 in fish, particularly in carp 7, 9, 10. Preliminary data suggest higher DI activity in red-eared sliders than in snapping turtles, suggesting that the DIs may be a hypothetical metabolic pathway for reductive debromination in red-eared sliders 44. Glutathione-S-transferase (GST), which is responsible for phase II conjugation processes through dehalogenation reactions, have also been reported to debrominate PBDEs 45 but are less likely to be involved in debromination in fish 9, 10. Although it is possible that there could be differences in the enzymatic systems of the embryonic red-eared slider and snapping turtle that would result in differential debromination capacities, further study is required to assess specifically the extent of debromination in the two species and the mechanisms by which it occurs.
Topical dosing techniques used in this study were successful in delivering BDE-47 and -99 into the egg contents of both red-eared sliders and snapping turtles. However, topical dosing of BDE-99 resulted in a concentration lower than the target in the egg contents, likely because of the greater molecular size of the compound relative to BDE-47. Concentrations of BDE-99 were lower in the red-eared sliders, compared with snapping turtles, perhaps because of metabolism of the parent compound by the red-eared slider embryos. Based on the low transfer efficiency of both BDE-47 and -99 across the eggshell and differences in the concentrations in the egg contents, it is difficult to predict an embryonic exposure from topical application. However, the small variation observed in the concentrations of BDE-47 and BDE-99 in the egg contents of both species suggests that the amount of BDE in the egg contents can be replicated for a single compound at a single concentration. Additional studies would be needed to determine whether transfer across the eggshell into the egg contents would remain constant over a range of topical doses and what portion of the concentration in the egg contents actually reaches the embryo.
The present study was supported by graduate fellowships granted to K. Eisenreich by the Society of Environmental Toxicology and Chemistry/Procter & Gamble and the U.S. Environmental Projection Agency STAR Graduate Fellowship Program (FP-91690301). We thank A. Sides for egg management and sampling, and J. Baker, H. Stapleton, and A. Heyes for analytical support. The information presented in the present study has not been subjected to review by the supporting agencies, and no official endorsement should be inferred. Publication charges were paid by the Gene Lane Fund for Turtle Research. This is contribution 4680 of the University of Maryland Center for Environmental Science.