Manufactured chemicals such as pesticides have been linked with reproductive or health ill effects in reptilian species (Alligator mississippiensis, Guillette et al. 1994; Caretta caretta, Keller et al. 2004, 2006). Many pesticides are persistent compounds, and their rate of accumulation is influenced by various factors such as age, sex, or diet (Milton and Lutz 1997). Some lipophilic pesticides can bioaccumulate at high concentrations in the body of animals (e.g., fat and adipose tissues) and can persist for a long time (e.g., dichlorodiphenyltrichloroethane [DDT] and its metabolites) (Milton and Lutz 1997). Hence, animals that are high in the trophic chain, such as carnivorous marine turtles, are more likely to be affected by environmental toxicants, which bioaccumulate via feeding.

Pollutants such as DDT and dichlorodiphenyldichloroethylene (DDE) have been identified as endocrine disrupters in different species (Brandt et al. 1997). A correlation between exposure to environmental toxicants and reproductive abnormalities such as reduced plasma testosterone levels and small penis sizes has been observed in American male alligators (A. mississippiensis) in Lake Apopka, Florida (Vonier et al. 1996; Guillette et al. 2000). Similarly, high concentrations of a mixture of organochlorines (OCs) present in the freshwater snapping turtle, Chelydra serpentina, altered secondary sexual characteristics, such as the ratio of the precloacal length to the posterior lobe of the plastron (de Solla et al. 1998). In addition, both OC and polychlorinated biphenyl (PCB) concentrations in eggs decreased hatching success and increased developmental abnormalities (tail, leg, digit, or claw deformities) in C. serpentina (Bishop et al. 1998).

There is also evidence that pesticides and heavy metals interfere with testosterone and/or estradiol binding protein properties in vitro in nesting green turtles, Chelonia mydas (Ikonomopoulou et al. 2009).

Marine turtles are aphagic during migration and nesting activities (Bjorndal 1985), and free fatty acids are mobilized from adipose tissue stores to provide nesting turtles with the necessary energy for the successful completion of metabolically demanding activities (Hamann et al. 2002). As a consequence of this mobilization, persistent organic pollutants (POPs) that have accumulated in the animals are also released, leading to increased pesticide concentrations in the blood (Keller et al. 2004) with detrimental reproductive or health effects. Furthermore, many POPs such as PCBs are transferred from the maternal bloodstream to eggs during the process of vitellogenesis (C. serpentina, Kelly et al. 2008; Dermochelys coriacea, Guirlet et al. 2010; Stewart et al. 2011) and therefore, turtle eggs reflect the maternal diet and contaminants that were consumed prior to nesting (reviewed in Alava et al. 2011).

This study shows evidence of small amounts of POPs in the blood and eggs of a Natator depressus population that inhabits the Great Barrier Reef, a unique marine environment worldwide, which is generally considered pristine. These data could assist in conservation management plans of this species in the Queensland nesting population and in interpreting data from other populations worldwide.

Methods

Whole blood and egg samples were collected from N. depressus during November and December 2006 at South End, Curtis Island, Gladstone, Queensland, Australia (lat 23°45′S, long 151°18′E; nap figure is provided in Ikonomopoulou et al. [2011]). Three eggs were selected at random from each of 20 individuals while they were engaged in oviposition. Once oviposition had finished and within 2 min of handling the turtle, whole blood samples (10 ml) from 20 individuals were collected using the method described in Owens and Ruiz (1980). Blood was collected into glass tubes with lithium heparin anticoagulant (Labtek PTY) for OC and PCB analysis. Blood and egg samples were stored on ice for a maximum of 2 hrs and then stored frozen at −20°C until analysis.

To assess recovery, 1 egg was injected with 200 µl of an OC spiking solution (mix A containing: hexachlorobenzene [HCB], lindane, heptachlor, aldrin, dieldrin at concentrations of 0.2 mg/l; heptachlor epoxide and DDE at concentrations of 0.1 mg/l; and DDT at a concentration of 0.3 mg/l) and 20 µl of internal standard (2-nitro-m-xylene, dibromobiphenyl, d10-pyrene, triphenylphosphate, decachlorobiphenyl). A second egg was injected with 200 µl of a polychlorinated biphenyl spiking solution (100 µl of 1 mg/kg Arochlor 1248 and 100 µl hexane) and 20 µl of the same internal standard.

The eggs were defrosted, the shells were removed and the whole contents were homogenized and refrozen at −20°C until extraction. The extraction protocol for OC and PCB analysis was originally obtained from Queensland Health Forensic and Scientific Services Organic Laboratory (QHFSSOL) and subsequently modified by QHFSSOL and van de Merwe et al. (2009). This is described in detail in Ikonomopoulou (2008) and is given in outline here.

The first extraction step was accelerated solvent extraction using an ASE 100 extractor (DIONEX). Approximately 8 g of homogenized egg was weighed into a clean beaker and the volume was made up to approximately 100 ml with Hydromatrix™/anhydrous sodium sulfate (50∶50 w/w). About 5 g of polyacrylate-polyalcohol (HPLC grade, Sigma-Aldrich) was then added to a 100-ml accelerated extraction cell and the diluted egg sample was poured in. To each sample, 100 µl of internal standard was added and the extraction cycle was run. The lipid content of the extracted egg samples was determined gravimetrically by weighing the extract, allowing the solvent to evaporate overnight and then reweighing. After lipid determination, the extract was redissolved in dichloromethane and the lipid portion was removed by pumping it though a gel permeation column with 100-Å pore size (Purex®) with a 0.45-µm teflon membrane prefilter (Millipore®). All samples, duplicate aliquots of dichloromethane only (blanks), and a quality control sample (pooled from extracts from the sea turtle C. caretta) were run through the gel permeation column and the individual effluent collected. Finally, any residual lipid was removed by adsorption chromatography using Florisil™ columns (Ikonomopoulou 2008; van de Merwe et al. 2009), producing a concentrated extract of organic pollutants in hexane. All samples including hexane blanks (n  =  2), spiked standards (n  =  2) and dichloromethane controls were analyzed for OCs and PCBs using gas chromatography–mass spectrometry (GC-MS) (Shimadzu), GCMS-QP5050. The lower limit of quantitation (LLOQ) for individual OCs and PCBs was 0.5 and 0.2 ng/g of lipid, respectively. The LLOQ was defined as 10 times the signal to noise ratio of the matrix blank. In contrast to the lower limit of detection (LLOD), which is defined as 3 times the signal to noise ratio (Shrivastava and Gupta 2011).

An approved NATA (National Association of Testing Authorities Australia) accredited method was used to extract and quantify the pollutants in blood samples. This method was described in detail in Ikonomopoulou (2008) and is given in outline here. Five grams of whole blood were weighed out for each sample, control, and a sample spiked with an internal standard, and fortified aliquots were homogenized and extracted as follows. Two milliliters of 4% sodium sulfate solution was added and vortex mixed for 15 min. Seven milliliters of hexane was then added and the sample was rotated for 40 min and placed into an ultrasonic bath for 20 min to facilitate OC extraction. All samples were then centrifuged at 800 × g (IEC centra-8R centrifuge, International Equipment Company, USA) for 20 min at 18°C. The hexane layer was carefully separated, transferred, and evaporated down to less than 2 ml over a nitrogen stream on a hot block (40°C). The extracts were combined and concentrated down via nitrogen steam to 1 ml final volume. Any residual biological macromolecules were removed by adsorption chromatography using micro Florisil™ columns (Ikonomopoulou 2008). The eluted extracts were then evaporated to 1 ml and analyzed by gas chromatography with electron capture detection (ECD). In order to confirm the identity of the compounds detected by ECD, GC-MS was also used.

In addition to the test samples, samples, a reagent blank (deionized water), and a control blood sample (pesticide-free human blood from QHFSSOL stocks) were extracted as previously described. Recovery was assessed by extracting a human pesticide-free stock blood sample spiked with 100 µl of an OC solution of similar composition to the one used for egg analysis. The recovery of PCBs was assessed by spiking a pesticide-free blood sample with 1 ml of Arochlor 1016 mix (concentration: 1.5 mg/l). An internal standard sample was also prepared in another sample for the ECD by diluting 100 µl internal standard solution to 900 µl of hexane (HPLC grade). The LLOQs were 1.5 ng/g for PCBs and 0.15 ng/g for all other OC compounds.

Validation of the analysis in the blood and eggs of N. depressus was carried out according to the method of Bishop et al. (1991). The mean recovery from whole blood samples was 80% for OCs (lindane, HCB, dieldrin, aldrin, heptachlor, and the isomer p,p-DDE) and 92% for total PCBs. In eggs, the mean recoveries were lindane, 18%; heptachlor, 62%; BZ 52, 67%; BZ 44, 41.4%; and BZ 66, 61% (nomenclature of Ballschmiter and Zell 1980). The main target analytes for this analysis in both blood and eggs for OCs and PCBs are presented in Table 1.

Table 1.  The main target analytes (organochlorine and polychlorobiphenyl concentrations) investigated in the blood and egg tissues in Natator depressus.

View Fullscreen

Results

Eleven of the blood samples analyzed had OC compounds present in amounts below the LLOQ, and the other 9 had lindane (0.35 ± 0.04 ng/g), dieldrin (0.23 ± 0.03 ng/g), HCB (0.16 ± 0.08 ng/g), the DDT breakdown product dichlorodiphenyldichloroethane (DDD) (0.24 ng/g), and total PCBs (0.26 ± 1.3 ng/g). The 9 samples with detectable PCB and OC concentrations were pooled, and the pool was analyzed by GC-MS to confirm the results. The identity of the OCs (dieldrin, DDT, DDD, lindane, HCB) was confirmed by GC-MS, but the quantitation by GC-MS was < LLOQ (< 0.15 ng/g).

However, the following PCB compounds were quantifiable by GC-MS in the pooled sample: BZ 153 (8 ng/g) and BZ 138 (6 ng/g). A peak corresponding to BZ 206 was also detected with < LLOQ (< 1.5 ng/g).

The following compounds were detectable in eggs: trifluralin, trans-nonachlor, p,p′-DDT, p,p′-DDE, dieldrin, endrin aldehyde, and several PCBs; however, all were present at amounts less than their respective lower limits of quantitation.

The lipid content of the eggs of N. depressus (n  =  30) was relatively constant, with a mean of 7.41% ± 0.09% (SE).

Discussion

In this study we detected PCBs and lindane in the blood of nesting N. depressus turtles at concentrations less than the LLOQ. The OC pesticide lindane is not commonly found in sea turtles (Milton and Lutz 1997; Pugh and Becker 2001), so its presence in this population of N. depressus (albeit in trace amounts) is of interest and needs to be further evaluated because of the potential reproductive effects. In pregnant cows, lindane causes a decrease in the ratio of progesterone to oxytocin concentration, influencing uterine contractions and enhancing a risk of abortion in pregnant females (Mlynarczuk et al. 2010). Furthermore, lindane has been reported to interfere and/or to disrupt pregnancy and embryonic development in mice and rabbits (Tiemann 2008). There have been some reports of lindane detection in other marine chelonians; for example, in C. mydas plasma (at a mean concentration of 0.78 ng/g) (Labrada-Martagón et al. 2011) and in tissues of Lepidochelys olivacea (e.g., liver 22.4 ng/g) (Gardner et al. 2003) off the coast of Baja California. In the Mediterranean Sea, 2 individuals of C. caretta were reported to have detectable plasma lindane concentrations (below 1 µg/kg) (Mckenzie et al. 1999).

The population of the carnivorous N. depressus that was studied is hypothesized to feed on the Great Barrier Reef (Limpus, pers. comm., March 2009), an area that is considered pristine. However, traces of metals and pesticides have been found in other species that feed in the area. For example, levels of DDE (45 µg/kg) and sum PCBs (171 µg/kg) have been measured in C. mydas (Vetter et al. 2001). Furthermore, dioxins (260–390 pg/g), traces of pollutant metals, including arsenic (1.4 mg/kg) and mercury (0.11 mg/kg), and pesticides, including dieldrin, DDT, and their metabolites (concentrations between 0.1 and 59 µg/kg), have been detected in the liver (metals) and blubber (pesticides) of dugongs (Dugong dugon) (Haynes et al. 1998, 1999, 2005). Metallic pollutants such as chromium (mean  =  0.17 mg/kg) and arsenic (mean  =  0.67 mg/kg) have also been detected in the eggs of N. depressus (Ikonomopoulou et al. 2011). Similarly, small quantities of polybrominated diphenyl ethers have been detected in various seafood samples in Moreton Bay, Queensland, Australia (Hermanussen et al. 2008). Overall, the data indicate that while the Great Barrier Reef region and Moreton Bay are relatively free from pollutants, low levels do exist, and it would be logical to expect to detect them initially in carnivorous species that are high in the food web such as N. depressus or long-lived species such as D. dugon.

As an example, C. caretta, which is high in the marine food web, has higher chlorobiphenyls levels (adipose tissue, 775–893 µg/kg wet weight) in comparison to C. mydas (39–261 µg/kg wet weight) and D. coriacea (47–178 µg/kg) from the Mediterranean Sea and the Atlantic waters, respectively (Mckenzie et al. 1999).

However, the concentrations of pesticides detected in this study do not approach levels that have been shown to disrupt sex steroid binding in C. mydas, which were circa 10⁻² mg/l (Ikonomopoulou et al. 2009), suggesting that reproductive impairment is unlikely at the current time. There is, however, a need for monitoring possible accumulation in the future.

The pesticide profile in the blood of these turtles was more complex than in eggs. PCBs were detected in the blood but could not be found in the eggs. This may be because of the low amounts found in all samples, which were all below the lower limit of quantitation. Analytical factors probably also contributed to the pattern as the recovery was much more variable in eggs (between 18% and 62%) than in whole blood (> 90%). The average lipid content in egg samples was within the range found (below 10%) in other sea turtle studies (Mckenzie et al. 1999; Kelly et al. 2008; van de Merwe et al. 2009), suggesting that lipids did not interfere with the main pesticide analysis. However, the worst recovery from eggs (18%) was for lindane. This highlights the difficulty of extracting organic pollutants, particularly lindane from complex matrices of lipids and proteins such as is present in egg yolk, and the method used probably requires modification. Kucklick et al. (2002) reported lindane detection in the blubber of ringed seals and polar bears (both species with high fat content) using a semipreparative aminopropylsilyl solid phase extraction column. Analytical recoveries were not reported for lindane or any other POP, but the quality control samples imply that recovery was circa 85%–90%. Repeating some of the assays with an extraction method based on Kucklick et al. (2002) would be worthwhile.

In summary, marine turtles often are used as biomarkers of environmental contamination (Milton and Lutz 1997), and this study gives basic data on a population that is considered to feed in a relatively clean marine environment. The fact that low levels of pollutants can be found in such a population, however, does suggest that there may be a need to reassess pollution management plans for this unique area, especially since low levels of lindane and DDT and its breakdown metabolites have also been measured in fish, air, and seawater samples in the Great Barrier Reef (Haynes and Johnson 2000).