Introduction

Large numbers of organic compounds have been released into the marine environment due to industrial, agricultural, and other anthropogenic activities (Dachs and Méjanelle 2010). As a result, persistent organic pollutants (POPs) have become a significant concern among these organic compounds because of their potential hazard to marine species (Jakimska et al. 2011). POPs, such as organochlorine pesticides (OCPs) and polycyclic aromatic hydrocarbons (PAHs), are recognized for their toxicity, persistence in the environment, resistance to degradation, and capability to bioaccumulate in tissues of living organisms (Devi 2020; Weltmeyer et al. 2021).

The southern Gulf of Mexico region hosts an IUCN Index site for the green turtle, Chelonia mydas (Seminoff 2004). Three of these nesting sites occur along the coast of Campeche, namely Carmen, Isla Aguada, and Sabancuy (Guzmán and García 2016). Nevertheless, the southern Gulf of Mexico accounts for 76% of Mexican oil and gas production (Schifter et al. 2005). Furthermore, elevated PAHs and OCPs have been detected in the marine environment and some species on the southern coast of Campeche (Ponce Vélez and Botello 2005). These POPs could have entered this marine environment through oil extraction and agriculture activities (Carvalho et al. 2009; Rosano-Hernandez et al. 2012).

Green turtles are exposed to POPs via sediments, water, and food and can accumulate these pollutants in their tissues (Jakimska et al. 2011; Gallen et al. 2019). However, there are still gaps in this species’ status of POPs exposure. Adverse effects of POPs have been suggested on sea turtles (Finlayson et al. 2016). During vitellogenesis, lipids and pollutants move from the maternal bloodstream to the developing eggs (Miller and Limpus 2003; Guirlet et al. 2010; van de Merwe et al. 2010). Reactive oxygen species (ROS) are produced during normal cellular metabolism; however, oxidative stress occurs when there is an imbalance between the production of ROS and antioxidant defense (Valavanidis et al. 2006). As a defense mechanism against oxidative stress and damage, organisms have developed systems that involve enzymatic and non-enzymatic antioxidants (Valko et al. 2007). Evidence shows that exposure to POPs, including PAHs and OCPs, may induce oxidative stress, causing alterations in antioxidant defense and damage to lipids, proteins, and DNA (van der Oost et al. 2003). The antioxidant system can be altered by exposure to different stressors, including chemical pollutants; their quantification and the presence of cellular damage due to oxidative stress has been used as a biomarker in many studies (Bal et al. 2022; Morão et al. 2022).

Although green turtles are regarded as species of conservation concern and are considered sentinel organisms in marine biomonitoring (Aguirre and Lutz 2004), few studies have reported concentrations of POPs in green turtles from the Gulf of Mexico and the relationship between POPs exposure and indicators of oxidative stress (Swarthout et al. 2010; García-Besné et al. 2015; Salvarani et al. 2019). The present study measured the concentrations of 36 POPs in the plasma and eggs of green turtles nesting in three sites along the Campeche coast. In addition, it assessed the antioxidant response detected in the plasma of nesting C. mydas from the southern Gulf of Mexico as well as the variation of POP concentrations in female green turtles from three different nesting sites.

Materials and Methods

Study Site and Sample Collection

The fieldwork described in this study was conducted under Mexican regulations and permits provided by SEMARNAT (SGPA/DGVS/05845/13 and SGPA/DGVS/08106/14). From a conservation perspective, blood sampling is a non-invasive method of monitoring pollutants. On the other hand, using viable eggs is a destructive technique, and the number of eggs permitted is limited. Since one egg represents less than 1% of the total number of eggs that a female may lay in the entire season in the region, it is not possible to determine how POPs concentrations in the eggs vary.

From July to September 2014, blood and eggs were collected from 29 female Chelonia mydas on 3 major nesting beaches along the Campeche coast in the southern Gulf of Mexico: Carmen 18°41′N, 91°42′W, Isla Aguada (IA) 18°47′N, 91°29′W, and Sabancuy (SB) 18°59′N, 91°11′W (Fig. 1). These beaches are located within the Laguna de Terminos Reserve. They are included in the IUCN Index site for green turtles in the southern Gulf of Mexico—an IUCN Index site is used to refer to beaches where nesting activity is intense, and long‐term nesting monitoring programs are developed to provide population trends (Seminoff 2004). However, agricultural activities in the surrounding areas may originate the release of organochlorine and organophosphorus pesticides into the marine ecosystem (Carvalho et al. 2009). Furthermore, oil and gas production has chronically impacted almost 300 km of the southern coast of the Gulf of Mexico (Rosano-Hernandez et al. 2012).

Fig. 1
figure 1

Map of the study site. Three nesting sites included Carmen, Isla Aguada, and Sabancuy, in the southern Gulf of Mexico

Full size image

Nesting sites were surveyed at night to locate female green turtles in oviposition activity. One egg was randomly collected at oviposition or immediately after the egg-laying process, rinsed with deionized water, and wrapped in aluminum foil. Once the female laid eggs, blood collection was performed on each female, and two morphometric characters were measured: curved carapace length (CCL) and curved carapace width (CCW). Blood samples (6–10 mL) were collected from the dorsal cervical sinus (Owens and Ruiz 1980) using a disposable syringe and blood collection tubes containing sodium heparin. Blood samples were centrifuged at 2000 g for 10 min, and plasma was immediately separated. All samples were frozen in liquid nitrogen until analysis at the Laboratory of Environmental Coastal Sciences (Facultad de Química, Universidad Nacional Autónoma de México, Sisal, Yucatan, Mexico).

Sample Preparation and Chemical Analysis

Concentrations of pesticides (20 OCPs and Chlorpyrifos) and 16 PAHs in maternal plasma and eggs were determined by ultrasound-assisted extraction (USE) and solid-phase extraction (SPE) based on EPA methods 3550C, 3535, and 1699 (US EPA 2007a, b, c) and analytical procedures described in Trejo-Acevedo et al. (2009). For plasma analysis, 500 to 2000 μL were used, 2 mL of a saturated solution of ammonium sulfate (0.5 g/mL), and 2 mL of ethanol were added to achieve the precipitation and denaturation of proteins. USE was conducted twice with an ultrasonic processor (Cole Parmer model CPX500) at 70% amplitude for 2 min using 10 mL of hexane as an extraction solvent. The organic phase was separated by centrifugation (370 g for 8 min). Extracts were concentrated, and OCPs and PAHs were obtained by SPE using C18 500 mg/6 mL cartridges (Supelclean ENVI-18, 57064, Supelco). Cartridges were conditioned with 6 mL of dichloromethane, 6 mL of acetone, and 6 mL of hexane. Sample extracts were passed by gravity flow and finally eluted with 6 mL dichloromethane: hexane (3:7, v/v) by gravity flow and dried under vacuum for 1 min. The solvent retrieved was changed into ethyl acetate and concentrated using a gentle nitrogen flow, and samples were analyzed by GC–MS.

Egg content (yolk and albumen) was lyophilized for 72 h and then ground into powder. For each sample, 0.5 g of the lyophilized egg was extracted by USE with an ultrasonic processor (Cole Parmer model CPX500) at 70% amplitude for 2 min using 10 mL of hexane:acetone (1:1, v/v), and the organic phase was separated by centrifugation (1760 g for 10 min). Given the high concentration of lipids in turtle eggs, extracts were cleaned by dSPE using a kit roQ™ QhEChERS (KS0-8926, Phenomenex) to eliminate interferences caused by organic and acid compounds, pigments, and sugars. Cleaned extracts were concentrated, and OCPs and PAHs were obtained by SPE using Strata® Florisil 500 mg/6 mL cartridges (8B-S013-HCH, Phenomenex). First, cartridges were conditioned with 10 mL of hexane; then, sample extracts were passed by gravity flow. Finally, cartridges were eluted with 15 mL hexane:acetone (9:1, v/v) by gravity flow and dried under vacuum for 1 min. The solvent retrieved was changed into ethyl acetate, concentrated using a gentle nitrogen flow, and analyzed by GC–MS using deuterated PAHs as internal standards.

Plasma and egg extracts were analyzed on a gas chromatograph (Agilent Technologies, GC 6850 Series II) coupled to a mass selective detector (Agilent Technologies, 5975B VL MSD) with an automatic liquid sampler (Agilent Technologies, 7683B ALS). The injection was carried out in split-less mode at 280 °C. Chromatographic separation was performed by silica capillary column Zebron ZB-5MSi (Phenomenex). The carrier gas was He (ultra-pure grade) with a flow rate of 0.8 mL/min. Each sample was analyzed using two different chromatographic conditions. For OCPs, the oven temperature was initially set at 50 °C, increased 10 °C/min to 180 °C, then raised 1.5 °C/min to 200 °C (hold time 2 min), and finally increased 6 °C/min to 290 °C (hold time 1 min). For PAHs, the oven temperature was initially set at 60 °C, then increased 6 °C/min to 290 °C (hold time 11.67 min). Mass spectra (m/z 50–550) were recorded at a rate of five scans per second with an energy of 70 eV. Mass spectrometric analysis for quantitative measurement of the analytes was performed by electronic impact (EI) in selected ion monitoring (SIM) mode by use of two characteristic fragment ions for each chemical, one target and one qualifier ion (Supplementary Tables 1 and 2). Analytical quality control included procedural blanks, calibration curves using analytical standards (Chlorpyrifos PESTANAL® 45395-100MG, Pesticide 8081 Std Mix Supelco CRM46845, QTM PAH Mix Supelco CRM47930), and internal standards.

Antioxidant Activity

Oxidative stress biomarkers measured in this study included three antioxidant enzymes (SOD, GPx, and GST), total GSH, and lipid hydroperoxides. SOD was evaluated using a Sigma-Aldrich assay kit (19160), the reaction of which produces a water-soluble formazan dye upon the reduction of radical superoxide. The reduction rate of the superoxide radical is linearly related to the xanthine oxidase activity of the sample and is inhibited by SOD. GPx activity was determined indirectly by the oxidation of GSH to oxidized glutathione (GSSG) catalyzed by GPx (Paglia and Valentine 1967). GST activity was measured by the conjugation of reduced glutathione (GSH) and 1-chloro-2,4-dinitrobenzene, leading to the formation of glutathione-dinitrobenzene conjugate at 340 nm (Habig et al. 1974). Total GSH was measured using the Sigma-Aldrich Glutathione assay kit (CS0260). This kit uses an enzymatic recycling method with glutathione reductase (Baker et al. 1990); the GSH sulfhydryl group reacts with Ellman’s reagent and produces a yellow-colored compound read at 405 nm. Lipid hydroperoxides (LH) were measured using a modified ferrous oxidation-xylenol orange (FOX) method using the PeroxiDetect Kit (PD1, Sigma-Aldrich) following the manufacturer’s instructions (Rodríguez-Fuentes et al. 2017). The procedure is based on iron oxidation by peroxide that forms a coloring component with xylenol orange at acidic pH measured at 560 nm.

Data Analysis

Given the high number of non-detected POPs concentrations in egg and plasma samples, data were analyzed using statistics for censored data (Helsel 2012). Data were arranged in a censored matrix to compute summary statistics and totals. For each pollutant, an extra column indicates whether the included value was censored (below the method detection limit). A robust regression on order (RROS) method using the cenros function in NADA for R (Helsel 2012) was used to estimate all sites and per site mean, median, and standard deviation (SD). ANOVA was used to evaluate differences in morphometric variables between sites.

For multivariate analysis of pollutants, permutational MANOVA (PERMANOVA) and principal coordinate analysis (PCO) were performed, including as variables pollutants that presented less than 80% of censored data (below the method detection limit, Helsel 2012) using Primer v 7.0 + PERMANOVA add on. Data was set as an ordinal rank matrix, and then resemblance was calculated using each set’s Euclidean distance of samples (Legendre and Legendre 1998). Finally, permutational multiple paired t tests were applied to compare the centroids of the different sites in each data set; 9999 unrestricted permutations were used to generate the empirical F and t distributions (Anderson 2001).

For multivariate analysis of biomarkers, PERMANOVA and PCO were used. Data were transformed using the function log10(x+1) and normalized. The resemblance was calculated using the Euclidean distance of samples (Legendre and Legendre 1998). As described above, permutational multiple paired t tests were used to compare the centroids of the different sites in each data set.

Censored regressions were done using the cenreg command in NADA for R between a censored Y value (pollutant concentration) and an uncensored X variable (biomarker). The site was included in the regression as a factor. For correlation between data that had uncensored values, Spearman rank-order correlation was used using Statistica v 7.0.

Results

The mean values of CCL and CCW of nesting female C. mydas from the southern Gulf of Mexico are shown in Table 1. The size of females did not show statistically significant differences between nesting sites.

Table 1 Mean, standard deviation, and range of curved carapace length (CCL) and curved carapace width (CCW) of female Chelonia mydas nesting in the southern Gulf of Mexico
Full size table

Pesticides detected in the plasma of female C. mydas nesting in three sites along the Campeche coast in the southern Gulf of Mexico are presented in Supplementary Table 3; 11 of 21 analyzed pesticides in plasma showed less than 80% of censored data. Endosulfan sulfate and cis-Chlordane were not detected in any plasma sample. PERMANOVA indicated significant differences between sites (Supplementary Table 4), demonstrating differences in the composition of the prevalent pesticides. OCPs content in turtles from Carmen was significantly different from turtles from Isla Aguada and Sabancuy (Supplementary Table 5). PCO analysis explained 50.5% of the total variation in the first two axes; PCO1 was strongly associated with heptachlor concentrations (Fig. 2).

Fig. 2
figure 2

Principal coordinate analysis (PCO) of organochlorine pesticides present in the plasma of green turtle (Chelonia mydas) nesting in Carmen (red triangle), Isla Aguada (blue circle), and Sabancuy (green square), Campeche, Mexico

Full size image

The analysis of eggs indicated that 12 of 21 analyzed pesticides showed less than 80% of censored values (Supplementary Table 6). Endosulfan sulfate and cis-Chlordane were not detected in any egg. PERMANOVA indicated significant differences between sites (Supplementary Table 4). Figure 3 shows the result of the PCO that explained 65% of the total variation in the first two axes. Endrin, α+β+γ-HCH, and Aldrin were the prevalent pesticides in Isla Aguada and Carmen turtles. The rest of the pesticides are more associated with turtles sampled at Sabancuy. Permutational multiple paired t tests indicated statistical differences between the three sites (Supplementary Table 5).

Fig. 3
figure 3

Principal coordinate analysis (PCO) of organochlorine pesticides present in the egg of green turtle (Chelonia mydas) nesting in Carmen (red triangle), Isla Aguada (blue circle), and Sabancuy (green square), Campeche, Mexico

Full size image

Spearman correlation in OCs present in plasma and eggs with less than 80% censored data; α+β+γ-HCH, trans-chlordane, 4,4′-DDE, 4,4′-DDT, and methoxychlor were the pesticides with these characteristics. No significant correlations were found with the concentration found in plasma and eggs. Some OCPs (δ-HCH, heptachlor, and endrin ketone) were detected with less than 80% censored values in plasma but not in eggs. In contrast, 4,4′-DDD, aldrin, dieldrin, endrin, and endrin aldehyde were detected with less than 80% of censored values in eggs but not in plasma (Supplementary Tables 3 and 6).

Concerning PAHs concentrations in plasma, all 15 tested compounds were present with less than 80% of censored values (Supplementary Table 7). PERMANOVA indicated significant differences between sites demonstrating differences in the composition of the predominant PAHs in plasma (Supplementary Table 4). In addition, permutational multiple paired t tests indicated significant differences between the three sampling sites (Supplementary Table 5). PCO analysis was able to explain in the first two axes 38.6% of the total variation (Fig. 4).

Fig. 4
figure 4

Principal coordinate analysis (PCO) of PAHs present in the plasma of green turtle (Chelonia mydas) nesting in Carmen (red triangle), Isla Aguada (blue circle), and Sabancuy (green square), Campeche, Mexico

Full size image

PAH concentrations in eggs showed that 10 of 15 tested compounds were present with less than 80% of censored values (Supplementary Table 8). PERMANOVA indicated significant differences between sites (Supplementary Table 4). Permutational multiple paired t tests indicated significant differences between the three sampling sites (Supplementary Table 6); PCO explained 48.9% of the total variation in the first two axes (Fig. 5).

Fig. 5
figure 5

Principal coordinate analysis (PCO) of PAHs present in the egg of green turtle (Chelonia mydas) nesting in Carmen (red triangle), Isla Aguada (blue circle), and Sabancuy (green square), Campeche, Mexico

Full size image

Naphthalene, acenaphthylene, 2-Bromo naphthalene, acenaphthene, fluorene, phenanthrene, anthracene, fluoranthene, pyrene, and dibenz[a,h]anthracene were present in both eggs and plasma with less than 80% censored data. Spearman correlations between the concentration of these PAHs in plasma and eggs indicated that they were not statistically significant. Benz[a]anthracene + chrysene, benzo[b]fluoranthene, benzo(a)pyrene, indeno[1,2,3-cd]pyrene, and benzo[ghi]perylene were present in plasma with less than 80% of censored values, but not in the eggs (Supplementary Tables 7 and 8).

The results of the measured biomarkers are presented in Table 2. PERMANOVA indicated significant differences between nesting sites (Supplementary Table 3), with Carmen as the site that was different from the other sites (Supplementary Table 4). PCO analysis explained 69% of the total variation in the first two axes. PCO1 was associated with GPx activities and total GSH concentration; PCO2 was associated with SOD and GST activities (Fig. 6).

Table 2 Maximum likelihood correlation coefficients between organochlorine pesticides and PAHs present in plasma with biomarkers and morphometric data of green turtle (Chelonia mydas) nesting in Campeche, Mexico. Only significant correlations are presented, indicating p value < 0.05 (*) or < 0.001 (**)
Full size table
Fig. 6
figure 6

Principal coordinate analysis (PCO) of biomarkers present in the plasma of green turtle (Chelonia mydas) nesting in Carmen (red triangle), Isla Aguada (blue circle), and Sabancuy (green square), Campeche, Mexico

Full size image

Results of the significant likelihood-r correlation between biomarkers, pollutants, and morphometric variables are presented in Tables 3 and 4 for plasma and egg, respectively. The correlations between morphometric variables and biomarkers were not significant.

Table 3 Maximum likelihood correlation coefficients between organochlorine pesticides and PAHs present in egg with biomarkers and morphometric data of green turtle (Chelonia mydas) nesting in Campeche, Mexico. Only significant correlations are presented, indicating p value < 0.05 (*) or < 0.001 (**)
Full size table
Table 4 Maximum likelihood correlation coefficients between persistent organic pollutants (pesticides and polycyclic aromatic hydrocarbons) present in plasma with biomarkers and morphometric data of green turtle (Cheloniamydas) nesting in Campeche, Mexico. Only significant correlations are presented indicating p-value < 0.05 (*) or <0.001 (**)
Full size table

Discussion

Few studies report the concentrations of OCPs in the plasma of C. mydas worldwide. In the present study, endrin ketone and α+β+γ-HCH were the OCPs detected at the highest level in plasma. García-Besné et al. (2015) analyzed OCPs in the blood of 32 nesting C. mydas sampled in Sabancuy and Punta Xen, Campeche; mirex, endrin ketone, γ-HCH, 4,4′-DDT, and 2,4-DDE were detected in more samples (34 to 53% of the turtles) and mean levels of these compounds ranged from 0.9 to 8.6 ng/g lipid fraction. Reports of OCPs in blood or plasma worldwide indicate the presence of 4,4′-DDE, 4,4′-DDT, dieldrin, α-HCH, β-HCH, γ-HCH, heptachlor, and trans-chlordane at the highest concentrations (Swarthout et al. 2010; van de Merwe et al. 2010; Komoroske et al. 2011; Labrada-Martagón et al. 2011; Camacho et al. 2014). Some OCPs are restricted or banned in many countries, Mexico included, because of their harmful environmental effects.

Results in eggs indicated that endrin and endrin aldehyde were the OCPs detected at the highest concentrations. García-Besné et al. (2015) reported 1250 ng/g lip of endrin ketone in one sample of C. mydas eggs collected in Campeche. On the other hand, Salvarani et al. (2019) mention that the sum of Drins (∑Drins) and iciathe sum of HCHs (∑HCHs) were the OCPs with the highest mean levels in C. mydas eggs collected in 2014 and 2015 in Isla Aguada and Punta Xen, Campeche (0.51 and 2.64 ng/g dw of ∑Drins and 0.49 and 4.93 ng/g dw of ∑HCHs, respectively).

To date, few studies have been performed to assess PAHs concentrations in the blood of sea turtles worldwide; these studies have been conducted on C. mydas, loggerhead, Caretta caretta, hawksbill turtles, Eretmochelys imbricata (Camacho et al. 2012, 2013, 2014; Bucchia et al. 2015; Casini et al. 2018; Vijayasarathy et al. 2019). In this study, indeno[1,2,3-cd]pyrene, benzo[a]pyrene, and benzo[b]fluoranthene were the PAHs with the highest concentration in the plasma of C. mydas. Indene[1,2,3-cd]pyrene is primarily found in gasoline and diesel exhaust, coal tar, and petroleum asphalt. Benzo[a]pyrene and benzo[b]fluoranthene are mainly found in gas and diesel exhaust, coal tar, amino acids, fatty acids, carbohydrate pyrolysis products, creosote oil, petroleum asphalt, and shale oils.

Reports of PAHs worldwide present different distributions of the individual compounds found at the highest concentration; phenanthrene and low molecular weight PAHs were detected at the highest levels in plasma and total blood of sea turtles C. caretta sampled in the Canary Islands and Cape Verde (Camacho et al. 2012) and Mediterranean Sea (Casini et al. 2018), respectively.

POPs found in the present study have been reported in water, sediment, and biota collected on the coast of Campeche and the Gulf of Mexico (Vázquez-Botello et al. 1993; Noreña-Barroso et al. 1999; Díaz-González et al. 2005; Ponce Vélez and Botello 2005; Rendón von Osten et al. 2005; Carvalho et al. 2009; Quetz et al. 2009).

Although these nesting sites are located within the protected Laguna de Terminos Reserve, they are influenced by several sources of pollution. Laguna de Terminos is the second-largest estuarine system in the Gulf of Mexico, receiving discharges from four main rivers (Usumacinta, Palizada, Chumpan, and Candelaria). Agricultural activities are also carried out in the surrounding areas, releasing pesticides into the estuarine system. Moreover, oil extraction facilities are located 100 km from the nesting sites, resulting in intense oil-related activities between offshore platforms and the study area. OCPs enter via runoff from agricultural areas where these pesticides were used; sources of PAHs may include oil spills and oil-related activities (transportation, storage, and use of crude oil) and combustion of fossil fuels and biomass. Although comparisons between reported POPs concentrations are complicated by the numerous factors that may differ between studies, the composition and concentration of POPs in C. mydas in this study may reflect differences in the contaminant profiles of their foraging grounds since pollutants are ingested primarily through food (Finlayson et al. 2016) and green turtles exhibit high levels of fidelity to foraging areas (Broderick et al. 2007). Differences in the present study concerning the nesting site may be related to the specific influence of the sources of pollutants, for example, proximity to the rivers, oil-related activities, or the urban nucleus. The currents and other oceanographic variables may also affect the transport and distribution of pollutants.

Some studies on sea turtles prove that POPs are transferred from maternal blood to eggs (Guirlet et al. 2010; van de Merwe et al. 2010; Stewart et al. 2011). In this study, few OCPs and PAHs detected in the plasma of nesting females were found in their corresponding egg, and no statistically significant relationship was found. The lack of a significant correlation between POP concentrations in maternal plasma and those in egg samples in this study could be explained because POPs in female plasma reflect recent exposure to contaminated sources. In contrast, the concentrations of POPs in their corresponding egg comprise POPs that have already been mobilized and transferred during vitellogenesis several months earlier. There have been other studies on turtles where no significant correlations were found between POP concentrations in female plasma and eggs (Rauschenberger et al. 2004; García-Besné et al. 2015), like the findings in the present work. Low molecular weight PAHs in eggs were more abundant and frequent than high molecular weight PAHs. Very few compounds of 5 or 6 rings, except for Dibenz[a,h]anthracene, were detected in the eggs. The abundance and frequency of low molecular weight PAHs in eggs may suggest preferential transfer of less lipophilic POPs from the maternal blood to eggs. Selective removal of less lipophilic POPs from the organism to eggs has been previously reported in sea turtles (Mckenzie et al. 1999; Guirlet et al. 2010; Stewart et al. 2011). Differences in the speed of POPs diffusion from the bloodstream to the different tissues may explain the lower content of highly lipophilic compounds (higher Kow values) in eggs compared to plasma levels (Muñoz and Vermeiren 2020).

Several studies involving sea turtles have assessed growth-related variations in POPs concentrations using carapace lengths (Mckenzie et al. 1999; Komoroske et al. 2011; Camacho et al. 2014) since no reliable age determination methods exist for sea turtles (Bjorndal et al. 1998). The present study used censored regression using the maximum likelihood correlation coefficient to evaluate the correlation between the concentration of POPs found in plasma and eggs and the morphometric data. Like the findings in the present work (where approximately 30% of the pollutants in plasma and 60% of contaminants in egg presented a negative correlation with female morphometry), some authors have reported negative relationships between body size and POPs concentrations in green turtles from other regions (Mckenzie et al. 1999; Richardson et al. 2010; Malarvannan et al. 2011).

A shift in feeding ecology might explain the negative correlations. For example, green turtles change from omnivorous as juveniles to herbivorous diets during adulthood (Bjorndal 1997). This change reduces the ingestion of POPs as the individual grows (Mckenzie et al. 1999; Keller 2013).

Despite the conservation status of C. mydas, there needs to be more information regarding the oxidative stress response against exposure to pollutants in this species. Few studies have been conducted on the antioxidant response of C. mydas performed in Baja, California, Mexico (Valdivia et al. 2007; Richardson et al. 2010; Labrada-Martagón et al. 2011). One more study reported oxidative stress indicators conducted on the hawksbill turtle, E. imbricata, nesting in Punta Xen, Campeche (Tremblay et al. 2017). In the present study, there was no relationship between morphometric variables and antioxidant response or oxidative damage in the plasma of nesting C. mydas. Similar results had been reported for C. mydas (Labrada-Martagón et al. 2011) and E. imbricata (Tremblay et al. 2017) and may be related to the similar size of the sampled organisms. Exposure to OCPs and PAHs may induce the production of ROS, leading to oxidative stress. The oxidative stress and excess ROS are counteracted by antioxidative defense mechanisms, which include the enzymes GPx, GST, and SOD, and the non-enzymatic antioxidants, i.e., GSH, vitamin C, and vitamin E, among others. GPx, SOD, GST, and lipid peroxidation levels have been proposed as valuable biomarkers for oxidative stress in aquatic organisms (Valavanidis et al. 2006). Multivariate analysis of POPs demonstrated differences in plasma and egg concentration in the different nesting sites that were studied in Campeche, Mexico; the regional separation of the sea turtles by their biomarkers had been previously related to differences in the habitat and environmental conditions (Labrada-Martagón et al. 2011).

In this study, GSH levels detected in the plasma of nesting C. mydas were negatively correlated with 29% of the pollutants contaminants in the plasma and 60% in the egg. GSH concentrations were strongly correlated with trans-Chlordane, endrin ketone, and Indeno 1,2,3-cd pyrene in plasma and to DDT and its metabolites, methoxychlor, anthracene, and fluoranthene in eggs. These results suggest a decrease in the total GSH level in the plasma of C. mydas exposed to pollutants. The reduction in GSH concentration may be attributed to the utilization of this compound as an antioxidant and an essential conjugant in the phase II metabolism of these pollutants. The decrease in GSH levels due to exposure to contaminants has also been reported in other aquatic species (Sanchez et al. 2007; Karadag et al. 2014). GPx activity in the plasma of nesting C. mydas was negatively correlated to 33% of the pollutants present in plasma and 70% of the contaminants present in the egg; it was strongly correlated to concentrations of 4,4′ DDE and endrin ketone in plasma, and DDT and its metabolites, methoxychlor, anthracene, and fluoranthene in eggs. This result would indicate that the presence of these pollutants reduces GPx activity detected in plasma C. mydas. On the contrary, in C. mydas from Baja California, GPx activity was positively correlated with trans-chlordane (Labrada-Martagón et al. 2011). Still, this activity was monitored in erythrocytes, not plasma, an extracellular source of antioxidant enzymes.

SOD activity in the plasma of nesting C. mydas was negatively correlated to the concentration of 29% of the pollutants in the plasma and 65% of the contaminants in the egg. It was strongly correlated with 4-4′ DDE, endrin ketone in plasma, DDT metabolites, methoxychlor, anthracene, and fluoranthene in plasma. In accordance, Kovacik et al. (2019) negatively correlated the presence of contaminants with SOD and lipid peroxides in fish. This result would suggest reducing SOD activity in C. mydas exposed to these pollutants. In other studies, no relationship was detected between SOD activity and the concentration of contaminants in C. mydas from Baja California (Labrada-Martagón et al. 2011) and in E. imbricata from Campeche (Tremblay et al. 2017). However, other studies on fish species have shown an increase in SOD activity in organisms exposed to pollutants (Monteiro et al. 2010; Carvalho et al. 2012).

GST activities were negatively correlated to the concentration of 45% of the pollutants in the plasma and 60% in the egg. It was strongly related to 4-4′DDE, endrin ketone in plasma, and DDT and its metabolites, aldrin, methoxychlor, anthracene, and fluoranthene in eggs. In previous reports in C. mydas from Baja California, GST activity was positively correlated to the concentrations of α-, β-, γ-HCH, ƩHCHs, heptachlor, and ƩHeptachlors (Labrada-Martagón et al. 2011).

In the plasma of nesting C. mydas, a negative correlation was detected between lipid peroxidation and 25% of the pollutants contaminants in the plasma and 60% in the egg. Exposure to pollutants, including OCPs and PAHs, may cause an increase in lipid peroxidation in fish species, as has been previously reported elsewhere (Avci et al. 2005; Otitoloju and Olagoke 2011). The results of biomarkers of oxidative stress are difficult to interpret and compare since there are tissue, concentration, and time dependency. It is not rare that for the same biomarker, some reports present an increase, others a no change, and even others have a reduction in enzyme activities or the concentration of antioxidants or oxidative damage indicators. As previously reported (van der Oost et al. 2003 and references therein), oxidative stress biomarkers are unspecific; they may or may not be related to the presence of pollutants since not all the contaminants (or their metabolites) produce ROS or oxidative stress. In addition, studies are showing an abnormal increase in ROS content under the activation of POPs in wildlife species (Tian et al. 2021, González-Mille et al. 2019). Oxidative stress biomarkers are also related to respiration, and any stressor may increase oxidative damage (for example, changes in oxygen concentration or temperature). It has been previously reported that other hypoxia-tolerant reptiles must deal with a significant amount of ROS generated due to ischemia-reperfusion associated with diving (Valdivia et al. 2007). Therefore, the measurement of oxidative stress is a valuable tool for the health assessment of endangered species, but these biomarkers should be interpreted carefully. The use of multivariate methods to assess the antioxidant status of organisms exposed to pollutants may be a valuable tool to understand what is happening in the sampled organism. For example, in the present study, PCO shows that the PCO1 axis is related to GPx and GSH and the presence of LH; PCO2 is related to Sand but also LH. It has been reported that SOD and CAT (the latter was not measured in the present study) are the first lines of defense against ROS, inhibiting superoxide produced during the metabolism of some pollutants or cellular respiration (Dobal et al. 2022). On the other hand, GSH plays a double role as an antioxidant (working as an electron donor for GPx) and as a conjugant of phase II metabolism reactions catalyzed by GST. GST also has an essential role in the antioxidant system, reducing lipid hydroperoxides to alcohol (Regoli et al. 2014). LH acts as an indicator of oxidative stress. In the present study, turtles nesting in Sabancuy were associated with the concentration of LH; it is interesting to add that turtles nesting in this site also presented higher concentrations of some OCPs and PAHs that may be related to this finding. On the other hand, higher levels of SOD and GPx were present in turtles nesting in Carmen. This increase in activity may be a defense against oxidative damage with the consequent lower concentration of LH.

Conclusions

In the present study, concentrations of OCPs, PAHs, and chlorpyrifos were analyzed in the plasma and eggs of nesting female Chelonia mydas from the southern Gulf of Mexico, providing baseline data on POP concentrations in this species. The use of censored statistics allowed us to do a realistic calculation of summary statistics, multivariate procedures, and correlations incorporating non-detects in the analysis. Few compounds were detected in plasma and their corresponding eggs, and correlations were non-significant. Concentrations of POPs in the plasma of C. mydas showed a negative correlation with female morphometry. Spatial differences in concentrations of POPs among nesting sites may reflect differences in foraging areas. Detection of POPs in C. mydas indicated that plasma and eggs are suitable matrices to assess POP concentrations and multivariate analysis for biomarkers to determine the relationship between POPs and the antioxidant system and oxidative damage biomarkers.