Enantio-alteration of gene transcription associated with bioconcentration in adult zebrafish (Danio rerio) exposed to chiral PCB149

Enantioselective enrichment of chiral PCB149 (2,2’,3,4’,5’,6-hexachlorobiphenyl) was analysed in adult zebrafish (Danio rerio) exposed to the racemate, (−)-PCB149, and (+)-PCB149. Greater enrichment of (−)-PCB149 compared to (+) PCB149 was observed following 0.5 ng/L exposure; however, as the exposure time and concentration increased, racemic enrichment was observed in adult fish exposed to the racemate. No biotransformation between the two isomers was observed in fish exposed to single enantiomers. When zebrafish were exposed to different forms of chiral PCB149, enantioselective expression of genes associated with polychlorinated biphenyls (PCBs) was observed in brain and liver tissues and enantioselective correlations between bioconcentration and target gene expression levels were observed in brain and liver tissues. The strong positive correlations between expression levels of target genes (alox5a and alox12) and PCB149 bioconcentration suggest that prolonged exposure to the racemate of chiral PCB149 may result in inflammation-associated diseases. Prolonged exposure to (−)-PCB149 may also affect metabolic pathways such as dehydrogenation and methylation in the brain tissues of adult zebrafish. Hepatic expression levels of genes related to the antioxidant system were significantly negatively correlated with bioconcentration following exposure to (+)-PCB149.

Non-racemic enrichment. Significant differences in enantiomer fraction (EF) values (EFs > 0.5) were observed at a dose concentration of 0.5 ng/L (Fig. 1), indicating preferential enrichment of (− )-PCB149 and consequently, enantioselective enrichment. By contrast, enantioselectivity was not observed at exposure of 0.1 μ g/L and 0.25 μ g/L. The concentrations of (− )/(+ )-PCB149 were analysed and the data are presented in Table 2.
Hepatic gene expression. The transcription of genes related to the antioxidant system (sod, cat, and Gpx; Fig. 2A) increased significantly with increasing exposure concentrations at 7 days post exposure (dpe) for the PCB racemate. Upon prolonged exposure to the racemate, the expression levels of the three target antioxidant genes in liver tissue decreased at the three exposure concentrations, respectively. By contrast, the expression levels of this genes were not altered by exposure to either (− )-PCB149 or (+ )-PCB149. The tendency towards upregulation of Gpx expression levels declined with prolonged exposure time to the racemate at 0.25 μ g/L (Fig. 2a), but the  tendency towards downregulation of Gpx expression levels was enhanced by prolonged exposure to (− )-PCB149 and (+ )-PCB149 at 0.25 μ g/L (Fig. 2a). Enantioselective hepatic expression was observed for exposure to 0.5 ng/L of each form of PCB149 at 7, 14, and 28 dpe. At higher exposure concentrations (0.1 μ g/L and 0.25 μ g/L), clear changes in the expression of genes related to the lipid-peroxidative pathway (apoa1a, alox12, and alox5a) were observed in liver tissue (Fig. 2B). For 0.25 μ g/L at 14 dpe, hepatic transcription of apoa1a, alox12, and alox5a tended to be normal (similar to β -actin expression). No enantioselective expression of the three genes was observed in fish exposed to 0.1 μ g/L (− )-PCB149 or (+ )-PCB149.
Hepatic transcription of genes related to the dehydrogenase system (sdhaf2, gapdh, and ldha) (Fig. 2C) was clearly enantioselective for exposure to 0.5 ng/L at 7, 14, and 28 dpe: the fold inhibition of sdhaf2 and gapdh expression increased to 20-fold at 0.5 ng/L and to 25-fold at 0.25 μ g/L at 14 dpe for the racemate. Downregulation of sdhaf2 and ldha was observed when (− )-PCB149 and (+ )-PCB149 were present separately at 0.25 μ g/L at 7 dpe; however, upregulation was observed when both isomers were present. (A) adult zebrafish exposed to 0.5 ng/L; (B) adult zebrafish exposed to 0.1 μ g/L; (C) adult zebrafish exposed to 0.25 μ g/L. Asterisks denote significant differences between treatments and control (determined by Dunnett post hoc comparison, P < 0.01, **). Error bars indicate standard deviations.  Table 2. Concentrations of (−)-PCB149/(+)-PCB149 in adult zebrafish when exposed to enantiomers. Exposure to increasing concentrations of the PCB racemate resulted in increased expression levels of hemk1 and comta, genes relating to methylation in liver tissues (Fig. 2D) at 7 dpe for the racemate. By contrast, for (− )-PCB149, decreases in hemk1 and comta expression were observed at increasing exposure concentrations at 7 dpe. Enantioselectivity was also observed at 7 dpe. Hepatic transcription (Fig. 2E) of genes related to hydroxylation (cyp2k6, cyp2aa4, and cyp19a1b) in an enantioselective manner was observed under certain exposure conditions; for example, the hepatic expression of cyp19a1b increased by exposure to 0.1 μ g/L of all three forms at 7 dpe.
Cerebral gene expression. Small fluctuations in the levels of the three antioxidant genes were observed in brain tissues (Fig. 3A), with the exception of Gpx at 14 dpe (Fig. 3a). The expression levels of Gpx expression increased with exposure to increasing concentrations of the racemate and cerebral transcription of Gpx increased, while for (+ )-PCB149, cerebral transportation of Gpx decreased with increasing exposure concentration.
Significant differences in the expression levels of three lipid-peroxidative pathway-related genes were observed in (− )-PCB149 and (+ )-PCB149-treated fish. Downregulation of apoa1a, upregulation of alox12, and near-normal expression of alox5a were observed in the brain tissue of fish exposed to 0.25 μ g/L of all three forms at 14 dpe (Fig. 3B).
No fluctuations in ldha expression were observed in brain tissue ( Fig. 3C) during the exposure period; however, with increased exposure time, the initial downregulation of gapdh expression was followed by upregulation for each form of PCB149 at all three concentrations of each form.
Small fluctuations in the cerebral transcription levels of hemk1 and comta ( Fig. 3D) were observed and cerebral expression of cyp2k6 was inhibited at 7 dpe for the racemate (Fig. 3E). Exposure to increasing concentrations of (− )-PCB149 resulted in a decrease in the cerebral expression of cyp2k6, whereas exposure to (+ )-PCB149 resulted in enhanced expression of cyp2k6. Significant downregulation of cyp2aa4 and up-regulation of cyp19a1b were observed at 7 dpe. The expression of cyp19a1b in brain tissue differed depending on the exposure time and concentration. At increasing concentrations of all three forms, for example, cyp19a1b expression was either enhanced (7 dpe) or inhibited (14 dpe and 28 dpe). Moreover, at the same concentration, cerebral transcription of cyp19a1b decreased with prolonged exposure time.

Chemometrics.
To assess enantioselective enrichment, EF values were calculated using bioconcentration data. To better understand the relationship between concentration levels and fluctuations in the expression of target genes with various functions in liver and brain tissues, we performed correlation analyses (Table 3) and principal component analysis (PCA) on the loading Bi plots (Fig. 4).

Discussion
Racemic PCB149 was previously detected in groupers (Epinephelus marginatus) collected between 1994 and 1995 from the northwest African Atlantic Ocean 25 as well as in Arctic cod (Boreogadus saida) 26 . The results of the exposure experiments at higher concentration of PCB149 described here are consistent with these reports. An enantioselective capacity for chiral PCB bioprocessing has also been observed in Lake Superior trout 27 . Fish species are considered to have low PCB biotransformation capacities 28 and in the present study, no biotransformation between the two isomers was observed.
Interestingly, we previously demostrated that embryo-larvae have a higher accumulation capacity for PCB149 than adult zebrafish: at the same exposure concentration (0.1 μ g/L), the concentration of (+ )-PCB149 in embryo-larvae reached 3.69 mg/L at 11 dpe, whereas the concentration of (+ )-PCB149 reached only 0.57 mg/L after 28 dpe in adult zebrafish. This difference may be due to the ability of PCB149 to permeate the embryonic membrane 29 ; the chorion is composed of lipid and thus may facilitate the passage of lipophilic PCB149 into the embryo. During the endogenous feeding period, nutrition is provided by the yolk sac before the yolk sac completely disappears. Endogenous feeding and a fully developed metabolic pathway 30 thus may also enhance PCB149 bioaccumulation in larvae. However, PCB149 mainly enters adult zebrafish tissues through gill respiration and the skin, which may result in slower bioaccumulation.
PCAs analysis on loading Bi plots (Fig. 4) was performed to investigate the relation between target genes with different functions and bioconcentration. As shown in Fig. 4A-C, expression of the target genes in brain tissues exhibited a relatively scattered trend, whereas the expression levels of the target genes in liver tissues were relatively concentrated in Fig. 4D-F. The PCAs plots also suggested that the changes in the expression levels of the target genes with different functions were more closely correlated with bioconcentration in liver tissues than in brain issues. This finding was confirmed by the correlation analyses (Table 3).
Enantioselective correlations between bioconcentration and the target genes were observed in the brain and liver tissues of adult zebrafish exposed to different forms of chiral PCB149. The correlation coefficient and PCA results indicated that the cerebral expression levels of alox5a, alox12, and cyp2aa4 were significantly positively correlated with bioconcentration following zebrafish exposure to the racemate. In zebrafish exposed to (− )-PCB149, the cerebral expression of cyp2aa4, comta, and hemk1 was significantly negatively correlated with bioconcentration, and sdhaf2, ldha, and cyp19a1b expression was highly negatively correlated with   bioconcentration. The cerebral expression of sod and cyp2k6 was significantly negatively correlated with bioconcentration following exposure to (+ )-PCB149. The bioactive compounds generated by genes related to the lipoxygenase pathway may result in a variety of biological activities important for carcinogenesis 31 . Arachidonate lipoxgenase (alox) enzymes metabolize arachidonic acid and thus generate potent inflammatory mediators, thereby leading to inflammation-associated diseases 32 . Both alox5a and alox12 enzymes have been described as pro-carcinogenic: For example, the alox12 enzymes may increase the expression of genes encoding proinflammatory cytokines such as tumour necrosis factors 33 and aberrant expression of alox5 has been reported in patients with breast cancer 34 . We observed highly positive correlations between the expression of these genes (alox5a, alox12) and PCB149 bioconcentration, suggesting that prolonged exposure to the racemate of PCB149 may result in inflammation-associated diseases. Furthermore, prolonged exposure to (− )-PCB149 may affect metabolic pathways in the brain tissue of adult zebrafish. The catechol-O-methyltransferase (COMT) enzyme, which forms reactive adducts with DNA, is considered to be a potentially important contributor to cancer development. Garner and colleagues have determined that polymorphic variants of the COMT gene are associated with an increased risk of hormone-related cancer 35 . Succinate dehydrogenase (SDH), which is complex II of the respiratory chain, is situated at the centre of two essential energy-producing cell processes: the tricarboxylic acid cycle and mitochondrial oxidative phosphorylation. Decreased SDH levels may result in a limited capacity for oxidative phosphorylation and consequently ATP production 36 and sdhaf2 plays a role in hereditary disease 37 . The LDHA enzyme, a metabolic enzyme in the glycolysis pathway, plays a critical role in cancer progression 38 and dysregulation of LDHA expression has been observed in various cancer cell types including squamous cell carcinoma 39 and breast cancer 40 . The negative correlations between bioconcentration and the expression of these genes in brain tissue of zebrafish exposed to (− )-PCB149 observed in the present study suggest that prolonged exposure to (− )-PCB149 could cause related cancer.
No obvious correlation between PCB149 bioconcentrations and the changes in the hepatic expression levels of target genes were observed in zebrafish exposed to the racemate and (− )-PCB149; however, the hepatic expression levels of cat, Gpx, and cyp2aa4 exhibited significant negative correlations with bioconcentration for (+ )-PCB149 exposure. Oxidative stress occurs when there is an imbalance in the biological oxidant-to-antioxidant ratio, which resulted in oxidative damage to lipids, proteins, carbohydrates, and nucleic acids 41 . In fish, this oxidative stress in a defensive mechanism involving the activation of the antioxidant system to detoxify reactive oxygen species (ROS) 42 . The enzyme SOD converts ROS to H 2 O 2 and the enzymes CAT and Gpx then catalyse the transformation of H 2 O 2 to oxygen and water 43 . In this study, adult zebrafish exposed to (+ )-PCB149 exhibited high negative correlations between the cerebral expression levels of sod and PCB bioconcentration as well as between the hepatic expression levels of cat and Gpx genes and PCB concentration, suggesting that increases in (+ )-PCB149 bioconcentrations may result in change in the expression levels of cat and Gpx in liver tissue following changes in the expression levels of sod in brain tissue.
The P450 enzymes typically catalyse mono-oxygenase reactions that involve oxygen molecular and an equivalent number of electrons. These enzymes are encoded by different members of the CYP gene superfamily 44 . The CYP2 gene family in zebrafish is a novel subfamily with no orthology to human and mammalian CYP genes 45 . We observed different correlations between the expression levels of cyp2aa4 and PCB149 bioconcentration: a significant positive correlation was observed following exposure to the racemate, whereas a significant negative correlation with cerebral gene expression levels was observed following exposure to (− )-PCB149. A significant negative correlation was also observed for hepatic gene expression levels in liver tissue following (− )-PCB149 exposure. However, the regulation and function of CYP2s remains poorly understood. CYP19, known as aromatase, catalyses the final step in the conversion of androgens to oestrogens 44 . Cyp19a1b (generally named cyp19b) is primarily expressed in brain tissue, and brain cyp19a1b is physiologically significant for oestrogen-dependent neurogenesis 46 and the regulation of the gonadal-pituitary-hypothalamic axis 47 . We did not observe correlation between the hepatic expression levels of cyp19a1b and bioconcentration; onlyfor (− )-PCB149 exposure did expression levels of cyp19a1b in the brain exhibit a strong negative correlation with bioconcentration.

Method
Zebrafish husbandry. Juvenile AB strain zebrafish (Danio rerio) were obtained from Beijing Hongdagaofeng Aquarium Department and cultured in a fish facility (Esen Corp.) at 26 °C with a photoperiod of 14/10 (light/ dark) 31 . Six-month-old adult zebrafish were selected for this study and were fed dried brine shrimp (equivalent to 2% of the fish body weight) daily.
The racemate was separated and prepared on a Lux Cellulose-2 column (250 × 4.6 mm, 5 μ m, Phenomenex, Torrance, CA) using an Agilent 1200 series high performance liquid chromatography (HPLC) instrument (Wilmington, DE) with 100% n-hexane as the mobile phase at a flow rate of 1.0 mL/min. The enantiomers (− )-PCB149 and (+ )-PCB149 48 were repeatedly collected separately, concentrated to dryness using a nitrogen-evaporator (Hangzhou Allsheng Instruments company, China) and then dissolved in acetone (Fisher). The purities and concentrations of the isomers were determined by gas chromatography-mass spectrometry (GC-MS).
Scientific RepoRts | 6:19478 | DOI: 10.1038/srep19478 Exposure and sample collection. This study was performed in accordance with Chinese legislation and was approved by the independent animal ethics committee at China Agricultural University. Adult zebrafish were exposed to racemate, (− )-PCB149, and (+ )-PCB149 at 0.5 ng/L, 0.1 μ g/L, and 0.25 μ g/L plus solvent control. Three tanks were used for each exposure concentration of each target compound; one tank represented one repetition. A total of 40 six-month-old adult fish were assigned at random to one tank (25 L) containing 20 L test solution was assigned to randomly contain forty six-month-old adult fish. During the test, 50% of test solution was changed every day until the end of experiment at 28 days. At 7, 14, and 28 days post exposure (dpe), twelve adult fish from one tank were randomly selected and assigned to two groups: six for gene expression analysis and six for bioconcentration analysis. For analysis, the fish were anesthetized with MS-222 on ice. Liver and brain tissues were obtained by dissection, and stored overnight in RNA storage solvent at 4 °C before being removed from the RNA storage solvent and stored at − 80 °C until RNA extraction. During the exposure experiments described above, the exposure environment including temperature, humidity, and light cycle was identical to the culture environment.
Quantification of PCB149 levels in water and biological samples. For exposure analysis, the tank water was sampled for GC-MS analysis at five time points: 0, 1, 7, 14, and 28 dpe. Water samples (1 L) were subjected to five liquid-liquid extractions with n-hexane (100 mL each time) in a separatory funnel with violent shaking. The n-hexane layer was transferred to heart-shaped flasks and concentrated to near-dryness by rotary evaporation (Shanghai Ailang Instruments, Shanghai, China) at 35 °C. The concentrated solutions were then blown to dryness by nitrogen evaporation and the residue was again dissolved in 0.1 mL of isooctane for GC-MS analysis. Six adult zebrafish weighed for follow-up in a previous study 49 and were also processed for GC-MS analysis.  Gene expression studies. Total RNA was extracted from liver/brain tissue using an RNAprep Pure Tissue Kit (Tiangen Biotech, China). The quality of the isolated RNA was evaluated based on the quality of the 28 s and 18 s rRNAs by 2% agarose gel electrophoresis and the purity of the RNA preparations was assessed by OD 260 / OD 280 ratio. The concentration of the RNA was determined by OD 260 using a UV1240 spectrophotometer (Perkin Elmer, USA). First-strand complementary DNA (cDNA) was synthesized from 0.5 μ g of total RNA using a FastQuant RT Kit (Tiangen Biotech). Quantitative real-time polymerase chain reaction (qPCR) was performed using a SuperReal PreMix Plus Kit (Tiangen Biotech) and an ABI 7500 qPCR system (Applied Biosystems, USA). Primers were designed using the Primer 6.0 tool and are shown in Table 4. The housekeeping gene β -actin was used as an internal standard to eliminate variations in mRNA and cDNA quantity and quality. Three-step qPCR was used to quantify the relative levels of housekeeping and target genes: 95 °C for 15 min; 40 cycles of 95 °C for 10 s, 60 °C for 20 s, and 72 °C for 32 s. A melting curve analysis was also performed to confirm the specificity of PCR product as a single peak. Relative quantification of target genes normalized to β -actin levels was performed by the 2 −ΔΔCt method. Statistical analysis. EFs were evaluated in samples to determine enantiomeric compositions 50 ; the EFs was defined as the enantiomeric concentration ratio (− )/[(− )+ (+ )] for PCB149. EF values were in the range of 0-1, where 0.5 represents a racemate.
Statistical analyses were performed using SPSS16.0 software. Differences were determined by one-way ANOVA followed by post hoc Dunnett tests. All data are expressed as mean ± standard error of the mean and differences with P < 0.05 were considered statistically significant. Relationships between bioconcentrations and target genes transcription levels were analysed by PCA, followed by the calculation of correlation coefficients to further confirm these relationships.