Marine Pollution Bulletin 85 (2014) 352–362 Contents lists available at ScienceDirect Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul The occurrence and ecological risks of endocrine disrupting chemicals in sewage effluents from three different sewage treatment plants, and in natural seawater from a marine reserve of Hong Kong Elvis G.B. Xu a, Shan Liu b, Guang-Guo Ying b, Gene J.S. Zheng c, Joseph H.W. Lee d, Kenneth M.Y. Leung a,⇑ a The Swire Institute of Marine Science and School of Biological Sciences, The University of Hong Kong, Pokfulam, Hong Kong, China State Key Laboratory of Organic Geochemistry, CAS Centre for Pearl River Delta Environmental Pollution and Control Research, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Guangzhou, China c Department of Chemistry, Hong Kong Baptist University, Kowloon, Hong Kong, China d Department of Civil and Environmental Engineering, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China b a r t i c l e i n f o Article history: Available online 17 March 2014 Keywords: Endocrine disrupting chemicals Sewage treatment plant Marine reserve Marine protected areas Environmental risk assessment Ecotoxicology a b s t r a c t We determined the concentrations of 12 endocrine disrupting chemicals (EDCs) in sewage effluents collected from three different sewage treatment plants (STPs) in Hong Kong, and found 4-nonylphenol (NP) and bisphenol A (BPA) were the most abundant EDCs. Effluent concentrations of NP and BPA were higher in dry season than in wet season, but opposite seasonal changes of NP were observed in receiving waters, probably due to the surface runoff. The two secondary STPs showed higher removal efficiency for these compounds than the preliminary STP, while having higher removal efficiency in wet season. Therefore, it is necessary to upgrade the preliminary STP and improve the EDC removal efficiency in dry season. Seawaters from the Cape D’ Aguilar Marine Reserve adjacent to these STPs also exhibited elevated NP levels with a hazard quotient >1. Furthermore, diluted effluents from the STPs elicited significant transcriptional responses of EDC-related genes in the marine medaka fish. Ó 2014 Elsevier Ltd. All rights reserved. 1. Introduction Endocrine disrupting chemicals (EDCs) are of global concern and broadly defined as chemicals that interfere with the normal function of endocrine systems in wildlife and humans (Burkhardt-Holm, 2010). Numerous laboratory experiments indicate that EDCs can cause negative health effects (e.g. growth, behaviour, reproduction and immune function) in fishes through disrupting their endocrine systems (Mills and Chichester, 2005). Estrogenic EDCs can adversely affect male fishes through induction of vitellogenin and inhibition of the development of secondary sexual characteristics at very low exposure concentrations (Länge et al., 2001). Most EDCs are man-made organic chemicals being introduced to the marine environment through anthropogenic inputs such as contaminated sewage effluents and surface runoff. Typical representatives of synthetic EDCs, 4-nonylphenol (NP) and bisphenol A (BPA) are the major contributors to the endocrine-disrupting activities in aquatic environments (Auriol et al., 2006). NPs are the main degradation products of alkylphenols polyethoxylates which have ⇑ Corresponding author. Tel.: +852 22990607; fax: +852 25176082. E-mail address: kmyleung@hku.hk (K.M.Y. Leung). http://dx.doi.org/10.1016/j.marpolbul.2014.02.029 0025-326X/Ó 2014 Elsevier Ltd. All rights reserved. been widely used as surfactants in household, agriculture, and industrial processes (White et al., 1994). At an exposure concentration as low as 10 ppb, NP can cause an increase of vitellogenin mRNA and a decrease in the growth rate of testes in male rainbow trout (Lech et al., 1996). BPA is an industrial raw material mainly used in plastic, rubber, adhesive, and cable industries, and known to cause a decrease in sperm production in mice (Von Saal et al., 1998), and lead to a delay in hatching of eggs and a suppression of growth in juvenile rainbow trout (Aluru et al., 2010). It has been widely recognized that effluent discharges from sewage treatment plants (STPs) are the major source of the EDCs to aquatic environments (Zhang and Zhou, 2008). A growing body of research has indicated that sewage effluents and even their receiving waters can introduce estrogen-like effects in fishes (Gibson et al., 2005). There are limited documented studies examining the composition and concentrations of EDCs in sewage effluents and natural seawaters of Hong Kong (Li et al., 2007; Kueh and Lam, 2008). Li et al. (2007) discovered that concentrations of NP ranged from 29 to 2591 ng/L in surface water samples collected from Mai Po Marshes Nature Reserve, northwest of Hong Kong and its levels were higher in winter (dry season; November and December) than in late summer (moderately wet season; September and October). E.G.B. Xu et al. / Marine Pollution Bulletin 85 (2014) 352–362 Kueh and Lam (2008) surveyed the ambient occurrences of selected EDCs, such as nonylphenol and nonlyphenol ethoxylates (NPEO), in coastal waters, rivers, sediments and biota, and their results suggested that sewage effluents acted as primary sources for these chemical contaminants. However, little is known about (1) the composition of EDCs in sewage effluents, and (2) the removal efficacy of EDCs from raw sewage by different types of STPs in Hong Kong. Since sewage effluents often comprise of a complex mixture of EDCs, it is essential to examine the composition of EDCs in local sewage effluents and identify the dominated chemical contaminants. Furthermore, the seasonal variability of EDC concentrations in STPs and receiving waters in sub-tropical Hong Kong are still largely unknown. This knowledge is important to decide appropriate measures for minimizing ecological risks from EDC emissions to sensitive receivers such as marine reserves in the marine environment. EDCs can alter the expression of estrogenic-related (ER) genes, such as cyp19a and cyp19b, which may result in developmental and reproductive abnormalities in fishes (Kortner et al., 2009). EDCs can also cause disruptive endocrine effects through aryl hydrocarbon receptors (AHRs) and the peroxisome proliferatoractivated receptors (PPARs) in fishes. The AHR pathway mainly regulates the activation of several genes that encode phase I and phase II xenobiotic metabolism enzymes, while the PPAR pathway intermediates receptors and genes involved in the regulation of energy homeostasis, cell proliferation, differentiation and survival (Fang et al., 2012). In this study, we used embryos and larvae of the marine medaka fish (Oryzias melastigma) to assess their transcriptional response to sewage effluents and receiving waters, involving 13 genes in the ER, AHR and PPAR signalling pathways. There were three main objectives in this study. First, an attempt was made to quantify, for the first time, the concentrations of 33 common EDCs and identify the most abundant ones in sewage effluents collected from the three STPs which are located at south of Hong Kong Island and close to the Cape D’ Aguilar Marine Reserve (Fig. 1). The 33 EDCs include natural and synthetic estrogens, androgens, progestagens and glucocorticoids. Second, since both NP and BPA were identified as the most abundant EDCs in our study, we further monitored the concentrations of NP and BPA in influents and effluents, as well as in the receiving waters from the Cape D’ Aguilar Marine Reserve during both dry and wet seasons. Based on the measured concentrations of NP and BPA in the sewage effluents and receiving waters and their corresponding predicted no effect concentrations, ecological risks of these two compounds were assessed. Third, we investigated the effect of diluted sewage effluents and natural seawaters from the marine reserve on the mortality, hatching success, and gene expression in embryos of the marine medaka O. melastigma through a laboratory study. The results would provide supplementary information for evaluating the current ecological risk of NP and BPA to marine organisms in the marine reserve. 2. Materials and methods 2.1. Sampling The three Sewage Treatment Plants (i.e., Shek O, Stanley, and SWIMS STP) are located in south of Hong Kong Island, Hong Kong, within 3 km distance to the Cape D’ Aguilar Marine Reserve (Fig. 1). Shek O STP is a preliminary treatment facility (i.e., screening plant) designed for removal of suspended matters with a diameter larger than 6 mm and this plant only treats about 864 m3/day of raw sewage from Shek O Village District (Table 1; Fig. 2). Stanley STP is a secondary treatment plant and daily treats about 8100 m3 of raw sewage from Stanley and Tai Tam Districts. It has the largest treatment capacity among the three STPs. Its primary treatment 353 consists of a screen and a grit chambers, and its secondary treatment includes an aerobic bioreactor followed by a secondary sedimentation and disinfection (Table 1; Fig. 2). The sludge is returned from the secondary sedimentation chamber for biological treatment, while the surplus activated sludge is dried before being taken to landfill. The SWIMS STP, located within the Cape D’ Aguilar Marine Reserve, is designed to mainly treat the wastewater generated from the Swire Institute of Marine Science (SWIMS) within the marine reserve (Table 1; Fig. 2). As a trickling biofilter treatment plant, SWIMS STP consists of septic tank, biofilter tank, sand leach tank, disinfection tank and sedimentation chamber. The three STPs represent different levels of treatment efficiency: a screening treatment plant (Shek O STP), a secondary biological treatment plant (Stanley STP), and a trickling filter plant (SWIMS STP). We sampled both influents and effluents from the three STPs, and receiving seawaters within the marine reserve three times during wet season (April, May, and June 2012) and three times during dry season (December 2012, January and February 2013). Hence, three replicates were used for each STP in each season. For both influents and effluents, 15-min wastewater samples were composited at the Stanley STP to prepare 24-h flow-weighted samples, and the 8-h composite samples with 1-h interval were collected at the SWIMS STP. For Shek O STP, influent and effluent samples were grabbed in triplicates between 11:00 and 13:00, with a 0.5 h interval between each influent or effluent sample. Natural seawater samples were collected at 0.5 m below sea surface in the Cape D’ Aguilar Marine Reserve. In investigation of EDC composition in effluents and bioassay, we grabbed effluent samples in duplicates from each STP with a pre-cleaned stainless-steel bucket in April 2012. After collection, samples were transported on ice to laboratory and immediately filtered through GF/C glass fiber filter papers, and then stored at 4 °C. All samples were extracted for EDCs within a week. Blank water was taken to field as control to monitor any contamination during the transport. 2.2. Chemical analysis Thirty-three phenolic EDCs, including androgens, estrogens, xenoestrogens, glucocorticoids, and progestagens, were analysed with RRLC–MS/MS as described by Liu et al. (2011). For each of the effluent samples, 1 L of water sample was filtered through a glass fiber filter (Whatman GF/F, 0.7 lm, UK). 100 lL each of 1 mg/L of E1-d4, E2-d4, T-d3, S-d3, CRL-d2 and P-d9 were added as the internal standards. Solid Phase Extraction (SPE) cartridges (Oasis HLB, 6 mL and 500 mg each) were preconditioned with methanol and HPLC water. The filtered water samples passed through the SPE cartridges at 5–10 mL/min. The target compounds were eluted using ethyl acetate. The extracts were dried and re-dissolved in 1 mL of methanol for clean-up. The glass cartridge was filled with glass wool, silica gel and anhydrous sodium sulphate from bottom to top. Each extract was added to the cartridge, which was preconditioned with methanol, ethyl acetate/methanol (90:10, v/v), and hexane. After the cartridge was rinsed with hexane, the target compounds were eluted with ethyl acetate/methanol. The eluate was then dried and reconstituted. The target compounds were analysed by RRLC–MS/MS with EI. Liquid chromatography was performed on an Agilent 1200 series RRLC system (Agilent Technologies). The chromatographic separation was performed on an Agilent Zorbax SB-C18 (100 mm 3 mm, 1.8 lm) column with pre-column filter (2.1 mm, 0.2 lm). The column oven temperature was 40 °C and the injection volume was 10 lL. Mass spectrometry was performed using an Agilent 6460 Triple Quadrupole detector with ESI in both negative and positive modes (Agilent Corporation, USA). The quantitative analysis was performed in MRM mode. The analytical procedure for NP and BPA was based on Zhao et al. (2009). Briefly, 1 L of each of the influent, effluent or seawater 354 E.G.B. Xu et al. / Marine Pollution Bulletin 85 (2014) 352–362 Fig. 1. A map showing the outfall locations of the three sewage treatment plants and the Cape D’ Aguilar Marine Reserve. samples was filtered through a glass fiber filter (Whatman GF/F, 0.7 lm, UK). Methanol was used to elute non-filterable particles on the filter and combined with the filtered sample. For solid phase extraction, 100 lL of 1 mg/L of 4-n-NP were added to each sample as internal standards. SPE cartridge (Oasis HLB, 6 mL, 500 mg) was preconditioned with methanol/HPLC water (1:1, v/v). The filtered water samples passed through the SPE cartridges at 10 mL/min. The target compounds were eluted from the cartridges using methanol and DCM. The extracts were dried and then re-dissolved in methanol. For derivatization, 100 lL of an extract were removed to a tube and dried. 2 mL of 1 M NaHCO3 solution and 1 mL of 1 M NaOH solution were added to the tube. 2 mL of n-hexane, 50 lL of 10% pyridine in toluene and 50 lL of 2% PFBOCl in toluene were added in sequence. After separation, the supernatant of n-hexane phase was transferred and dried. The final extract was re-dissolved in n-hexane, which was ready for GC–MS analysis. The MS was operated in the selected-ion monitoring mode with electron-impact ionization (ionization voltage, 70 eV). The target compounds were separated by gas chromatography (Agilent 6890N) with a DB-5MS capillary column (length: 30 m; i.d.: 0.25 mm; coated film thickness: 0.25 lm). A mass spectrometry (Agilent 5973) was used as the detector. The oven temperature program was set as follows: the initial temperature was 70 °C for 1 min, then increased to 170 °C at 20 °C/min, to 230 °C at 6 °C/min, to 280 °C at 12 °C/min for 6 min, and held at 300 °C for 2 min. The injector was set at 280 °C. The GC–MS interface and the ion-source temperature were at 280 °C and 250 °C, respectively. Helium was used as the carrier gas at 1 mL/min. A 1 lL sample was injected in splitless mode with solvent delay of 3 min. The characteristic ions and retention times of the target compounds were obtained and identified with full scan mass spectra from m/z 50 to 500. The compounds of interest were identified in SIM mode. 355 E.G.B. Xu et al. / Marine Pollution Bulletin 85 (2014) 352–362 Table 1 Description of the sewage treatment system and treatment capacity of the three STPs considered in this study (information from personal contact with Drainage Service Department, HKSAR, in 2012). STP Population Stanley Shek O SWIMS Hydraulic retention time (h) BOD (ton/year) Suspended solid (ton/year) Sources of the catchment area Treatment process 11,600 (design) 8100 (Average daily flow at present) 21.9 (design) 13.6 at present 9 9 Main areas of Stanley and Tai Tam Districts Secondary treatment processes: 1,180 (design) 864 (Average daily flow at Present) Not applicable 3 0.5 Treatment capacity About 27,000 at present About 3050 at present About 30 Flow (m3/day) 1. 2. 3. 4. 100 80 Main areas of Shek O village District Screening of coarse material; Settlement of grit particles; Biological treatment of sewage; Disinfection Preliminary treatment processes: Screening suspended matter with diameter >6 mm and grit removal. 0.04 0.02 Area of Cape D’ Aguilar Marine Reserve Trickling filter treatment processes: 1. 2. 3. 4. Analyses of blank controls showed no contamination during the sampling, transportation and extraction processes. The recoveries of surrogate standard 4-n-NP were more than 65% for all the samples. Recoveries determined were 44.2–127.0% for wastewater samples and 71.7–113.0% for seawater samples. Ranges of the limit of detection (LOD) and limit of quantitation (LOQ) for 4-n-NP were 0.01–0.39 and 0.02–1.63 ng/L for wastewater samples and 0.01–0.24 and 0.03–0.8 ng/L for seawater samples, respectively. 2.3. Medaka embryo-larval bioassay and gene expression Fertilized eggs of the marine medaka O. melastigma were cultured and acclimatized in artificial seawater at a salinity of 30‰ and a temperature of 28 ± 1 °C with a 14 h-light/10 h-dark photoperiod for 2 days. The embryos of 2 day post fertilization (dpf) were exposed to 1% and 10% (v/v) of effluents from each STP and surface natural seawaters from the Cape D’ Aguilar Marine Reserve, and to artificial seawater as the control. The effluents of different dilutions (i.e., 1:100 and 1:10 dilution) were selected to represent environmentally relevant concentrations that animals living close to the outfall of effluent might experience. Each experimental group contained 50 embryos, which were randomly distributed into petri dishes containing 30 mL of exposure solution. The media were daily renewed. Three replicates were conducted for each experimental group. The mortality rates of the embryos from 2 to 10 dpf, the hatchability of eggs, and the size of juveniles after 21-d exposure were recorded. For each replicate, 10 embryos at 4 dpf, 10 embryos at 10 dpf, and two juveniles at the first fry stage were collected for quantitative real time polymerase chain reaction (qRT-PCR) analysis. The primers of 13 endocrine disruption related genes, including ERa, ERb, ERc, VTG1, VTG2, AHR, ARNT, cyp1a, cyp19a, cyp19b, PPARa, and PPARb and PPARc are presented in Table S1 in Appendix A. The procedures of qRT-PCR followed the methods described in Fang et al. (2012). The embryos or juveniles were randomly collected and homogenized in 1 mL RNA-Solv reagent (Omega) using a glass homogenizer. The total RNA was extracted from the homogenates using Omega kits according to the manufacturer’s instructions. Equal amounts of RNA were applied to qRT-PCR using SYBR Premix Ex TaqTM kit (TaKaRa) on a Bio Red CFX 96 Real-Time System. The PCR thermal profile was as follows: an initial denaturation step at 95 °C for 30 s, followed by 40 cycles at 95 °C for 5 s, and 60 °C for 34 s, and ending with a dissociation curve analysis. Septic tank; Bio-filter tank; Sand leach tank; Disinfection and sedimentation Gene expression levels were normalized to the 18 s rRNA expression levels. The fold change of the tested genes was analysed by the 244Ct method (Livak and Schmittgen, 2001). 2.4. Data analysis Two-way analysis of variance (ANOVA) followed by post hoc Tukey test was used to examine statistical differences of removal rates of NP or BPA from wastewater samples among the different STPs (3 levels) and between different seasons (2 levels). Student’s t test was used to examine the seasonal differences of concentrations of NP or BPA in natural seawater samples. For the expression results of each target gene, one-way ANOVA followed by post hoc Tukey test was used to examine statistical differences amongst different treatment and control groups. For all statistical tests, p < 0.05 was considered significant. 3. Results 3.1. Composition of EDCs in sewage effluents Twelve different EDCs were detected in effluent samples from the three STPs with the mean concentrations ranging between 5.25 ng/L (progesterone) and 4510 ng/L (NP) (Fig. 3). In general, xenoestrogens showed higher concentrations than other steroids, and estrogens were detected more frequently than androgens. Eight common EDCs including NP, BPA, 4-tert-octylphenol, triclosan, triclocarban, estrone (i.e., E1), 4-androstene-3,17-dione, and progesterone were universally detected in all samples from the three STPs. The effluent samples of Stanley STP only contained these eight compounds. In effluent samples obtained from Shek O STP, three more EDCs namely, 17a-boldenone, androsterone and testosterone were found. Norgestrel was only detected in effluent samples collected from the SWIMS STP, in addition to the eight common EDCs. Glucocorticoid group was not detected in any sample. Due to varying sources of influents and different treatment processes of the three STPs, the composition profiles of target EDCs from the three STPs were slightly different, while on average over 90% of the total EDC concentration was contributed by NP and BPA (Fig. 3). To further investigate the removal efficiency of the EDCs detected in STPs, we chose these two most abundant compounds (i.e., NP and BPA) as the targeted model EDCs for our subsequent investigations. 356 E.G.B. Xu et al. / Marine Pollution Bulletin 85 (2014) 352–362 Fig. 2. Schematic diagrams to illustrate various sewage treatment processes at each of the three different sewage treatment plants. 3.2. Removal efficiency of NP and BPA in the three STPs Both NP and BPA were continually detected in all sewage influent and effluent samples during both wet and dry seasons (Table 2). Concentrations of NP and BPA in the effluents and influents measured in the current study were within the ranges reported for other STPs in China (Jin et al., 2008; Nie et al., 2012) and Hong Kong (Kueh and Lam, 2008). Amongst the three STPs, mean influents concentrations of NP ranged from 646.3 to 2235.3 ng/L and from 906.9 to 1467.5 ng/L in wet and dry season, respectively (Table 2). Mean influent concentrations of BPA ranged from 159.3 to 617.2 ng/L and from 454.1 to 1141.2 ng/L in wet and dry season, respectively (Table 2). Mean effluent concentrations of NP and BPA were between 441.4–762.5 ng/L and 57.0–159.4 ng/L in wet season, and between 520.3–1562.0 ng/L and 258.0–713.0 ng/L in dry season, respectively (Table 2). Clearly, both NP and BPA showed significantly higher effluent concentrations in dry season than in wet season (Student’s t test: t P 2.03, df = 52, p < 0.001). Nonetheless, such a seasonal trend was not observed in the influent. The removals of NP and BPA in the STPs were dependent on the type of the sewage treatment facility and also dependent on the season (Table 2). There was a significant interaction of STP type and season (two-way ANOVA: F2, 48 = 5.52, p < 0.05). The mean removal rates for NP and BPA were 13.0% and 42.5% at Shek O STP, 69.2% and 73.0% at Stanley STP, and 61.4% and 53.7% at SWIMS STP, respectively. Higher removal rates were observed at the two biological treatment plants (i.e. Stanley and SWIMS STPs) in wet season than in dry season (Table 2). We also found significantly higher removal efficiencies of NP and BPA in the two biological treatment plants than in the preliminary treatment plant at Shek O (two-way ANOVA: F2,48 = 103.59, p < 0.01; Table 2). 3.3. NP and BPA in the receiving environment NP and BPA were also detected within the Cape D’ Aguilar Marine Reserve, which is relatively close to the three STPs (Fig. 1; Table 2). The mean concentrations of NP and BPA were 392.5 and 64.5 ng/L in wet season, and 109.4 and 69.5 ng/L in dry season, respectively. Interestingly, the mean concentration of NP in the marine reserve was significantly higher in wet season than that in dry season (Student’s t test: t = 7.70, df = 45, p < 0.001) and this was opposite to the seasonal pattern observed for sewage effluents. In contrast, the concentration of BPA varied insignificantly between the two seasons. 3.4. Implications from the medaka embryo-larval bioassay The results of mortality, hatchability and growth of the marine medaka embryos after 21 days of exposure to effluents and seawaters are summarized in Fig. 4. During exposure from 2 to 10 dpf, mortalities of the embryos were significantly higher in the treatments of SO10% (10% effluent from Shek O STP) and SW10% (10% effluent from SWIMS STP) than in the control groups, with average mortality rates of 59% and 18%, respectively (One-way ANOVA: F7,16 = 328.00, p < 0.05). A significantly lower E.G.B. Xu et al. / Marine Pollution Bulletin 85 (2014) 352–362 hatchability of embryos was also observed in SO10% and SW10% treatment groups, having 13% and 54% hatching success, respectively (F7,16 = 328.00, p < 0.05). The effects of the effluents on growth rate were assessed by measuring the interorbital distance (i.e., width) and the total length of larvae after 21-day exposure. We observed no significant difference of interorbital distance among all treatment groups, however, juveniles exposed to ST10% (i.e., 10% effluent from Stanley STP) and SO1% (i.e., 1% effluent from Shek O STP) were significantly smaller than the control fish in terms of their total length (F6,41 = 5.22, p < 0.05). Effects of the natural seawater treatment on mortality, hatchability and growth of medaka were not significantly different when compared with those in the artificial seawater control group. Overall, the results indicated that the 10 times diluted effluents from the STPs which were contaminated with various EDCs negatively affected the health of medaka embryos. The genes related to endocrine disruption responded differently to different exposures (i.e., effluents with different dilutions, seawater collected from the marine reserve, and artificial seawater; see Figs. S1–S3 in Appendix A), yet some general trends were observed as follows. Firstly, magnitudes of the mRNA expression in the medaka fish responding to the different treatment groups generally followed an order as: diluted effluents > natural seawaters > artificial seawater. For example, exposure to natural seawaters from the marine reserve caused no significant alteration in expression of VTG1 in 4 dpf fish, but VTG1 was significantly upregulated by 6-fold in 4 dpf fish after exposure to ST1% treatment (One-way ANOVA: F7,16 = 41.99, p < 0.05). The largest fold change amongst all gene expressions was observed in 4 dpf fish after exposure to Shek O effluents, with 50-fold of up-regulation in cyp1a (Fig. 5a). Shek O effluents also elicited significant up-regulations of VTG1 and VTG2 at the first fry stage (F6,14 P 46.15, p < 0.05; Fig. 5b and c). In contrast, the mRNA expression levels of ERb, ARNT, cyp19b, PPARa, and PPARb were all significantly down-regulated (F6,14 P 13.20, p < 0.05), while no significant difference was observed for ChgH, ChgL and PPARc (see Figs. S1–S3 in Appendix A). At 10 dpf, a significant up-regulation was observed for cyp1a (F6,14 = 129.95, p < 0.05) (Fig. 5d), while ERc, VTG1, VTG2, ChgH, ChgL, ARNT, cyp19b, and subtypes of PPAR were significantly down-regulated (F6,14 P 5.91, p < 0.05). At the first fry stage, ChgL, ARNT, cyp1a, cyp19a, cyp19b, PPARa, and PPARb were all significantly down-regulated (F6,14 P 15.66, p < 0.05). For fish exposed to the natural seawater collected from the marine reserve, their expression profiles for certain genes were significantly different from those of the control animals. The genes 357 of ERb, cyp19b, ARNT and PPARb were significantly downregulated for fish embryos at 4 dpf after exposure to the natural seawater (Fig. S1). For fish embryos at 10 dpf, the genes ERa, ERb and ChgH were significantly down-regulated whereas cyp19a gene was significantly up-regulated (Fig. S2). The genes of ERb, cyp19a, cyp19b, ChgH, ChgL, PPARa, PPARb were all significantly downregulated in the fish at the first fry stage (Fig. S3). The pattern of down-regulation of various genes was somewhat similar to those exposed to diluted sewage effluents. 4. Discussion 4.1. Composition of EDCs in sewage effluents Our results agree with a previous study that NP and BPA are the mostly detected phenolic EDCs in sewage effluents and their concentrations are consistently higher than the other EDC compounds (Wang et al., 2012). The total unit loads for NP and BPA were calP culated based on the following equation: ULt = 365 Cifi, where ULt is total EDC unit load (g/year) of the three STPs, Ci is the mean EDC concentration in effluent from each STP, and fi is the mean flow of each STP (m3/day). Based on this rough estimation, about 3500 g of NP and 1300 g of BPA would be discharged into the marine environment from the three STPs every year. We found the highest concentrations of most EDCs (7 out of 12 samples) from Shek O effluents, suggesting that it is an important source of EDC pollution in this area. In Shek O effluents, for example, the mean concentrations of NP (3595.03 ng/L) and E1 (24.14 ng/L) were around one order of magnitude higher than their corresponding predicted no effect concentration (PNEC) at 330 ng/L (EC, 2001) and 3 ng/L (UKEA, 2002), respectively. The occurrence of the EDCs at such high concentrations might have disrupted the endocrine system in marine organisms living nearby the sewage outfall of Shek O STP. 4.2. Removal efficiency of NP and BPA in the three STPs The main objective of wastewater treatment systems in Hong Kong is to remove organic substances, phosphorus, nitrogen, ammonia and E. coli from wastewater. Although EDCs can also be reduced by STPs, incomplete removal of EDCs would be largely attributable to the processes of the STPs (i.e., physical, chemical and/or biological treatment) and operational conditions (e.g. retention time of the sewage) (Birkett and Lester, 2003). In the present study, the three STPs were chosen to represent different types of Fig. 3. Composition profiles of the target EDCs in effluent samples collected from each of the three sewage treatment plants. 258.0 (65.7–395.2) 64.5 (11.0–407.5) Not applicable 69.5 (25.1–243.7) Not applicable 57.0 (44.7–81.0) 267.5 (147.7–472.7) 159.3 (40.8–342.3) 64.2b 454.1 (343.2–576.7) 43.2b 159.4 (64.0–265.4) 713.0 (356.4–1035.3) 617.2 (262.3–915.6) 74.2b 1014.6 (243.2–2262.1) 73.6b 140.9 (61.7–373.0) 537.2 (249.4–930.1) 1562.0 (391.7–2916.1) Wet season Removal rate (%) Dry season Removal rate (%) BPA (ng/L) 268.7 (35.2–1204.1) 47.6a 1141.2 (404.2–1836.7) 37.5a 520.3 (176.3–747.7) 441.4 (246.8–623.5) 2235.3 (448.5–4443.9) 80.3b 906.9 (401.9–1220.8) 42.6b 459.2 (185.8–1167.5) 1835.9 (1526.9–2042.6) 75.0b 1467.5 (772.8–1818.8) 63.4b 762.5 (541.9–1144.2) Influent 646.3 (530.0–959.5) 18.0a 1446.9 (469.5–2465.0) 8.0a Wet season Removal rate (%) Dry season Removal rate (%) NP (ng/L) Influent Stanley STP Influent Effluent Shek O STP Effluent SWIMS STP Effluent Seawater 392.5 (139.1–496.8) Not applicable 109.4 (60.8–327.9) Not applicable E.G.B. Xu et al. / Marine Pollution Bulletin 85 (2014) 352–362 Table 2 Summary of mean concentrations and their ranges (in brackets) of NP and BPA in influents and effluents from the three STPs and in seawaters from the Cape D’ Aguilar Marine Reserve. The mean removal rates of NP and BPA from each of the STP are given as bold numbers. For removal rates at the same row, the rates with different letters are significantly different (ANOVA followed by post hoc Tukey test: p < 0.05). 358 treatment processes (Fig. 2). The mechanical treatment consisted of two steps were used in all the investigated STPs. Screening was to remove objects such as rags and pieces of over 6 mm in diameter. The sewage then was passed through a sand trap where main solid organic material with lipophilic compounds would settle. We also detected NP and BPA at micrograms per gram dry weight levels in solid samples collected in the sand trap (data not shown). Stanley STP represented the activated sludge treatment, which was the most common type of biological treatment in Hong Kong. The sewage was pumped into large open-air basins containing suspended bacteria, where degradation or transformation of EDCs occurred (Auriol et al., 2006). Mixing was carried out by aeration, and the residence time for the treatment was 13.5 h. Trickling filters was used at the SWIMS STP. After settled in the septic tank, sewage was passed through a system of meandering biofilter tank. Such tank provided varying depths to create aerated and anoxic zones for biodegradation or transformation of EDCs. The residence time for the SWIMS STP was only about 0.5 h. However, such a short residence time can be a negative influence on the EDC removal (DEPA, 2003). The removal efficiency of BPA found in the present study (see Table 2) was in agreement with literature values (Auriol et al., 2006). The removal efficiency of NP in Shek O STP was much lower than the reported literature values of up to 99% (Auriol et al., 2006). This may be attributed to the fact that Shek O STP is only a preliminary treatment plant, and the degradation of NPEO in the sewer may increase the concentration of NP in the final effluent. In contrast, the removal efficiency of NP at Stanley and SWIMS STPs was high and consistent with the literature data (Auriol et al., 2006). Besides, the SWIMS STP, which applies biological filter technology, was generally less efficient in removing NP and BPA when compared with the activated sludge process adopted at Stanley STP. This could be due to a short hydraulic retention time in the biofilter STP (Clara et al., 2005). Svenson et al. (2003) also observed that the activated sludge process showed higher estrogenic removal (81%) than trickling filters (28%). Therefore, it is necessary to upgrade the treatment facilities at Shek O and SWIMS STPs to better remove EDC residues from the raw sewage. Both NP and BPA showed significantly higher effluent concentrations in dry season than in wet season. Such seasonal patterns can be partially explained by the dilution effect associated with high rainfall and elevated water utilization during the summer, wet season in Hong Kong. The flow of sewage (m3/day) was approximately 30% and 115% higher during the samplings in wet season at Stanley and Shek O STP, respectively. Wang et al. (2010) also reported two to fivefold lower concentrations of 7 selected EDCs from 3 STPs in wet season than in dry season, which was mainly due to the dilution by rain water. Furthermore, when temperature is high during the summer, wet season, both NP and BPA may undergo a faster physicochemical and biological degradation, and perhaps STPs with biological treatment may have higher removal efficiency for these compounds in summer due to higher microbial metabolisms and activities (Manzano et al., 1999). Ko et al. (2007) and Nie et al. (2012) also demonstrated that concentrations of NP and BPA in the effluents were higher in winter and spring than in summer and autumn, which was closely related to the microbial activity and concentrations of mixed liquor suspended solids. Higher removal rates for NP and BPA were observed at Stanley and SWIMS STPs in wet season than in dry season. As discussed above, the faster degradation rates for NP and BPA in summer were probably attributable to the presence of more active microorganisms (e.g. more arylsulfatase enzymes) at higher temperature (Manzano et al., 1999). In the present study, the average temperature of sewage was 16 °C during the samplings in dry season and that was 26 °C in wet season. Some researchers also reported E.G.B. Xu et al. / Marine Pollution Bulletin 85 (2014) 352–362 Fig. 4. Average mortality, hatchability and growth of O. melastigma after exposure to treatments with different proportions of effluents and to the natural seawater collected from the marine reserve [ST: Stanley STP; SO: Shek O STP and SW: Swire Institute of Marine Science STP]. Mean and SD are shown; ⁄ denotes a significant difference between the treatment and control groups, p < 0.05. varying concentrations and different removal rates of EDCs in different seasons, and concluded that temperature can affect the biodegradation process to a large degree (Jin et al., 2008). Lian et al. (2009) investigated the fate of NPEO and their metabolites (including NP) in 4 STPs. They found that NP was not only biodegraded but was also produced from its parent compound (NPEO) during wastewater treatment process, and higher removal efficiencies of both NP and NPEO in summer were probably due to high temperature. 4.3. NP and BPA in the receiving environment The recent knowledge of NP and BPA occurrence and behaviour in river water is well established, but such knowledge in marine environments is still limited (Sharma et al., 2009). The measured concentration range of NP in the present study was slightly higher than that reported in marine waters of Hong Kong by Kueh and Lam (2008), which ranged from less than 100 up to 270 ng/L. For BPA, although no data is available for the comparison with our study in the marine environment of Hong Kong, our reported BPA concentrations are one order of magnitude lower than those quantified in the Pearl River, China (Gong et al., 2012). The higher concentrations of NP in summer, wet season in the receiving waters could be caused by several factors. Firstly, this could be due to high temperature and associated microbial activity, leading to an enhanced degradation of NPEO in marine sediment and hence an increased NP concentrations in water column during summer (Li et al., 2004). Levels of NP in both natural surface water and in suspended particles were found decreasing with decreasing water temperature (Xu et al., 2006). Fu et al. (2007) also reported that a higher concentration of NP in summer in coastal waters of 359 Qingdao, China, was mainly due to the higher degradation rate of NPEO and re-suspended sediment under strong wind. Some phenomena might also occur in the Cape D’ Aguilar Marine Reserve, though further investigation would be required. It is also postulated that the other major reason for the elevated NP concentrations in natural seawater during wet season may be attributable to an increased input of NP from the increased surface runoff (Zhao et al., 2009). In the present study, surface runoff and storm water discharges may have also increased the input of NP during wet season. Kueh and Lam (2008) found the storm water in Hong Kong containing 80–12,000 ng/L of NP and 260– 29,200 ng/L of NPEO. To better understand its environmental behaviour, the concentrations of NP should be further measured in suspended particle, sediment, and surface runoff samples collected around and within the marine reserve. Investigations should also involve its parent compounds (e.g. NPEO) and its major aerobic metabolites (e.g. NP1EC, NP2EC) and anaerobic metabolites (e.g. NP1EO, NP2EO) (Ahel et al., 1994). Seasonal variation of BPA in the seawaters of the marine reserve was not observed in the present study. The level of BPA in seawater might have reached an equilibrium condition of leaching from chemical products (e.g. epoxy and polycarbonate plastics), solution in seawater, sorption to suspended particles, incorporation into organic matters, aerobic degradation with hydroxyl radicals, bioaccumulation by marine organisms, and mineralization by bacteria (Cousins et al., 2002). These processes can be affected by various factors, including seawater temperature, pH, inorganic ions and phytoplankton in seawater, and reactive oxygen species (Sajiki and Yonekubo, 2002). However, Patrolecco et al. (2006) found different seasonal patterns of EDCs in different rivers in Italy, and concluded that the levels of BPA in aquatic compartments were affected by differences in hydrological conditions between different sampling campaigns, and that the process of re-suspension and re-dissolution from sediment was an important source of EDCs. Thus, the environmental fate of BPA and NP in the Cape D’ Aguilar Marine Reserve should be further investigated in more detail. The ecological risks from the NP and BPA in the marine reserve were assessed using the risk quotient (RQ) approach, i.e., a ratio between the measured environmental concentration (MEC) and PNEC (RQ = MEC/PNEC). The RQ must remain below 1 to ensure an acceptable risk to the environment (EU, 1994). Given that the proposed PNECs of NP and BPA were 330 and 150 ng/L for marine water (EC, 2001; EU, 2010), mean RQs for NP and BPA in the marine reserve were calculated as 1.1 and 0.4 in wet season, and 0.4 and 0.5 in dry season, respectively. The risk of BPA was consistently low, but that for NP was high with the RQ exceeding 1 during wet season. Therefore, organisms in the marine reserve were likely adversely influenced by the elevated level of NP during wet season. 4.4. Implications from the medaka bioassay The overall changes in expression of various genes in O. melastigama after exposure to the diluted sewage effluents are summarized in Fig. 6. The mRNA expression profiles were dependent on the developmental stages of medaka embryos and effluent concentrations in a gene-subtype-specific manner. For instance, VTG1 and VTG2 were significantly up-regulated at 4 dpf but then inhibited after 6 days of further exposure to the effluents (i.e., 10 dpf). PPARa and PPARb exhibited similar expression profiles in the fish with significant down-regulation upon the exposure regardless of the developmental stage, but PPARc did not change in its expression among all developmental stages and among the treatments. Overall, more genes (11 out of 13) were up or down regulated at the late embryonic stage (10 dpf) than those at the early embryonic stage (4 dpf with 7 genes) and the 1st fry stage 360 E.G.B. Xu et al. / Marine Pollution Bulletin 85 (2014) 352–362 Fig. 5. Examples showing mean expression levels (i.e., relative fold changes) of selected genes at different developmental stages of O. melastigma: (a), cyp1a at 4 dpf; (b), VTG1 at 1st fry stage; (c), VTG2 at 1st fry stage; and (d), cyp1a at 10 dpf after exposure to the control (artificial seawater), five different effluent samples (A: ST10%, B: ST1%, C: SO10%; D: SO1%, F: SW10% and G: SW1%, respectively) and a natural seawater sample obtained from the marine reserve (E) [ST: Stanley STP; SO: Shek O STP and SW: Swire Institute of Marine Science STP]. The data in triplicate are presented as the mean and SD, relative to the control; ⁄p < 0.05. Fig. 6. A summary of the gene expression profile of O. melastigma at 4 dpf, 10 dpf and 1st fry stage, respectively, after exposure to diluted sewage effluents. The relative expression levels of genes in the treatment vs. the control were indicated as follows: significant up-regulation, up arrow; significant down-regulation, down arrow; no effect, horizontal line, based on the results from one-way ANOVAs at p < 0.05. (with 7 genes). These results indicated that the expression of subtypes of ER or PPAR genes were dependent on the developmental stage of the fish, which was consistent with results reported from other studies (Seo et al., 2006; Cocci et al., 2013). However, cyp19b mRNA expression levels in the medaka embryo and juvenile were significantly reduced at all developmental stages. Similarly, NP also exhibited potent inhibitory effects on cyp19 genes and significantly reduced brain aromatase activity in Atlantic salmon (Kortner et al., 2009). Reduced ovarian aromatase activity in red mullet was suggested to be also caused by NP (Martin-Skilton et al., 2006). On the contrary, NP was found to induce cyp19A2 gene in dose-dependent manner in zebrafish juveniles (Kazeto et al., 2004), and cause significant induction of cyp19 isomers in immature Atlantic salmon (Meucci and Arukwe, 2006). The differential abundance and expression of cyp19 genes in different fish species after exposure to estrogenic compounds have been reported previously (Trant et al., 2001; Cheshenko et al., 2008). Cytochrome aromatase as well as estrogen receptor genes isotypes showed differential organ-specific, NP and BPA concentration- and time-dependent expression patterns after exposure to environmental relevant concentrations of NP and BPA (Lee et al., 2006). 5. Conclusion In this study, we first screened 33 common EDCs and found that there were twelve EDCs present in effluents from three STPs located at south of Hong Kong Island, and NP and BPA were the most abundant EDCs. Afterwards, this study comprehensively investigated the occurrence, seasonal variation and biological effects of NP and BPA in influent and effluent samples collected from the three STPs, and in seawaters collected from the Cape D’ Aguilar Marine Reserve adjacent to these STPs. We discovered that concentrations of NP and BPA in influents were comparable to those in effluents from the preliminary STP in Shek O, indicating its poor removal efficiency for these compounds. In contrast, concentrations of the two compounds were significantly decreased at Stanley and SWIMS STPs following more efficient biological treatments. Effluent concentrations of NP and BPA were higher in dry season than in wet season, but opposite seasonal changes of NP were observed in receiving waters (i.e., the Cape D’ Aguilar Marine Reserve), probably because of the increased input of NP from the increased surface runoff during the wet season. Our results also showed that Stanley STP using an activated sludge process was E.G.B. Xu et al. / Marine Pollution Bulletin 85 (2014) 352–362 more effective to remove NP and BPA from wastewater than the biological filter adopted at SWIMS STP. Lower removal rates were observed at these two biological STPs in dry, winter season than in wet, summer season, suggesting that the EDC removal process is temperature dependent. Natural seawater samples taken from the marine reserve also exhibited elevated levels of NP with a risk quotient greater than one in wet season, indicating potential hazards of this compound to marine organisms. In addition, our laboratory experiment further confirmed that diluted effluents from the three STPs and natural seawaters from the marine reserve can elicit transcriptional responses of genes related to endocrine disruption pathways in the marine medaka fish. Overall, our results demonstrated that sewage effluents can act as the major source for the continuous input of estrogenic compounds into the marine environment. The existing sewage treatment facilities at Shek O and SWIMS STPs should be upgraded as a means to reduce the discharge of EDCs into the marine environment and hence lower their ecological risks to marine organisms living in the receiving waters including those inhabiting the Cape D’ Aguilar Marine Reserve. Acknowledgements This work is jointly supported by the Area of Excellence (AoE) Scheme under the University Grants Committee of the Hong Kong Special Administration Region (HKSAR), China (Project No. AoE/P04/2004), and by a research grant from the Swire Educational Trust. The authors thank the Drainage Services Department of the HKSAR Government for granting us a permission to collect the sewage influent and effluent samples for this study. Elvis Xu would also like to thank John Swire & Sons Limited and the Swire Educational Trust for providing him a James Henry Scott (Hong Kong) PhD Scholarship. The authors also thank Andy Yi and Karen Villarta for their valuable comments on early drafts of this manuscript, and staff and postgraduate students at the Swire Institute of Marine Science for assisting this project. The authors are grateful to the Agriculture Fisheries and Conservation Department for granting a permit for taking samples from the Cape D’ Aguilar Marine Reserve. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.marpolbul.2014. 02.029. References Ahel, M., Giger, W., Koch, M., 1994. Behaviour of alkylphenol polyethoxylate surfactants in the aquatic environment—Occurrence and transformation in sewage treatment. Wat. Res. 28 (5), 1131–1142. Aluru, N., Leatherland, J.F., Vijayan, M.M., 2010. Bisphenol A in oocytes leads to growth suppression and altered stress performance in juvenile rainbow trout. PLoS One 5 (5), e10741. Auriol, M., Filali-Meknassi, Y., Tyagi, R.D., Adams, C.D., Surampalli, R.Y., 2006. Endocrine disrupting compounds removal from wastewater, a new challenge. Process Biochem. 41, 525–539. Birkett, J.W., Lester, J.N., 2003. Endocrine disrupters in wastewater and sludge treatment processes. IWA Publishing, London, UK. Burkhardt-Holm, P., 2010. Endocrine disruptors and water quality, a state-of-theart review. Int. J. Water Resour. D. 26, 477–493. Cheshenko, K., Pakdel, F., Segner, H., Kah, O., Eggen, R.I.L., 2008. Interference of endocrine disrupting chemicals with aromatase CYP19 expression or activity, and consequences for reproduction of teleost fish. Gen. Comp. Endocrinol. 155, 31–62. Cocci, P., Mosconi, G., Palermo, F.A., 2013. Effects of 4-nonylphenol on hepatic gene expression of peroxisome proliferator-activated receptors and cytochrome P450 isoforms (CYP1A1 and CYP3A4) in juvenile sole (Solea solea). Chemosphere, available online 15.07.13. Cousins, I.T., Staples, C.A., Klecka, G.M., Mackay, D., 2002. A multimedia assessment of the environmental fate of bisphenol A. Hum. Ecol. Risk Assess. 8 (5), 1107– 1135. 361 Clara, M., Kreuzinger, N., Strenn, B., Gans, O., Kroiss, H., 2005. The solids retention time – a suitable design parameter to evaluate the capacity of wastewater treatment plants to remove micropollutants. Water Res. 39, 97–106. Danish Environmental Protection Agency, (DEPA), 2003. Evaluation of analytical chemical methods for detection of estrogens in the environment. Working Report No. 44. Danish Environmental Protection Agency, Danish Ministry of the, Environment. EC, 2001. European Union Risk–Assessment Report Vol.10, 2002 on 4-nonylphenol (branched) and nonylphenol, European Chemicals Bureau, Joint Research Centre, European Commission, Ispra, Italy. ISBN 92-827-801. <http://ecb.jrc.it/ existing-chemicals(under,existing-chemicals/risk-assessment/report)>. EU, 1994. Ad Hoc Working Party, III/5504/94 Draft 4. Assessment of potential risks to the environment posed by medicinal products for human use, excluding products containing live genetically modified organisms. EU, 2010. Updated European Union Risk Assessment Report. 4,40–Isopropylidenediphenol (Bisphenol–A). European Commission, EUR 24588 EN. Fang, C., Wu, X.L., Huang, Q.H., Liao, Y.Y., Liu, L.P., Qiu, L., Shen, H.Q., Dong, S.J., 2012. PFOS elicits transcriptional responses of the ER, AHR and PPAR pathways in Oryzias melastigma in a stage–specific manner. Aquat. Toxicol. 106–107, 9–19. Fu, M., Li, Z., Gao, H., 2007. Distribution characteristics of nonylphenol in Jiaozhou Bay of Qingdao and its adjacent rivers. Chemesphere 69, 1009–1016. Gong, J., Ran, Y., Chen, D.Y., Yang, Y., Zeng, E.Y., 2012. Association of endocrinedisrupting chemicals with total organic carbon in riverine water and suspended particulate matter from the Pearl River, China. Environ. Toxicol. Chem. 31 (11), 2456–2464. Gibson, R., Smith, M.D., Spary, C.J., Tyler, C.R., Hill, E.M., 2005. Mixtures of estrogenic contaminants in bile of fish exposed to wastewater treatment works effluents. Environ. Sci. Technol. 39, 2461–2471. Jin, S.W., Yang, F.X., Liao, T., Hui, Y., Xu, Y., 2008. Seasonal variations of estrogenic compounds and their estrogenicities in influent and effluent from a municipal sewage treatment plant in China. Environ. Toxicol. Chem. 27 (1), 146–153. Kazeto, Y., Place, A.R., Trant, J.M., 2004. Effects of endocrine disrupting chemicals on the expression of CYP19 genes in zebrafish (Danio rerio) juveniles. Aquat. Toxicol. 69, 25–34. Ko, E.J., Kim, K.W., Kang, S.Y., Kim, S.D., Bang, S.B., Hamm, S.Y., Kim, D.W., 2007. Monitoring of environmental phenolic endocrine disrupting compounds in treatment effluents and river waters, Korea. Talanta 73, 674–683. Kortner, T.M., Mortensen, A.S., Hansen, M.D., Arukwe, A., 2009. Neural aromatase transcript and protein levels in Atlantic salmon (Salmo salar) are modulated by the ubiquitous water pollutant, 4-nonylphenol. Gen. Comp. Endocrinol. 164, 91–99. Kueh, C.S.W., Lam, J.Y.C., 2008. Monitoring of toxic substances in the Hong Kong marine environment. Mar. Pollut. Bull. 57 (6), 744–757. Länge, R., Hutchinson, T.H., Croudace, C.P., Siegmund, F., Schweinfurth, H., Hampe, P., Panter, G.H., Sumpter, J.P., 2001. Effects of the synthetic estrogen 17 alphaethinylestradiol on the life-cycle of the fathead minnow (Pimephales promelas). Environ. Toxicol. Chem. 20 (6), 1216–1227. Lech, J.J., Lewis, S.K., Ren, L., 1996. In vivo estrogenic activity of nonylphenol in rainbow trout. Fund. Appl. Toxicol. 30, 229–232. Lee, Y.M., Seo, J.S., Kim, I.C., Yoon, Y.D., Lee, J.S., 2006. Endocrine disrupting chemicals (bisphenol A, 4-nonylphenol, 4-tert-octylphenol) modulate expression of two distinct cytochrome P450 aromatase genes differently in gender types of the hermaphroditic fish Rivulus marmoratus. Biochem. Biophys. Res Co. 345, 894–903. Lian, J., Liu, J.X., Wei, Y.S., 2009. Fate of nonylphenol polyethoxylates and their metabolites in four Beijing wastewater treatment plants. Sci. Total Environ. 407, 4261–4268. Li, D.H., Kim, M.S., Oh, J.R., Park, J.N., 2004. Distribution characteristics of nonylphenols in the artificial Lake Shihwa, and surrounding creeks in Korea. Chemosphere 56, 783–790. Li, X.L., Luan, T.G., Liang, Y., Wong, M.H., Lan, C.Y., 2007. Distribution patterns of octylphenol and nonylphenol in the aquatic system at Mai Po Marshes Nature reserve, a subtropical estuarine wetland in Hong Kong. J. Environ. Sci. China 19, 657–662. Liu, S., Ying, G.G., Zhao, J.L., Chen, F., Yang, B., Zhou, L.J., Lai, H.J., 2011. Trace analysis of 28 steroids in surface water, wastewater and sludge samples by rapid resolution liquid chromatography-electrospray ionization tandem mass spectrometry. J. Chromatogr. A. 1218, 1367–1378. Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative gene expression data using real–time quantitative PCR and the 2(-Delta Delta C (T)) Method. Methods 25, 402–408. Manzano, M.A., Perales, J.A., Sales, D., Quiroga, J.M., 1999. The effect of temperature on the biodegradation of a nonylphenol polyethoxylate in river water. Water Res. 33, 2593–2600. Martin-Skilton, R., Lavado, R., Thibaut, R., Minier, C., Porte, C., 2006. Evidence of endocrine alteration in the red mullet, Mullus barbatus from the NW Mediterranean. Environ. Pollut. 141 (1), 60–68. Meucci, V., Arukwe, A., 2006. The environmental estrogen, 4-nonylphenol modulates brain estrogen-receptor- and aromatase (CYP19) isoforms gene expression patterns in Altantic salmon (Salmo salar). Mar. Environ. Res. 62, S195–S199. Mills, L.J., Chichester, C., 2005. Review of evidence, are endocrine-disrupting chemicals in the aquatic environment impacting fish populations? Sci. Total Environ. 343, 1–34. Nie, Y.F., Qiang, Z.M., Zhang, H.Q., Ben, W.W., 2012. Fate and seasonal variation of endocrine-disrupting chemicals in a sewage treatment plant with A/A/O process. Sep. Purif. Technol. 84, 9–15. 362 E.G.B. Xu et al. / Marine Pollution Bulletin 85 (2014) 352–362 Patrolecco, L., Capri, S., De Angelis, S., Pagnotta, R., Polesello, S., Valsecchi, S., 2006. Partition of nonylphenol and related compounds among different aquatic compartments in Tiber River (Central Italy). Water Air Soil Pollut. 172, 151–166. Sajiki, J., Yonekubo, J., 2002. Degradation of bisphenol A (BPA) in the presence of reactive oxygen species and its acceleration by lipids and sodium chloride. Chemosphere 46, 345–354. Seo, J.S., Lee, Y.M., Jung, S.O., Kim, I.C., Yoon, Y.D., Lee, J.S., 2006. Nonylphenol modulates expression of androgen receptor and estrogen receptor genes differently in gender types of the hermaphroditic fish Rivulus marmoratus. Biochem. Biophys. Res. Co. 346 (1), 213–223. Sharma, V.K., Anquandah, G.A., Yngard, R.A., Kim, H., Fekete, J., Bouzek, K., Ray, A.K., Golovko, D., 2009. Nonylphenol, octylphenol, and bisphenol–A in the aquatic environment: a review on occurrence, fate, and treatment. J. Environ. Sci. Health A 44 (5), 423–442. Svenson, A., Allard, A.S., Ek, M., 2003. Removal of estrogenicity in Swedish municipal sewage treatment plants. Water Res. 37, 4433–4443. Trant, J.M., Gavasso, S., Ackers, J., Chung, B.C., Place, A.R., 2001. Developmental expression of cytochrome P450 aromatase genes (CYP19a and CYP19b) in zebrafish fry (Danio rerio). J. Exp. Zool. 290 (5), 475–483. UK Environment Agency, (UKEA), 2002. Proposed Predicted-No-Effect-Concentrations (PNECs) for Natural and Synthetic Steroid Oestrogens in Surface Waters. Research and Development Technical. Report P2–T04/1. Von Saal, F., Cooke, P.S., Buchanan, D.L., Palanza, P., Thayer, K.A., Nagel, S.C., Parmigiani, S., Welshons, W.V., 1998. A physiologically based approach to the study of bisphenol A and other estrogenic chemicals on the size of reproductive organs, daily sperm production, and behavior. Toxicol. Ind. Health 14, 239–260. Wang, L.Y., Zhang, X.H., Tam, N.F.Y., 2010. Analysis and occurrence of typical endocrine–disrupting chemicals in three sewage treatment plants. Water Sci. Technol. 62 (11), 2501–2509. Wang, L., Ying, G.G., Chen, F., Zhang, L.J., Zhao, J.L., Lai, H.J., Chen, Z.F., Tao, R., 2012. Monitoring of selected estrogenic compounds and estrogenic activity in surface water and sediment of the Yellow River in China using combined chemical and biological tools. Environ. Pollut. 165, 241–249. White, R., Jobling, S., Hoare, S.A., Sumpter, J.P., Parker, M.G., 1994. Environmentally persistent alkylphenolic compounds are estrogenic. Endocrinology 135, 175– 182. Xu, J., Wang, P., Guo, W., Dong, J., Wang, L., Dai, S., 2006. Seasonal and spatial distribution of nonylphenol in Lanzhou Reach of Yellow River in China. Chemosphere 65, 1445–1451. Zhang, Y., Zhou, J.L., 2008. Occurrence and removal of endocrine disrupting chemicals in wastewater. Chemosphere 73, 848–853. Zhao, J.L., Ying, G.G., Wang, L., Yang, J.F., Yang, X.B., Yang, L.H., Li, X., 2009. Determination of phenolic endocrine disrupting chemicals and acidic pharmaceuticals in surface water of the Pearl Rivers in South China by gas chromatography–negative chemical ionization–mass spectrometry. Sci. Total Environ. 407, 962–974.