Photolytic Degradation Products of Two Highly Brominated Flame

Sep 15, 2014 - Doug Crump , Amani Farhat , Suzanne Chiu , Kim L. Williams , Stephanie P. Jones , and Valerie S. Langlois. Environmental Science ...
0 downloads 0 Views 772KB Size
Download PDF
Article pubs.acs.org/est

Photolytic Degradation Products of Two Highly Brominated Flame Retardants Cause Cytotoxicity and mRNA Expression Alterations in Chicken Embryonic Hepatocytes Guanyong Su,†,‡ Robert J. Letcher,*,†,‡ Doug Crump,† Reza Farmahin,†,§ John P. Giesy,∥,⊥,# and Sean W. Kennedy†,§ †

Ecotoxicology and Wildlife Health Division, Environment Canada, National Wildlife Research Centre, Carleton University, Ottawa, Ontario K1A 0H3, Canada ‡ Department of Chemistry, Carleton University, Ottawa, Ontario K1S 5B6, Canada § Centre for Advanced Research in Environmental Genomics, Department of Biology, University of Ottawa, Ottawa, Ontario Canada ∥ Department of Veterinary Biomedical Sciences and Toxicology Centre, University of Saskatchewan, Saskatoon, Saskatchewan S7N 5B3, Canada ⊥ Department of Zoology and Center for Integrative Toxicology, Michigan State University, East Lansing, Michigan 48824, United States # Department of Biology & Chemistry, State Key Laboratory in Marine Pollution, City University of Hong Kong, Kowloon, Hong Kong, SAR China S Supporting Information *

ABSTRACT: Tetradecabromo-1,4-diphenoxybenzene (TeDB-DiPhOBz) and 2,2′,3,3′,4,4′,5,5′,6,6′-decabromodiphenyl ether (BDE-209) are photolytically unstable flame retarding chemicals. Here, photocatalyzed byproducts of TeDB-DiPhOBz and BDE-209 (i.e Br8- to Br11-PB-DiPhOBz congeners from TeDB-DiPhOBz, and Br6- to Br8-BDE congeners from BDE-209), formed after 21 days of natural sunlight irradiation (SI), were assessed for exposure effects on cytotoxicity and mRNA expression levels of selected genes in chicken embryonic hepatocytes (CEH). CEHs were exposed for 36 h to concentrations of SI- and nonirradiated (NI)TeDB-DiPhOBz and BDE-209. Cytotoxic effects were observed only in CEH exposed to 50 μM SI-BDE-209. Results from a custom-designed Avian ToxChip polymerase chain reaction array showed that NI-TeDB-DiPhOBz and NIBDE-209, up to maximum concentrations of 1.9 and 9 μM, respectively, caused limited changes in mRNA levels of 27 genes from toxicologically relevant pathways, including phase I/II metabolism, the thyroid hormone pathway, lipid/cholesterol metabolism, oxidative stress, immune response, and cell death. In contrast, 12 and 14 of the 27 genes were altered after exposure to 25 μM SI-TeDB-DiPhOBz or 10 μM SI-BDE-209, respectively. Aryl hydrocarbon receptor (AhR)-related CYP1A4 mRNA levels were the most altered on the PCR array with an induction of 560- and 5200-fold after exposure to 1 or 25 μM SI-TeDB-DiPhOBz, respectively, and 2500- and 2300-fold after exposure to 1 or 10 μM SI-BDE-209, respectively. A dioxin-responsive luciferase reporter gene assay confirmed that the CYP1A4 inductions were independent of the dissolution solvents used (tetrahydrofuran/n-hexane, n-hexane, or methanol) during photolysis. Overall, degradation of TeDB-DiPhOBz and BDE-209 by natural sunlight generates byproducts that affect in vitro expression of genes, especially the AhR-mediated CYP1A4.

■

INTRODUCTION

Due to their ability to reduce flammability and hinder fire ignition in the products that contain them, brominated flame retardants (BFRs) have been widely used in various commercial products such as furniture, textiles, plastics, paints, and electronic appliances.1,2 Considering their bioaccumulation, © 2014 American Chemical Society

Received: Revised: Accepted: Published: 12039

July 14, 2014 September 12, 2014 September 15, 2014 September 15, 2014 dx.doi.org/10.1021/es503399r | Environ. Sci. Technol. 2014, 48, 12039−12046

Environmental Science & Technology

Article

following exposure to TeDB-DiPhOBz and BDE-209 or their phototransformed byproducts, generated in situ as a result of natural sunlight irradiation.

long-range transport, and adverse biological effects, two of three commercial mixtures of polybrominated diphenyl ethers (PBDEs)Penta- and Octa-BDE formulationswere recently (May, 2009) listed in Annex A of the persistent organic pollutant (POP) Stockholm Convention, indicating that these chemicals are destined for elimination.3 Deca-BDE mixtures contain >90% of 2,2′,3,3′,4,4′,5,5′,6,6′-decaBDE (BDE-209), and its commercial use remains unregulated, although several manufacturers have announced voluntary phase-outs. BDE-209 is normally added to manufactured products without being chemically bonded with product polymers and therefore can be easily released from the products and enter the environment. Tetradecabromo-1,4-diphenoxybenzene (TeDB-DiPhOBz), also abbreviated as 4′-PeBPO-BDE208, is a halogenated polyphenyl ether that is a replacement for BDE-209 and is the main constituent of several commercial technical formulations. A 1973 U.S. Patent (US3760003A) to the Dow Chemical Company detailed the production of halogenated polyphenyl ethers that included TeDB-DiPhOBz. In 1986, Albermarle Chemical Corporation acquired Dow Chemical’s bromine chemical business. Although Albermarle has claimed that the production of their major TeDB-DiPhOBz-based BFR, SAYTEX-120, was phased-out and discontinued commercially in January 2011, SAYTEX 120-related mixtures continue to be produced and marketed in some regions around the world, especially Asia.4−6 To our knowledge, there are no published reports on the presence of TeDB-DiPhOBz in any environmental compartments worldwide. Several novel methoxylated polybrominated diphenoxybenzene (MeO-PB-DiPhOBz) congeners were reported in herring gull eggs from sites across the Great Lakes of North America and have been present for at least the past 30 years.5 MeO-PB-DiPhOBz’s were hypothesized to be degradation products of TeDB-DiPhOBz.5 TeDB-DiPhOBz and BDE-209 are highly brominated, generally nonvolatile, and, due to their log octanol−water partition coefficients of >10, have low bioavailability or potential to bioaccumulate. When exposed to UV-A, -B, -C, or natural sunlight, TeDB-DiPhOBz can undergo rapid photolysis and degrade via stepwise, reductive debromination.6 For example, when exposed to natural sunlight, half-lives of TeDB-DiPhOBz during photolysis ranged from 4.9 to 7.4 min. Similarly, when irradiated by sunlight, BDE-209 can be degraded by photolysis,7−9 with a half-life in n-hexane (1% THF) of 5.3 min.6 BDE-209 has also been shown to degrade to less brominated, more bioavailable, bioaccumulative, and potentially more toxic compounds when fed to Amercian kestrels10 or European starlings via silastic implants.11 To our knowledge, there are no published reports on biological effects in biota as a result of exposure to TeDB-DiPhOBz or its degradation byproducts. Polymerase chain reaction (PCR) array technology is a relatively new and powerful toxicogenomic approach that combines real-time PCR (RT-PCR) performance with the high throughput ability of microarrays.12 Recently, 16 organic flame retardants (OFRs) were screened using a chicken embryonic hepatocyte (CEH) assay in combination with a customdesigned Avian ToxChip PCR array that simultaneously measures mRNA expression levels of 27 genes from various toxicologically relevant pathways, including phase I and II metabolism, the thyroid hormone pathway, lipid homeostasis, oxidative stress, immune response, and cell death.13 In the present study, an in vitro CEH assay and an Avian ToxChip PCR array were used to examine mRNA alterations

■

MATERIALS AND METHODS Chemicals. To our knowledge, pure standards for TeDBDiPhOBz or any less-brominated PB-DiPhOBzs are not yet commercially available. Technical SAYTEX-120 (TeDB-DiPhOBz; Lot# 0GN01-$I0) and BDE-209, in solid powder form, were kindly supplied by Wellington Laboratories (Guelph, ON, Canada). The purity of BDE-209 was reported to be greater than 98% by Wellington Laboratories. Chemical structures of both BFRs are provided in Figure S1. Organic solvents used in this research were provided by Caledon Laboratories Ltd. (Georgetown, ON, Canada) with the exception of dimethyl sulfoxide (DMSO) and tetrahydrofuran (THF), which were purchased from Sigma-Aldrich (St. Louis, MO, USA). Sample Preparation. TeDB-DiPhOBz and BDE-209 powder was dissolved in 30% THF/n-hexane solution to achieve a final, nominal concentration of 300 μM. Aliquots of 10 mL of the resulting solutions were transferred into borosilicate glass tubes (16 × 125 mm; Fisher thermo scientific Inc.) in four replicates each for sunlight irradiation. Borosilicate glass vessels were the same as those used in irradiation studies by Chen et al.6 Borosilicate glass effectively transmits radiation from the infrared down to approximately λ = 300 nm. Therefore, as in Chen et al.,6 the present irradiation experiments were based on the percent transmittance of wavelength ranges through the borosilicate glass, which was 100% for UV-A, ∼50% for UV-B, and ∼10% for UV-C. For the present study, the solvent control (n = 4 replicates) had the same volume of 30% THF/n-hexane. On day 0, for the solvent comparison assessments, two of the four replicates of the TeDB-DiPhOBz, BDE-209, and control solutions were blown down to dryness under a gentle nitrogen flow and redissolved in methanol (TeDB-DiPhOBz) or n-hexane (BDE-209) until the solvent became clear, which indicated complete dissolution of the BFRs. For nonirradiated (NI) samples, half the sample volume was transferred to a 2 mL amber GC vial and stored in the dark at −20 °C for subsequent quantification. The remaining sample volume was blown down to dryness under a gentle flow of nitrogen and redissolved in DMSO for subsequent administration to CEH. The remaining two replicates of the original TeDB-DiPhOBz, BDE-209, and control solutions were exposed for 21 days to natural sunlight irradiation (SI) in the natural outside environment. On day 21, these samples were taken into the laboratory and prepared the same way as the NI solutions for quantification and in vitro dosing of CEH. From this point forward, samples collected on day 0 are referred to as NI-TeDB-DiPhOBz and NI-BDE-209, and those collected after 21 days of sunlight irradiation are referred to as SI-TeDB-DiPhOBz and SI-BDE-209. The sunlight irradiation was conducted from December 24, 2013 to January 14, 2014 in Ottawa, and the location coordinates were 45° 40′ 06″ N and 75° 74′ 22″ W. Since all the TeDB-DiPhOBz or BDE-209 was degraded to byproducts after 21 days of exposure to sunlight, the “μM” units of SI-TeDB-DiPhOBz or SI-BDE-209 do not represent actual concentrations of degraded byproducts. They are the “complex” concentration of byproducts relative to the initial concentration of their precursor compounds, NI-TeDBDiPhOBz or NI-BDE-209. For in vitro experiments, DMSO stock solutions of 5000 μM and 10000 μM were prepared for 12040

dx.doi.org/10.1021/es503399r | Environ. Sci. Technol. 2014, 48, 12039−12046

Environmental Science & Technology

Article

Table 1. Pathways, RefSeq Accession, and Description of 32 Genes on the Avian ToxChip PCR Array pathways phase I and II metabolism

immune function

glucose and fatty acid metabolism oxidative stress lipid/cholesterol metabolism

thyroid hormone pathway

FXR and LXR cell death steatosis steroid metabolism control

gene symbol

RefSeq accession

description

CYP3A37 CYP1A4 UGT1A9 SULT1B1 SULT1E1 BATF3 IL16 HSP90AB1 PDK4 MT4 TXN ACSL5 HMGCR SLCO1A2 LBFABP CD36 SCD TTR DIO1 THRSP IGF1 NCOA3 CYP7B1 CASP1 LOC100859733 HSD3B1 ALAS1 EEF1A1 RPL4 GGDC RTC PPC

NM_001001751 NM_205147 XM_001234353 NM_204545 NM_420616 XM_419428 NM_204352 NM_206959 NM_001199909 NM_205275 NM_205453 NM_001031237 NM_204485 XM_416421 NM_204634 NM_001030731 NM_204890 NM_205335 NM_001097614 NM_213577 NM_001004384 XM_417385 XM_418276 NM_204924 XM_003641931 NM_205118 NM_001018012 NM_204157 NM_001007479 SA_00517 SA_00104 SA_00103

cytochrome P450 A 37 cytochrome P450 1A4 UDP glucuronosyltransferase 1 family, polypeptide A9 sulfotransferase family, cytosolic, 1B, member 1 sulfotransferase family 1E, estrogen-preferring member1 basic leucine zipper transcription factor, ATF-like 3 interleukin 16 (lymphocyte chemoattractant factor) heat shock 90 kDa protein 1, beta pyruvate dehydrogenase kinase, isozyme 4 metallothionein 4 thioredoxin Acyl-CoA synthetase long-chain family member 5 3-hydroxy-3-methylglutaryl-coenzyme A reductase solute carrier organic anion transporter family, member 1A2 fatty acid binding protein 1, liver CD36 molecule (thrombospondin receptor) stearoyl-CoA desaturase (delta-9-desaturase) transthyretin deiodinase, iodothyronine, type I thyroid hormone responsive (SPOT14 homologue, rat) insulin-like growth factor 1 (somatomedin C) nuclear receptor coactivator 3 cytochrome P450, family 7, subfamily B, polypeptide 1 caspase 1, apoptosis-related cysteine peptidase (interleukin 1, beta, convertase) cell death activator CIDE-3-like hydroxy-delta-5-steroid dehydrogenase, 3 beta- and steroid delta-isomerase 1 aminolevulinate, delta, synthase1 eukaryotic translation elongation factor 1 alpha 1 ribosomal protein L4 chicken genomic DNA contamination reverse transcription control positive PCR control

(Ottawa, Canada) and incubated for 19 days at 37.5 °C with 60% relative humidity. At day 19, incubated embryos were euthanized by decapitation, and livers were removed, pooled, and treated with Percoll (GE Healthcare, Little Chalfont, UK) and DNase I (Roche Applied Science, Penzberg, Upper Bavaria, Germany). The resulting cell pellet was suspended in 32 mL of Medium 199 (Life Technologies, Burlington, Canada), supplemented with 1 μg/mL insulin (Sigma-Aldrich) and thyroxine (Sigma-Aldrich), per gram of pellet. Twenty-five μL of the cell suspension was distributed into 48-well plates containing 500 μL of fresh supplemented medium and incubated for 24 h (37.5 °C and 5% CO2) prior to dosing. The final nominal concentration ranges for the NI- and SIBFRs in the medium were 0.01−1.9 μM (NI-TeDB-DiPhOBz), 0.1−25 μM (SI-TeDB-DiPhOBz), 0.01−9 μM (NI-BDE-209), and 0.1−50 μM (SI-BDE-209; cell viability, n = 3; PCR array, n = 3). Cell Viability. The viability of CEH was evaluated by measuring adenosine triphosphate (ATP) using ViaLight Plus kits (Lonza Group Ltd., Basel, Switzerland). Tris(1,3-dichloro2-propyl) phosphate (TDCIPP; 300 μM nominal) was used as a positive control in the present study to determine relative viability of CEH based on a previous study.14 After the initial 24 h incubation, CEHs were exposed to NI-TeDB-DiPhOBz (0.01, 0.1, 1, or 1.9 μM), SI-TeDB-DiPhOBz (0.1, 1, 10, or 25 μM), NI-BDE-209 (0.01, 0.1, 1, or 9 μM), and SI-BDE-209 (0.1, 1, 10, or 50 μM) for 36 h. The culture plates were then

SI-TeDB-DiPhOBz and SI-BDE-209, respectively, and 380 and 1800 μM for NI-TeDB-DiPhOBz and NI-BDE-209, respectively. Stock and serial dilutions in DMSO were administered to CEH such that the DMSO concentration was 0.5% in the aqueous medium. Determination of the Debrominated Byproducts of Sunlight Irradiated TeDB-DiPhOBz and BDE-209. Similar to our previous studies,6 determination of the debrominated products of SI-TeDB-DiPhOBz solutions was carried out using an Agilent 1200 liquid chromatographic (LC) system, coupled with an Agilent 6250A quadrupole-time-of-flight mass spectrometer (Q-TOF)-MS, with atmospheric pressure photoionization in the negative ion mode (APPI(−)). Determination of the debrominated byproducts in the SI-BDE-209 solutions was carried out using an Agilent gas chromatograph (GC) 6890 coupled with a 5973 quadrupole mass spectrometer (MS) detector. The detailed instrumental parameters for LCAPPI(−)-Q-TOF-MS and GC-MS(ECNI) analysis can be found in the Supporting Information. Representative mass chromatograms are illustrated in Figures S2 and S3, which show the TeDB-DiPhOBz/PB-DiPhOBz and BDE-209/PBDE congener patterns, respectively, formed by natural sunlight irradiation. Chicken Embryonic Hepatocyte Assay. Methods for the CEH cell culture have been described previously.13 In brief, 32 white leghorn chicken (Gallus gallus domesticus) eggs were purchased from the Canadian Food Inspection Agency 12041

dx.doi.org/10.1021/es503399r | Environ. Sci. Technol. 2014, 48, 12039−12046

Environmental Science & Technology

Article

March 10, 2014 in Ottawa. TeDB-DiPhOBz was not included in the solvent assessment because of its poor solubility in methanol or n-hexane. The AhR-mediated transcriptional activity of SI-BDE-209 byproducts from the three different organic solvents was compared by using a luciferase reporter gene (LRG) assay. A detailed description of transfection methods and the LRG assay is provided elsewhere.16,17 Briefly, monkey kidney cells (COS7) were transiently transfected with 8 ng of chicken AHR1 expression construct,16 1.55 ng of cormorant ARNT, and 7.5 ng of pGL4-ccCYP1A5 (both were kindly provided by Dr Hisato Iwata, Ehime University),17 0.75 ng of Renilla luciferase vector (phRL-CMV; Promega), and 32.2 ng of salmon sperm DNA (Invitrogen). Transfected cells were incubated for 5 h prior to treatment with DMSO or DMSO solutions of SI-BDE-209 (0.001 to 50 μM). A nominal concentration of 300 nM 2,3,7,8tetrachlorodibenzo-p-dioxin (TCDD) was included as a positive control. The concentration of the stock TCDD solution was determined by isotope dilution following EPA method 1613 (U.S. EPA, 1994) by high-resolution gas chromatography high-resolution mass spectrometry, and its actual concentration was 72.9 μg/mL as described elsewhere.18 Luciferase activity was measured after 20 h of incubation. Data Analysis. For CEH viability assessment, three technical replicates per concentration of each NI-BFR and SIBFR were included. After correction for background (positive control, 300 μM of TDCIPP), the luciferase intensity of each treatment was normalized to a percent response value expressed relative to the response elicited by the DMSO control. The data were fit to a nonlinear regression curve (log(agonist) vs response) by use of GraphPad software (version 5, San Diego, CA). PCR array data analysis was conducted using MxPro v4.10 software (Agilent Technologies, Santa Clara, CA, USA), and the cycle threshold (Ct) was set to 0.1. The fold change of target gene mRNA abundance relative to the vehicle control was calculated using the 2−ΔΔCt method, and significant differences in fold change compared to the DMSO vehicle control were determined using one-way ANOVA. For visualization, nonsignificant fold changes (p > 0.05) and those less than 2 were set to 0 to minimize noise, and the gene expression profiling was performed on R 3.0.2 version using “gplots” package. Data analysis for the LRG assay was described in detail elsewhere.16,19 In brief, luminescence values, expressed as a ratio of firefly luciferase units to Renilla luciferase units, are presented as the percent response relative to a 300 nM TCDD positive control. For each treatment, four LRG curves were generated from data originating from four wells. The visualization of LRG assay data and calculation of median effect concentration (EC50) was conducted by fitting the data to a four parameter logistic model using GraphPad software.

removed from the incubator and kept at room temperature for at least 5 min. The medium was aspirated, and 100 μL of fresh medium was added to each well prior to the addition of 50 μL of cell lysis reagent. An aliquot of 100 μL of cell lysate/medium mixture was transferred to a white walled luminometer plate (Corning Incorporated 3610, Corning, NY, USA). Immediately, 100 μL of ATP Monitoring Reagent (AMR plus) was added to each well, and the plate was incubated for 2 min at room temperature. The intensity of the luciferase luminescence was monitored with a 1-s integrated reading time using the Luminoskan Ascent (Thermo Fisher Scientific, Wilmington, DE, USA). RNA Isolation and cDNA Synthesis. RNA was isolated from CEH (36 h exposure) using RNeasy 96 kits according to the manufacturer’s protocol with a slight modification (Qiagen, Valencia, CA, USA). Here, 50% ethanol in water was used instead of 70% ethanol, and this modification increased RNA yields from CEH similar to previous studies.15 The quality and concentration of extracted RNA was determined using a NanoDrop 2000 (Thermo Scientific, Wilmington, DE, USA), and 200 ng of RNA was reverse-transcribed following the protocols of the QuantiTect Reverse Transcription kit (Qiagen, Valencia, CA, USA). cDNA was placed on ice until PCR array processing. Avian ToxChip PCR Array. The custom chicken RT2 Profiler PCR Array was built by SABiosciences (Qiagen, Valencia, CA, USA) according to our specifications (Table 1). Each 96-well array contained three identical sets of 27 target genes and five control genes, allowing three technical replicates to be screened per plate. The five control genes included two internal control genes, a positive PCR control, a reverse transcription control, and a well to test for genomic DNA contamination. The cDNA was added directly to the RT2 SYBR Green Mastermix (Qiagen, Valencia, CA, USA), and 25 μL of this mixture was added to each well containing a set of primers at preoptimized concentrations. All arrays were run using the Stratagene MX3005P PCR system (Agilent Technologies, Santa Clara, CA, USA) with the following thermal profile: 95 °C for 10 min followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min and ending with a dissociation curve segment of 95 °C for 1 min, 55 °C for 30 s, and 95 °C for 30 s. The SITeDB-DiPhOBz and SI-BDE-209 byproduct samples were assayed at two concentrations, the greatest noncytotoxic dose (cell survival rate >85%) and 1 μM. Their NI-TeDB-DiPhOBz and NI-BDE-209 precursors were tested at their greatest noncytotoxic concentrations. No amplification was observed in the genomic DNA contamination control, and the positive PCR control and RT control met the appropriate quality control guidelines, which ensured the robustness of the observed gene expression profiles from the PCR array. Assessment of Solvent Effects. For the SI-TeDBDiPhOBz and SI-BDE-209 (irradiation in 30% THF/nhexane), and based on the PCR array, mRNA expression levels of the Aryl hydrocarbon receptor (AhR)-responsive CYP1A4 gene were up-regulated 560- to 5200-fold following exposure to the sunlight irradiated flame retardants. This strongly suggested that some AhR-agonists were formed from the photodegradation of TeDB-DiPhOBz or BDE-209. To clarify whether the observed alterations in mRNA levels were not due to interactions of THF and photoproducts, BDE-209 was prepared in three different organic solvents (30% THF/nhexane [same as original exposure], n-hexane, or methanol) and exposed to sunlight for 21 days between February 17 and

■

RESULTS Determination of Debrominated Products of TeDBDiPhOBz and BDE-209. After a 21-day exposure to sunlight irradiation in 30% THF/n-hexane, TeDB-DiPhOBz was depleted to nondetectable concentrations (Figure S2). Br8- to Br11-PB-DiPhOBz homologue groups of congeners were the major debrominated products in the SI-TeDB-DiPhOBz solution, with the Br10-PB-DiPhOBz homologue group showing the greatest APPI(−)-MS-TOF response. GC-MS(ECNI) analysis for 47 PBDE congeners in the SI-BDE-209 solution revealed that there was no detectable BDE-209; however, 18 PBDE congeners were quantifiable, including five hexa-BDEs 12042

dx.doi.org/10.1021/es503399r | Environ. Sci. Technol. 2014, 48, 12039−12046

Environmental Science & Technology

Article

BDE-209. IGF1, TTR, CYP7B1, MT4, LBFABP, and THRSP were consistently down-regulated after exposure to the byproduct mixtures of SI-TeDB-DiPhOBz and SI-BDE-209 and were classified into one subcluster. The other eight genes, CYP3A37, TXN, HSD3B1, LOC100859733, PDK4, UGT1A9, ACSL5, and ALAS1 showed a consistent up-regulation trend following exposure to SI-TeDB-DiPhOBz and SI-BDE-209. The 27 genes on the PCR array were grouped into 11 biological pathways according to their specific functions (Table 1). After exposure to 1 or 25 μM SI-TeDB-DiPhOBz, mRNAs associated with eight of the pathways were affected including phase I and II metabolism (CYP3A37, CYP1A4, and UGT1A9), glucose and fatty acid metabolism (PDK4), oxidative stress (TXN), lipid/cholesterol metabolism (SLCO1A2 and LBFABP), the thyroid hormone pathway (THRSP, IGF1, and NCOA3), cell death (LOC100859733), steatosis (HSD3B1), and steroid metabolism (ALAS1). Genes from 8 of the 11 pathways were also altered by 1 or 10 μM SI-BDE-209, including phase I and II metabolism (CYP3A37, CYP1A4, and UGT1A9), oxidative stress (MT4 and TXN), lipid/cholesterol metabolism (ACSL5 and LBFABP), the thyroid hormone pathway (TTR, THRSP, and IGF1), FXR and LXR (CYP7B1), cell death (LOC100859733), steatosis (HSD3B1), and steroid metabolism (ALAS1). Effects of Irradiation in Different Solvents. As discussed above, mRNA levels of the AhR-responsive CYP1A4 gene were up-regulated 560- to 5200-fold. This finding suggested that AhR-agonists (dioxin-like compounds) were formed from the photodegradation of TeDB-DiPhOBz or BDE-209 by sunlight irradiation. To assess any potential solvent-related effects that could have resulted in a false positive for BFR degradation byproducts, irradiation experiments with BDE-209 were conducted in two additional solvent systems, n-hexane and methanol. A luciferase reporter gene assay was then used to compare the AhR-mediated transcriptional activity of SI-BDE209 irradiated in three different organic solvents. At concentrations ranging from 0.001 to 50 μM, the maximal responses caused by SI-BDE-209 in 30% THF/n-hexane, nhexane, and methanol relative to a 300 nM TCDD positive control were similar, 31 ± 4.8%, 37 ± 3.4%, and 45 ± 6.5%, respectively (Figure 3). On the basis of the fitted curves, the EC50 values for the SI-BDE-209 byproducts were 3.8 ± 1.1, 4.4 ± 0.3, and 2.6 ± 0.5 μM in 30% THF/n-hexane, n-hexane, and methanol, respectively. These results clearly demonstrated that the observed AHR-mediated reporter gene activity was altered by photodegradation byproducts of BDE-209 and not solvent reaction products.

(BDE-138, BDE-139, BDE-140, BDE-153, BDE-154), eight hepta-BDEs (BDE-171, BDE-180, BDE-181, BDE-183, BDE184, BDE-188, BDE-190, BDE-191), and five octa-BDEs (BDE-196, BDE-197, BDE-201, BDE-202, BDE-203; Figure S3). On the basis of the total ion MS chromatogram of SI-BDE209, down to the earliest retention times of 5 to 10 min, there were also numerous peaks representing compounds containing bromide anions, which were not BDE congeners (Figure S3). Cell Viability. Any form of cell injury results in a rapid decrease in ATP concentrations in the cytoplasm, but none of the tested concentrations of NI-TeDB-DiPhOBz or NI-BDE209 caused significant (one way ANOVA; Dunnett’s p < 0.05) cytotoxic effects to the CEH relative to the DMSO solvent control (Figure S4). Extensive cytotoxic effects were observed following exposure to 50 μM SI-BDE-209 with a cell survival percentage of 3.3 ± 0.21%. On the basis of a fitted curve, the inhibitory concentration at 50% (IC50) of the byproduct mixture in the SI-BDE-209 was 26 ± 3.1 μM (Figure 1).

Figure 1. Concentration-dependent effects of sunlight irradiated BDE209 byproducts (SI-BDE-209) on cell viability of chicken embryonic hepatocytes following 36 h of exposure. Data were log-transformed and the percent viability data (n = 3 wells per treatment group) were fit to a nonlinear regression curve (log(agonist) vs response) to determine the IC50 (±SEM). Error bars for each point represent the SD.

Although significant differences (p < 0.05) were also observed for SI-TeDB-DiPhOBz at nominal concentrations of 10 and 25 μM relative to DMSO control, the % CEH viability compared to the DMSO control was 92.7 ± 1.3% and 88.9 ± 2.9%, respectively. Avian ToxChip PCR Array. After exposure to 1.9 μM NITeDB-DiPhOBz or 9 μM NI-BDE-209, mRNA expression trends of the 27 genes in CEH were different than with 1 and 25 μM SI-TeDB-DiPhOBz and 1 and 10 μM SI-BDE-209, respectively (Figure 2, Table S1). Only one gene, UGT1A9, was up-regulated (fold-change = 2.7) in CEH exposed to 1.9 μM NI-TeDB-DiPhOBz, and one gene, CYP1A4, was upregulated by 9 μM of NI-BDE-209 with a fold-change of 2.6 (Figure 2). In contrast, the number of altered genes was 12 and 14 at the greatest concentration of the photolytically degraded SI-TeDB-DiPhOBz and SI-BDE-209 mixtures, respectively. At 1 μM of each mixture, only 2 and 5 genes were altered significantly. Hierarchical clustering of the 27 genes was conducted based on mRNA expression levels, and one gene, CYP1A4, was separated into an independent group from all other genes due to its extremely large fold-change, ranging from 560- to 5200-fold after exposure to SI-TeDB-DiPhOBz and SI-

■

DISCUSSION The solvent mixture (THF/n-hexane) used in this study was also used previously in studies that established that TeDBDiPhOBz and BDE-209 photolytically debrominate.6,20 In previous studies,6 1% and 10% THF was spiked in methanol or n-hexane as the solvent system for sunlight irradiation of TeDBDiPhOBz and BDE-209. The greater percentage (30%) of THF was used in the present study in order to dissolve a greater amount of TeDB-DiPhOBz and generate more concentrated degradation byproducts for in vitro CEH exposures. In this study, reductive debromination of TeDB-DiPhOBz was observed after sunlight irradiation, findings similar to those from our previous study using a different solvent system;6 however, different debrominated byproducts were generated. For example, after irradiation with sunlight, photolytic 12043

dx.doi.org/10.1021/es503399r | Environ. Sci. Technol. 2014, 48, 12039−12046

Environmental Science & Technology

Article

Figure 2. Transcriptional profiles of 27 target genes on the Avian ToxChip PCR array following exposure to 1.9 μM TeDB-DiPhOBz, 25 μM SITeDB-DiPhOBz, 1 μM SI-TeDB-DiPhOBz, 9 μM BDE-209, 10 μM SI-BDE-209, and 1 μM SI-BDE-209. Hierarchical clustering was conducted based on mRNA expression fold-changes derived from a mean of three replicates. Genes with a p > 0.05 or fold-change lower than 2 were set to 0 to minimize noise. Blue, white, and yellow represent down-, no, and up-regulation. The raw mRNA fold-change data are available in Table S1 for all treatment groups.

21 day exposure period, generally the debromination products were more brominated, and specifically the Br8- to Br11-PBDiPhOBz homologue groups of congeners. This is likely due in part to the season during which the natural sunlight exposures were conducted and the ambient temperature during sunlight irradiation. The irradiation was performed during the winter in Ottawa in this study, whereas the previous irradiation period was during the summer in Ottawa, meaning substantially longer daylight hours and higher ambient temperatures. The initial concentrations of the BFRs might have also played a role in the TeDB-DiPhOBz debromination. For example, the initial concentration of TeDB-DiPhOBz in this study was 300 μM, which is much greater than the concentration used previously.6 Studies of BDE-209 photolysis have been conducted in organic solvents such as THF/methanol/water mixtures20 and toluene,9 as well as in house dust8 and sediment.9 The time course of BDE-209 photolytic reactions varied among matrices, but the different matrices had little effect on ultimate degradation.9 The debromination pathway of BDE-209 begins with the loss of one bromine atom, relative to the ether linkage between the two phenyl groups, which results in the formation of all three nona-BDEs (BDE-206, -207, and -208) followed by subsequent formation of octa-BDEs and so on.8,9,20 In the present study, hexa-BDEs were observed; however, BDE-209 and nona-BDE congeners were not detected, indicating their complete degradation (Figure S3) after a 21 day period of sunlight irradiation in 30% THF/n-hexane. CEH viability was affected after exposure to 50 μM SI-BDE209, whereas NI-BDE-209 (0.01−9 μM), NI-TeDB-DiPhOBz (0.01−1.9 μM), or SI-TeDB-DiPhOBz (0.1−25 μM) did not alter CEH viability (Figure S4). To our knowledge, this is the

Figure 3. Concentration-dependent effects of sunlight irradiated BDE209 byproducts from the three different organic solvents (30% THF/ n-hexane, n-hexane, and methanol) on AHR1-mediated luciferase reporter gene activity in COS-7 cells. A full-length chicken AHR1 construct was transfected into COS-7 cells with cormorant ARNT, CYP1A5 reporter construct, and Renilla luciferase reporter vector. Data are presented as the percent response relative to a 300 nM TCDD positive control. Each data point represents the mean positive control-normalized luciferase ratio (firefly luminescence/Renilla luminescence) determined from four wells; bars represent SE of the mean values. Raw luciferase activity data are provided in Table S2.

byproducts of TeDB-DiPhOBz contained numbers of bromine atoms ranging from three to eight in the Br3- to Br8-DiPhOBz homologue groups, whereas Br4- to Br11-DiPhOBz and Br4- to Br7-DiPhOBz homologue groups were generated previously using THF, methanol (1% THF), or n-hexane (1% THF) solvents systems.6 In the present study, however, after the same 12044

dx.doi.org/10.1021/es503399r | Environ. Sci. Technol. 2014, 48, 12039−12046

Environmental Science & Technology

Article

first report on the in vitro cytotoxicity of TeDB-DiPhOBz, SITeDB-DiPhOBz, and SI-BDE-209 on any cell system. Previous reports have mainly focused on cytotoxicity of BDE-209 itself. For example, the relative growth rates of two microalgae specimens (Heterosigma akashiwo and Kareenia mikimotoi) decreased dramatically with increasing concentrations of BDE209 with 96 h-EC50 values of 23 and 120 mg/L, respectively.21 When RTG-2 cells were exposed for up to 72 h to concentrations of BDE-209 ranging from 12.5 to 100 μM, cell viability was inhibited in a time- and concentrationdependent manner.22 BDE-209 (10−100 μM) exposure also inhibited the growth of human hepatoma cells.23 In the present study, the highest NI-BDE-209 exposure concentration of 9 μM did not elicit any significant (p > 0.05) cytotoxic effects and was lower than exposure concentrations used in the aforementioned studies with other cell systems. Among the altered genes identified in CEH using the PCR array, IGF1, ALAS1, and CYP1A4 were of interest due to the magnitude of their fold changes (i.e., >10) following exposure to SI-TeDB-DiPhOBz or SI-BDE-209. IGF1 is associated with the thyroid hormone pathway, which is critically important to normal central nervous system development, growth, and metabolism in avian species.24 Previous studies reported the down-regulation of IGF1 following in ovo exposure of chicken embryos to technical hexabromocyclododecane (HBCDD)25 and in vitro exposure of CEH to five other organic flame retardants.13 ALAS1 controls the production of delta-aminolevulinate synthase 1 or ALA-synthase, which is the first rate controlling enzyme that controls cellular heme biosynthesis.26 Heme is essential in oxygen transport and metabolism in living systems, and ALAS1 expression is increased in vivo by xenobiotic chemical inducers of cytochrome P450 hemoproteins through mechanisms that are poorly understood.27 Expression of CYP1A4 is regulated by the AhR, which is the main receptor that interacts and binds to TCDD and other dioxin-like chemicals.28 Induction of CYP1A4 mRNA ranged from 560- to 5200-fold following exposure of CEHs to photolytic byproducts of SI-TeDB-DiPhOBz and SI-BDE-209. This level of induction was much greater than that observed for 16 organic flame retardants screened in a previous study.13 In contrast to the significant up-regulation of CYP1A4 elicited by SI-TeDB-DiPhOBz and SI-BDE-209, NI-TeDB-DiPhOBz and NI-BDE-209 had minimal to no effect. The induction of CYP1A4 observed following SI-BFR exposure is similar to the induction of hepatic cytochrome P450-dependent arachidonic acid epoxygenation in diverse avian orders by TCDD.29 This strongly suggests that dioxin-like compounds are formed photolytically from TeDB-DiPhOBz and BDE-209 as a result of natural sunlight exposure. Toxicological information regarding TeDB-DiPhOBz or its degradation byproducts is currently not available. TeDB-DiPhOBz itself contains 14 bromine atoms, making it relatively heavy and nonvolatile and with less potential for bioaccumulation in the environment; however, it is also found to undergo rapid photolysis following exposure to UV radiation or natural sunlight.6 Interestingly, a very recent study from our lab found that TeDB-DiPhOBz and its debrominated products comprising four homologue groups, Br10- to Br13-PB-DiPhOBz, were not detectable in any surficial sediment samples from several sites in Lakes Huron and Erie. The sampling sites included one near the mouth of the highly FR-contaminated Saginaw River, near the confined disposal facility (CDF) located in Saginaw Bay at Channel-Shelter

Island, which receives dredged sediment from the Saginaw River.30 In the present study, photolysis of TeDB-DiPhOBz resulted in products that activated the AhR pathway evidenced by the significant up-regulation of CYP1A4 mRNA, suggesting that more concern should be focused on the degraded byproducts rather than TeDB-DiPhOBz itself. For SI-BDE-209, 18 lesserbrominated PBDE congeners were quantified as a result of photolysis of BDE-209 (Figure S3). PBDEs have been reported to interact with the AhR but do not bind with sufficient affinity to initiate AhR-mediated signaling and gene expression31 and might not elicit the observed up-regulation of CYP1A4 mRNA. Previous studies demonstrated the formation of polybrominated dibenzofurans (PBDFs) in sand, sediment, soil, or plastics spiked with BDE-209 after UV/sunlight irradiation.9,32 However, PBDFs were only detected in some of the samples, and some researchers also suggested that the PBDFs were probably quickly degraded.9 Regardless, based on the findings of the present study, PBDFs, which are AhR agonists, may be among the photodegradation byproducts of BDE-209 that led to such large mRNA alterations of CYP1A4 in CEH and warrant further investigation.

■

ASSOCIATED CONTENT

* Supporting Information S

Further details on the analytical methods and additional tables and figures as noted in the text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Tel.: 1-613-998-6696. Fax: 1-613-998-0458. E-mail: Robert. [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Environment Canada’s Chemicals Management Plan (CMP) (to R.J.L., S.W.K., D.C.) provided major funding for this project. Supplemental funding was from the Natural Science and Engineering Research Council (NSERC) of Canada (to R.J.L. and J.P.G.). Prof. Giesy was supported by the Canada Research Chair program, a Visiting Distinguished Professorship (Dept. of Biology and Chemistry, State Key Laboratory in Marine Pollution, City University of Hong Kong), the 2012 “High Level Foreign Experts” (#GDW20123200120) program funded by the State Administration of Foreign Experts Affairs, the P.R. China to Nanjing University, and the Einstein Professor Program (Chinese Academy of Sciences). We would like to thank Suzanne Chiu and Dr. Shaogang Chu at NWRC for their technical support.

■

REFERENCES

(1) Betts, K. S. Rapidly rising PBDE levels in North America. Environ. Sci. Technol. 2002, 36 (3), 50A−52A. (2) U.S. PBDE milestones:. Environ. Sci. Technol. 2003, 37 (21), 384A. (3) Stockholm Convention, The new POPs under the Stockholm Convention, 2009 website: http://chm.pops.int/Convention/ ThePOPs/TheNewPOPs/tabid/2511/Default.aspx. (4) Chemical Register Website; http://www.chemicalregister.com/ Tetradecabromodiphenoxy_Benzene/Suppliers/pid58673.htm.

12045

dx.doi.org/10.1021/es503399r | Environ. Sci. Technol. 2014, 48, 12039−12046

Environmental Science & Technology

Article

(5) Chen, D.; Letcher, R. J.; Gauthier, L. T.; Chu, S.; McCrindle, R. Newly discovered methoxylated polybrominated diphenoxybenzenes have been contaminants in the Great Lakes herring gull eggs for thirty years. Environ. Sci. Technol. 2012, 46 (17), 9456−63. (6) Chen, D.; Letcher, R. J.; Gauthier, L. T.; Chu, S. Tetradecabromodiphenoxybenzene flame retardant undergoes photolytic debromination. Environ. Sci. Technol. 2013, 47 (3), 1373−80. (7) Hua, I.; Kang, N.; Jafvert, C. T.; Fabrega-Duque, J. R. Heterogeneous photochemical reactions of decabromodiphenyl ether. Environ. Toxicol. Chem. 2003, 22 (4), 798−804. (8) Stapleton, H. M.; Dodder, N. G. Photodegradation of decabromodiphenyl ether in house dust by natural sunlight. Environ. Toxicol. Chem. 2008, 27 (2), 306−12. (9) Soderstrom, G.; Sellstrom, U.; de Wit, C. A.; Tysklind, M. Photolytic debromination of decabromodiphenyl ether (BDE 209). Environ. Sci. Technol. 2004, 38 (1), 127−32. (10) Letcher, R. J.; Marteinson, S. C.; Fernie, K. J. Dietary exposure of American kestrels (Falco sparverius) to decabromodiphenyl ether (BDE-209) flame retardant: uptake, distribution, debromination and cytochrome P450 enzyme induction. Environ. Int. 2014, 63, 182−90. (11) Van den Steen, E.; Covaci, A.; Jaspers, V. L.; Dauwe, T.; Voorspoels, S.; Eens, M.; Pinxten, R. Accumulation, tissue-specific distribution and debromination of decabromodiphenyl ether (BDE 209) in European starlings (Sturnus vulgaris). Environ. Pollut. 2007, 148 (2), 648−53. (12) Arikawa, E.; Sun, Y.; Wang, J.; Zhou, Q.; Ning, B.; Dial, S. L.; Guo, L.; Yang, J. Cross-platform comparison of SYBR Green real-time PCR with TaqMan PCR, microarrays and other gene expression measurement technologies evaluated in the MicroArray Quality Control (MAQC) study. BMC Genomics 2008, 9, 328. (13) Porter, E.; Crump, D.; Egloff, C.; Chiu, S.; Kennedy, S. W. Use of an avian hepatocyte assay and the avian toxchip polymerse chain reaction array for testing prioritization of 16 organic flame retardants. Environ. Toxicol. Chem. 2013, 33, 573−582. (14) Crump, D.; Chiu, S.; Kennedy, S. W. Effects of tris(1,3-dichloro2-propyl) phosphate and tris(1-chloropropyl) phosphate on cytotoxicity and mRNA expression in primary cultures of avian hepatocytes and neuronal cells. Toxicol. Sci. 2012, 126 (1), 140−8. (15) Head, J. A.; Kennedy, S. W. Differential expression, induction, and stability of CYP1A4 and CYP1A5 mRNA in chicken and herring gull embryo hepatocytes. Comp. Biochem. Physiol., Part C: Toxicol. Pharmacol. 2007, 145 (4), 617−24. (16) Farmahin, R.; Wu, D.; Crump, D.; Herve, J. C.; Jones, S. P.; Hahn, M. E.; Karchner, S. I.; Giesy, J. P.; Bursian, S. J.; Zwiernik, M. J.; Kennedy, S. W. Sequence and in vitro function of chicken, ring-necked pheasant, and Japanese quail AHR1 predict in vivo sensitivity to dioxins. Environ. Sci. Technol. 2012, 46 (5), 2967−75. (17) Lee, J. S.; Kim, E. Y.; Iwata, H. Dioxin activation of CYP1A5 promoter/enhancer regions from two avian species, common cormorant (Phalacrocorax carbo) and chicken (Gallus gallus): association with aryl hydrocarbon receptor 1 and 2 isoforms. Toxicol. Appl. Pharmacol. 2009, 234, 1−13. (18) Herve, J. C.; Crump, D.; Jones, S. P.; Mundy, L. J.; Giesy, J. P.; Zwiernik, M. J.; Bursian, S. J.; Jones, P. D.; Wiseman, S. B.; Wan, Y.; Kennedy, S. W. Cytochrome P4501A induction by 2,3,7,8-tetrachlorodibenzo-p-dioxin and two chlorinated dibenzofurans in primary hepatocyte cultures of three avian species. Toxicol. Sci. 2010, 113 (2), 380−91. (19) Zhang, R.; Manning, G. E.; Farmahin, R.; Crump, D.; Zhang, X.; Kennedy, S. W. Relative potencies of aroclor mixtures derived from avian in vitro bioassays: comparisons with calculated toxic equivalents. Environ. Sci. Technol. 2013, 47 (15), 8852−61. (20) Eriksson, J.; Green, N.; Marsh, G.; Bergman, A. Photochemical decomposition of 15 polybrominated diphenyl ether congeners in methanol/water. Environ. Sci. Technol. 2004, 38 (11), 3119−25. (21) Zhang, X.; Tang, X.; Zhou, B.; Wang, Y. Effects of Decabromodiphenyl Ether (BDE-209) on Inter-Specific Competition between Two Species of Marine Bloom-Forming Microalgae. PloS one 2013, 8 (3), e56084 DOI: 10.1371/journal.pone.0056084.

(22) Jin, S.; Yang, F.; Hui, Y.; Xu, Y.; Lu, Y.; Liu, J. Cytotoxicity and apoptosis induction on RTG-2 cells of 2,2′,4,4′-tetrabromodiphenyl ether (BDE-47) and decabrominated diphenyl ether (BDE-209). Toxicol. In Vitro 2010, 24 (4), 1190−6. (23) Hu, X. Z.; Xu, Y.; Hu, D. C.; Hui, Y.; Yang, F. X. Apoptosis induction on human hepatoma cells Hep G2 of decabrominated diphenyl ether (PBDE-209). Toxicol. Lett. 2007, 171 (1−2), 19−28. (24) Maglich, J. M.; Watson, J.; McMillen, P. J.; Goodwin, B.; Willson, T. M.; Moore, J. T. The nuclear receptor CAR is a regulator of thyroid hormone metabolism during caloric restriction. J. Biol. Chem. 2004, 279 (19), 19832−8. (25) Crump, D.; Egloff, C.; Chiu, S.; Letcher, R. J.; Chu, S.; Kennedy, S. W. Pipping success, isomer-specific accumulation, and hepatic mRNA expression in chicken embryos exposed to HBCD. Toxicol. Sci. 2010, 115 (2), 492−500. (26) Tian, Q.; Li, T.; Hou, W.; Zheng, J.; Schrum, L. W.; Bonkovsky, H. L. Lon peptidase 1 (LONP1)-dependent breakdown of mitochondrial 5-aminolevulinic acid synthase protein by heme in human liver cells. J. Biol. Chem. 2011, 286 (30), 26424−30. (27) Fraser, D. J.; Podvinec, M.; Kaufmann, M. R.; Meyer, U. A. Drugs mediate the transcriptional activation of the 5-aminolevulinic acid synthase (ALAS1) gene via the chicken xenobiotic-sensing nuclear receptor (CXR). J. Biol. Chem. 2002, 277 (38), 34717−26. (28) Denison, M. S.; Nagy, S. R. Activation of the aryl hydrocarbon receptor by structurally diverse exogenous and endogenous chemicals. Annu. Rev. Pharmacol. Toxicol. 2003, 43, 309−34. (29) Gilday, D.; Bellward, G. D.; Sanderson, J. T.; Janz, D. M.; Rifkind, A. B. 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) induces hepatic cytochrome P450-dependent arachidonic acid epoxygenation in diverse avian orders: regioisomer selectivity and immunochemical comparison of the TCDD-induced P450s to CYP1A4 and 1A5. Toxicol. Appl. Pharmacol. 1998, 150 (1), 106−16. (30) Trouborst, L.; Chu, S.; Chen, D.; Letcher, R. J. Methodology and Determination 1 of Tetradecabromo-1,4-diphenoxybenzene Flame Retardant and Breakdown By-Products in Sediments from the Laurentian Great Lakes. Chemosphere accepted. (31) Oberg, M.; Westerholm, E.; Fattore, E.; Stern, N.; Hanberg, A.; Haglund, P.; Wiberg, K.; Bergendorff, A.; Hakansson, H. Toxicity of Bromkal 70−5DE, a technical mixture of polybrominated diphenyl ethers, following 28 d of oral exposure in rats and impact of analysed impurities. Chemosphere 2010, 80 (2), 137−43. (32) Kajiwara, N.; Noma, Y.; Takigami, H. Photolysis studies of technical decabromodiphenyl ether (DecaBDE) and ethane (DeBDethane) in plastics under natural sunlight. Environ. Sci. Technol. 2008, 42 (12), 4404−9.

12046

dx.doi.org/10.1021/es503399r | Environ. Sci. Technol. 2014, 48, 12039−12046