Article pubs.acs.org/est
Photochemical Formation of Polybrominated Dibenzo-p-dioxins from Environmentally Abundant Hydroxylated Polybrominated Diphenyl Ethers Kristina Arnoldsson, Patrik L. Andersson, and Peter Haglund* Department of Chemistry, Umeå University, S-901 87 Umeå, Sweden S Supporting Information *
ABSTRACT: High levels of polybrominated dibenzo-p-dioxins (PBDDs) have been found in Baltic Sea biota, where the toxic load owing to, for example, polychlorinated dibenzo-p-dioxins and other organic pollutants is already high. The levels and geographic pattern of PBDDs suggest biogenic rather than anthropogenic origin, and both biotic and abiotic formation pathways have been proposed. Photochemical formation from hydroxylated polybrominated diphenyl ethers (OH-PBDE) is a proposed pathway for PBDDs in marine environments. Ultraviolet radiation-initiated transformations of OH-BDEs 47, 68, 85, 90, 99, and 123, which all are abundant in the environment, were investigated. It was shown that the most abundant PBDDs in the environment (1,3,7-triBDD and 1,3,8triBDD) can be formed from the most abundant OH-BDEs (OH-BDE 47 and OH-BDE 68) at high rates and with percentage yields. In fact, most of the PBDDs that have been identified in the Baltic Sea environment were formed with high yield from the six studied OH-PBDE, through initial cyclization and subsequent debromination reactions. The high formation yields point to this route as an important source of PBDDs in biota. However, congeners showing relatively high retention in fish, specifically 1,3,6,8- and 1,3,7,9-tetraBDD, were not formed. These are likely formed by enzymatic coupling of brominated phenols.
■
of PBDDs.13 These are found at substantial levels in marine organisms, and it has been suggested that they are primarily from natural production, 14,15 although formation from anthropogenic PBDEs have also been proposed: via metabolism,16−18 via atmospheric hydroxyl radical reactions,19,20 and via oxidation in sewage treatment plants.20,21 Photochemical formation of OH-PBDEs from 2,4-bromophenol has also been reported.22 The reported levels of OH-PBDEs in Baltic biota are in the range 12−25 μg/g extractable organic material for red alga (Ceramium tenuicorne) and 5−260 ng/g lipid for blue mussel (Mytilus edilus).3,5 The most abundant congener in red alga was 6-hydroxy-2,2′,4,4′,5-pentaBDE (OH-BDE 99) at 4.4 μg/g, while 6-hydroxy-2,2′,4,4′-tetraBDE (OH-BDE 47) predominated in mussel at 70 ng/g lipid. For comparison, the concentrations of PBDDs reported from the same area and species were about 45 ng/g for alga and 2−360 ng/g for mussel.3,5 Blue mussel would reflect the water composition, as the two classes of pollutants have similar lipophilicity and the filter-feeding mussels have low metabolic capacity. A trans-
INTRODUCTION Polybrominated dibenzo-p-dioxins (PBDDs) are a group of compounds of emerging interest as potential environmental stressors.1−6 This interest is due to a toxic response, mediated via the aryl hydrocarbon (Ah)-receptor, that is similar to that of the polychlorinated dibenzo-p-dioxins (PCDDs).7−10 Levels of PBDDs as high as 4 ng/g wet weight have been found in mussels from a remote area in the Baltic Sea, and levels of PBDDs in coastal fish generally exceeds those of PCDDs.1 This is of particular concern, as the total toxic equivalency concentrations of PCDDs in certain fish species in the Baltic Sea frequently exceed the maximum residue levels for PCDDs and polychlorinated dibenzofurans (PCDFs) (4 pg/g fresh weigh) set by the European Commission.11 To establish exposure pathways, identify possible ecological effects, and take actions to minimize environmental impact, identification of the sources of pollutants is essential. A predominantly natural origin for PBDDs in the Baltic Sea have been proposed based on their temporal and spatial distribution, and both biotic and photochemical pathways have been suggested.1,2,6 Enzymatic formation of PBDDs through the action of bromoperoxidase on 2,4,6-tribromophenol, abundant in the marine environment, has recently been shown.12 Hydroxylated polybrominated diphenyl ethers (OH-PBDEs) have been proposed as precursors for photochemical formation © 2012 American Chemical Society
Received: Revised: Accepted: Published: 7567
March 30, 2012 June 18, 2012 June 21, 2012 June 21, 2012 dx.doi.org/10.1021/es301256x | Environ. Sci. Technol. 2012, 46, 7567−7574
Environmental Science & Technology
Article
experiments. The OH-PBDEs were dissolved in MeOH and added to the media to reach 10 μM (final MeOH level approximately 0.5%). No precipitate was observed despite low water solubility, likely due to the high dissociation degree of the OH-PBDEs. Spectra were collected between 220 and 400 nm (Supporting Information, Figure S2a). Preparation of Artificial Coast Water. For details, see Supporting Information. Artificial seawater of 35‰ salinity was prepared according to Kester et al.,33 with the exception that hydrogen borate and strontium chloride were excluded. The artificial seawater was then diluted and sodium hydrogen carbonate was added, to obtain a salinity of 6.3‰ and pH of 8.4 (after aeration) to reflect Baltic Sea conditions (data from SHARK34). Hereafter, this solution is termed artificial coast water (ACW). For some of the experiments, fulvic acid (Nordic Lake Reference Fulvic Acid35) was added at one of two levels: 5 mg/L DOC (DOC 5) and 10 mg/L DOC (DOC 10), to represent open sea and coastal conditions, respectively.34,36 OH-PBDE Irradiation Experiments. UV exposure experiments were performed both in the laboratory with artificial UV light (experiments 1−16) and under natural solar conditions outdoors (experiment 17; Table 1). For irradiation setup details, see Supporting Information.
formation of OH-PBDEs to PBDDs, even at low yields, could therefore theoretically result in the detected levels of PBDDs. Brominated compounds, such as PBDEs and OH-PBDEs, are susceptible to reaction via ultraviolet (UV)-irradiation because the bromine−carbon bond is weak. The main process is debromination to lower brominated congeners,23−26 but other reactions, such as cyclization, are possible.13,27 Intramolecular coupling to form halogenated dibenzo-p-dioxins is possible when the diphenyl ether is substituted with an o-hydroxyl group and has a halogen substituent in ortho position of the nonhydroxylated phenyl ring. Triclosan, the environmentally abundant OH-BDE 47, and their chlorinated derivatives, have been shown to cyclize to the corresponding halogenated dioxin after UV−irradiation.13,28,29 All reported OH-PBDEs of natural origin are substituted with the hydroxyl group in the ortho position and with bromine substitution in the 2′- or 2′,4′positions of the nonphenolic ring.30,31 Thereby, the prerequisites for photochemical-induced intramolecular coupling to form PBDDs are fulfilled. The present study aimed at assessing the importance of photochemical transformation as a route of PBDD formation. The process was studied under conditions mimicking the Baltic Sea environment, using OH-PBDEs that all have been reported to be abundant in Baltic Sea algae, shellfish, and fish.30 The PBDD product pattern of selected OH-PBDEs over time was investigated and compared to the PBDD congener profiles found in biota. The influence on PBDD product formation due to light absorption by dissolved organic carbon (DOC) was explored because the DOC concentrations differ significantly between the different basins of the Baltic Sea, and recent climate change scenario models indicate that the levels of allochtonous (terrestrial) DOC will increase in the future. Finally, the experimental setup was verified by comparing the PBDD product patterns obtained using artificial and natural sunlight. The environmental relevance of UV-initiated transformations as a formation pathway vs enzymatic coupling of phenolic compounds for the occurrence of PBDDs in Baltic Sea biota was considered.
Table 1. Design of Experiments Including Substrate and Mixtures of Substrates, Matrix Composition, and Exposure Timea experiment
substrate
Artificial UV Light 1 OH-BDE 47 2 OH-BDE 68 3 OH-BDE 85 4 OH-BDE 90 5 OH-BDE 99 6 OH-BDE 123 7 mix 47 + 90c 8 mix 47 + 90 9 mix 47 + 90 10 mix 68 + 85f 11 mix 68 + 85 12 mix 68 + 85 13 mix 47 + 90 14 mix 68 + 85 15 mix 47 + 90 16 mix 68 + 85 Solar Light 17 mix 47 + 90
■
MATERIALS AND METHODS Chemicals. The following OH-PBDE congeners were used in the present study: 6-OH-2,2′,4,4′-BDE (OH-BDE 47), 2′OH-2,3′,4,5′-BDE (OH-BDE 68), 6-OH-2,2′,3,4,4′-BDE (OHBDE 85), 6-OH-2,2′,3,4′,5-BDE (OH-BDE 90), 6-OH2,2′4,4′,5-BDE (OH-BDE 99), and 2-OH-2′,3,4,4′,5-BDE (OH-BDE 123). The structures are shown in Supporting Information, Figure S1. The OH-PBDEs were a gift from Dr. G. Marsh, Stockholm University.31 The other chemicals used are described in Supporting Information. Characterization of OH-PBDEs. The absorption spectra of OH-PBDEs are influenced by the proportion of phenol to phenolate. The phenolate has a shift in absorptivity toward longer wavelengths because of electron delocalization and, thus, will have a greater overlap with the UV irradiation spectrum. The pKa values of the OH-PBDEs were calculated32 and ranged from 5.8 to 7.2 for the studied compounds (Supporting Information, Table S1). Experiments were performed at pH 8.4 (reflecting the Baltic Sea conditions) where the OH-PBDEs will mostly be in the phenolate form, having a dissociation degree of 93.8% to 99.7% (Supporting Information, Table S1). UV spectra of individual OH-PBDEs were measured in artificial coast water, pH 8.4, to find out the OH-PBDE spectra overlap with the sunlight and artificial sunlight emissions used in the
matrix
exposure
ACWb ACW ACW ACW ACW ACW ACW DOC 5d DOC10e ACW DOC 5 DOC 10 ACW ACW ACW ACW
2h 2h 2h 2h 2h 2h 4h 4h 4h 4h 4h 4h 9h 9h 18 h 18 h
ACW
3 h/9 h
a
Fulvic acid (Nordic Lake Reference Fulvic Acid35) was used for addition of organic carbon (DOC). bArtificial coast water, 6.3‰ salinity and pH 8.4. cEquimolar mixture of OH-BDEs 47 and 90. d Dissolved organic carbon (DOC) at 5 mg C/L ACW. eDOC at 10 mg C/L ACW. fEquimolar mixture of OH-BDEs 68 and 85.
In the first set of experiments (1−6; Table 1), all six OHPBDEs were irradiated as single components to find out which PBDDs were formed from each OH-PBDE. As the product patterns were different depending on congener, the number of experiments were reduced in the second set of experiments (7− 16) by using mixtures of OH-PBDEs with different PBDD product outcome, viz., OH-BDEs 47 and 90, and OH-BDEs 68 and 85, respectively. Experiments 7−12 were performed with ACW, DOC 5, or DOC 10 to study the influence of DOC on 7568
dx.doi.org/10.1021/es301256x | Environ. Sci. Technol. 2012, 46, 7567−7574
Environmental Science & Technology
Article
Figure 1. PBDD product yields in mmol/mol starting material for (a) mixtures of hydroxylated polybrominated diphenyl ether (OH-BDE) 47 and OH-BDE 90, and (b) mixtures of OH-BDE 68 and OH-BDE 85, after 4 h of irradiation. The OH-PBDEs were dissolved in artificial coastal water (ACW) with dissolved organic carbon (DOC) addition of 0 mg C/L (ACW; white bars), 5 mg C/L (DOC5; striped bars), or 10 mg C/L (DOC10; cross-hatched bars). Starting concentration of each OH-PBDE congener was 330 nM. Error bars represent ±1 SD (n = 3). *, **, and *** denote significant differences between DOC5/10 and ACW (p < 0.1, p < 0.05, and p < 0.01 respectively).
homologue group were detected, and the isotopic ratios of the ions for each group were calculated. For a positive identification, the ratio had to be within 25% of the theoretical value and the GC retention time had to be within 0.1 min of the corresponding native standard. The following congeners lack native standards and were tentatively identified: 1,3-/1,7-/ 1,8-DBDD and 1,4,7-/1,2,7-/1,2,8-TrBDD according to Haglund.38 The attribution of the coeluting 1,2,4,7- and 1,2,4,8- TeBDD to OH-BDEs 90 and 99, respectively, was based on the expected cyclization products and the identity of the primary debromination products. The quantification was performed using the internal standard method, and the results were expressed as mmol/mol starting material. Recoveries of the internal standards were generally good, ranging between 71% and 105%. For the tentatively identified congeners, the quantification was performed using 2,7-/2,8-DBDD and 1,3,8-TrBDD, assuming the same response factors for isomers (the response factors of available standards differ no more than a factor of 2). Mono-BDDs were detected but not reported, as evaporative losses during workup procedure cannot be excluded. Quantification of the individual congeners of 2,7- and 2,8-DBDD was impossible in the mixture experiments owing to poor chromatographic resolution. For the single compound experiments where either isomer was formed, attribution was not a problem because the apparent resolution was 0.5. In the evaluation of the time-trend experiments where both congeners were formed, the total 2,7/2,8-yield was attributed to each OH-PBDE congener based on the relative yields in the single-compound experiments. Blank levels were always below the level of detection. Some control samples exhibited levels of PBDDs at a few μmol/mol. Because the contributions were less than 1% of the levels in the samples, they remained uncorrected.
the PBDD yields and patterns. The experiments were performed in triplicate. Two of these experiments (7 and 10) were also used together with experiments 13−16 to investigate how the PBDD product patterns evolve over time. A final control experiment (exp. 17) was performed using natural sunlight to ensure that natural and artificial light produce similar PBDD yields and product patterns. Details of the experimental setup are presented in the Supporting Information. For each experiment, 10 mL of the selected aqueous media (ACW, DOC 5, or DOC 10) was added to a Petri dish (Duran glass), giving a media height of 5 mm. The OH-PBDEs were dissolved in methanol and a portion (∼50 μL) added to the vessel to obtain 330 nM final concentrations. The addition was performed under the water surface to avoid crystal formation on the surface due to methanol evaporation; thereafter, they were covered with a Duran glass lid (UV cutoff approximately 305 nm) and transferred to a shaking table (∼60 rpm) in the irradiation chamber. For the outdoor experiment, the vessels were left open without lids so that the sample was exposed to the full spectral range of sunlight. To compensate for evaporative losses of water in the outdoor experiment, the vessels were weighed each hour and pH-adjusted water (NaHCO3; pH 8.4) was added to maintain constant weight. Evaporation of OH-PBDEs or PBDDs was not anticipated because of their high boiling points. In each set of experiments, one matrix blank and one control sample were run. The control samples contained the same solutions as the samples but were kept dark by covering the test vessels with aluminum foil. Workup. All work was performed with as little direct light as possible to avoid debromination. Extraction with hexane:diethyl ether (9:1) and removal of phenolics with 0.5 M KOH partitioning were performed according to Hovander et al.37 with minor modifications. For details, see Supporting Information. Analysis. For details, see Supporting Information. Gas chromatography−high-resolution mass spectrometry (GCHRMS) analysis of mono- to tetra-BDD and BDFs was performed as described in Haglund et al.1 using a Micromass Ultima Autospec Ultra (Waters Corp., Milford, MA) operated in the electron ionization mode at 34 eV with a resolution of 10 000. The two most intense molecular ions for each
■
RESULTS AND DISCUSSION Photochemical-induced PBDD formation from OH-PBDEs yielded products in the percent range with regard to starting material with the highest yields for the tetrabromo OH-BDEs 47 and 68, at 12 and 9 mmol/mol, respectively (Figure 1 and Table 2). Penta-BDDs were screened for in experiments involving pentabromo OH-BDEs, and PeBDD yields were
