Uptake and biotransformation of 2,2′,4,4′-tetrabromodiphenyl ether (BDE-47) in four marine microalgae species

Hydroxylated- and methoxylated- polybrominated diphenyl ethers (OH-PBDEs and MeO-PBDEs) are more toxic than PBDEs and occur widely in the marine environment, and yet their origins remain controversial. In this study, four species of microalgae (Isochrysis galbana, Prorocentrum minimum, Skeletonema grethae and Thalassiosira pseudonana) were exposed to BDE-47, which is synthetic and is the predominant congener of PBDEs in the environment. By chemical analysis after incubation of 2 to 6 days, the efficiency of uptake of BDE-47 and, more importantly, the potential of undergoing biotransformation to form OH-PBDEs and MeO-PBDEs by the microalgae were investigated. Growth rates of these axenic microalgae were not affected upon exposure to environmentally relevant concentrations (0.2–20 μg BDE-47 L−1), and accumulation ranged from 0.772 ± 0.092 μg BDE-47 g−1 lipid to 215 ± 54 μg BDE-47 g−1 lipid within 2 days. Debromination of BDE-47 and formation of BDE-28 occurred in all microalgae species (0.01 to 0.87%), but biotransformation to OH-PBDEs was only found in I. galbana upon exposure to extremely high concentration. The results of this study showed that biotransformation of microalgae species is unlikely an explanation for the OH-PBDEs and MeO-PBDEs found in the marine environment.

to produce various organohalogens including MeO-PBDEs, either by their own tissue or by the associated cyanobacteria within them 8,25 . The great diversity of brominated organic compounds existing in the wild tends to support natural origins of these compounds because there is a lack of relevant PBDE precursors. Industrial processes would not have produced, or led to formation of, all of the discovered brominated compounds. Natural origins of MeO-PBDEs are further supported by the identification of related radiocarbon compounds from True beak's whales 26 and in sponge-cyanobacteria associations 27 , since radiocarbon could only be formed in nature but not in industrial processes. Disproval of biotransformation as the likely origin was also provided by the temporal trends of 6-MeO-BDE-47 and 2′-MeO-BDE-68 in pike from Swedish waters from 1967 to 2000, which were different from the trend of PBDEs production. Greatest concentrations of the MeO-PBDEs were detected in fish collected before 1970 while concentrations of PBDEs showed an increasing trend only up to the mid-1980s 28 .
Alternatively, results of other studies have shown the possibility of biotransformation of PBDEs to OH-PBDEs and MeO-PBDEs. For instance, human CYP2B6 has been found to be able to transform anthropogenic BDE-47 into 6 congeners of OH-PBDEs 29 . Likewise, in vivo study using mice dosed with DE-71 (a commercial mixture of PBDEs) have also found OH-PBDEs as metabolites in blood plasma 30 , and seven congeners of MeO-PBDEs have been detected as metabolites in blood and liver of rainbow trout (Oncorhynchus mykiss) after exposure to deca-PBDEs 31 . It has been suggested that bacterial O-methylation of PBDEs, or the OH-PBDEs from hepatic metabolism of PBDEs, might explain the formation of MeO-PBDEs 32 because such a mechanism was the major route of biodegradation of phenols in the environment 33 . In addition, there could be interconversion between OH-PBDEs and MeO-PBDEs. For instance, Japanese medaka has been shown to be able to transform OH-PBDEs to MeO-PBDEs and vice versa, and that demethylation of MeO-PBDEs took place, at a faster rate than hydroxylation of PBDEs, to form OH-PBDEs 34 . Since such biotransformation should result in lesser concentrations of OH-PBDEs and MeO-PBDEs and since biogenic production of these analogs has only been found in a few marine species, the high concentrations of OH-PBDEs and MeO-PBDEs in animals at high trophic level is difficult to explain.
It has been suggested that an unidentified source of planktonic organisms with high turn-over rates could be responsible for the high concentrations of OH-PBDEs and MeO-PBDES found in the ocean, because the enriched 14 C radiocarbon accumulated along the trophic level could only be acquired from surface water rather than from local hydrodynamics exposed by sessile marine sponges as suggested by many researchers 27 . Phytoplankton (or microalgae), which are abundant in surface waters and have a large surface to volume ratio, can be expected to accumulate PBDEs from water. They also play an important role in bioaccumulation of organic pollutants along the aquatic food chain. Despite this, uptake of PBDEs by microalgae has only been reported by a single study which investigated the efficiency of removal of PBDEs by freshwater microalgae in sewage treatment effluents 35 . The mechanism of passive uptake of POPs into microalgae can be described by a two-compartment model -immediate adsorption onto cell wall of the microalga followed by absorption (diffusion into the cell matrix with a mechanism similar to partitioning) 36,37 . However, whether microalgae could metabolize PBDEs and transform them into OH-PBDEs and MeO-PBDEs remain unknown. The significant correlations found between alkenones (a phytoplankton biomarker) and OH-PBDEs and MeO-PBDEs in surface sediment might suggest coccolithophorid species is a production source of 6-OH-BDE-47 and 6-MeO-BDE-47 38 .
This study was conducted to determine whether microalgae could synthesize OH-PBDEs and MeO-PBDEs by a) de novo production or b) biotransformation of PBDEs acquired from the environment. Four species of marine microalgae from different taxonomic groups were selected for experiment, including the Haptophyceae Isochrysis galbana, the Dinophyceae Prorocentrum minimum, the Bacillariophyceae Skeletonema grethae and the Bacillariophyceae Thalassiosira pseudonana. The uptake efficiencies of BDE-47 in these four different microalgae were also determined. To the best of our knowledge, this is the first attempt to investigate whether microalgae, as primary producers in the aquatic environment, could be a source of OH-PBDEs and MeO-PBDEs (from de novo synthesis and/or biotransformation of anthropogenic PBDEs) in the marine environment.

Uptake and bioaccumulation of BDE-47 into microalgae.
In all species tested, uptake of BDE-47 was positively related to the exposure concentrations. For example, uptake was 5 μg g −1 lipid in the 2 μg L −1 treatment and 50 μg g −1 lipid in the 20 μg L −1 treatment in S. grethae on day 2 (Fig. 2). It was only in I. galbana exposed to 2000 μg L −1 where uptake concentration was not as expected due to the lack of intact cells collected for measurement. Concentrations of BDE-47 in all species decreased from day 2 to day 6. As the algal cells multiplied and biomass increased from day 2 to day 6, the lipid content was likely to increase. If the total uptake amount of BDE-47 reached the maximum by day 2, increases in biomass and lipid content on subsequent days would result in reduced lipid-normalized concentration in algal cells. The background concentration of BDE-47 in filtered seawater used in this study was 0.087 ± 0.023 μg L −1 , whereas targeted OH-PBDEs and MeO-PBDEs were not detected.

Discussion
The results from this study show that BDE-47 can be readily taken up by all four species of microalgae, either by adsorption and absorption. In just two days, uptake efficiencies of BDE-47 into microalgae had reached 76 ± 19% for the flagellates I. galbana and 68 ± 6% for the dinoflagellates P. minimum, while the diatoms S. grethae and T. pseudonana exhibited lesser uptake efficiencies (ca. 31%). This result was consistent with the finding that bioaccumulation factors (BAF) of PBDEs by P. minimum being greater than that of Thalassiosira species 39 . This could have been due to polysaccharides in cell walls of diatoms 40 which are produced in larger amounts than by the other phytoplankton species 41 . The hydrophilic polysaccharides could reduce the amount of lipophilic BDE-47 adsorbed to cells, thus resulting in lesser bioaccumulation. There is only limited information on concentrations of PBDEs in marine microalgae, except that 0.023 μg BDE-47 g −1 lipid was reported in Thalassiosira species from the South Pole 42 . The concentration of BDE-47 accumulated (4.7 μg g −1 lipid) by T. pseudonana in this experiment was 100-fold greater than the reported concentration by Chiuchiolo et al. 42 . However, a more appropriate comparison with contaminated area is not possible since no other studies have reported concentration of PBDEs in microalgae from marine environments (also reviewed in Kosek et al. 43 ). The log BAFs for BDE-99 obtained during a 2-day incubation for P. minimum and Thalassiosira weissflogii were approximately 8.5 39 , which means BAFs were several orders of magnitude greater than those observed in this study for BDE-47.
Both 6-MeO-BDE-47 and 6-OH-BDE-47 were absent from the controls of all species used in this study, showing that these four species of microalgae could not produce these PBDE analogs naturally in the absence of PBDE as the precursor. Neither OH-PBDEs nor MeO-PBDE was detectable in filtered seawater used for the experiments. The result of this study therefore did not support the self-production or biotransformation of 6-MeO-BDE-47 by phytoplankton. However, 2′-MeO-BDE-68 and 6-MeO-BDE-47 were detected at pg g −1 dry mass concentrations in axenic cultures of Chaetoceros curvisetus, Prorocentrum donhaiense, and Emiliania huxleyi 44 , which suggested that biosynthesis may be possible, at least in some species. The difference in findings might be due to the differences in algal species, experimental protocol (in which seawater was only filtered through 0.45 μm, allowing passing of certain bacteria that may perform biotransformation 45 ) and source of seawater (in which the background levels of PBDEs and their metabolites in the seawater obtained from the East China Sea were not analyzed in their study). Concentrations of PBDEs in seawater in the East China Sea has not been measured previously 46 , but the same MeO-PBDE congeners in the microalgae cultures in Fan et al. 44 were also found at concentrations of 0.001-0.022 μg g −1 lipid in shellfish from the same area 47 . Although this might not directly confirm the presence of MeO-PBDEs or PBDEs in seawater because these pollutants might be transported through trophic transfer, their presence in natural seawater used for algal culture in their study cannot be ruled out. Mean concentrations of 2′-MeO-BDE-68 and 6-MeO-BDE-47 in seawater in Queensland were estimated to be 41 and 58 pg L −1 (by passive samplers) 48 , and the presence of MeO-PBDEs in water cannot be ruled out because the concentration might be too low to be detected by solvent-extraction.
Since algal cells were actively dividing during incubation, the presence of BDE-28 in some of the cultures would support the assumption that BDE-47 had been available for metabolism within algal cells. The detection of BDE-28 in cells exposed to greater concentrations of BDE-47 (1 μg L −1 for P. minimum and 10 μg L −1 for others) demonstrated that debromination had taken place during the incubation. BDE-28 was only detected on day 4 and 6 in P. minimum while it occurred at all sampling time points for the other three species. In nature, debromination is a common process to breakdown PBDEs, and light could accelerate debromination rate 49 . The tetra-brominated BDE-47 itself in the environment was assumed to be formed by debromination from more brominated congeners, such as the decabrominated BDE-209, derived from flame retardants. However, in both field and laboratory situations, light must be available to support growth and metabolism of microalgae and as such the possibility of photo-debromination could not be eliminated. A recent study demonstrated that tolerant freshwater microalgae isolated from wastewater treatment plants can degrade 68-86% of tetra-brominated BDE-47 in 7 days 35 , suggesting that removal of BDE-47 by marine microalgae could also be possible.
The presence of OH-PBDEs at the extreme high concentration of BDE-47 (2000 μg L −1 ) in I. galbana indicated potential capability for this species to metabolize PBDEs into OH-PBDEs by biotransformation. The metabolite 6-OH-BDE-47 has also been found in bile of crucian carp exposed to the hexa-brominated BDE-153 50 . Although the tetra-brominated 5-OH-BDE-47 and tri-brominated 2′-OH-BDE-28 were not congeners targeted Uptake percentage in their study, OH-tetra-BDE and OH-tri-BDE with unknown structures were shown in their GC-MS data 50 .

I. galbana P. minimum S. grethae T. pseudonana
Hydroxylation of PBDEs could be catalyzed by enzymes of the CYP family in mammals 29 , and genes encoding for CYPs have also been found in single-celled diatoms 51 . Nonetheless, further mechanistic studies such as pathway analysis are necessary to explain why biotransformation only occurred at such high concentration and when growth rate was inhibited. The finding of OH-PBDEs in this study must be interpreted with caution due to very low concentrations of metabolite found. In contrast to this study, it was found that common sole probably produced 4′-OH-BDE-49 and 4′-OH-BDE-101 as metabolite products after PBDE exposure, while 6-OH-BDE-47 in the fish was mainly of biogenic nature 52 . Nonetheless, the fact that no OH-PBDE was found in both controls and treatments at lower concentrations showed that OH-PBDEs were not produced by biogenesis in I. galbana. The majority of OH-PBDEs found in biota were believed to be transformed by demethylation from MeO-PBDEs within the organisms 23,34,53 . However, no MeO-PBDEs could be detected in any microalgae species in this study. Direct biotransformation (hydroxylation) from BDE-47 to OH-BDE-47 with debromination before or after the transformation might therefore be the possible mechanism. Hydroxylation is generally the first step of xenobiotic detoxification, and it was suggested to serve as a mechanism for PBDE removal in tolerant microalgae 35 . Yet, one cannot rule out the possibility that MeO-PBDEs, as intermediate metabolite, became too low to be detected since the biotransformation from MeO-PBDEs to OH-PBDEs was very efficient. The extreme high BDE-47 concentration used in this experiment shed light on the biological response at unrealistically great concentration. Notably, strong growth inhibition was found at the 2000 μg L −1 treatment within 2 days. The great BDE-47 concentration might have triggered certain atypical cell mechanism to produce OH-PBDEs, and thereby exerted toxic effects and inhibited cell growth. It has been shown that ortho-tetra-BDE radical was formed readily by photolysis of BDE-47, and that the formation of hydroxyl radical (•OH) was essential for the photo-formation of OH-PBDEs from PBDEs 21 . If the microalgae had formed hydroxyl radical during the exposure, this might as well be the cause of the observed growth inhibition.
In conclusion, 0.2-20 μg BDE-47 L −1 did not inhibit growth of the four microalgae species studied within 6 days, but biotransformation of BDE-47 to BDE-28 was clearly evident in all species. Only under extreme high BDE-47 concentration that small amount of OH-PBDEs metabolites (i.e. 2′-OH-BDE-28, 5-OH-BDE-47, and 6-OH-BDE-47) were produced in I. galbana associated with growth inhibition. Besides, there was no natural production of the targeted OH-PBDEs nor MeO-PBDEs found in the microalgae species tested here. It should be noted that these findings were limited to the exposure conditions and analytical methods used in this study. However, with the exposure to environmentally-relevant concentration of BDE-47, this study has found that it is unlikely for the four microalgae species to produce the high concentrations of OH-PBDEs and MeO-PBDEs observed in the marine environment by either de novo production or biotransformation.     Table 4. Result of a preliminary experiment showed that algal absorption of BDE-47 was below 60% within 1 day. As a result, algae were collected on day 2, 4 and 6 in subsequent experiments, to ensure that the algal cells can multiply at least once and also that significant amount of BDE-47 could be up-taken by the microalgae. Since log phase was reached by day 6, algal cells should have completed their life cycle, and biotransformation, if any, should have occurred. The concentration of the solvent DMSO at 0.01% is below the NOEC values (≥0.058%) 55 . Except for P. minimum which ordinarily had a lower density at log-phase and had shown growth inhibition at 2 μg BDE-47 L −1 , the nominal exposure concentrations for the other 3 species were 2 μg BDE-47 L −1 (the greatest environmental concentration reported) and a magnitude greater at 20 μg BDE-47 L −1 (greater confidence of measurement above detection limit). Exposure concentration for P. minimum was 0.2 and 2 μg BDE-47 L −1 . 200 and 2000 μg BDE-47 L −1 were used in an additional experiment for I. galbana to test for response under extreme high concentrations. Chemical analysis. The background level of PBDEs in FSW was measured in four replicates. Organic chemicals were extracted by same volume of hexane for three times in a separation funnel. The extract was dried and then re-dissolved in 100 μL of hexane before measurement.

Methods
For microalgae, the extraction and clean-up were based on Wang et al. 56 with modifications. The flow chart in Fig. 4 shows the procedures for the chemical analysis of microalgae. Firstly, samples were spiked with surrogate standard of 13 C-BDE-47, 13 C-6-MeO-BDE-100 and 13 C-6-OH-BDE-100 (Wellington Laboratories, Guelph, ON, Canada) in acetone. For lipid extraction, 5 mL of chloroform/methanol (1:1) was added and homogenized with algal cells by sonication at 40 °C for 30 min followed by centrifugation for three times. The solvent was dried by a stream of nitrogen gas, and the lipid weight was determined gravimetrically. After re-dissolving by 4 mL of hexane, phenolic phase was separated by adding 2 mL of 0.5 M potassium hydroxide in 50% ethanol and then mixed by vortex. The upper neutral layer was transferred to new tubes and extracted once again by hexane. The phenolic fraction that partitioned into the aqueous layer was then acidified by 2 mL of 2 M hydrochloric acid followed by two extractions to 4 mL of hexane/MTBE (9:1; v/v). Both fractions were then blow dried by nitrogen before purification.
The neutral fraction was purified by passing through a column with acidified silica gel, aluminum oxide and anhydrous sodium sulphate. Chemicals were eluted by 2 mL of hexane followed by 2 mL of DCM.    The column oven programme was as follow: from 60 °C (2 min) increased at a rate of 15 °C min −1 to 250 °C, then increased at a rate of 5 °C min −1 to 280 °C which was held for 10 min, and lastly increased at a rate of 30 °C min −1 to 290 °C and held for 10 min. The standard curves were obtained by injecting at 1 to 500 ng mL −1 for PBDEs and MeO-PBDEs, or 0.5 to 500 ng mL −1 for OH-PBDEs.
Quality control. Procedural blanks which consisted of the three surrogate standards were used for background reduction for every 12 to 15 samples. The average recovery percentages as determined by the mass of surrogate standards in each sample was 89% (69-113%) for 13 C-BDE-47, 77% (70-95%) for 13 C-6-MeO-BDE-100 and 86% (71-110%) for 13 C-6-OH-BDE-100. To confirm the purities of the BDE-47 stock used for treatment, the stock at 2 × 10 5 μg L −1 was analyzed by the same method and no BDE-28, MeO-PBDEs nor OH-PBDEs could be detected.
Calculations and statistical analysis. For  Bioaccumulation is defined as the accumulation of chemical as a result from uptake from all sources (including surrounding water), and bioaccumulation factor (BAF, L kg −1 ) is defined as the ratio of analyte in the biota and water 58 . For this study, its calculation is essentially the same as the bioconcentration factor:   Statistical analysis was performed using SigmaPlot 11.0. One-way ANOVA was used to find difference in algal density in different treatments at the same time point. Two-way ANOVA was used to find difference in the uptake percentage (arcsin square root transformed) of BDE-47 by number of days of exposure and by species. Only the concentration which had been tested in all species (2 μg L −1 ) was subjected to test for difference among species. Data passed normality test (Shapiro-Wilk) and had equal variability (both p > 0.05). Holm-Sidak test was performed and each treatment was compared to the control if significant difference was found by ANOVA (p < 0.05).