Degradation and Metabolism of Tetrabromobisphenol A (TBBPA) inSubmerged Soil and Soil−Plant SystemsFeifei Sun,† Boris Alexander Kolvenbach,‡ Peter Nastold,‡ Bingqi Jiang,§ Rong Ji,*,†,∥

and Philippe Francois-Xavier Corvini†,‡

†State Key Laboratory of Pollution Control and Resource Reuse, School of the Environment, Nanjing University, 163 Xianlin Avenue,210023 Nanjing, People’s Republic of China‡Institute for Ecopreneurship, School of Life Sciences, University of Applied Sciences and Arts Northwestern Switzerland,Grundenstrasse 40, Muttenz CH-4132, Switzerland§Fujian Provincial Academy of Environmental Science, No. 10, Huan Bei San Cun, Fuzhou 350013, People’s Republic of China∥Institute for Marine Sciences & Institute for Climate and Global Change Research, Nanjing University, 22 Hankou Road, 210093Nanjing, People’s Republic of China

*S Supporting Information

ABSTRACT: Contamination by tetrabromobisphenol A (TBBPA),the most widely used brominated flame retardant, is a matter ofenvironmental concern. Here, we investigated the fate andmetabolites of 14C-TBBPA in a submerged soil with an anoxic−oxic interface and planted or not with rice (Oryza sativa) and reed(Phragmites australis) seedlings. In unplanted soil, TBBPAdissipation (half-life 20.8 days) was accompanied by mineralization(11.5% of initial TBBPA) and the substantial formation (60.8%) ofbound residues. Twelve metabolites (10 in unplanted soil and 7 inplanted soil) were formed via four interconnected pathways:oxidative skeletal cleavage, O-methylation, type II ipso-substitution,and reductive debromination. The presence of the seedlings stronglyreduced 14C-TBBPA mineralization and bound-residue formationand stimulated debromination and O-methylation. Considerableradioactivity accumulated in rice (21.3%) and reed (33.1%) seedlings, mainly on or in the roots. While TBBPA dissipation washardly affected by the rice seedlings, it was strongly enhanced by the reed seedlings, greatly reducing the half-life (11.4 days) andincreasing monomethyl TBBPA formation (11.3%). The impact of the interconnected aerobic and anaerobic transformation ofTBBPA and wetland plants on the profile and dynamics of the metabolites should be considered in phytoremediation strategiesand environmental risk assessments of TBBPA in submerged soils.

■ INTRODUCTION

Tetrabromobisphenol A (TBBPA) is one of the most widelyused flame retardants in the world, with applications in printedcircuit boards and the production of several types of polymers.1

TBBPA can be released into the environment during theproduction, use, and disposal of flame-retardant-containingproducts.1,2 Importantly, it has been detected in variousenvironmental media, e.g., air, sewage sludge, sediment, andsoil, and in organisms such as mussels, birds, and even humanadipose tissue, breast milk, and plasma.1,3−9 The undesirableeffects of TBBPA include its activity as a potential endocrinedisrupter and as a source of oxidative stress in a wide variety oforganisms.3,10

Once released into the soil environment, TBBPA tends toaccumulate because of its high lipophilicity and poor watersolubility. In anoxic soils, TBBPA may be reductivelydebrominated to bisphenol A (BPA), with a half-life (T1/2) of36−430 days,11,12 accompanied by the formation of large

amounts of bound residues.11 In oxic clay soils, TBBPA has aT1/2 of 65−93 days.12 Under oxic conditions, it may bebacterially O-methylated to mono- and dimethyl ethers (MeO-TBBPA and diMeO-TBBPA, respectively) and then mineral-ized.13,14 However, the fate and metabolism of TBBPA in soilwith an anoxic−oxic interface, such as flooded soils, isunknown.Sequential anoxic−oxic treatment has been proposed for the

removal of TBBPA and other halogenated compounds found incontaminated sites, based on the debromination of TBBPAunder anoxic conditions and the rapid degradation of BPAunder oxic conditions.11,15 However, because of bound-residueformation during its anoxic incubation, TBBPA cannot be

Received: July 13, 2014Revised: November 12, 2014Accepted: November 17, 2014

Article

pubs.acs.org/est

© XXXX American Chemical Society A dx.doi.org/10.1021/es503383h | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

completely removed from the soil; instead, repeated anoxic−oxic incubations have been suggested.11 Yet, whether thepresence of adjacent anoxic and oxic layers in flooded soilsaccelerates the transformation of TBBPA is unknown.Wetland plants transport O2 to their roots from above-

ground, thereby forming oxic zones within anoxic soilcompartments.16,17 In the rhizosphere, root exudates providesubstrates for microbial growth and cometabolism but alsosurfactants that increase the bioavailability of pollutants as wellas inducers of degrading enzymes.18 Thus, at anoxic−oxicinterfaces the transformation of pollutants may be stimulated.Indeed, the common reed (Phragmites australis) and cord-grass(Spartina alternif lora), two plants frequently found in wetlands,were shown to accelerate the debromination of TBBPA in saltmarshes,19 whereas the effects of freshwater wetland plants onthe fate of TBBPA in soil have yet to be examined.The objectives of the present study were (1) to investigate

the degradation, metabolites, and residue distribution ofTBBPA in both submerged soil and a soil−plant system and(2) to elucidate the effects of the anoxic−oxic interface andwetland plants on the fate and metabolism of TBBPA in soil. Apaddy rice soil and soil−plant system were chosen becausepaddy fields constitute one of the most important floodedecosystems in the world, especially in Asia,20 and both TBBPAand BPA have been detected in paddy soils in China.21 Inaddition to paddy rice (Oryza sativa), P. australis, anotherwetland plant, was chosen, as it is commonly used for the insitu remediation of saltwater and freshwater sedimentscontaminated by organic pollutants such as polyaromatichydrocarbons and phenols.19,22,23

■ MATERIALS AND METHODSChemicals. Uniformly 14C-ring-labeled TBBPA (14C-

TBBPA) with a specific activity of 1.48 × 109 Bq/mmol anda radiochemical purity of 99% was synthesized from uniformly14C-ring-labeled phenol via BPA.11 Nonlabeled TBBPA andother chemicals were purchased from Sigma-Aldrich (Switzer-land).Soil. A silty clay loam soil with a pH 8.3, organic matter

content of 6.7%, and cation exchange capacity 46.2 cmol(+)/kgwas sampled from a paddy rice field located in Jiangning,Nanjing, China. The soil was air-dried, separated from plantsresidues and stones, and then sieved through a 2 mm sievebefore use.Plants. Seeds of O. sativa and P. australis were obtained

from Nanjing Agriculture University, China. The sterilizedseeds were germinated at 30 °C in the dark and the seedlingswere grown in the soil for 1 week and 3 weeks, respectively.Seedlings with similar biomasses were used in the incubationexperiments.Incubation Experiments. Incubation experiments were

carried out as follows: 100 mg P (KH2PO4)/kg and 100 mg N(NH4NO3)/kg were added as nutrients to 600 g of soil, whichwas then flooded with 780 mL of H2O. This water-saturated(submerged) soil was stored in the dark for 1 week to activateresident soil microorganisms. About 20 g of dried soil wasspiked with 14C-TBBPA in 0.2 mL of methanol. After removalof the methanol by evaporation in a fume hood for 30 min, themixture was thoroughly combined with the submerged soil,resulting in a final TBBPA concentration of 5 mg/kg soil (dryweight, dw) and a radioactivity of 3 MBq/kg soil (dw).Three treatments were tested: soil planted with O. sativa

growth (rice soil), soil planted with P. australis growth (reed

soil), and soil without plant growth (unplanted soil). Fortreatments with plant growth, four rice or reed seedlings ofsimilar sizes were transplanted into 50 g of soil (dw) in 100 mLflasks. For treatment without plant growth, 20 g of soil (dw)was placed into flasks of smaller diameter, thus allowing theformation of an oxic−anoxic interface in the flasks during theincubation. Three replicates were prepared per treatment. Eachflask was placed in a 1000 mL beaker connected successively tobottles containing ethylene glycol (200 mL, to trap radioactivevolatile organic compounds) and NaOH (2 M, 200 mL, to trap14CO2) and to a vacuum pump (Supporting Information, SI,Figure S1). The flow-through incubation systems were set up ina growth chamber with cycles of light and temperature thatmimicked field conditions: 16 h of light and an incubationtemperature of 34 °C followed by 8 h of darkness and anincubation temperature of 26 °C. A water layer of ∼1 cm wasmaintained above the soil in all the flasks during the incubation.At incubation days 0, 10, 22, 35, 48, and 66, aliquots of thethree soil types were sampled to monitor TBBPA degradationand metabolite formation. On days 10 and 66 (reed seedlings)and 10, 35, and 66 (rice seedlings), the seedlings were sampledto analyze the amount and distribution of the radioactivity inthe plants.

Soil Fractionation. The soil samples were centrifuged(10 000g, 5 min) to remove the water and then freeze-dried.Radioactivity in the water supernatant was determined by liquidscintillation counting (LSC) (see SI). About 1.5 g of the soilpellet was extracted three times using 15 mL of methanol undercontinuous shaking (150 rpm, 1 h). This extraction procedureled to a recovery of 96.2 ± 0.4% of the applied TBBPA. Aftercentrifugation (3600g, 10 min) of the suspension, thesupernatants were combined and the radioactivity wasmeasured by LSC. The methanol-extracted soil was air-driedat room temperature, and the amount of radioactivity wasdetermined by combusting 150 mg in an oxidizer (see SI). Afterorganic solvent extraction of the soil, the remaining,nonextractable radioactivity was defined as the boundresidues.24 These were further fractionated into humic acid(HA)-bound, fulvic acid (FA)-bound, and humin-boundresidues by alkaline extraction and acidification, as describedpreviously.11 The radioactivity in these fractions wasdetermined by LSC (see SI). The concentrations of TBBPAand its metabolites in organic extracts of the soil samplesprepared on various days during the experiment were analyzedusing high-performance liquid chromatography (HPLC) withan online radio flow detector (see SI). To identify the TBBPAmetabolites, the extracts at day 66 were purified by HPLC andanalyzed using gas chromatography−mass spectrometry(GC−MS) (see SI).

Uptake and Distribution of Radioactivity in Plants.The distribution of radioactivity was analyzed by auto-radiography of freeze-dried plant material, carried out by itsexposure for 1 week to imaging plates. The plates were thenscanned by a Fuji scanner (Fuji FLA-9000, Japan). Total plantradioactivity and the radioactivity in the roots and above-ground were determined by an oxidizer (see SI).

Data Analysis. SigmaPlot 12.0 was used to fit the TBBPAdegradation data to the first-order kinetics equation C = C0 e

−kt,where C0 and Ct are the concentration at time 0 and t,respectively, and k is the degradation rate constant. The half-life(T1/2) was calculated using T1/2 = ln 2/k. Significance wasanalyzed using an ANOVA or Student’s t-test. Differences wereconsidered significant if P < 0.05.

Environmental Science & Technology Article

dx.doi.org/10.1021/es503383h | Environ. Sci. Technol. XXXX, XXX, XXX−XXXB

■ RESULTS AND DISCUSSION

Both the rice and the reed seedlings were able to grow inTBBPA-contaminated soil (5 mg TBBPA/kg soil). The totalradioactivity recovered from the various soil and plant fractionsduring the experiments was in the range of 93−103%, 95−97%,and 96−105% for unplanted soil, rice soil, and reed soil,respectively (Figure 1).

Mineralization of TBBPA. In a previous study, the rate ofanaerobic cleavage of the benzene ring of TBBPA was very low,such that after 195 days of incubation in anoxic soil only 1.3%of the original amount had become mineralized.11 In anotherstudy, only 1.1% of the TBBPA in an oxic sandy soil slurryamended with bacterial growth medium was mineralized after20 days of incubation.13 By the end of the incubation period ofour study (66 days), a significant amount of 14C-TBBPA in theunplanted soil was mineralized (11.5 ± 0.4%) (Figure 1A),likely reflecting its higher rate of aerobic degradation in the oxic

layer of the top soil. In the soils with plant growth, only ∼5% ofthe 14C-TBBPA was mineralized (Figure 1B,C). This reducedmineralization in planted submerged soils could be attributed to(1) plant-mediated alteration of TBBPA metabolism in the soiland (2) a reduction in the amount of mineralizable 14C-labeledresidues (parent and metabolites) because of their substantialuptake by the plants (see below). Because the headspace of theincubation system was replaced continuously by fresh air,14CO2 reassimilation was likely to be negligible.In addition, the amount of volatile radioactive organic

compounds was not significant in any of the treatments duringthe incubation, ruling out the phyto-volatilization of TBBPAand its metabolites.

Degradation and Metabolites of TBBPA. During theincubation, organic-solvent-extractable radioactivity in theunplanted soil decreased to 20.2 ± 3.3% (SI Figure S2A).The amount was slightly lower in soil planted with reedseedlings (16.2 ± 1.6%), whereas there was no effect on theamount of extractable radioactivity in soil planted with riceseedlings (20.8 ± 0.6%, SI Figure S2A). Radio-HPLC analysisof the organic extracts showed the formation of 14C-labeledTBBPA metabolites (SI Figure S3). At the end of theincubation, >90% the labeled TBBPA was removed from theunplanted soil and the half-life (T1/2) of TBBPA under thiscondition was 20.8 ± 0.1 days (Figure 1A), which was fasterthan the reported dissipation of TBBPA under anoxicconditions in a rice paddy soil (T1/2 36 days),11 an estuarinesediment (T1/2 30−40 days),25 salt marshes (T1/2 70 to >130days)19 and a heavy clay soil (T1/2 430 days).

12 Loss of TBBPAin the unplanted soil was also faster than the dissipation ofTBBPA under oxic conditions in a freshwater sediment (T1/240 days),14 a sandy soil slurry (T1/2 41 days),

13 and a heavy claysoil (T1/2 65−93 days)12 but slower than the dissipation in ariver sediment under anoxic (T1/2 11 days)26 or oxic (16.8days)27 conditions. Although the types of soil and sediment canaffect the transformation of TBBPA in environmental matrices,our results suggest that the oxic−anoxic interface of submergedsoils provides a favorable environmental niche for the rapidtransformation of TBBPA.TBBPA dissipation was slightly enhanced in rice soil after 35

days of plant growth but was strongly accelerated in reed soil(T1/2 11.4 ± 0.4 days) throughout the incubation. In fact, underthe latter conditions, TBBPA had completely dissipated in thesoil by day 48 of the incubation (Figure 1C). The differenteffects of the two wetland plants on TBBPA transformation insoil can be explained by differences in root exudatecomposition, root morphology, and biomass, and thedeveloping rhizosphere microbial community, as suggested ina previous study examining the dissipation of TBBPA in saltmarshes containing reed and cord-grass.19 In that study, cord-grass significantly stimulated the dissipation of TBBPA, whilereed, in contrast to our results, had either no or negative effects.This discrepancy between our results and those reported byRavit et al.19 suggests that in addition to the particular plantspecies, rhizosphere effects on TBBPA transformation dependon environmental matrix properties and the initial soilcommunity from which the rhizosphere community isrecruited.The metabolites in the soil extracts at the end of the

incubation were identified by GC−MS, according to theirbromine isotope patterns and the resulting characteristic massfragments of the metabolites (the mass spectra of TBBPA andits metabolites are shown in SI Figures S4−S15). Twelve

Figure 1. Total recovered radioactivity (left y-axis) and theradioactivity recovered from CO2, water-soluble fraction (right y-axis), TBBPA, metabolites, and bound residues (left y-axis) during theincubation of 14C-TBBPA in a submerged soil either not planted (A)or planted with rice (B) or reed (C) seedlings. Data are the means ofthree individual experiments ± one standard deviation.

Environmental Science & Technology Article

dx.doi.org/10.1021/es503383h | Environ. Sci. Technol. XXXX, XXX, XXX−XXXC

metabolites (10, 7, and 5 metabolites from the unplanted soil,rice soil, and reed soil, respectively) were identified. Theirstructure and the identification of their fragments aresummarized in SI Table S1. The structures are also shown inFigure 2. These metabolites included debromination products(compounds 1−4 and 9), O-methylation products of TBBPAand its debromination and cleavage metabolites (compounds5−8, 10, and 11), and compounds with a single benzene ring(compounds 9−12). Among the O-methylation products, fourwere methyl ethers of brominated BPAs (bromoBPAs;compounds 5−8) and two were methyl ethers of single-ringmetabolites of TBBPA (compounds 10 and 11). Only oneether of bromoBPA (MeO-TBBPA; compound 7) was detectedin the reed soil whereas three ethers of bromoBPAs (MeO-TBBPA, diMeO-TBBPA, and MeO-triBBPA; compounds 7, 8,and 5, respectively) were identified in the rice soil and fourethers (compounds 5−8) in the unplanted soil.Dynamics of Metabolite Formation. In both the

unplanted soil and the rice soil, no metabolites were detected

before 22 days (Figure 1A,B), and all the radioactivity in theorganic extracts was from TBBPA rather than from itsmetabolites (SI Figure S2). After 35 days of incubation inunplanted soil, about 29.5 ± 6.0% of the initial 14C-TBBPA wastransformed to extractable metabolites, an amount thatessentially remained constant during the following month ofincubation (Figure 1A). By contrast, as shown in Figure 1C, inthe reed soil, metabolites were already detected on day 10 (10.6± 1.2% of the initial 14C), with the maximum amount reachedon day 22 (52.3 ± 8.1% of initial 14C) after which there was asteady decrease over the course of the incubation to 16.2 ±1.6%. While considerable amounts of extractable 14C-TBBPAremained in the unplanted soil and in the rice soil, nosignificant amount of 14C-TBBPA could be extracted from thereed soil (SI Figure S2). This rapid formation of metabolites inthe reed soil was in agreement with the stimulated trans-formation of TBBPA by the reed seedlings (Figure 1); at thesame time, the mineralization of 14C-TBBPA in reed soil wasmarkedly slower than in unplanted soil (Figure 1).

Figure 2. Identified metabolites of TBBPA transformation and the proposed pathways leading to their formation in a submerged soil with an oxic−anoxic interface in the absence and presence of plant growth. The compounds in dashed brackets were not detected in the soil extract and representhypothetical intermediates. The blue and red lines indicate the stimulated transformation pathways and the metabolites in soil planted with rice andreed seedlings, respectively.

Environmental Science & Technology Article

dx.doi.org/10.1021/es503383h | Environ. Sci. Technol. XXXX, XXX, XXX−XXXD

The dynamics of the formation of the different metabolitesare shown in Figure 3. In the unplanted soil, the lower-

brominated BPAs and the O-methylation products ofbromoBPAs were detected on day 35 of the incubation (Figure3C,D), indicating that anaerobic and aerobic transformation ofTBBPA took place simultaneously in the soil. These O-methylated metabolites were the dominant metabolites at theend of the incubation, accounting for 8.8 ± 0.3% of the initialTBBPA. After their initial occurrence in the soil their totalamount remained nearly stable (Figure 3D), which suggeststhat, together with debromination, O-methylation of bromo-BPAs was an important process in the soil. These findings arealso consistent with the recalcitrance of O-methylatedbromoBPAs to aerobic degradation processes, unlike lower-brominated BPAs, the amounts of which decreased after day 35of the incubation (Figure 3C). Bacteria capable of O-methylating halogenated phenolic compounds are ubiquitousin the environment.28 Indeed, 10% of aerobic heterotrophicmicroorganisms in a freshwater marsh sediment were shown toO-methylate TBBPA.14 The O-methylation of bromoBPAsdetermined in the present study underlines the significance ofthe O-methylation of halogenated phenols in soil environments.Given that we did not detect further debromination products ofdiBBPA (i.e., monoBBPA and BPA) in the unplanted soil, thedecrease in triBBPA and diBBPA (Figure 3C) suggests thatdebromination was slow. Single-ring compounds were detectedin the unplanted soil only at concentrations too low to monitortheir formation/dissipation dynamics (Figure 3A,B).Plant-mediated stimulation altered the amounts of the

metabolites in the soil as well as the time course of theiroccurrence during the incubation. Before day 48, the amountsof unknown polar metabolites (Figure 3A), single-ringmetabolites, BPA (Figure 3B), and O-methylated bromoBPAs

(Figure 3D) were considerably higher in the planted soils(especially reed) than in the unplanted soil. The amounts of theunknown polar metabolites and lower brominated BPAs(Figure 3A,C) decreased after their appearance in the reedsoil, while the levels of BPA and the single-ring metabolitesremained consistently high up to day 48 of the incubation,suggesting the temporal accumulation of both BPA and thesingle-ring metabolites in the reed soil. At the end of theincubation there were fewer metabolites (compounds 9−12) ineither of the planted soils than in the unplanted soil (SI TableS1), indicating a faster transformation of these metabolites inthe former incubations. The abundance of polar metaboliteswas in agreement with the higher amounts of water-solubleradioactivity in the reed soil (Figure 1). In addition,debromination and single-ring metabolites were present inhigher amounts than the O-methylation metabolites ofbromoBPAs (Figure 3B−D), but at the end of the incubationthe latter became dominant, accounting for 7.9 ± 0.8% and11.3 ± 2.1% of the initial TBBPA in rice soil and reed soil,respectively (Figure 3D).The presence of the reed and rice plants apparently

stimulated anaerobic reduction (Figure 3), such that evenmonoBBPA and BPA were detected in these soils (SI TableS1). In both of the planted soils, and especially in the reed soil,the unidentified polar metabolites (Figure 3A) might havederived from the further aerobic degradation of BPA in therhizosphere. During the stimulated transformation of TBBPAby reed growth, O-methylated bromoBPAs (10.6 ± 1.2% of theinitial TBBPA) occurred already on day 10 of the incubation(Figure 3D), followed by single-ring metabolites anddebromination metabolites, both of which likewise occurredearlier in this soil than in the unplanted or rice soil (Figure 3).This time course of metabolite development suggests that inthe reed soil aerobic O-methylation occurred more rapidly thananaerobic debromination. The presence of the reed plantsstrongly stimulated the transformation of TBBPA in the soil.Root exudates have been shown to support the growth of

dehalogenating microorganisms19and heterotrophic O-methyl-ating microorganisms in the rhizosphere; however, dependingon their origin they may differentially affect the functions of theassociated microbial communities.29,30 This would explain thedifferent rates of debromination and O-methylation ofbromoBPAs in the rhizosphere of rice vs reed seedlings andtherefore the temporal differences in the appearance of thedebromination and methylation metabolites in the two plantedsoils. Further studies on the exudate composition, oxygencontent, and microbial community in the rhizosphere mayprovide detailed insights into the different behaviors of rice andreed seedlings in stimulating TBBPA transformation.

Transformation Pathways. On the basis of the detectedmetabolites and the sequence of their occurrence under thethree conditions of the study (unplanted, rice-planted, andreed-planted; Figure 3; SI Table S1) and because the state ofthe unplanted soil during the incubation changed from oxic toanoxic, we propose the following pathways for TBBPAtransformation in submerged soil with and without plantgrowth (Figure 2).In the unplanted soil, four pathways likely participate in the

transformation of TBBPA: (I) oxidative skeletal cleavage, (II)O-methylation, (III) type II ipso-substitution, and (IV)reductive debromination (Figure 2). Pathways I and IIIproduce single-ring metabolites that can be further mineralizedin the soil under both oxic and anoxic conditions. Oxidative

Figure 3. Relative amounts of unidentified polar metabolites (A), BPAand single-ring metabolites (compounds 9−12) (B), lower-bromi-nated BPAs (compound 1−4) (C), and O-methylation metabolites ofbromoBPAs (compounds 5−8) (D) in organic extracts of soilincubated with 14C-TBBPA without and with plant growth. Thestructures of the metabolites are shown here and in Figure 2. Fordetails, see SI Table S1. Values are the means of three individualexperiments ± one standard deviation.

Environmental Science & Technology Article

dx.doi.org/10.1021/es503383h | Environ. Sci. Technol. XXXX, XXX, XXX−XXXE

skeletal cleavage (pathway I) is similar to the aerobic bacterialdegradation of BPA,31 which results in the cleavage of thiscompound into two single-ring compounds via oxidativeskeletal rearrangement (SI Figure S16). TBBPA and its O-methylation metabolite MeO-TBBPA (compound 7) have aBPA skeleton and are transformed to 1,3-dibromo-2-methoxy-5-vinylbenzene (compound 10) via sequential skeletal cleavage,O-methylation, reduction, and dehydration (SI Figure S16).Type II ipso-substitution is involved in the degradation of

alkylphenols with an α-quaternary carbon on the alkyl chain(e.g., nonylphenol and BPA) by sphingomonads (Sphingomonasxenophaga Bayram, Sphingomonas sp. TTNP3, Sphingobium sp.IT-4, and Sphingobium sp. IT-5) and by a nonsphingomonadstrain (Stenotrophomonas sp. IT-1).22,32−34 Analogous to thedegradation of BPA by type II ipso-substitution, generatinghydroquinone and carbocationic para-isopropylphenol,34

TBBPA degradation by type II ipso-substitution would yield2,6-dibromo-hydroquinone (compound 14) and the carboca-tion 2,6-dibromo-para-isopropylphenol (compound 15) (Fig-ure 2);13 O-methylation of the former resulted in the formationof 2,6-dibromohydroquinone monomethyl ether (compound11), and the latter was rapidly deprotonated to 2,6-dibromo-4(propen-2-yl)-phenol (compound 12), which was debromi-nated to 2-bromo-4(propen-2-yl)-phenol (compound 9)(Figure 2) under anoxic conditions. The carbocationicintermediates may be protonated during BPA degradation bystrain TTNP334 and hydroxylated during TBBPA degradationin soil slurry,13 but in the present study there was no evidenceof either reaction in the degradation of TBBPA in thesubmerged soil.In the aerobic metabolism of TBBPA in the environment, O-

methylation was shown to be an important reaction.14 In theunplanted soil, we detected not only mono- and dimethylethers of TBBPA (MeO- and diMeO-TBBPA) but also mono-and dimethyl ethers of debrominated TBBPA (MeO- anddiMeO-triBBPA) (Figure 2; SI Table S1). We attribute theformation of triBBPA ethers to combined O-methylation anddebromination at the oxic−anoxic interface within the soil.These compounds would also be detected in environmentscharacterized by temporally alternating oxic and anoxic states,such as soils or sediments with water fluctuation zones anddry−wet cycles.Debromination of TBBPA under anoxic conditions (pathway

IV) may generate BPA, via less-brominated BPAs (triBBPA,diBBPA, and monoBBPA).11,15,25 Since in our study debromi-nation in the soil was slower than O-methylation, only twodebromination products of TBBPA (triBBPA and diBBPA)were identified in the unplanted soil, while mono- and dimethylethers of triBBPA presumably formed via the reduction of theTBBPA ethers (Figure 2).The stimulated transformation pathways of TBBPA in the

soil with rice and reed growth are shown with blue and redlines, respectively, in Figure 2. The plants stimulated theanaerobic debromination and the aerobic O-methylation ofTBBPA. In both planted soils, TBBPA was completelydebrominated to BPA (compound 4) in a stepwise process,via triBBPA, diBBPA, and monoBBPA. While both plantsenhanced the formation of MeO-TBBPA (compound 7), riceseedlings stimulated further O-methylation and debrominationof MeO-TBBPA to diMeO-TBBPA (compound 8) and MeO-triBBPA (compound 5), respectively (Figure 2), in reactionsthat may have been carried out by aerobic and anaerobic

bacteria, respectively, in the rhizosphere of the rice soil, whereoxic and anoxic zones coexisted.

Uptake of TBBPA and Its Metabolites by Plants. Bothrice and reed seedlings showed a high potential to accumulateTBBPA and its metabolites from the soil. Reed seedlingsaccumulated significantly (P < 0.05) more radioactivity thanrice seedlings (33.1 ± 10.5% and 21.3 ± 1.5%, respectively,Figure 4A). This difference was consistent with the formation

of larger amounts of polar metabolites in the reed rhizosphere(Figure 3A), since plants more readily take up polar thannonpolar compounds.35 A previous study also showed that theaccumulation of TBBPA in terrestrial vegetables was plant-species-specific,36 probably reflecting the microbial communityand its activity in the respective rhizosphere. The considerableaccumulation of radioactivity detected in the plants was inagreement with the lower amounts of both bound-residueformation (see below) and TBBPA mineralization in theplanted than in the unplanted soil (Figure 1).During the incubation, most of the radioactivity that

accumulated in the plants was localized to their roots (Figure4B). 14C radio-imaging of the plants (SI Figure S17) showedthe distribution of radioactivity within the roots and in theabove-ground parts (stems and leaves) of the plants. At the endof the incubation, 18.4 ± 1.7% and 28.1 ± 5.9% of the initialradioactivity was recovered, respectively, from the roots of therice and reed seedlings and 1.2 ± 0.2% and 5.0 ± 1.5% from theabove-ground parts (Figure 4B). The radioactivity translocationfactor (TF), defined as the ratio of radioactivity in the above-ground to that in the below-ground plant tissue, was 0.21 onday 35 in the rice seedlings, but 0.067 and 0.18 on day 66 in therice and reed seedlings, respectively, indicating: (1) adecreasing TF with increasing plant growth, attributable tothe formation of the hydrophobic O-methylated bromoBPAs inthe soil during the incubation (Figure 3) and the tendency of

Figure 4. Radioactivity in the total plants (A) and in the leaves androots (B) of rice and reed seedlings grown in soil containing 14C-TBBPA, assayed on days 35 and 66 of incubation. Values are themeans of three individual experiments ± one standard deviation.

Environmental Science & Technology Article

dx.doi.org/10.1021/es503383h | Environ. Sci. Technol. XXXX, XXX, XXX−XXXF

these compounds to adsorb onto plant roots, and (2) thegreater accumulation of TBBPA and its metabolites in theabove-ground parts of reed rather than rice seedlings. Invegetable cabbage, TBBPA accumulation was higher in theroots (2−3 times) than in the shoots, while in radish it wasalmost the same.36 The chemical nature of the radioactivity thataccumulated in the rice and reed seedlings remains to beinvestigated.Distribution of Bound Residues in Soil. In all treat-

ments, the residual radioactivity bound to the soils increasedover time such that by the end of the incubation it accountedfor the majority of the radioactivity in the soils (Figure 1).From the initially applied radioactivity in the unplanted soil,60.8 ± 3.0% was associated with bound (nonextractable)residues (Figure 1A). A previous study showed that theformation of TBBPA-derived bound residues in soil was low(<35%) under anoxic conditions but increased under oxicconditions.11 In the present study, the large amount of boundresidues in the unplanted soil can thus be attributed to the oxiclayer in the upper soil. In the soils with plant growth, boundresidues accounted for 50.1 ± 4.0% and 40.0 ± 5.8% of theradioactivity in rice and reed seedlings, respectively (Figure1B,C). The lower amount of labeled bound residues in theplanted than the unplanted soil likely reflects either theaccumulation of labeled TBBPA residues in the plants or therelease of bound residues by the plant roots. Bound residuesderived from TBBPA or its metabolites and formed underanoxic conditions are released into the soil environment whenthe soil becomes oxic.11 Since oxic microsites surround theroots of the seedlings, the release of bound residues of TBBPAinto the soil can be expected. Bound residues may be formedvia several mechanisms, including physicochemical interaction,physical entrapment, and chemical bonding, or it may bepresent as biomass.37,38 The quantification of differentmechanisms in bound-residue formation is important in theevaluation of the stability of bound residues and indetermination of the risk posed by their presence.The bound residues in the soils were further fractionated into

FA-, HA-, and humin-bound residues (SI Figure S18). In allthree soil treatments, the amounts of residual radioactivitywithin the FA and HA fractions were initially low (<11%) butgenerally increased, albeit slowly, during the incubation (SIFigure S18A,B). Most of the bound radioactivity was located inthe humin fraction, and it increased markedly over time (SIFigure S18C). However, in both of the planted soils, thisfraction was smaller than in the unplanted soils, suggesting thatplant growth, and especially that of reed seedlings (SI FigureS18C), hindered the formation of humin-bound residues.O2 diffuses into the soil via the aerenchyma.17 In the

presence of O2, bromoBPAs otherwise bound to soil underanoxic conditions may be released11 and then undergodebromination, yielding diBBPA, monoBBPA, and BPA. Thiswould partly explain the stimulated debromination of TBBPAin the planted soils and the detection of these compounds(Figure 2; SI Table S1).Environmental Implications. Ours is the first study of the

metabolism of TBBPA in submerged soils with an oxic−anoxicinterface. Anaerobic and aerobic transformation of TBBPA tookplace simultaneously in these soils, with further transformationof the metabolites by either route, evidence of the complexity ofTBBPA metabolism in flooded soil. The transformation ofTBBPA in the flooded soil was faster than that reported inpurely oxic or anoxic soils,11−13 suggesting that the presence of

an oxic−anoxic interface, such as in wetland soils and paddyrice fields, provides a suitable environment for the rapidremoval of TBBPA.The pattern and timing of TBBPA transformation in the

submerged soils could be explained by several interconnectedpathways. Of these, oxidative skeletal rearrangements and typeII ipso-substitution seemed to be responsible for itsmineralization. Bacterially mediated O-methylation also playedan important role in TBBPA metabolism, resulting in theformation of relatively stable metabolites in the soil. In additionto MeO-TBBPA and diMeO-TBBPA, two metabolitespreviously reported in freshwater and marine sediments14,39,40

and in soil slurry,13 our study is the first to detect methyl ethersof TBBPA debromination products (MeO-triBBPA anddiMeO-triBBPA) in environmental samples. These O-methy-lated derivatives are more lipophilic and more persistent in theenvironment than the parent compound.4,14 DiMeO-TBBPAwas shown to be toxic to zebra fish embryos,41 suggesting thatthe formation of O-methylated derivatives poses a risk to theenvironment and to human health.The accelerated transformation of TBBPA and the reduction

in bound-residue formation in planted vs unplanted submergedsoil demonstrate the potential of wetland plants, especially reed,in the remediation of TBBPA-contaminated soils. However,before phytoremediation can be effectively implemented, themechanisms by which plants alter the dynamics of TBBPA, andthe amounts of its metabolites in the soil must be betterelucidated. Also unclear are the effects of the markedlyincreased amounts of lipophilic MeO-TBBPA, whose environ-mental behavior is not clear. In addition, the plantsaccumulated considerable quantities of uncharacterized residuesin their below- and above-ground tissues. The nature of thesecompounds, including their environmental safety, must bedetermined in further studies.

■ ASSOCIATED CONTENT

*S Supporting InformationGC−MS analysis, radioactivity determination, autoradiographyof the plants, HPLC radio-chromatograms, identification of theTBBBPA metabolites, mass spectra of its trimethylsilylatedmetabolites, proposed oxidative skeletal cleavage pathways, andthe distribution of bound residues within the humic fractions.This material is available free of charge via the Internet athttp://pubs.acs.org/.

■ AUTHOR INFORMATION

Corresponding Author*Tel.: +86-25-8968 0581; fax: +86-25-8968 0559; e-mail: ji@nju.edu.cn.

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTS

This work was supported by the National Natural ScienceFoundation of China (NSFC) (Grant Nos. 21237001,21177057), Sino Swiss Science and Technology Cooperation(SSSTC) (Grant Nos. IZLCZ2_138846, IP07_092011), andthe ear-marked cooperation project of the State Key Laboratoryof Pollution Control and Resource Reuse (grant no.PCRRF12022).

Environmental Science & Technology Article

dx.doi.org/10.1021/es503383h | Environ. Sci. Technol. XXXX, XXX, XXX−XXXG

■ REFERENCES(1) de Wit, C. A. An overview of brominated flame retardants in theenvironment. Chemosphere 2002, 46, 583−624.(2) Birnbaum, L. S.; Staskal, D. F. Brominated flame retardants:Cause for concern? Environ. Health Persp. 2004, 112, 9−17.(3) Covaci, A.; Voorspoels, S.; Abdallah, M. A. E.; Geens, T.; Harrad,S.; Law, R. J. Analytical and environmental aspects of the flameretardant tetrabromobisphenol-A and its derivatives. J. Chromatogr. A2009, 1216, 346−363.(4) Hakk, H.; Letcher, R. J. Metabolism in the toxicokinetics and fateof brominated flame retardants - a review. Environ. Int. 2003, 29, 801−828.(5) He, M. J.; Luo, X. J.; Yu, L. H.; Liu, J. A.; Zhang, X. L.; Chen, S.J.; Chen, D.; Mai, B. X. Tetrabromobisphenol-A and hexabromocy-clododecane in birds from an e-waste region in South China: Influenceof diet on diastereoisomer-and enantiomer-specific distribution andtrophodynamics. Environ. Sci. Technol. 2010, 44, 5748−5754.(6) He, M. J.; Luo, X. J.; Yu, L. H.; Wu, J. P.; Chen, S. J.; Mai, B. X.Diasteroisomer and enantiomer-specific profiles of hexabromocyclo-dodecane and tetrabromobisphenol A in an aquatic environment in ahighly industrialized area, South China: Vertical profile, phasepartition, and bioaccumulation. Environ. Pollut. 2013, 179, 105−110.(7) Johnson-Restrepo, B.; Adams, D. H.; Kannan, K. Tetrabromobi-sphenol A (TBBPA) and hexabromocyclododecanes (HBCDs) intissues of humans, dolphins, and sharks from the United States.Chemosphere 2008, 70, 1935−1944.(8) Luo, X. J.; Zhang, X. L.; Chen, S. J.; Mai, B. X. Free and boundpolybrominated diphenyl ethers and tetrabromobisphenol A infreshwater sediments. Mar. Pollut. Bull. 2010, 60, 718−724.(9) Zhang, X. L.; Luo, X. J.; Chen, S. J.; Wu, J. P.; Mai, B. X. Spatialdistribution and vertical profile of polybrominated diphenyl ethers,tetrabromobisphenol A, and decabromodiphenylethane in riversediment from an industrialized region of South China. Environ.Pollut. 2009, 157, 1917−1923.(10) Sun, Y. Y.; Guo, H. Y.; Yu, H. X.; Wang, X. R.; Wu, J. C.; Xue, Y.Q. Bioaccumulation and physiological effects of tetrabromobisphenolA in coontail Ceratophyllum demersum L. Chemosphere 2008, 70,1787−1795.(11) Liu, J.; Wang, Y. F.; Jiang, B. Q.; Wang, L. H.; Chen, J. Q.; Guo,H. Y.; Ji, R. Degradation, metabolism, and bound-residue formationand release of tetrabromobisphenol A in soil during sequential anoxic-oxic incubation. Environ. Sci. Technol. 2013, 47, 8348−8354.(12) Nyholm, J. R.; Lundberg, C.; Andersson, P. L. Biodegradationkinetics of selected brominated flame retardants in aerobic andanaerobic soil. Environ. Pollut. 2010, 158, 2235−2240.(13) Li, F. J.; Wang, J. J.; Nastold, P.; Jiang, B. Q.; Sun, F. F.; Zenker,A.; Kolvenbach, B. A.; Ji, R.; Corvini, P. F. X. Fate and metabolism oftetrabromobisphenol A in soil slurries without and with theamendment with the alkylphenol degrading bacterium Sphingomonassp. strain TTNP3. Environ. Pollut. 2014, 193, 181−188.(14) George, K. W.; Haggblom, M. M. Microbial O-methylation ofthe flame retardant tetrabromobisphenol-A. Environ. Sci. Technol.2008, 42, 5555−5561.(15) Ronen, Z.; Abeliovich, A. Anaerobic-aerobic process formicrobial degradation of tetrabromobisphenol A. Appl. Environ.Microbiol. 2000, 66, 2372−2377.(16) Meade, T.; D’Angelo, E. M. [C-14]Pentachlorophenolmineralization in the rice rhizosphere with established oxidized andreduced soil layers. Chemosphere 2005, 61, 48−55.(17) Brix, H.; Schierup, H. H. Soil oxygenation in constructed reedbeds: The role of macrophyte and soil-atomsphere interface oxygentransport. In Constructed Wetlands in Water Pollution Control; Cooper,P. F., Findlater, B. C., Eds.; Pergamon Press: Oxford, 1990; pp 53−66.(18) Shaw, L. J.; Burns, R. G. Biodegradation of organic pollutants inthe rhizosphere. Adv. Appl. Microbiol. 2003, 53, 1−60.(19) Ravit, B.; Ehrenfeld, J. G.; Haggblom, M. M. Salt marshrhizosphere affects microbial biotransformation of the widespreadhalogenated contaminant tetrabromobisphenol-A (TBBPA). Soil Biol.Biochem. 2005, 37, 1049−1057.

(20) Xiao, X. M.; Boles, B.; Frolking, S.; Li, C. S.; Babu, J. Y.; Salas,W.; Moore, B., III Mapping paddy rice agriculture in South andSoutheast Asia using multi-temporal MODIS images. Remote Sens.Environ. 2006, 100, 95−113.(21) Huang, D. Y.; Zhao, H. Q.; Liu, C. P.; Sun, C. X. Characteristics,sources, and transport of tetrabromobisphenol A and bisphenol A insoils from a typical e-waste recycling area in South China. Environ. Sci.Pollut. Res. 2014, 21, 5818−5826.(22) Toyama, T.; Murashita, M.; Kobayashi, K.; Kikuchi, S.; Sei, K.;Tanaka, Y.; Ike, M.; Mori, K. Acceleration of nonylphenol and 4-tert-octylphenol degradation in sediment by Phragmites australis andassociated rhizosphere bacteria. Environ. Sci. Technol. 2011, 45, 6524−6530.(23) Jouanneau, Y.; Willison, J. C.; Meyer, C.; Krivobok, S.; Chevron,N.; Besombes, J. L.; Blake, G. Stimulation of pyrene mineralization infreshwater sediments by bacterial and plant bioaugmentation. Environ.Sci. Technol. 2005, 39, 5729−5735.(24) Barriuso, E.; Benoit, P.; Dubus, I. G. Formation of pesticidenonextractable (bound) residues in soil: magnitude, controlling factorsand reversibility. Environ. Sci. Technol. 2008, 42, 1845−1854.(25) Voordeckers, J. W.; Fennell, D. E.; Jones, K.; Haggblom, M. M.Anaerobic biotransformation of tetrabromobisphenol A, tetrachlorobi-sphenol A, and bisphenol A in estuarine sediments. Environ. Sci.Technol. 2002, 36, 696−701.(26) Chang, B. V.; Yuan, S. Y.; Ren, Y. L. Aerobic degradation oftetrabromobisphenol-A by microbes in river sediment. Chemosphere2012b, 87, 535−541.(27) Chang, B. V.; Yuan, S. Y.; Ren, Y. L. Anaerobic degradation oftetrabromobisphenol-A in river sediment. Ecol. Eng. 2012, 49, 73−76.(28) Allard, A. S.; Remberger, M.; Neilson, A. H. Bacterial O-methylation of halogen-substituted phenols. Appl. Environ. Microbiol.1987, 53, 839−845.(29) Grayston, S. J.; Wang, S. Q.; Campbell, C. D.; Edwards, A. C.Selective influence of plant species on microbial diversity in therhizosphere. Soil Biol. Biochem. 1998, 30, 369−378.(30) Walker, T. S.; Bais, H. P.; Grotewold, E.; Vivanco, J. M. Rootexudation and rhizosphere biology. Plant Physiol. 2003, 132, 44−51.(31) Spivack, J.; Leib, T. K.; Lobos, J. H. Novel pathway for bacterialmetabolism of bisphenol-ARearrangements and stilbene cleavage inbisphenol-A metabolism. J. Biol. Chem. 1994, 269, 7323−7329.(32) Gabriel, F. L. P.; Heidlberger, A.; Rentsch, D.; Giger, W.;Guenther, K.; Kohler, H. P. E. A novel metabolic pathway fordegradation of 4-nonylphenol environmental contaminants bySphingomonas xenophaga Bayramipso-Hydroxylation and intra-molecular rearrangement. J. Biol. Chem. 2005, 280, 15526−15533.(33) Corvini, P. F. X.; Hollender, J.; Ji, R.; Schumacher, S.; Prell, J.;Hommes, G.; Priefer, U.; Vinken, R.; Schaffer, A. The degradation ofα-quaternary nonylphenol isomers by Sphingomonas sp. strain TTNP3involves a type II ipso-substitution mechanism. Appl. Microbiol.Biotechnol. 2006, 70, 114−122.(34) Kolvenbach, B.; Schlaich, N.; Raoui, Z.; Prell, J.; Zuhlke, S.;Schaffer, A.; Gugengerich, F. P.; Corvini, P. F. X. Degradation pathwayof bisphenol A: Does ipso substitution apply to phenols containing aquaternary α-carbon-structure in the para position. Appl. Environ.Microbiol. 2007, 73, 4776−4784.(35) Dettenmaier, E. M.; Doucette, W. J.; Bugbee, B. Chemicalhydrophobicity and uptake by plant roots. Environ. Sci. Technol. 2009,43, 324−329.(36) Li, Y. N.; Zhou, Q. X.; Wang, Y. Y.; Xie, X. J. Fate oftetrabromobisphenol A and hexabromocyclododecane brominatedflame retardants in soil and uptake by plants. Chemosphere 2011, 82,204−209.(37) Shan, J.; Jiang, B. Q.; Yu, B.; Li, C. L.; Sun, Y. Y.; Guo, H. Y.;Wu, J. C.; Klumpp, E.; Schaeffer, A.; Ji, R. Isomer-specific degradationof branched and linear 4-nonylphenol isomers in an oxic soil. Environ.Sci. Technol. 2011, 45, 8283−8289.(38) Kastner, M.; Nowak, K. M.; Miltner, A.; Trapp, S.; Schaffer, A.Classification and modelling of non-extractable residue (NER)

Environmental Science & Technology Article

dx.doi.org/10.1021/es503383h | Environ. Sci. Technol. XXXX, XXX, XXX−XXXH

formation of xenobiotics in soil - a synthesis. Crit. Rev. Environ. Sci.Technol. 2014, 44, 1−65.(39) Sellstrom, U.; Jansson, B. Analysis of tetrabromobisphenol A ina product and environmental samples. Chemosphere 1995, 31, 3085−3092.(40) Watanabe, I.; Kashimoto, T.; Tatsukawa, R. The flame retardanttetrabromobisphenol A and its metabolite found in river and marinesediments in Japan. Chemosphere 1983, 12, 1533−1539.(41) McCormick, J. M.; Paiva, M. S.; Haggblom, M. M.; Cooper, K.R.; White, L. A. Embryonic exposure to tetrabromobisphenol A and itsmetabolites, bisphenol A and tetrabromobisphenol A dimethyl etherdisrupts normal zebrafish (Danio rerio) development and matrixmetalloproteinase expression. Aquat. Toxicol. 2010, 100, 255−262.

Environmental Science & Technology Article

dx.doi.org/10.1021/es503383h | Environ. Sci. Technol. XXXX, XXX, XXX−XXXI