Hexabromocyclododecane in Wastewater Sludge ... - ACS Publications

Jul 27, 2006 - AND. JOHN M. ARIANO §. Toxicology and Environmental Research and Consulting, ..... publications/index.php). (2) Hunziker, R. W.; Gonsi...
0 downloads 0 Views 336KB Size
Environ. Sci. Technol. 2006, 40, 5395-5401

Biodegradation and Product Identification of [14C]Hexabromocyclododecane in Wastewater Sludge and Freshwater Aquatic Sediment J O H N W . D A V I S , * ,† S T A N L E Y J . GONSIOR,† DAN A. MARKHAM,† URS FRIEDERICH,‡ RENE W. HUNZIKER,‡ AND JOHN M. ARIANO§ Toxicology and Environmental Research and Consulting, The Dow Chemical Company, 1803 Building, Midland, Michigan 48674, Dow Europe GmbH, Bachtobelstrasse 3, CH-8810 Horgen, Switzerland, and Great Lakes Chemical Corporation, A Chemtura Company, 1801 Highway 52 West, West Lafayette, Indiana 47906

In a previous study the biodegradation of hexabromocyclododecane (HBCD) was reported to occur under realistic environmental concentrations in soils and freshwater aquatic sediments with biotransformation half-lives ranging from approximately 2 days to 2 months. In this study we extend our knowledge as to the environmental behavior of HBCD with respect to the fate of the three major diastereomers of HBCD (R, β, and γ) as well as to the identification of major intermediate metabolites formed during degradation. Substantial biological transformation of the R-, β-, and γ-[14C]HBCD diastereomers was observed in wastewater (i.e., digester) sludge and in freshwater aquatic sediment microcosms prepared under aerobic and anaerobic conditions. Concomitant with the loss of [14C]HBCD in these matrixes there was a concurrent production of three [14C]products. Using a combination of high performance liquid chromatography atmospheric pressure photoionization mass spectrometry and gas chromatography electron impact ionization mass spectrometry these metabolites were identified as tetrabromocyclododecene, dibromocyclododecadiene, and cyclododecatriene. We propose that HBCD is sequentially debrominated via dihaloelimination where at each step there is the loss of two bromines from vicinal carbons with the subsequent formation of a double bond between the adjacent carbon atoms. These results demonstrate that microorganisms naturally occurring in aquatic sediments and anaerobic digester sludge mediate complete debromination of HBCD.

Introduction Hexabromocyclododecane (HBCD) is used as a flame retardant mainly in building insulation composed of extruded or expanded polystyrene foam (over 85% of total volume) and some minor uses in back-coats of upholstery textiles (1). * Corresponding author phone: (989)636-8887; fax: (989)6389863; e-mail: [email protected]. † The Dow Chemical Company. ‡ Dow Europe GmbH. § Great Lakes Chemical Corporation. 10.1021/es060009m CCC: $33.50 Published on Web 07/27/2006

 2006 American Chemical Society

The environmental behavior of HBCD is influenced by its low aqueous solubility (65.6 µg/L, sum of the individual solubilities of the three diastereomers), high log Kow of 5.62, and low vapor pressure (6.3 × 10-5 Pa) (2). In most of today’s uses of HBCD it is blended in a polymer matrix and one would not expect significant emission to the environment. However, HBCD has been detected in various environmental matrixes and in biota. Morris et al. (3) have reported levels in river and coastal sediments ranging from 9 to 360 µg/kg dry weight. The latter value refers to a sampling location in the vicinity to a HBCD production site. Levels in soil ranged from 140 to 1300 µg/kg dry weight and were reported in proximity to facilities where HBCD was used. Maximum levels in the Lake Ontario food web range from 0.03 µg/kg wet weight for plankton to 4.5 µg/kg wet weight in lake trout (4). In recent work it has been reported that a shift occurs along the food chain, from γ-HBCD, the predominant diastereomer in the technical product and in soil/sediment samples, to the R-HBCD, which appears to be prevalent in the biota. Zegers and co-workers (5) reported the presence of HBCD in the blubber of harbor porpoises and dolphins from different European seas. The diastereomer composition in these mammals was exclusively the R-HBCD. Similar results were reported by Janak et al. (6) where they observed a predominance of the R-HBCD diastereomer in muscle and liver of several fish species from the Western Scheldt estuary. The predominance of the R-HBCD in wildlife may be partially explained by the results of Zegers et al. (5) where they reported the selective biotransformation (in vitro) of the β- and γ-HBCD by cytochrome P450 systems in microsomal preparations of liver from rats and harbor seals. In contrast, the R-HBCD did not appear to be metabolized in their test system. In previous work by our group, we have investigated the biodegradation of a technical composite of HBCD using typical environmental concentrations in soils and freshwater aquatic sediments (7). Biotransformation half-lives ranging from approximately 2 days to 2 months were determined for the γ-HBCD. No brominated degradation products were observed nor could the biodegradation be confirmed for the minor diastereomers, R and β. The failure to detect degradation products was likely due to the analytical sensitivity and the low levels of parent material used in this study (∼2-5 nM). The purpose of the present study was (a) to extend the knowledge on the degradation of HBCD with respect to the identification of major intermediate products formed during degradation and (b) in view of the reported diastereomer specific accumulation in biota to evaluate the biological degradation of the three diastereomers of HBCD (R, β, and γ).

Experimental Section Studies examining the degradation of [14C]HBCD under aerobic and anaerobic conditions were conducted with activated sludge and digester sludge, respectively. Parallel studies were performed with soil or freshwater aquatic sediments in microcosms prepared under aerobic and low redox conditions. Sufficient levels of [14C]HBCD (180 to ∼1700 nM, calculated) were added to the reaction mixtures to ensure that [14C]products formed at 10% yield (based on total radioactivity added) could be detected. This approach resulted in amounts of [14C]HBCD that exceeded the water solubility of the test material by greater than one to 2 orders of magnitude. Chemicals. The identity of the nonlabeled HBCD was confirmed by Fourier transform infrared spectroscopy. Great VOL. 40, NO. 17, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5395

Lakes Chemical Corporation (West Lafayette, IN) provided the R and β diastereomers of HBCD, cyclododecane, cyclododecene, tetrabromocyclododecene, 1,2-dibromocyclodedecane, and 1,5,9-cyclododecatriene (CDT). The mixed CDT diastereomers were obtained from GFS Chemicals, Inc. (Columbus, OH). [14C]HBCD (24.5 mCi/mmol, in acetone) was supplied by Wellington Laboratories and was labeled at minimum at the 1, 5, and 9 positions. The test chemical contained the R (∼8%), β (∼8%), and γ (∼82%), diastereomers of HBCD, similar in composition to the industry composite test material that was also used in the previous studies (7). Test Sample Preparation. Collection and storage of wastewater sludge and freshwater aquatic sediment were similar to the procedure outlined in our previous work (7). Activated and digester sludge were collected from a domestic wastewater treatment plant (WWTP). [14C]HBCD was added to all sludge or sediment reaction mixtures using acetone as the carrier solvent. Wastewater Sludge. The activated (i.e., aerobic) sludge study design was based on the OECD Zahn-Wellens/EMPA test for assessing the inherent biodegradability of compounds (8). [14C]HBCD was added to sealed 1 L glass vessels containing 300 mL of activated sludge (suspended solids (SS) ) 1000 ( 100 mg/L) and a defined mineral medium (9). [14C]HBCD was added at a nominal concentration of 3.6 mg/L (∼1.7 µM). Biologically inhibited control mixtures were amended with 250 mg/L mercuric chloride prior to the addition of [14C]HBCD. All reaction mixtures were equipped with caustic (i.e., CO2) traps and were incubated at 20-22 °C. The concentrations of [14C]HBCD and [14C]products were determined by withdrawing subsamples from each glass reaction vessel. The anaerobic digester sludge study was based on the ISO Standard 11734 test for assessing anaerobic biodegradation (10). Digester sludge in mineral salts medium (2130 mg/L SS) was dispensed (30 mL) into 60 mL serum bottles. The reaction mixtures (duplicates) were prepared under an anaerobic atmosphere (78% N2, 20% CO2, and 2% H2). [14C]HBCD was added to test mixtures at a nominal concentration of 4.2 mg/L (∼220 nM) and then incubated at 35 °C. Biologically inhibited controls were prepared by autoclaving sealed sludge mixtures for 30 min at 121 °C and 103 kPa. Sediment Microcosms. Aquatic sediment and overlying river water were collected from the Schuylkill River (Valley Forge, PA) as previously described (7). The study design was based on OECD method 308 for assessing aerobic and anaerobic biodegradation in aquatic sediments (11). Viable and biologically inhibited microcosms (duplicates) were prepared as previously described (7). Microcosms were prepared and incubated under either aerobic or low redox conditions. For the low redox microcosms, sediment reaction mixtures were prepared under an anaerobic atmosphere (78% N2, 20% CO2, 2% H2). For the aerobic microcosms, the oxygen levels in headspace gases were routinely monitored, and oxygen gas was added to the headspace if the oxygen concentrations fell below 12%. [14C]HBCD was added at a level of 4-5 mg/kg sediment dry weight (∼190-224 nM), and then the microcosms were incubated statically at 20-22 °C. Sampling. Acetonitrile was added to the sediment and sludge samples to analyze for [14C]HBCD. Wastewater sludge samples were mixed overnight with acetonitrile, and the mixtures were filtered (45 µm) prior to analysis. For the sediment microcosms, the water and sediment layers were separated, and each phase was mixed overnight with acetonitrile. All mixtures were assayed by liquid scintillation counting (LSC) and high performance liquid chromatography coupled with radiochemical detection (HPLC-RAM). The 5396

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 17, 2006

headspace gases of selected reaction mixtures (sediment and sludge) were routinely examined for volatile [14C]products using gas chromatography (GC) RAM. Data Analyses. Concentrations of HBCD or the individual diastereomer were normalized to the initial concentration and examined over time. The experimental data were fitted to the first-order decay model

Ct ) e-kt C0

(1)

where C is the concentration, k is the first-order rate constant, and t is time. Microsoft Excel Solver was utilized to determine the optimal rate constant (k) for each set of reaction mixtures by the least-squares method for both HBCD disappearance (viable and biologically inhibited controls) and disappearance of each of the diastereomer (two replicates each). [14C]HBCD Degradation Products. The acetonitrile added to the sediment or sludge microcosms was concentrated using Strata C18-E solid-phase extraction (SPE) columns (Phenomenex, Torrance, CA). Discrete eluent fractions containing each of the individual [14C]products were also collected from repeat HPLC analyses and concentrated on the SPE columns. Radioactivity recovered from the SPE columns for the three samples using the concentration procedure was shown to be 96.3% ( 0.7%. Analytical Methods. Total radioactivity in aqueous samples was determined using LSC. The distribution of radioactivity between [14C]HBCD and [14C]products was determined using HPLC-RAM (Agilent 1100) with a mobile phase containing 78% acetonitrile and 22% water delivered at 1.0 mL/min. Separation was achieved on a ZORBAX Rx-C18 column (4.6 mm × 250 mm, 5 µm), and detection was achieved with a UV detector (200 nm) and a Berthold model LB-509 HPLC radioflow detector arranged in series. Under the described conditions, the R, β, and γ diastereomers of [14C]HBCD eluted with approximate retention times of 13, 14.5, and 20 min, respectively (Supporting Information Figure SI 1). Radioactivity bound to the sludge or sediment was determined by combustion in a Biological Oxidizer model OX-300. Product Identification. Identification of [14C]products was conducted using HPLC atmospheric pressure photoionization mass spectrometry (APPI-MS) or GC electron impact ionization mass spectrometry (EI-MS). An Agilent 1100 liquid chromatograph with a mass-selective detector (MSD) and an APPI interface or an Agilent 5973 gas chromatograph with a MSD was used for confirmation of identity. Separations for the LC-MS-APPI were achieved on a Dupont ZORBAX Rx-C18 column (250 mm × 4.6 mm). Operational description of the LC-MS-APPI analysis conditions is provided in the Supporting Information.

Results and Discussion Primary Degradation of HBCD. Significant degradation of HBCD was noted in digester sludge and freshwater aquatic sediment mixtures with the consecutive appearance of three major [14C]products (Figure 1), which were designated as products I, II, and III based on the order of their elution in the HPLC assay (SI Figure 1 in the Supporting Information). After 28 days the concentration of [14C]HBCD (total of the three diastereomers) in the viable digester sludge decreased from 225 nM to 22 nM with the production of ∼93 nM of [14C]products. Mass balance determinations after 28 days indicated that ∼63% (SI Table 1 in the Supporting Information) of the initial radioactivity could be recovered as either [14C]HBCD or degradation [14C]products. It was found later that a significant amount of the radioactivity in the form of product III (>25%) was strongly sorbed to the butyl rubber

FIGURE 1. Loss of [14C]HBCD and production of [14C] degradation products in digester sludge and freshwater aquatic sediments. Closed symbols (symbols 9,[) are the viable reaction mixtures with asterisks representing the abiotic controls. The column at the far right represents the proposed degradation pathway for HBCD. stoppers used to seal the sludge reaction mixtures (SI Table 1 in the Supporting Information). Others have reported similar findings of rapid degradation of HBCD under low redox conditions in digester sludge (12). Degradation of [14C]HBCD was also noted in the freshwater aquatic sediments with similar trends observed in microcosms prepared and incubated under both aerobic and low redox conditions. Total [14C]HBCD decreased from 195 to 109 nM in the aerobic sediment and from 192 to 74 nM in the anaerobic sediment, during 113 days (Figure 1). The three major degradation products apparent in the digester sludge were detected in both sets of sediment preparations amounting to 90 and 120 nM of [14C]products, in the aerobically and anaerobically prepared sediments, respectively. After 113 days mass balance determinations for both sets of sediments indicated ∼97-98% of the initial radioactivity could be recovered as either [14C]HBCD or degradation [14C]products (SI Tables 2 and 3 in the Supporting Information). Assessment of biodegradation of HBCD in the soil and activated sludge was complicated due to the high loss of [14C]HBCD in both the abiotic and the viable reaction mixtures. In the viable activated sludge there was an ∼21% loss of HBCD over 56 days (data not shown) while formation of [14C]products, including 14CO2, was negligible and remained below 2%. There were even greater losses of [14C]HBCD in the abiotic controls (∼60%). In soil, the loss of [14C]HBCD was similar in both the viable and the abiotic reaction mixtures with less than 2% of the radioactivity recovered as degradation products (data not shown). Biological activity in the sediment and digester sludge enhanced HBCD transformation as the relative rates of HBCD

degradation in the viable reaction mixtures were approximately 2.7-10 times faster than those observed in the biologically inhibited (i.e., abiotic) controls (SI Table 4 and SI Figure 2 in the Supporting Information). Loss of HBCD and appearance of [14C]products were also noted in the abiotic controls of the digester sludge and sediment reaction mixtures (Figure 1). For the sediment abiotic controls there was some loss of [14C]HBCD over the course of the study and the levels of [14C]products reached 16% and 33% of the initial HBCD concentration in the aerobic and anaerobic sediments, respectively. Loss of HBCD was most noticeable in the abiotic anaerobic digester sludge mixtures where there was a ∼72% loss (138 nM) of [14C]HBCD with a concomitant production of 130 nM of [14C]products after 21 days. Much of the transformation of HBCD in the digester sludge controls was likely the result of abiotic transformation processes. Abiotic conditions in the digester sludge were confirmed by microbial plate counts and by monitoring methane concentrations in the headspace. No viable organisms were detected by culturing techniques in the abiotic digester sludge controls, and methane levels in the headspace remained static ( γ- > R-HBCD) with R-HBCD degradation rates (k) ranging from 28% to 47% slower than those observed for γ-HBCD (SI Table 4 in the Supporting Information). However, due to the level of variance in replicate observations in the sediment reaction mixtures these differences in degradation rates were not significant. A two-way statistical analysis of variation (ANOVA) was conducted for the biodegradation rate for different reaction mixtures. Within each sediment mixture (e.g., aerobically and anaerobically prepared) the degradation rates for R-, β-, and γ-HBCD were not significantly different from each other at an R-level of 0.05. In digester sludge, the degradation rate for β-HBCD was significantly different from those of R-HBCD (p ) 0.03) and γ-HBCD (p < 0.01). However, the sludge degradation rate for R-HBCD was not significantly different from that of γ-HBCD (p ) 0.42). At present the biological relevance of these findings is not evident. These results are consistent with the observation reported by Gerecke et al. (12) where they examined the degradation of HBCD under anaerobic conditions in digester sludge. They reported a half-life of 0.66 days for a technical HBCD mixture with the transformation of R-HBCD being slightly slower compared to those of γ- and β-HBCD, although rapid degradation was noted for all three diastereomers. They also reported preliminary evidence suggesting that degradation of all three HBCD diastereomers occurred in full-scale anaerobic digesters as well. It is interesting to note that the degradation half-lives of HBCD reported by Gerecke et al. (12) were faster than the degradation of HBCD observed in our current study where it took approximately 15 days to reach 50% loss of HBCD in the digester sludge (Figure 1). In contrast we previously reported (7) biotransformation half-lives for HBCD in aquatic sediment systems ranging from 11 to 32 days and from 1.1 to 1.5 days under aerobic and anaerobic conditions, respectively. In the current investigation, the initial HBCD concentration was about 90-fold higher than the concentration reported by Gerecke et al. This discrepancy in degradation rates between these two studies can be attributed to the observation that biodegradation kinetics are significantly impacted by the concentration of the test chemical with rates often changing from first order to zero order at higher concentrations (20, 21). At higher concentrations, for poorly soluble substances such as HBCD, biodegradation rates are more dependent on mass transfer limitations than on true biodegradation kinetics (22). Available data strongly suggest that meaningful kinetic data should only be determined at environmentally relevant concentrations (23, 24, 25). Therefore, apparent degradation rates from biodegradation studies conducted at high substrate concentrations may not accurately reflect degradation rates at lower concentrations, as they normally occur in the environment. Product Identification. Three major [14C]products were generated in the digester sludge and sediment microcosms. In the viable digester sludge mixtures products II and III constituted the greater part of the [14C]products with product I never achieving levels of more than 3% over the 28 day incubation period (Figure 1). In the sediment microcosms there was a sequential appearance of products I, II, and III. The degradation products were isolated and concentrated using SPE, and the concentrates were analyzed using HPLCAPPI-MS. Product I, with an HPLC relative retention time 15.2 min, was the first product formed at high levels in the sediment

FIGURE 3. HPLC-APPI-MS ion chromatograms of (A) product I and (B) product I fortified with an authentic standard of tetrabromocyclododecene (TBCDe). samples. This compound represented the dominant degradation product formed during the first month of the study, reaching levels ranging from approximately 14% to 20%. The proposed structure for product I was assigned as tetrabromocyclododecene (TBCDe, MW ) 482) and was likely the result of a dihaloelimination reaction of HBCD resulting in the loss of two bromines from vicinal carbons with the subsequent formation of a double bond between the adjacent carbon atoms. An authentic standard of tetrabromocyclododecene gave a negative-ion APPI mass spectrum containing an M + Br- ion cluster at m/z 561, an ion cluster at m/z 158, 160, and 162 with a relative abundance ratio of 1:2:1 (characteristic of two bromines [Br2]) and single bromine ion clusters at m/z 125 and 127 and m/z 79 and 81 with relative abundance ratios of 1:1 (SI Figure 3 in the Supporting Information). The ion cluster at m/z 125 and 127 is most likely due to a rearrangement product. A similar analysis was conducted for a concentrated extract of product I, but we were unable to generate sufficient levels of the material to provide a full scan response. Subsequent analyses were conducted with the MS operating in the single-ion-monitoring (SIM) mode to monitor the eight ions of interest that were present in the TBCDe standard. The negative-ion APPI ion chromatograms of product I showed an appropriate response at each of the ions monitored at the identical HPLC retention time corresponding to the same response observed for the authentic TBCDe standard (Figure 3A). Further confirmation that product I was indeed TBCDe was provided by ion chromatograms of a combined extract containing product I and the TBCDe standard (co-chromatography). The four major ions of each ion pair are presented in Figure 3B. This ion chromatogram, containing product I fortified with the TBCDe standard, resulted in an increase in

the relative response of approximately 3-fold for each of the ions at the identical HPLC retention times. Finally, the HPLC retention time of the tetrabromocyclododecene standard matched the retention time of the corresponding [14C]products in digester sludge and in both the aerobically and the anaerobically prepared sediments (∼15.2 min). The second degradation product, with a HPLC retention time of 18.6 min, was likely the result from the further dihaloelimination of TBCDe (product I) resulting in the loss of two additional adjacent bromines and the formation of a further double bond between the vicinal carbons. The proposed structure for this product is assigned as dibromocyclododecadiene (DBCDi, MW ) 322). Concentrated extracts of product II were analyzed using GC-EI-MS and had a retention time of 14.4 min. However, the EI mass spectrum of product II showed ions at m/z 241 and m/z 243 corresponding to M-H2Br+, a base peak ion at m/z 161 corresponding to M-H3Br2+, and an EI fragmentation pattern consistent with a cyclical C12 compound (Figure 4A). No authentic standard of DBCDi could be obtained for this study. However, an authentic standard of dibromocyclododecane (MW ) 326), which is similar to DBCDi except for the lack of two double bonds in the ring structure, showed ions at m/z 245 and m/z 247 corresponding to M-H2Br+ and a base peak at m/z 165 corresponding to M-H3Br2+. Note that these masses are 4 amu greater than the corresponding m/z assignments observed for DBCDi which account for the four additional hydrogen atoms this compound would have with no double bonds in the ring structure (data not shown). The structure of this second debromination product is consistent with the appearance of this compound during the course of the study. While product I was formed early in the sediment microcosms, only low levels (