Research Determination of Halogenated Natural Products in Passive Samplers Deployed along the Great Barrier Reef, Queensland/Australia W A L T E R V E T T E R , * ,†,‡ P A U L H A A S E - A S C H O F F , †,‡ NATALIE ROSENFELDER,† TATIANA KOMAROVA,‡ AND JOCHEN F. MUELLER‡ University of Hohenheim, Institute of Food Chemistry, Garbenstrasse 28, D-70599 Stuttgart, Germany, The University of Queensland, and National Research Centre for Environmental Toxicology (EnTox), 39 Kessels Road, Coopers Plains 4108, Australia
Received March 27, 2009. Revised manuscript received June 14, 2009. Accepted June 29, 2009.
Halogenated natural products (HNPs) have been increasingly reported to occur in marine wildlife from all oceans. Several HNPs, such as 2,3,3′,4,4′,5,5′-heptachloro-1′-methyl-1,2′-bipyrrole (Q1) and 4,6-dibromo-2-(2′,4′-dibromo)phenoxyanisole (2′-MeOBDE 68 or BC-2), were detected at particularly high concentrations in dolphins from Queensland/Australia. About half of the coastline of Queensland (∼2500 km) is covered by the Great Barrier Reef, a rich ecosystem hosting a huge variety of species, many of which are known to produce natural compounds. In this study, semipermeable membrane devices (SPMDs) were deployed as passive samplers for about 30 days at 12 marine and 2 nonmarine sites (i.e., rivers) along the Great Barrier Reef as part of a routine monitoring program during November 2007 and May 2008. Q1 and 2′-MeO-BDE 68 were detected at the marine sites with frequencies of about 65% but not in any sample from the two rivers. Further HNPs (2,4,6-tribromophenol, TBP; 2,4,6-tribromoanisole, TBA; 2,2′dimethoxy-3,3′5,5′-tetrabromobiphenyl, 2,2′-diMeO-BB 80 or BC1; 3,5-dibromo-2-(2′,4′-dibromo)phenoxyanisole, 6-MeO-BDE 47 or BC-3; and 3,5-dibromo-2-(3′,5′-dibromo,2′-methoxy)phenoxyanisole, 2′,6-diMeO-BDE 68 or BC-11) were detected as well with frequencies of 18-97% in the marine samples, but no polybrominated flame retardants were detected. The highest amount of a single HNP, 2.3 µg/SPMD, was determined for TBP, which had a frequency of detection of only 46%. The maximum (average) amount in the SPMDs from marine sites was 44 ng (12 ng) for Q1 and 115 ng (20 ng) for 2′-MeO-BDE 68. A first order kinetic model was used to estimate concentrations of the HNPs in the water phase. Based on the depuration of performance reference compounds obtained at one of the sites, we assumed a sampling rate of 16 L/day. We used this sampling rate to estimate that the highest and average available concentrations of Q1 in the water during the deployment of * Corresponding author e-mail:
[email protected]. † University of Hohenheim. ‡ The University of Queensland. 10.1021/es900928m CCC: $40.75
Published on Web 07/13/2009
2009 American Chemical Society
the SPMD were 97 and 25 pg/L, respectively. The estimated maximum water concentrations of 2′-MeO-BDE 68, 2,2′-diMeOBB 80, 6-MeO-BDE 47, and 2′,6-diMeO-BDE 68 were on average 2-5.5 fold higher than that of Q1. The results confirm that the HNPs are produced throughout the Great Barrier Reef,whichappearstobeasignificantsourceofthesecompounds.
Introduction To date, some 4,500 halogenated natural products (HNPs) have been isolated from their producers, i.e., algae, sponges, and other primarily marine organisms (1). During the past decade, a few of these HNPs were repeatedly detected as residues in higher marine organisms (2, 3). The structures of these HNPs (Figure 1) feature heterocyclic (HDBPs (4) and Q1 (5)), alicyclic (MHC-1 (6) and PBHDs (7)), and aromatic backbones (polybrominated phenols (8), anisoles (9), phenoxyanisoles (10-13), and dimethoxybiphenyls (14)), having up to seven halogen substituents. Both the structures and concentrations found in the marine environment are comparable with those of anthropogenic organohalogen compounds. It cannot be excluded that they pose a risk to the environment like man-made persistent bioaccumulative and toxic chemicals. Thus, their occurrence and fate need to be studied more thoroughly. In most occasions, HNPs have been detected in the mg/kg range in lipids of selected higher animals that do not synthesize the HNPs but have sequestered them as a result of bioaccumulation, including biomagnification along the food chain, similar to anthropogenic POPs (2). Frequently, the structures proved to be identical to that of compounds previously characterized by natural product chemists. For instance, 6-MeO-BDE 47 (BC3) and MHC-1 were found to be produced by algae (6, 15) and 2′-MeO-BDE 68 (BC-2) and PBHDs are produced by sponges (16, 17). Yet, the natural producer of Q1 has not been identified. Particularly high concentrations of Q1 were determined in the blubber of marine mammals from Queensland/Australia (13). These marine biota samples also contained a range of polybrominated natural products that was verified by the identification of one producer in the same habitat (16), which suggests that Q1 might also be biosynthesized in the waters off the Northeastern coast of Australia. The Great Barrier Reef of Queensland is a rich source of countless species, many of which have not yet been studied by natural product chemists. In this study we focused on the identification of HNPs in seawater samples collected from the Great Barrier Reef by means of passive sampling (PS) devices. PS is a suitable technique for the enrichment of lipophilic organohalogen compounds. Moreover, PS gives rise to a time-weighed accumulation of the analytes instead of a snapshot at the point of sample collection as in the case of direct water analysis. The semipermeable membrane device (SPMD) sampler is among the most frequently used PS device for organic contaminants (18-21). It has been suggested that SPMDs are suitable for the enrichment of lipophilic analytes with log KOW > 3 (20), however linear uptake is typically limited to chemicals with log KOW > 5.5 (22). Standard SPMDs consist of thin, low-density polyethylene tubes filled with triolein. This triacylglyceride (m.w. 885.5 Da) in an SPMD is considered a good bioconcentration surrogate when appropriately deployed in aquatic systems because only chemicals 4.5 - 5 (i.e., all compounds of interest but TBA and TBP) we may assume that the uptake is water side controlled and hence related to the thickness of the water boundary layer that is affected by flow/turbulences. Although performance reference compounds (PRC) (1000 ng of d10-anthracene and 100 ng of d10pyrene) were included, their interpretation was problematic potentially due to photodegradation (likely during deployment), and thus they could not be used in this study. As a result, sampler-specific sampling rates for surrogates such as PAHs were not available. However, in a related study, Komarova et al. (2009) determined that the sampling rates of PAHs ranged from 15 to 48 L/day per two SPMDs for PAHs with a molar mass between 178 and 252 g/mol and estimated log KOW of 4.7-6.8 at Magnetic Island using the same set up (open cage deployment) (30). The average value for one SPMD would thus be 16 ( 8 L/day. Similarly, Booij et al. determined BDE sampling rates of about 16 L/day in SPMDs deployed in the North Sea (31). Due to the lack of sample-specific in situ calibration data and the inclusion of Magnetic Island in the present study, which is in many respects similar to the other sites (i.e., proximate to an island and subject to continuous tidal flow and turbulence with a mean monthly water temperature of 23-29 °C), we assumed an intermediate value of 16 L/day to be the best estimate for a representative sampling rate for most of the deployments. For the purpose of this study, we further predicted KOW values for the chemicals of interest (see below) and SPMD-water partition coefficients from an empirical relationship where log KSW ) ao + 2.321 log KOW - 0.1618 (log KOW)2 (21). Since the chemicals of interest are overall nonpolar, we applied the intercept ao of -2.61, which is also based on the collated data in Huckins et al. (21). Gas Chromatography with Electron Capture Detection (GC/ECD). Samples were analyzed using an HP 5890 series II GC/ECD system fitted with an HP 7673A autosampler (Hewlett-Packard/Agilent Technologies). The GC column was a 27 m long × 0.25 mm internal diameter fused silica capillary coated with 0.25 µm HP-5 MS (Agilent J&W Scientific). The detector temperature was set at 300 °C. Nitrogen (purity 5.0, BOC Gases, Sydney/Australia) was used as the makeup gas at a velocity of 60 mL/min. One microliter of sample extract was injected splitless (275 °C) using a constant flow rate of 2.0 mL/min hydrogen (purity 5.0; BOC Gases, Sydney/ Australia) as the carrier gas. The GC oven was programmed as follows: start at 100 °C (1.5 min), then increase at 8 °C/min to 190 °C (0 min), at 3.5 °C/min to 275 °C (0 min), and finally at 50 °C/min to 315 °C (2.16 min). The total run time was 40 min. Gas Chromatography with Electron-Capture Negative Ion Mass Spectrometry (GC/ECNI-MS). Analyses were performed with a CP-3800/1200 system (Varian, Darmstadt, Germany) using previously published MS parameters except for using ammonia (purity 5.0, pressure 3.6 Torr) as the moderation gas instead of methane (28). Helium (purity 5.0, Sauerstoffwerke, Friedrichshafen, Germany) was used as the carrier gas at a constant flow of 1.2 mL/min. GC conditions (HP-5 MS column, 30 m × 0.25 mm i.d. × 0.25 µm film thickness, Agilent J&W Scientific) were reported elsewhere except for a final isothermal phase of 13 min instead of 38 min (28). In the full scan mode, m/z 30-800 was screened throughout the run. Quality Control. SPMD blanks did not contain peaks in the ECD chromatograms at the retention time of the analytes
except 2,4-dibromoanisole. So it was thus excluded from the evaluation. Detection limits were assigned to approximately 2 ng/SPMD (based on a 30-days deployment time). All analytes determined in this study by GC/ECD were verified by GC/ECNI-MS analyses in the full scan mode. Selected samples were analyzed by GC/ECNI-MS in the selected ion monitoring mode to assess whether anthropogenic brominated organic chemicals including >100 individual polybrominated flame retardants (polybrominated diphenyl ethers, hexabromocyclododecane, polybrominated biphenyls (28)) were also present in the samples. Analyses of SPMD blanks were virtually free of impurities.
Results and Discussion Initial Verification of the HNPs. To our knowledge, there was no information available on the capability of passive samplers for the accumulation of HNPs. Thus, we had to test whether HNPs are enriched in SPMDs similarly to anthropogenic POPs. Initial lab spiking experiments (see Experimental Section) verified that (nonphenolic) HNPs accumulate and can subsequently be coextracted similarly to anthropogenic BDE congeners. However, the recovery of bromophenols was significantly lower. Moreover, the phenolic nature of TBP (which is likely to be almost completely dissociated in marine systems) indicated that the levels of freely dissolved TBP in the water column could not be determined accurately (see below). GC/ECD chromatograms of the SPMD extracts cleaned with GPC were of sufficient purity to be easily analyzed (Figure S2, Supporting Information). The retention times of the characteristic peaks matched those of known halogenated natural products. GC/ECNI-MS full scan analysis was used to confirm the presence of TBA, TBP, Q1, 2′-MeO-BDE 68 (BC-2), 2,2′-diMeO-BB 80 (BC-1), 6-MeO-BDE 47 (BC-3), and 2′,6-diMeO-BDE 68 (BC-11) in the SPMD extracts (Figures S3 and S4). Verification of Q1 in the SPMD marked the first determination of this compound in the water phase. Owing to the nondestructive sample preparation procedure, this is clear evidence that Q1 occurs in the dissolved phase and is therefore not a metabolite formed in the marine food chain. Along with Q1, at least two abundant BrCl6 congeners of Q1 (Figure S3), traces of Br2Cl5-congeners, and ultratraces of Br3Cl4-congeners were qualitatively detected in SPMDs, and we may thus assume that they occur in water. Amounts of HNPs in the SPMD Samplers. Following their positive GC/ECD detection and GC/ECNI-MS verification, we quantified seven HNPs (TBA, TBP, Q1, 6-MeO-BDE 47, 2′-MeO-BDE 68, 2′,6-diMeO-BDE 68, and 2,2′-diMeO-BB 80) in 39 SPMD samplers from 12 marine sites from coastal Queensland (Figure S1, Table 1). Since HNPs are widely thought to originate from marine sources, we additionally analyzed six SPMD samplers from two rivers (see Figure S1). Samples were not available from all months, but, with the exception of one site, there was at least one sample from November or December, i.e., during the warm late spring/ summer season in Australia. Moreover, for logistic reasons, the deployment time of the samplers was varied (Table 1). To be better able to compare data for chemicals that were expected to be in the linear phase (or in the curvilinear phase, i.e., log KOW > 5), the mass of these chemicals in the samplers was normalized to a deployment period of 30 days (Table 1). Due to the realistic premise that the overall concentration of the lipophilic analytes in the water phase will be unchanged and that the capacity of the sampler was orders of magnitude higher than the amounts actually trapped, the accumulation of analytes in passive samplers can be treated as a first order process. Hence, the SPMD data were evaluated by assuming a linear enrichment of analytes thoughout the deployment period(s). VOL. 43, NO. 16, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6133
TABLE 1. Amounts (ng) Detected Per Standard SPMD Sampler Deployed at Different Sites of the Great Barrier Reef sitea sampling phase
date
deployed [d]
Pixies Garden Pixies Garden Lizard Island Lizard Island Lizard Island Low Island Low Island Low Island Low Island Low Island Low Island Fitzroy Island Fitzroy Island Fitzroy Island Fitzroy Island Fitzroy Island High Island High Island High Island Normanby Isl. Normanby Isl. Normanby Isl. Normanby Isl. Normanby Isl. Orpheus Isl. Orpheus Isl. Magnetic Isl. Magnetic Isl. Magnetic Isl. Magnetic Isl. Magnetic Isl. Daydream Isl. Hamilton Isl. Hamilton Isl. Outer Whitsundays North Keppel Island North Keppel Island North Keppel Island North Keppel Island
Dec 07 Mar 08 Nov 07 Dec 07 Mar 08 Nov 07 Dec 07 Jan 08 Feb 08 Mar 08 May 08 Nov 07 Dec 07 Feb 08 Mar 08 May 08 Nov 07 Dec 07 Feb 08 Dec 07 Jan 08 Feb 08 Mar 08 May 08 Feb 08 Apr 08 Nov 07 Dec 07 Feb 08 Mar 08 May 08 Dec 07 Dec 07 Mar 08 Nov 07 Dec 07 Jan 08 Feb 08 May 08
62 71 72 23 28 30 19 42 30 43 62 36 31 40 45 30 52 94 71 30 30 36 41 35 62 36 52 30 30 60 37 208 62 107 53 27 41 69 33
TBAb
TBPb
Q1
equi.
equi.
linear
linear
linear
linear
curvilinear
115 31 53
8
2
27 16 36
18 4 27
230 25
25 15
25
16 11 5 1
190 85 48 260 60 360 340 94 49
28 9.3 74 28 13
74 8.3 880 2300 12 3.1 230 230 6 57 130 38 22 43 71 41 22 24 43 9.6 20 97 160 7.2 70 110 140 48 16 8 53 54 35 81 29 140 79 95 33 81 28 36
17 30 11 22 20 7 9
44 20 15 35 11 18 15 20 11
4 23 28 36 21 11 10 11
2′-MeO-BDE 68 2,2′-diMeO-BB 80 6-MeO-BDE 47 2′,6-diMeO-BDE 68
41 75 97 19 8 18 46
10 20 84 15 1
1 4 50 15 11 24 16 12 16 34
2
27 4 34 65
11
16 6
11
60 41 32 56
10 8
4 270
3
3
10 8 2
36 51 6
1
3 11
2
1
8 29
a Sampling sites sorted from North to South. b For TBA and TBP we present the mass sequestered in the sample (assuming equilibrium) whereas for the other compounds we normalized the mass sequestered to a deployment time of 30 days assuming that these chemicals remained in a linear uptake phase.
TABLE 2. Frequency of Detection (n = 39), Amounts (ng) in SPMD Samplers from Marine Sites, Physico-Chemical Parameters, and Estimated log KSW and Water Concentration of HNPs compound
TBA
TBP
Q1
sites detected frequency of detection highest (mean)a amounts [ng]
38 97%
18 46%
25 64%
880
2300
44
aqueous solubility log KOW sampling phase after 30 d log KSW estimated highest (mean) water concentration [ng/L]
(104) (68) (12) 12.2 mg/L (34) 59 mg/L (36) 4.6 µg/L (34) 4.5 (35) 3.9 (36) 6.56 ( 0.9b
2′-MeO-BDE 68 2,2′-diMeO-BB-80 6-MeO-BDE 47 2′,6-diMeO-BDE 68 27 69%
7 18%
22 56%
17 44%
97
84
270
27
(20) n.a. 6.33 ( 0.75b
(4) n.a. 6.05 ( 0.8b
(28) n.a. 6.34 ( 0.8b
(5) n.a. 5.28 ( 0.8b
equilibrium
equilibrium
linear
linear
linear
linear
curvilinear
4.56
3.98
5.65
5.60
5.51
5.60
5.13
0.092 (0.025)
0.20 (0.041)
0.18 (0.008)
0.56 (0.058)
0.022 (0.004)
4.3 (0.54)
c
n.c.
a Nondetected was taken into account as 1 ng; values based on a 30-day deployment time. b Calculated using Advanced Chemistry Development (ACD/Laboratories) Software V9.04 for Solaris (via SciFinder). c Not calculated; the analytical recovery of TBP was low (see Text); in addition, TBP was a pKa of 5.97 and is likely almost completely dissociated in the water phase of the Great Barrier Reef. These facts suggest that any efforts would underrate the actual water concentrations.
Of the bromophenols and -anisoles, TBA was usually the most-concentrated compound, and its frequency of detection (97%) was the highest of any compound (Table 2). Surprisingly, natural producers of this compound are still unknown. By contrast, several sources are known for TBP, and it 6134
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 16, 2009
appeared that biomethylation of TBP would be the real source for TBA. Our results suggest that TBA is not solely formed in fish, which often contain high TBA concentrations, but already exists in the water phase. In agreement with that, TBA has been identified as a major volatile organic compound
FIGURE 2. Monthly normalized concentrations [ng] of Q1 and 2′-MeO-BDE 68 in SPMD samplers from Low Island (November 2007 to May 2008) and Q1 at Normanby Island. No values available for April 2008; 2′-MeO-BDE 68 values in December and February and Q1 (November 2007) were below the limit of detection (value assigned ) 2 ng). in coastal air from Norway and the Antarctic (9, 32). Amounts of TBA reached up to 880 ng/SPMD with an average of 92 ng/SPMD (Tables 1 and 2). TBA was also detected in SPMDs that were deployed in two rivers, but the amounts (
