Environ. Sci. Technol. 2005, 39, 3999-4004
Temperature and Congener Structure Affect the Enantioselectivity of Toxaphene Elimination by Fish K E I T H A . M A R U Y A , * ,† KELLY L. SMALLING,† AND WALTER VETTER‡ Skidaway Institute of Oceanography, 10 Ocean Science Circle, Savannah, Georgia 31411, and Institute of Food Chemistry, University of Hohenheim, Garbenstrasse 28, D-70599 Stuttgart, Germany
Recent advances in enantioselective separation techniques have enabled scientists to investigate environmental fate processes of chiral pollutants. In this study, congenerand enantiomer-specific toxaphene residues were monitored in captive, naturally contaminated fish (Fundulus heteroclitus) to characterize the effect of temperature and compound structure on the enantioselectivity of the elimination process. A previous study performed under warm water conditions (Tmean ) 25 °C) demonstrated relatively rapid (t1/2 ≈ 7-14 d) and enantioselective elimination of the reductive dechlorination metabolites 2-exo,3-endo,6-exo,8,9,10-hexachlorobornane (B6-923 or Hx-Sed) and 2-endo,3exo,5-endo,6-exo,8,9,10-heptachlorobornane (B7-1001 or HpSed). As expected, repetition of this experiment at cooler water temperatures (Tmean ) 15 °C) resulted in a decrease in overall (i.e., both enantiomers) first-order elimination rate constants. Enantiomer fractions or ratios (EFs/ ERs) during elimination, however, varied by congener, ranging from racemic for very rapidly eliminated Cl5 homologues to increasingly nonracemic for selected Cl6Cl8 homologues (including B6-923, several unknown Cl7 compounds, B8-1414, and B8-1945). As a result, we propose a classification to describe the environmental persistence of chiral toxaphene pollutants based on congenerspecific elimination kinetics and susceptibility to biotransformation as measured by EFs/ERs.
Introduction Toxaphene, a nonsystemic biocide, is a persistent organic contaminant with a large global inventory (1). Banned in many countries during the 1980s, toxaphene was used primarily in agriculture during the 1960-80s in the United States and abroad (2). After undergoing a Wagner-Meerwein rearrangement, the nonspecific chlorination of camphene leads mainly to chlorinated bornanes (3). Seven of the ten carbons that make up the bornane skeleton may carry zero, one, or two chlorine substituents, which explains the complexity (>670 compounds) of the technical products (4). In contrast to the production of technical toxaphene, the * Corresponding author present address: Southern California Coastal Water Research Project, 7171 Fenwick Lane, Westminster, California 92683. Phone: (714) 372-9214; fax: (714) 894-9699; e-mail:
[email protected]. † Skidaway Institute of Oceanography. ‡ University of Hohenheim. 10.1021/es048432n CCC: $30.25 Published on Web 04/21/2005
2005 American Chemical Society
synthesis of single components of technical toxaphene (CTTs) is a demanding and time-consuming task. To date, approximately 50 CTTs have been isolated/purified/synthesized (5-7); the remaining 90-95% of toxaphene constituents are structurally unknown. The challenges associated with the analysis of toxaphene residues in the environment are further complicated by the differential stability/persistence and hydrophobicity of individual CTTs. While the majority of CTTs are transformed to a large degree, several congeners exhibit extreme persistence (1). Weathered toxaphene associated with soils and sediments consists mainly of penta- to heptachlorobornaness transformation products of higher chlorinated CTTs. In aquatic organisms such as fish, a select group of Cl7-Cl9 homologues resist biotransformation (1). Because most CTTs are chiral (8), however, it has been suggested that a systematic change in enantiomer proportion in controlled experiments is indicative of biotransformation (9, 10). Thus, the recent application of chiral stationary phases (CSPs) to separate enantiomers of various contaminants has added a new dimension to investigations on their relative stability and, on a broader scale, on their overall global distribution and fate (9-11). Of particular interest are studies of toxaphene that have characterized enantioselectivity of uptake and elimination processes by model organisms (12-14). Recently, we observed that fish (Fundulus heteroclitus) from an area contaminated with weathered toxaphene eliminated 2-exo,3endo,6-exo,8,9,10-hexachlorobornane (B6-923, Figure 1) within a few days (t1/2 ) 7 d) under controlled, uncontaminated conditions (12). Moreover, the enantiomer ratio (ER) of B6923 increased throughout the several week elimination phase. These results were obtained under relatively warm conditions (water temperature range 22-28 °C, Tmean ) 25 °C) (12). In the present study, we repeated the elimination experiment with naturally contaminated Fundulus during the colder months (Tmean ) 15 °C) to investigate the influence of temperature on the elimination rates of B6-923 enantiomers. In addition, we tracked changes in ERs for other environmentally relevant chlorobornanes (e.g., B8-1414 or P-41) and devised a broad classification of toxaphene residues based on their environmental behavior in these and other experiments.
Materials and Methods Elimination Experiments. Mummichogs (F. heteroclitus), a small, nonmigratory fish commonly found in coastal Western Atlantic estuaries, were collected using baited minnow traps from the cooling water discharge canal of a former toxaphene manufacturing facility near Brunswick, GA (12). This site was impacted by toxaphene discharge for several decades, resulting in elevated toxaphene residue levels in sediment (15) and tidal creek invertebrates and fish (16). Toxaphene residue levels in mummichogs from this site range in the low to mid parts per million (17). Handling, acclimation, and experimental procedures were as documented previously (12). Briefly, fish were transported to the laboratory in coolers filled with aerated site water, acclimated for >3 d at ambient temperature in 0.3 m3 fiberglass tanks containing sandfiltered seawater under continuous flow-through conditions, and fed flake fish food (TetraMin, Tetra Holding Inc., Blacksburg, VA). Twenty-five fish were kept in each of three tanks, with duplicate tanks containing naturally contaminated fish and the third containing mummichogs from a tributary of the Skidaway River, Savannah, GA, that did not contain detectable levels of toxaphene. The pH (7.90 ( 0.07), VOL. 39, NO. 11, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3999
FIGURE 1. (a) 2-exo,3-endo,6-exo,8,9,10-Hexachlorobornane (B6923, Hx-Sed) and (b) its mirror image.
FIGURE 2. Time-dependent (a) water temperature and (b) enantiomer-specific and total B6-923 concentrations in mummichogs (F. heteroclitus) during the cold (Tmean ) 15 °C) elimination experiment. Concentrations (ng/g wet wt) were normalized to B9-1679 tissue concentrations. E1 and E2 represent the earlier and later eluting enantiomers, respectively. salinity (32 ( 3 ppt) and water temperature (mean ( sd, 15.0 ( 2.81 °C; Figure 2a) in each tank were monitored daily. Three fish were collected from each tank on days 0, 3, 11, 20, and 64 of the elimination phase. Individual weight and length measurements were taken before each fish was placed in a solvent-rinsed glass vial and stored at -20 °C. Sample Processing and Analysis. After homogenization with kiln-fired (550 °C for >8 h) Na2SO4, individual (whole) fish were Soxhlet-extracted with 400 mL of CH2Cl2 for >8 h and polar interferences removed using Florisil column chromatography (18). The CTT-containing Florisil fractions were further fractionated using silica gel column chromatography and HPLC (17). Fully fractionated and purified extracts were analyzed by gas chromatography-mass spectrometry in the electron capture negative ion mode (GCECNI-MS) using two instruments: (i) a Hewlett-Packard (HP) 5890 series II gas chromatograph interfaced to a 5989 mass 4000
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 11, 2005
spectrometer engine and (ii) an Agilent 6890 gas chromatograph coupled to a 5973 mass-selective detector (Agilent Technologies, Palo Alto, CA). For enantioselective analysis using instrument i, a 30 m × 0.25 mm i.d. CSP coated with 0.2 µm 25% randomly tert-butyldimethylsilylated β-cyclodextrin (β-BSCD) diluted with PS086 (BGB Analytik, Adliswil, Switzerland) was used (12, 17). Two fragment ions each were recorded for penta- to nonachlorobornane homologues using the selected ion monitoring (SIM) mode (17). A 30 m × 0.25 mm × 0.25 µm DB-XLB column (Agilent/J&W Scientific, Folsom, CA) was used for separations using instrument ii. Congener-specific concentrations were quantified for MSconfirmed analytes, and total toxaphene was estimated using a Varian 3400CX gas chromatograph with electron capture detection (Varian Inc., Palo Alto, CA) using previously published procedures (16, 18). Quality Assurance/Quality Control (QA/QC). QA/QC provisions for this study, including materials quality, instrument calibration, surrogate recovery, and the analysis of procedural blanks and toxaphene-spiked fish tissue, mirrored those published previously (12, 15, 16, 18). Briefly, all solvents were pesticide grade or equivalent (Optima, Fisher Scientific, Fair Lawn, NJ), and Florisil, silica gel, and Na2SO4 were chromatographic grade (Fisher Scientific). A 22-component mixture containing 17 chlorobornanes and 5 chlorocamphenes including 2-endo,3-exo,5-endo,6-exo,8,8,10,10-octachlorobornane (B8-1413 or P-26), 2-endo,3-exo,5-endo,6exo,8,9,10,10-octachlorobornane (B8-1414 or P-40), 2-exo,3endo,5-exo,8,9,9,10,10-octachlorobornane (B8-1945 or P-41), and 2-endo,3-exo,5-endo,6-exo,8,8,9,10,10-nonachlorobornane (B9-1679 or P-50) (TM2, Dr. Ehrenstorfer, Augsburg, Germany) was used for congener identification and quantification. In addition, 2-exo,3-endo,6-exo,8,9,10-hexachlorobornane (B6-923 or Hx-Sed) and 2-endo,3-exo,5-endo,6exo,8,9,10-heptachlorobornane (B7-1001 or Hp-Sed) purified/ isolated in environmental samples by the authors were quantified using a mean response factor for all TM2 components. Concentrations in F. heteroclitus were determined using six-point (external) GC-MS and gas chromatography with electron capture detection (GC-ECD) calibration curves (R2 g 0.99). No detectable toxaphene residues were found in procedural blanks (consisting of all reagents minus the fish sample), reference F. heteroclitus, unspiked fish food, or in supply or effluent seawater from either of the duplicate tanks with contaminated fish. Thus, the driving force for elimination into seawater was assumed to be unconstrained by aqueous toxaphene concentrations (100
>0.99
E2
ERb
EFb
760 442
460 138
1.7c 3.2
0.62 0.76
176
24
7.3
0.88
150
ndd
>100
>0.99
120
ndd
>100
>0.99
Normalized to B9-1679 from determinations on a DB-5 column (12). b ER ) enantiomer ratio () [E1]/[E2]); EF ) enantiomer fraction () [E1]/([E1] + [E2]). Estimated error ranges: ∼0.1 for ER < 2, 0.2-0.4 for ER > 2, 0.01-0.02 for EF. c Mean value (n ) 3, sd ) 0.16). d nd ) not detected. a
TABLE 2. Half-Lives (t1/2) and First-Order Elimination Rate Constants (ke) for B6-923 and B7-1001 in F. heteroclitus during Cool and Warm Water Conditions 15 °Ca
25 °Cb
15 °Ca
25 °Cb
ke(B6-923) -0.0503 -0.0983 ke(B7-1001) -0.02561 -0.0532 (d-1) (d-1) t1/2(B6-923) 14 7 t1/2(B7-1001) 27 13 (d) (d) a Mean water temperature (this study). (12).
b
Mean water temperature
racemic ERs, the ratio of enantiomers may alternatively be expressed using the EF:
EF ) [E1]/([E1] + [E2]) ) ER/(ER + 1)
(2)
Due to the rapidity of elimination and less than ideal resolution of B6-923 on the β-BSCD columnsthe only CSP known to resolve the enantiomers of this compoundsERs and EFs were calculated using chromatographic peak heights. The uncertainty in ERs associated with this approach was estimated to be (0.1 for racemic to moderately nonracemic values (1-2), and (0.2-0.4 for highly nonracemic values (e.g., 3-7). For EFs, the error decreased from (0.02 to (0.01 with increasing EFs, despite the wider variation observed. It is also known that coeluting compounds may render results determined on CSPs inaccurate (9-11). These potential issues were alleviated by exhaustive liquid chromatographic enrichment on silica gel (see Sample Processing and Analysis) that preseparated the majority of potentially coeluting compounds. Moreover, hepta- and octachloroCTTs found in the “B6-923” fraction eluted in a later GC time window (tR > 40 min) than the B6-923 enantiomers (tR < 38 min). Finally, nonchiral analyses revealed no indication of hexachloro isomers in concentrations that would interfere with our enantioselective measurements of B6-923.
Results and Discussion Temperature Effect on B6-923 Elimination. The concentration of B6-923 (both enantiomers) clearly decreased over the 64 d elimination period (Table 1, Figure 2). In comparison with the warm water elimination experiment (12), the concentration of B6-923 decreased at a lower rate for the same elimination time period. Assuming first-order kinetics for this elimination process, the rate constants (ke) and corresponding half-lives (t1/2) for both enantiomers of B6923 and B7-1001 were approximately doubled for the cold water elimination (Table 2). A rapid initial elimination was observed followed by a second slower elimination phase, ostensibly due to the nearly stepwise decrease in water temperature at ∼10 d (Figure 2).
As in the warm water experiment (12), the enantiomers of B6-923 were eliminated at different rates, with the later eluting enantiomer (E2) disappearing nearly twice as fast as the earlier eluting enantiomer (E1) (Figures 2b and 3). Whereas E2 was quantitatively and rapidly eliminated after 10 d, the elimination rate of E1 (and thus total B6-923) decreased dramatically after day 10, when water temperatures dipped below 15 °C (Figure 2). Interestingly, the ER increased more rapidly during the cold water elimination, doubling after only 3 d (Table 1). Differences in initial (t ) 0) ERs between the cold and warm experiments (1.6 vs 1.3) may also reflect the temperature-dependent processing by F. heteroclitus in the wild collected during different times of the year [April for the warm experiment (12), November for this study]. Taken together, these data suggest that the enantioselectivity of elimination was profoundly affected by the water temperature, and that E1 persisted in captive F. heteroclitus for a longer period of time during the cold experiment than was observed during the warm experiment. Chiral Signatures of Other Homologues. Two pentachlorobornanes referred to as “Penta-1” and “Penta-2” were previously detected in sediments (15) and fish (17). Thus, rapid elimination of Penta-1 and -2 in this experiment would support our previous hypothesis that elimination is offset by uptake for CTTs that remain elevated in fish from highly contaminated sites (12). Figure 4 demonstrates that both Penta-1 (enantiomers not resolved) and Penta-2 were quantitatively eliminated from fish within 3 d, a rate that is roughly double that of B6-923. Unfortunately, the rapid elimination of Penta-1 and -2 precluded the opportunity for detection of enantioselectivity. However, the racemic ratio of Penta-2 at t ) 0 indicated little or no enantioselectivity in terms of in situ uptake or elimination. Moreover, the ratios of Penta-1 and -2 were invariant in all three t ) 0 samples (data not shown). A similar effect was described by Skopp et al., who reported no enantioselectivity associated with the elimination of 2,2,5-endo,6-exo,8,9,10-heptachlorobornane (B7-515 or P-32), a nonpersistent CTT, in rats (14). Thus, it appears that, for certain lower chlorinated homologues, elimination rates not associated with enantioselective processes (e.g., passive diffusion via the gills) greatly exceed those that are enantioselective (i.e., biotransformation) and those of higher chlorinated homologues. Several hepta- and octachloro-CTTs were present in the investigated “B6-923” (125-150 mL) silica fraction of mummichog extracts (Figure 5). Although the relative abundance of these homologues differed between this fraction and the bulk sample, it is assumed that enantiomers fractionate nonselectively when subjected to physicochemical processing (e.g., silica gel fractionation). Thus, we can compare enantiomer distributions in Fundulus extracts at different times within this investigated fraction. For example, all enantiomers VOL. 39, NO. 11, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4001
FIGURE 3. GC-ECNI-MS-SIM chromatograms (m/z 307 and 309) of F. heteroclitus extracts illustrate enantioseparation of B6-923 using the β-BSCD chiral stationary phase for (a) t ) 0, (b) t ) 3 d, (c) t ) 11 d, and (d) t ) 64 d (end of experiment).
FIGURE 4. GC-ECNI-MS-SIM (m/z 273) analysis of F. heteroclitus extracts reveals pentachlorobornanes (Penta-1 and -2) at (a) t ) 0 were eliminated by (b) t ) 3 d. of four prominent unknown hepta-CTTs (“A-D”; 17) were detected in the B6-923 fraction. While compound A was racemic (or nearly so) at t ) 0, compounds B, C, and D were not (Figure 5). These initial conditions were also similar to previous findings for F. heteroclitus in their “naturally” contaminated environment (17). By the experiment’s end (t ) 64 d), however, compounds B, C, and D exhibited marked deviations from the racemate; more importantly, ERs were clearly shifted relative to the beginning of the elimination phase (t ) 0) (Figure 5). The results for octachloro-CTTs were similar to those observed for Cl7 homologues, although the time-dependent changes in ERs were less pronounced. Again, all isomers detected at t ) 0 were also detected at the end of the study (Figure 5). Although structures of several octachloro-CTTs remain unknown, the enantiomers of B8-1945 (Figure 5, peaks 1a and 1b) and B8-1414 (peaks 2a and 2b) were resolved in this study. At t ) 0, B8-1945 (peaks 1a and 1b) was nearracemic in contrast to B8-1414, where E1 (peak 2a) was more abundant than E2 (peak 2b). By day 64, however, ERs for both compounds were higher than at t ) 0. Two nonachlorobornanes were also detected in the investigated fraction; however, their ERs did not change over the 64 d elimination period (data not shown). The decreasing trend in ER shifts from Cl6 through Cl9 homologues suggests that, in addition 4002
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 11, 2005
to slower elimination due to increasing hydrophobicity (19), biotransformation potentials/rates decrease with increasing chlorine number. Effect of Temperature on Congener Recalcitrance. Coupled with our previous results (12), this study demonstrated that temperature plays a key role in determining the elimination kinetics of toxaphene congeners by a model poikilotherm (F. heteroclitus). Our temperature-dependent kinetic data for B7-1001 are comparable with the 32 d halflife reported for this same congener in exposed captive salmonids under slightly colder conditions (∼10 °C) (19). Moreover, our results indicate that enantiomer-specific elimination rates of the reductive dechlorination metabolite B6-923 were disproportionately different for warm and cold water conditions. Specifically, nonracemic elimination of B6923 was observed at both test temperatures (re,E2 > re,E1); however, as temperatures dropped below 15 °C in this study, the elimination of the remaining enantiomer (E1) became too slow to measure. Clearly, the mechanism(s) involved with enantioselective transformation and/or elimination of CTTs by fish is not presently understood. From these studies, however, we can conclude that individual CTTs exhibit differential recalcitrance in our test organism. It is plausible to assume that if biotransformation is occurring as is suggested by increasingly nonracemic proportions of B6-923 (and other homologues), enzymatic reactions favor one enantiomer relative to its mirror image. Enantiomer geometry and electron distribution may indeed result in a more favorable interaction with specific enzyme acceptors, resulting in selective depletion of one enantiomer. Temperature also influences the rate of chemical and biochemical reactions according to the Arrhenius equation (k ) Ae-Ea/RT), which states a doubling of the reaction rate (k) with every 10 °C increase (A is a constant of proportionality; see Table 3 for the definitions of other variables). Our comparison of first-order kinetics for B6-923 between our warm (∼25 °C) and cold (∼15 °C) experiments are consistent with Arrhenius’ postulate. The onset of a given reaction, however, is determined by its activation energy (Ea), which may also exhibit temperature dependence. It is conceivable that, at the warmer temperature, the activation energy was exceeded for both enantiomers of B6-923, and thus, reactions proceeded at different (enantiomer-specific) rates. This was apparent for both the warm experiment (12) and the initial phase (
