Environ. Sci. Technol. 1999, 33, 3458-3461
Persistency of Toxaphene Components in Mammals That Can Be Explained by Molecular Modeling W A L T E R V E T T E R * ,† A N D G E R D S C H E R E R ‡ Department of Food Chemistry, Friedrich-Schiller-University Jena, Dornburger Str. 25, D-07743 Jena, Germany, and Institut fu ¨ r Organische Chemie, Universita¨t Basel, St. Johanns-Ring 19, CH-4056 Basel, Switzerland
The insecticide toxaphene, a complex mixture of several hundred components, is readily degraded in the environment. Only very few hepta-, octa-, and nonachlorobornanes are accumulated in higher organisms. Of the 2626 isomers that are theoretically possible, only 2-endo,3-exo,5-endo,6exo,8,8,9,10,10-nonachlorobornane (B9-1679) is abundant in marine mammals. Using molecular models, we found that the nine bulky chlorine substituents of B9-1679 occupy conformations on the bornane backbone that result in a molecule that is free of steric hindrance and ring strain. The thermodynamic data derived from our modeling efforts proved to be predictive of the persistency of B9-1679 and 13 other hepta-, octa-, and nonachlorobornanes in mammals.
Introduction Toxaphene is one of the most heavily applied organochlorine insecticides (1). Toxaphene and toxaphene-like products consist of several hundreds of polychlorinated bornanes, camphenes, dihydocamphenes, and probably other hydrocarbons as well. The number of compounds detected in technical mixtures (Toxaphene, Strobane, Melipax, and other industrial products) varies from >200 to 670 chlorobornanes and other chlorinated bicyclic hydrocarbons (1). This number is high compared with that of technical PCB mixtures, but it is very small in comparison with the theoretically possible number of toxaphene components. For example, 32 768 congeners and enantiomers of chlorobornanes alone, the major substance class of toxaphene, are possible (2). The rapid degradation of less persistent chlorobornanes decreases the number of compounds of technical toxaphene found in the environment. Only a few chlorobornanes resist degradation and are thus bioaccumulated, especially at high trophic levels of marine food chains. In contrast to PCBs, the major compounds in the technical toxaphene mixtures, 2,2,5endo,6-exo,8,9,10-heptachlorobornane (B7-515 (3) or Parlar #32 (4)), 2,2,5-endo,6-exo,8,8,9,10-octachlorobornane, and 2,2,5-endo,6-exo,8,9,9,10-octachlorobornane (B8-806 and B8809 or Parlar #42), are not bioaccumulated. On the other hand, the major persistent chlorobornanes in marine mammals are not abundant in undegraded toxaphene or in sediment samples from Canadian lakes previously treated with toxaphene (5). This paper aims to explain these structure-dependent peculiarities with a combination of experimental results and * Corresponding author Phone: (Germany+) 3641 949 657; fax: (Germany+) 3641 949 652; e-mail:
[email protected]. † Friedrich-Schiller-University Jena. ‡ Universita ¨ t Basel. 3458
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 19, 1999
molecular modeling. Semiempirical calculations of chlorobornane structures were introduced in 1994 (6) and have since been used primarily to confirm NMR data (7-10). In a recent publication we reported evidence that the conformations of the chlorine atoms on the primary bornane carbons have a determining influence on the heats of formation (∆H°) of these compounds. We demonstrated that different conformations result in different ∆H° values and that those with the lowest values were most stable (10). Herein we present evidence that ∆H° is a good indicator of chlorobornane stability. Our results indicated that molecular modeling proved to be well-suited for explaining the stability and lability of chlorobornanes, previously recognized only by congener-specific analysis of biological samples.
Materials and Methods Semiempirical Calculations. Semiempirical calculations were performed using the Austin Model 1 (AM1) method (11) as implemented in the MOPAC package version 6.0 (12). The bornane backbone and the respective chloro and hydrogen substituents were imported into the AM1 program. To obtain structures with minimum ∆H° values, energy profiles were recorded by rotation of the methyl groups in steps of 120° about the C-8fC-7, C-9fC-7, and C-10fC-1 bonds, respectively. As a reference point, one bond was fixed and the other two (chlorinated) methyl groups were freely rotated until the optimal conformer, i.e., the one with the lowest ∆H°, was found with the AM1 method. With this technique, 33 ) 27 possible conformers were modeled. The combination of the 27 conformers resulted in rotational functions with the optimum conformer having the lowest ∆H°. The accuracy of the calculations was compared with published X-ray data of B7-515 (Parlar #32) (13). The calculated carbon skeleton of B7-515 (Parlar #32) reproduced the X-ray data within the margin of error, and the orientation of the chlorine atoms was equivalent. Furthermore, the C-H bond lengths of ∼0.8 Å in the X-ray structure were lengthened for the MOPAC calculations to a more realistic ∼1.1 Å. NMR spectra were simulated on the basis of the data set from the calculations which reproduced the measured spectra of the respective chlorobornanes (data not shown). Further details on our calculation method are presented elsewhere (6, 8, 10). Gas Chromatography/Mass Spectrometry. Chromatograms and mass spectra were recorded with an HP 5890 series II gas chromatograph connected to an HP 5989 B mass spectrometer (Hewlett-Packard, Waldbronn). The mass spectrometer was used in the electron capture negative ionization (ECNI-MS) mode. In the selected ion monitoring mode, m/z 411 and 413 were selected for nonachlorobornanes in the sample extracts. A stationary phase coated with CP-Sil 2 (Chrompack, Leipzig, Germany) was installed in the GC oven. All GC and ECNI-MS parameters have been recently presented in detail (8). Sample and Sample Cleanup Procedure. Blubber from a ringed seal (Phoca hispida) from the European Arctic (Spitsbergen) was analyzed for this study. Sample preparation was according to the method of Vetter et al. (14).
Results and Discussion Steric Conditions of Chlorobornanes and Conformations of Chlorine Substituents. The general synthesis pathway for toxaphene includes a Wagner-Meerwein rearrangement of the chiral camphene molecule and favors the formation 10.1021/es9901967 CCC: $18.00
1999 American Chemical Society Published on Web 08/28/1999
FIGURE 1. Structure and systematic numbering of the bornane backbone. The carbons C-1-C-6 are abbreviated as six-membered ring, C-7-C-9 as the bridge, and C-10 as the bridge head carbon. The bornane numbering represents the primary methyl group C-8 above C-5 and C-6 and the primary methyl group C-9 above C-2 and C-3 (6). Secondary carbons are C-2, C-3, C-5, and C-6; primary carbons are C-8, C-9, and C-10; exo is upward; and endo is downward. Conformations a, b, and c on the primary carbons C-8, C-9, C-10 are in accordance with ref 15. of chlorinated 1,7,7-trimethylbicyclo[2.2.1]heptanes or bornanes (Figure 1). The bridged bornane molecule consists of two exo-bicyclic five-membered rings that are relatively free of ring strain (6, 15). However, the degree of ring strain may change with the substitution of chlorine for hydrogen atoms. Bulky chlorine atoms require more space than hydrogen, and upon addition of a chlorine, the bornane backbone may be deformed. Under conditions where electron density is high and spatial restrictions exist, the resulting ring strain increases the ∆H° of the respective molecule. These factors prevent the generation of many thousand theoretically possible chlorobornanes. In the past, several empirical rules have been proposed to explain limitations on the variety of chlorobornanes (10, 16). For example, the maximum degree of chlorination on any given primary carbon atom is two and the maximum total number of chlorine substituents on primary carbons C-8 and C-9 is three (10, 16). These restrictions allow a maximum of five chlorine atoms on primary carbons. Using the AM1 method (11) as described previously (6), we demonstrated that a (hypothetical) trichloromethyl group does indeed deform the bornane backbone and causes a significant increase in ∆H° (10). To minimize steric hindrance with other bulky substituents, the chlorine atoms on the primary methyl groups C-8, C-9, and C-10 are preferably located in energetically favored conformations which may be distinguished by the letters a, b, and c as proposed by Parlar (see also Figure 1). However, rotations about the C-7fC-8, C-7fC-9, and C-1fC-10 bonds are not restricted at physiological temperatures (no atropisomers exist for the investigated chlorobornanes) (10). From the ∼20 hepta- to nonachlorobornanes we have modeled, a simple set of rules can be derived to identify the most stable conformer of chlorobornanes: energetically favored conformations of the chlorine atoms on the primary carbons on C-8, C-9, and C-10 are most likely directed by exo-chlorine atoms on C-2 and C-6 in the following manner:
2-exo f 8c, 9b, (9c), 10b, (10a) 6-exo f 8b, (8c), 9c, 10c, (10a) where the conformers in parentheses refer to a second chlorine atom in the case of a dichloromethyl group. Conformers without parentheses refer to the energetically favored conformer in the case of a chloromethyl group. These simple rules are very helpful to support the structure elucidation of chlorobornanes. Stern et al. isolated the two major components in sediment (Hx-Sed or B6-923 and HpSed or B7-1001) and elucidated the structures by GC/MS and 1H NMR spectroscopy (5). However, these techniques do not allow for an entire elucidation of the most stable conformations on the primary carbons (17). Although this is
FIGURE 2. GC/ECNI-MS chromatogram (m/z 411 and 413) of nonachlorobornanes in the blubber of a ringed seal (P. hispida) from Spitsbergen (European Arctic). not necessary for an exact description of a chemical structure, these conformers on the primary carbons seem to be the key for the understanding of the stability of chlorinated bornanes (10). Stern et al. suggested that for B7-1001, chlorine substituents on C-8 and C-9 were at 8c,9b or 8b,9c (5). The results presented herein clearly favor the latter conformer. Stability and Structure of Persistent Chlorobornanes in Biota. Several reports have produced evidence that only a few chlorobornanes are bioaccumulated in marine food chains (6, 18-20). Of the 2626 nonachlorobornanes that are theoretically possible, at least 51 are present in technical toxaphene (2, 21). In marine mammals, however, only 2-endo,3-exo,5-endo,6-exo,8,8,9,10,10-nonachlorobornane (B91679 or Parlar #50) is abundant (see Figure 2). Reductive dechlorination is a major degradation pathway for chlorobornanes, with geminal chlorine atoms being the most easily removed (6, 22). As a result, four of the nine chlorine atoms in B9-1679 can be found on the six-membered ring (i.e. one Cl atom each on C-2, C-3, C-5, and C-6). Staggered endo-exo-endo-exo conformation of the chloro-substituents on the six-membered ring was regarded as a determining factor leading to persistency of polychlorinated bornanes (21). For the following reasons, B9-1679 (Parlar #50) is the only plausible nonachlorobornane with this substructure. The remaining five chlorine atoms on the primary carbons may be in the 8,8,9,10,10- or 8,9,9,10,10position. Due to the “bridge-and-exo” rule (10), a dichloromethyl group on C-8 requires a 6-exo-chlorine atom, requirement for a dichloromethyl group on C-9 is a 2-exochlorine atom, and this excludes the 8,9,9,10,10-substitution on the primary carbons of B9-1679 (Parlar #50). The molecular model of B9-1679 (Parlar #50) deviates from the typical schematic presentation of chemical structures of chlorobornanes as shown in Figure 1. In reality, carbon C-10 is positioned upward with respect to the sixmembered ring (see Figure 3, left). In their preferable conformations (10a- and 10c-position, see above), chlorine atoms avoid proximity constraints with bulky chlorine substituents in the exo-position vicinal to C-1. The 10cchlorine atom has enough space between the hydrogens in the 2-exo- and 9a-position. On the other hand, the hydrogen atom on 10b causes no steric hindrance with the 6-exochlorine atom. Exchange of the substituents on 10a and 10b is obtained by rotation of 240° about the C-1fC-10 bond. In this conformation there is significant steric hindrance and ∆H° increases by ∼15 kJ/mol (Figure 3, right). This partly explains the conformations described above since a 2-exochlorine atom together with a 10c-chlorine atom and a 6-exochlorine atom together with a 10b-chlorine atom are unstable, due to steric hindrance. Our data predict that vicinal anti-chlorine atoms on the six-membered ring resist biodegradation. Under this premise, the following four combinations can be constructed: (1) 2-endo,3-exo,5-endo,6-exo; (2) 2-endo,3-exo,5-exo,6-endo; (3) 2-exo,3-endo,5-exo,6-endo; (4) 2-exo,3-endo,5-endo,6-exo. VOL. 33, NO. 19, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3459
FIGURE 3. (left) Molecular model of B9-1679 showing the 6-exochlorine atom close in space with the hydrogen on 10b and the 2-exo-hydrogen close in space with the chlorine atom on 10c. (right) Heat of formation (kJ/mol) of the conformers of the rotation about the C-1fC-10 bonding of B9-1679 (labels a, b, and c refer to positions substituted with chlorine atoms). The energetically favored conformation is a,c, which avoids bulky chlorine atoms in position close to the 6-exo-chlorine atom. However, #4 is not stable because of two exo-chlorines vicinal to C-1 (see below). Combination #2 is not formed since the synthesis procedure of toxaphene includes the intermediate 2-exo,10-dichlorobornane (or 6-exo,10-dichloromethane). Option #3 exists only as mirror image of #1 which includes mirroring of the substituents on C-8 and C-9 (according to the bridge-and-exo rule, a 8,8,9-substitution requires an exo-chlorine on C-6). Although the anti-conformation on C-2/C-3 as well as C-5/C-6 seem to be the prerequisite for persistence, the
staggered endo-exo-endo-exo conformation of chlorine atoms on the six-membered ring is the favorable conformation of chlorobornanes. Hence, B9-1679 meets all criteria for persistency. Table 1 lists all hepta-, octa-, and nonachlorobornanes with staggered endo-exo-endo-exo conformation that may exist. B9-1679 (Parlar #50) and the three octachlorobornanes have been identified in marine samples (5-6, 8, 14, 19). Of the heptachlorobornanes, however, only B7-1001 was detected so far in environmental samples (5, 18). Although persistency of organochlorines correlates with an increasing degree of chlorination (18), B7-1000 and B7-1002 may be among the few, still unknown heptachlorobornanes frequently identified in biological samples (14, 23). Heats of Formation for Chlorobornanes. After determination of the energetically favored conformations, ∆H° values were estimated using the AM1 method (see Table 2). Although there is no direct causality, an obvious association between ∆H° and biological stability was apparent for our example chlorobornanes. This is illustrated with B7-515 (Parlar #32), a very unstable chlorobornane in biota. B7-515 (Parlar #32) was neither detected in fish tissue (24, 25) nor in blubber of marine mammals (9, 14). As described previously, a 2-exo-chlorine atom directs an 8c,9b-conformation, and a 6-exo-chlorine atom results in the 8b,9c-conformation. Therefore, polychlorinated bornanes with two exo-chlorine atoms vicinal to C-1 such as B7-515 (Parlar #32) cannot have the optimal conformation since 8c,9b and 8b,9c exclude each other. To avoid this problem, B7-515 (Parlar #32) has to expend energy. The rotational profiles of B9-1679 (Parlar #50) demonstrated that a 9bconformation was at 15 kJ/mol less favorable than the optimal 9c-conformation. In contrast, the difference between an 8c,9b-conformation and an 8b,9c-conformation in the case of B7-515 (Parlar #32) was only ∼2 kJ/mol (10). On the basis of similar energetic balances for both chlorobornanes, the difference of ∼13 kJ/mol must be attributable to the additional strain energy in the B7-515 (Parlar #32) molecule. The data in Table 2 demonstrate that the ∆H° of B7-515 (Parlar #32) is significantly higher than that of B9-1679 (Parlar
TABLE 1. Structures of Hepta- Octa-, and Nonachlorobornanes code
structure
detected
B9-1679 (#50) B8-1412 (-) B8-1413 (#26) B8-1414 (#40) B7-1000 (-) B7-1001 (-) B7-1002 (-)
2-endo,3-exo,5-endo,6-exo,8,8,9,10,10-nonachlorobornane 2-endo,3-exo,5-endo,6-exo,8,8,9,10-octachlorobornane 2-endo,3-exo,5-endo,6-exo,8,8,10,10-octachlorobornane 2-endo,3-exo,5-endo,6-exo,8,9,10,10-octachlorobornane 2-endo,3-exo,5-endo,6-exo,8,8,10-heptachlorobornane 2-endo,3-exo,5-endo,6-exo,8,9,10-heptachlorobornane 2-endo,3-exo,5-endo,6-exo,8,10,10-heptachlorobornane
+ (19, 6) + (8) + (19, 6) + (14) + (5) -
TABLE 2. Structure and Heat of Formation (∆H°) of Chlorobornanes
3460
9
chlorobornane
structure
∆H° [kJ/mol] (most stable conformer)
B7-515 (#32) B7-1001 (-) B7-1453 (-) B8-786 (#51) B8-789 (#38) B8-806 (#42) B8-809 (#42) B8-1412 (-) B8-1413 (#26) B8-1414 (#40) B8-1945 (#41) B8-2229 (#44) B9-1025 (#62) B9-1679 (#50)
2,2,5-endo,6-exo,8,9,10 2-endo,3-exo,5-endo,6-exo,8,9,10 2-exo,3-endo,5-exo,9,9,10,10 2,2,5,5,8,9,10,10 2,2,5,5,9,9,10,10 2,2,5-endo,6-exo,8,8,9,10 2,2,5-endo,6-exo,8,9,9,10 2-endo,3-exo,5-endo,6-exo,8,8,9,10 2-endo,3-exo,5-endo,6-exo,8,8,10,10 2-endo,3-exo,5-endo,6-exo,8,9,10,10 2-exo,3-endo,5-exo,8,9,9,10,10 2-exo,5,5,8,9,9,10,10 2,2,5,5,8,9,9,10,10 2-endo,3-exo,5-endo,6-exo,8,8,9,10,10
-215.78 (8b,9c,10a) -243.51 (8b,9c,10c) -224.29 (9b,9c,10a,10b) -202.10 (8b,9c,10a,10b) -187.80 (9b,9c,10a,10b) -209.84 (8b,8c,9c,10a) -216.45 (8c,9b,9c,10a) -244.38 (8b,8c,9c,10c) -230.64 (8b,8c,10a,10c) -245.91 (8b,9c,10a,10c) -237.70 (8c,9b,9c,10a,10b) -206.78 (9b,9c,10a,10b) -187.69 (8c,9b,9c,10a,10b) -245.68 (8b,8c,9c,10a,10c)
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 19, 1999
#50) and several other chlorobornanes. The conformation of the chlorine atoms on C-8 and C-9 in combination with two exo-chlorine atoms vicinal to C-1 alone explain the low persistency of B7-515 (Parlar #32). It is noteworthy that ∆H° values for polychlorinated bornanes with two chloromethyl groups at the bridge are similar to those for chlorobornanes that have one additional chlorine atom in the bridge. Examples for this are the pairs B8-1414 (Parlar #40)/B9-1679 (Parlar #50), B7-515 (Parlar #32)/B8-809 (Parlar #42b), and B8-786 (Parlar #51)/B9-1025 (Parlar #62). The thermodynamic data in Table 2 provide clear evidence that low heat of formations (-∆H°) are found for chlorobornanes with exclusive endo-exo-endo-exo substitution on the six-membered ring. Furthermore, all the chlorobornanes identified as persistent congeners had low ∆H° values, except B8-2229 (Parlar #44). However, an explanation for the high ∆H° for B8-2229 of this component, which is abundant in seal blubber (14) and human adipose tissue (26), may be that B8-2229 (Parlar #44) is a metabolite of B9-1025 (Parlar #62) (27). In this case, reductive dechlorination of the 2-endochlorine of the latter compensates for the degraded B8-2229 (Parlar #44), which is ∼20 kJ/mol more stable than B9-1025 (Parlar #62). That B8-2229 (Parlar #44) is metabolizable to some extent by seals is supported by an enantiomeric ratio of B8-2229 (Parlar #44) in marine mammals which deviated strongly from the racemic composition (14, 28). Unfortunately, this explanation has not yet been proven with a feeding or a food chain study. Reductive Dehydrochlorination and Conclusions. In the 1970s, Casida et al. reported that chlorobornanes are mainly degraded by dechlorination or dehydrochlorination, and geminal chlorine atoms seemed to be the preferred place of attack (22). This was confirmed by our calculations because all chlorobornanes with geminal chlorine atoms had much higher ∆H° values than the persistent congeners. Dehydrochlorination leads to chlorobornenes. The double-bond of bornene leads to sp2 hybridized carbons C-2 and C-3 with C-1/C-2/C-3 and C-2/C-3/C-4 angles of 120°, respectively, instead of the sp3 hybridized carbons with tetrahedral angles of 109°28′. It is plausible that forcing the atoms into a plane and the shorter bond length of the methylene group cause ring strain. ∆H° values of several modeled chlorobornenes derived from B8-806/B8-809 (Parlar #42), B8-1413 (Parlar #26), and B9-1679 (Parlar #50) via elimination of HCl were drastically higher than those of the corresponding polychlorinated bornanes. Judging by the relative magnitude of estimated ∆H° values, elimination of HCl from B9-1679 (Parlar #50) or B8-1413 (Parlar #26) is energetically more favorable if the chlorine atom is removed from C-2 or C-6. This observation supports the conclusion that substituents on C-10 have an influence on the chlorine substituents vicinal to C-1 and by this on the stability of chlorobornanes. We conclude that molecular modeling of chemical structures and determination of thermodynamic data is a key step in understanding the fate of the many hundred components of toxaphene. With the help of empirical rules regarding the variety and the knowledge of the preferred
conformations of the substituents on the primary carbons, the overwhelmingly complex pesticide toxaphene can be reduced to a tangible and calculable problem. We are optimistic that this method will be applicable for other environmentally relevant xenobiotics as well, particularly those with sterically complex structures.
Acknowledgment W. V. wishes to thank K. A. Maruya, Skidaway Institute of Oceanography (Georgia, U.S.) for language revision in the final paper.
Literature Cited (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) (25) (26) (27) (28)
Saleh, M. A. Rev. Environ. Cont. Toxicol. 1991, 118, 1. Vetter, W. Chemosphere 1993, 26, 1079. Andrews, P.; Vetter, W. Chemosphere 1995, 31, 3879. Parlar, H.; Angerho¨fer, D.; Coelhan, M.; Kimmel, L. Organohalogen Compd. 1995, 26, 357. Stern, G. A.; Loewen, M. D.; Miskimmin, B. M.; Muir, D. C. G.; Westmore, J. B. Environ. Sci. Technol. 1996, 30, 2251. Vetter, W.; Scherer, G.; Schlabach, M.; Luckas, B.; Oehme, M. Fresenius J. Anal. Chem. 1994, 66, 552. Krock, B.; Vetter, W.; Luckas, B.; Scherer, G. Chemosphere 1996, 33, 1005. Vetter, W.; Klobes, U.; Krock, B.; Luckas, B.; Glotz, D.; Scherer, G. Environ. Sci. Technol. 1997, 31, 3023. Loewen, M. D.; Stern, G. A.; Westmore, J. B.; Muir, D. C. G.; Parlar, H. Chemosphere 1998, 36, 3119. Vetter, W.; Scherer, G. Chemosphere 1998, 37, 2525. Dewar, M. J. S.; Zo¨bisch, E. G.; Healy, E. F.; Stewart, J. J. P. J. Am. Chem. Soc. 1985, 107, 3902. Stewart, J. J. P. MOPAC 6.0: A Semiempirical Molecular Orbital Program, QCPE 455. Palmer, K.; Wong, R. Y.; Lundin, R. E.; Khalifa, S.; Casida, J. E. J. Am. Chem. Soc. 1975, 97, 408. Vetter, W.; Krock, B.; Luckas B. Chromatographia 1997, 44, 65. Parlar, H.; Fingerling, G.; Angerho¨fer, D.; Christ, G.; Coelhan, M. ACS Sympos. Ser. 1996, 671, 348. Hainzl, D.; Burhenne, J.; Parlar, H. Chemosphere 1994, 28, 245. Frenzen, G.; Hainzl, D.; Burhenne, J.; Parlar, H. Chemosphere 1994, 28, 2067. Fisk, A. T.; Norstrom, R. J.; Cymbalisty, C. D.; Muir, D. C. G. Environ. Toxicol. Chem. 1998, 17, 951. Stern, G. A.; Muir, D. C. G.; Ford, C. A.; Grift, N. P.; Dewailly, E.; Bidleman, T. F.; Walla, M. D. Environ. Sci. Technol. 1992, 26, 1838. Bidleman, T. F.; Walla, M. D.; Muir, D. C. G.; Stern, G. A. Environ. Toxicol. Chem. 1993, 12, 701. Zhu, J.; Mulvihill, M. J.; Norstrom, R. J. J. Chromatogr. A 1994, 669, 103. Casida, J. E.; Holmstead, R. L.; Khalifa, S.; Knox, J. R.; Ohsawa, T.; Palmer, K. J.; Wong, R. J. Science 1974, 183, 520. Buser, H.-R.; Mu ¨ ller, M. D. Environ. Sci. Technol. 1994, 28, 119. Alder, L.; Vieth, B. Fresenius J. Anal. Chem. 1996, 354, 81. Krock, B.; Vetter, W.; Luckas, B. Chemosphere 1997, 35, 1519. Gill, U. S.; Schwartz, H. M.; Wheatley, B.; Parlar, H. Chemosphere 1996, 33, 1021. Vetter, W. GIT Magazin 1998, 42/10, 992. Vetter, W.; Klobes, U.; Luckas, B.; Hottinger, G. Organohalogen Compd. 1997, 33, 63.
Received for review February 19, 1999. Revised manuscript received June 28, 1999. Accepted July 1, 1999. ES9901967
VOL. 33, NO. 19, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3461
