EVALUATION OF PERSISTENT CONTAMINANT LEVELS IN GREAT LAKES SALMON EGGS Colleen M. Alexander & James J. Pagano Environmental Research Center, Department of Chemistry, 310 Piez Hall, Chinook, coho, and steelhead eggs collected in 2006 from Lake Ontario, L. Erie, L. Michigan, and L. Superior were analyzed by gas chromatography for congener-specific polychlorinated biphenyls (PCBs), polybrominated diphenyl ethers (PBD Es), and select organochlorine pesticides (OCs). Across lake compar isons for total PCB, total DDT, and total PBDE indicated a consistent trend in contaminant concentrations in which Lake Michigan > Lake Ontario > Lake Erie > Lake Superior. The contaminants Mirex and octachlorostyrene (OCS) were found not to be consistent with this trend, but we re instead found to exist in higher concentrations in Lake Ontario in re lation to the other Great Lakes. PCB chlorination patterns, as indicated by the average number of chlorines per biphenyl (Avg Cl/BP) were compared within species. The PCB chlorination pattern in chinook eggs wa s found to be similar for both Lake Michigan (5.47) and Lake Ontario (5.4 3). The PCB chlorination pattern in coho eggs was notably different in La ke Superior (5.78) than in Lake Ontario (5.41). I. Introduction The USEPA Great Lakes Fish Monitoring Pr ogram (GLFMP) is designed to evaluate contaminant concentrations in Great Lakes s port and top predator fish, evaluate human exposure to contaminants, and obtain new information on persistent and emerging contaminants. As a part of the GLFMP sport fish program, the purpose of this preliminary research was to investigate the utilization of salmon eggs as bioindicators of persistent and emerging contaminants in the Great Lakes. The utilization of bioindicators in persiste nt contaminant research is widely discussed in the literature. Species such as snapping tur tles and tree swallows have been widely used as reliable bioindicators of persistent contaminants in the environment (Pagano et al., 1999; Nichols et al., 1995). The utilization of fish e ggs as biomonitors of persistent contaminants has advantages including: ease and cost effectiveness of collection and analysis; large composite egg batch samples reflecting entir e spawning populations; and reduced adverse effects to the general salmon population (Miller, 1993; Miller, 1994; Pagano, 2005; Pagano, 2007).
C. Alexander & J. Pagano 16 II. Materials and Methods The Salmon River Fish Hatchery located in the Village of Altmar, Oswego County is operated by the New York State Department of Environmental Conservation (NYSDEC) to meet the sport fishery stocking needs of Lake Ontario and tributaries. Lake Ontario coho (Oncorhynchus kisutch ) and chinook ( Oncorhynchus tshawytscha ) salmon eggs were randomly sampled by NYSDEC personnel at th e Salmon River Fish Hatchery. Sampling locations for Lake Superior, Lake Michigan, and Lake Erie eggs are noted in Fig. 1. Egg samples were collected during the Fall 2006 spawning run. For the Lake Ontario samples, paired egg and muscle fillet (skin off) samples were collected from individual coho and chinook salmon before fertilization as prev iously described (Pagano, 2005b). Fig. 1: Great Lakes 2006 fish egg collection sites. Map obtained from USEPA GLNPO, 2008. Egg samples were extracted at Clarkson Univ ersity by Accelerated Solvent Extraction (ASE) using a Dionex ASE 300 and dichlorome thane solvent. The sample extract was condensed to 2 mL in a TurboVap II for Gel Permeation Chromatography (GPC) cleanup. Sample cleanup followed EPA method 3640A (G PC cleanup pesticide option) using a Waters GPC system (binary pump, Envirogel column, UV detector, and fraction collector) followed by silica gel column for separation of PCBs/OCs/PBDEs from other interferences. Adsorption column chromatography clean-up u tilized 5.5 grams of 4% deactivated silica (Sigma-Aldrich, grade 923, 100-200 mesh) placed in a chromatography column and sample extract sequentially eluted with 25 mL of hexane (fraction 1) and 25 mL hexane:dichloromethane (1:1, fraction 2). Each fraction was then concentrated in a TurboVap II to 2 mL for gas chromatographic analysis. Congener-specific PCB, hexachlorobenzene, p-p' DDE, and Mirex analyses were conducted based on capillary column procedures previously described (Pagano et al., 1995; Pagano, 2005a; Pagano, 2007b). Briefly, analytical instruments were recalibrated every five samples, with a system blank, instru ment blank, and mid-level calibration check
Evaluation of Persistent Contaminan t Levels in Great Lakes Salmon Eggs 17 solution analyzed during each analytical r un. A Hewlett-Packard (HP) Model 5890II GC with an electron capture detector (ECD Ni63) and autosampler were used for primary data acquisition. The capillary column utilized was a HP Ultra II, 25 meter with 0.22 mm id and 0.33 u m film thickness. The calibration standard used was a 1:1:1:1 mixture of Aroclors 1221, 1016, 1254, and 1260 each at 200 pg/uL, he xachlorobenzene (HCB) at 5 pg/uL, and p-p' DDE and Mirex each at 10 pg/uL (Custom Mixed Fraction #3, AccuStandard, Inc.), which allows for the analysis of 99 chromatographic zones of 132 congeners/co-eluters. PCB analyses were confirmed with a HP Model 5890 II gas chromatograph with an electron capture detector (Ni63) and autosampler using a 60 meter DB-XLB capillary column with 0.25 mm id and 0.25 u m film thickness. The calibration standard used was a 1:1:1:1:1 mixture of congener mixture sets C-CSQ-SET 1-5; 10 pg/uL per individual congener (AccuStandard, Inc., New Haven, CT ). PCB data was processed using HP ChemStation software and Microsoft Excel spreadsheet procedures such that the mole percent (congener specific and homolog), and average chlorine/biphenyl (Avg Cl/BP) values were generated. Congener-specific PBDE analyses were conducted with HP Model 5890 II gas II. Results and Discussion A comparison of select contaminant and percent lipid data for five egg sample groups is chromatograph with an electron capture detector (Ni63) and autosampler using a 60 meter DB-XLB capillary column with 0.25 mm id and 0.25 u m film thickness. The PBDE calibration standard used was an 800 pg/uL (total PBDE = 12 components) solution using the original Great Lakes Chemical Corporation (Great Lakes DE-71, CAS 32534-819) technical formulation. The DE-71 techni cal formulation mass fractions and congener identifications were confirmed with pure PBDE congener standards (BDE-MXE) purchased from Wellington Laboratories, Guelph, ON, Canada, and by mass spectrometric confirmation. presented in Table 1. As noted, total PCB c oncentrations ranged from 863.1 ng/g in Lake Michigan chinook eggs to 105.3 ng/g in Lake Superior coho eggs. Similar across-lake concentration trends are noted for total PBDE and total DDT. Concentrations in these three contaminants groups were found to be greatest in Lake Michigan, followed by L. Ontario, L. Erie, and L. Superior. In comparison, la ke trout whole fish data from the USEPA Great Lakes National Program Office (GLNPO) Great Lakes Environmental Database (GLENDA) is reported in Fig. 2. Similar to the fish egg trend noted in Table 1, total PCB concentrations for 2003 lake trout whole fish were found to be greatest in Lake Michigan, followed by L. Ontario, L. Huron, L. Erie, and L. Superior. The similarity of fish egg and lake trout whole fish tr end data appears to indicate that fish eggs have the potential to act as bioindicators of contaminants in Great Lakes fish.
C. Alexander & J. Pagano 18 Table 1. Select contaminant data fo r five egg sample groups (concentrations expressed in ng/g wet weight). LM Chinook LO Chinook LO Co ho LE Steelhead LS Coho (N=10) (N=3) (N=3) (N=7) (N=3) Total PCB 863.1 663.7 428.8 311.8 105.3 Ave Cl/BP (PCB) 5.47 5.43 5.41 5.49 5.78 % Lipid 10.3% 5.2% 3.7% 11.5% 8.9% Total PBDE 55.2 40.3 25.6 23.2 23.2 Total DDT 291.9 267.3 154.9 40.8 34.7 Mirex 1.2 26.0 19.4 0.6 0.6 OCS 0.4 4.9 3.4 0.4 0.4 ry to total PCB, PBDE, and DDT acrosstrends, Mirex is present in significantly Contra lake higher concentrations in Lake Ontario than any of the other Great Lakes. For example, L. Ontario chinook egg Mirex concentrations (26.0 ng/g wet) are significantly higher than chinook eggs from L. Michigan (1.2 ng/g wet). Mirex was released into Lake Ontario by the Hooker Chemical company from the Niagara River and, to a lesser extent, the Oswego River (Pagano, 2004). Our fish egg results are consistent with the established fact that Mirex is a contaminant associated with Lake Ontario and not significantly present in the other Great Lakes. 0 200 400 600 800 1000 1200 1400 MichiganOntarioHuronErie*Superior Great Lake Total PCB concentration (ng/g wet) Michigan > Ontario > Huron > Erie > Superior Fig. 2. Across-lake total PCB concentrati ons in 2003 lake trout whole fish. Data obtained from EPA GLNPO GL ENDA database. Lake Er ie samples are walleye. n order tovali t to I date use of eggs as envi ronmental biomonitors, it is importan demonstrate if quantitative and qualitative relati onships exist between contaminants found in eggs and contaminants found in muscles. In Fig. 3, a linear regression analysis of PCB congener concentrations in 2006 eggs vers us muscle tissue was performed. Results indicate (R2 = 0.97 for coho, and R2 = 0.98 for chinook) a very strong quantitative fit
Evaluation of Persistent Contaminan t Levels in Great Lakes Salmon Eggs 19 between egg and muscle PCB congener concentrations for both species. Although the egg to muscle relationships are different for coho and chinook, both demonstrate a very strong quantitative relationship indicating that e gg PCB congener concentrations have the potential to be highly predictive of mu scle PCB congener concentrations. y = 0.0836x + 0.0168 R2 = 0.97 y = 0.2395x 0.0582 R2 = 0.98 0.0 5.0 10.0 15.0 20.0 25.0 30.0 0.010.020.030.040.050.060.070.0 PCB congeners eggs (ng/g)PCB congeners muscle (ng/g) Coho 2006 Chinook 2006 Fig. 3. Comparative linear regression analys is for 2006 coho and chinook average muscle and average egg PCB congener concentrations Data adapted from Pagano, 2007a. PCBs in the envir ich the number of chlorines substituted per biphenyl ma y vary from 1-10. In Fig. 4, PCB assigned onment are a complex mixture of 209 possible congeners, in wh congener numbers are displayed by mole perc ent in order to illustrate PCB chlorination patterns, independent of concentration in each fish-egg group. The lower assigned PCB congener numbers correspond to lower-chlor inated PCB congeners, and the higher assigned PCB congener numbers correspond to high er-chlorinated PCB congeners. In Fig. 4a, visual observation of L. Ontario and L. Mi chigan chinook egg data demonstrates very similar PCB chlorination patterns. As contrasted in Fig. 4b, a comparison of the L. Ontario and L. Superior coho data demonstrates a ma rkedly different (and higher-chlorinated) PCB chlorination pattern in L. Superior eggs.
C. Alexander & J. Pagano 20 0 2 4 6 8 10 12 14 16 15913172125293337414549535761656973778185899397 Assigned PCB Congener #Mole % Lake Michigan Lake Ontario (4a.) 0 2 4 6 8 10 12 14 16 15913172125293337414549535761656973778185899397 Assigned PCB Congener # Mole % Lake Superior Lake Ontario (4b.) Fig. 4. Congener-specific P CB mole percent comparison of L. Ontario versus L. Michigan chinook eggs (4a.) and L. Ontario versus L. Superior coho eggs (4b.). In Fig. 5, PCB homologs are displayed by mo le percent and average number of chlorines per biphenyl in order to represent the PCB ch lorination for each sample group. Homologs are subcategories of PCB congeners having equa l numbers of chlorine substituents. As noted in Fig. 5a, similar mid-chlorinate d homolog patterns and Avg Cl/BP values are observed in both L. Ontario (5.43) and L. Michigan (5.47) chinook eggs. As contrasted in Fig. 5b, notably different homolog chlori nation patterns and Avg Cl/BP values are observed for L. Superior (5.78) and L. Ontario (5.41).
Evaluation of Persistent Contaminan t Levels in Great Lakes Salmon Eggs 21 0 5 10 15 20 25 30 35 40 45 Cl 1Cl 2Cl 3Cl 4Cl 5Cl 6Cl 7Cl 8Cl 9 PCB homologsMole % Lake Ontario (5.43 Cl/BP) Lake Michigan (5.47 Cl/BP) (5a.) 0 5 10 15 20 25 30 35 40 45 Cl 1Cl 2Cl 3Cl 4Cl 5Cl 6Cl 7Cl 8Cl 9 PCB homologsMole % Lake Ontario (5.41 Cl/BP) Lake Superior (5.78 Cl/BP) (5b.) Fig. 5. Mole percent PCB homolog comparis on for L. Ontario versus L. Michigan chinook eggs (5a.) and for L. Ontario versus L. Superior coho eggs (5b.). Determination of the qualititativ e relationship between PCB c ongeners found in fish eggs and muscle tissue is important for evaluating the reliability of fish eggs as biomonitors. In Fig. 6, congener-specific PCB mole pe rcent values for 2006 coho and chinook were submitted to a linear regression analysis for egg and muscle tissue data. A strong qualitative fit between egg and muscle tissue is indicated for both coho (R2 = 0.98) and chinook (R2 = 0.98). These results qualitatively indicate that the contaminants found salmon eggs are highly predictive of the contaminants found in muscle tissue.
C. Alexander & J. Pagano 22 R2 = 0.98 R2 = 0.98 0.0 2.0 4.0 6.0 8.0 10.0 12.0 0.01.02.03.04.05.06.07.08.09.010.0 PCB congeners eggs (mole %)PCB congeners muscle (mole %) Coho 2006 Chinook 2006 Fig. 6. Comparative linear regression analysis for 2006 coho and chinook average muscle and average egg PCB mole pe rcent. Data adapted from Pagano, 2007a. IV. Conclusions Results from preliminary data indicate an acro ss-lake contaminant concentration trend for total PCB, PBDE, and DDT in which concen trations are highest in Lake Michigan, followed by L. Ontario, L. Erie, and L. Supe rior. Our data suggests that fish eggs are highly predictive both quantitatively and qualitativ ely of congener-specific PCBs found in muscle tissue. Although preliminary, our results to date suggest that fish eggs have the potential to be utilized as bioindicators of persis tent contaminants in the Great Lakes. V. Acknowledgements The authors would like to thank the USEPA Great Lakes Fish Monitoring Program for funding this research. We also thank Phil Hulbert and Andy Gruelich, NYS Department of Environmental Conservation for providing samp les and technical advice. We would also like to thank Beth Murphy EPA Project Manager, Greg Sumner SUNY Oswego ERC for assistance in preparing samples, and Be rnie Crimmins Clarkson University for extracting samples. VI. References Miller, M. A. (1993). Maternal transfer of organochlorine compounds in salmonines to their eggs. Can J. Fish. Aquat. Sci. 50:1405-1413. Miller, M. A. (1994). Organochlorine concentration dynamics in Lake Michigan chinook salmon. Arch. Environ. Contam. Toxicol. 27: 367-374.
Evaluation of Persistent Contaminan t Levels in Great Lakes Salmon Eggs 23 Nichols, J. W., Larsen, C. P., McDonald, M. E., Niemi, G. J., and Ankley, G.T. (1995). Bioenergetics based model for accumulation of polychlorinated biphenyls by nestling tree swallows, Tachycineta bicolor Environ. Sci. Technol 29(3): 604-612. Pagano, J., Scrudato, R., Roberts, R. and Bemis, J. (1995). Reductive dechlorination of PCB contaminated sediments in an anaerobic bioreactor system. Environmental Science & Technology 29:10 2584-2589. Pagano, J. J., Rosenbaum, P., Roberts, R., Sumner, G., and Williamson, L. (1999). Assessment of maternal contaminant burde n by analysis of snapping turtle eggs. Journal of Great Lakes Research 25(4):950-962. Pagano, J., (2004). Identification of a s ource of Aroclor 1268 and mirex to the Oswego River Area of Concern. In: Proceedings of the 228th National Meeting American Chemical Society Philadelphia, PA, USA, Vol. 44 No. 2: 525-530. Pagano, J. (2005a). Deposition and ambient c oncentrations of PBTs (PCBs, OCs, PBDEs, and Dioxins/Furans) in support of the La ke Ontario air deposition study (LOADS), Quality Assurance Project Plan (QAPP). Great Lakes Commission-Great Lakes Air Deposition (GLAD) Program, Ann Arbor, MI. June 28, 2005 Version 01. Pagano, J. (2005b). Utilization of salm onid eggs as bioindicators of organohalogen pollutants in Lake Ontario. Proceedings of the 230th National Meeting American Chemical Society Washington, DC, USA, Vol. 45 No. 2: 337-342. Pagano, J. (2007a). Polybrominated diphenyl ethers in Lake Ontario salmonid fillets and eggs. Proceedings of the 234th National Meeting American Chemical Society Boston, MA, USA, Vol. 47 No. 2: 223-229. Pagano, J. (2007b). Deposition and ambient concentrations of PBTs (PCBs, OCs, PBDEs, and Dioxins/Furans) in support of the La ke Ontario air deposition study (LOADS). Final Report: Great Lakes Commission-Great Lakes Air Deposition (GLAD) Program, Ann Arbor, MI. December 20, 2007. United States Environmental Protection Agency GLNPO, (2008). Great Lakes Monitoring. http://www.epa.gov/glnpo/monitoring/fish/index.html accessed March 11, 2008.