THE ANTAGONISTIC EFFECT OF SELENIUM ON MERCURY UPTAKE BY FRESHWATER FISH. Nelson BELZILE¹, Yu-Wei CHEN¹ & John GUNN², ¹Department of Chemistry & Biochemistry, ²Cooperative Freshwater Ecology Unit, Laurentian University, Sudbury, Ontario, P3E 2C6 Canada. E-mail: belzile@nickel.laurentian.ca

ABSTRACT

The concentrations of Se and Hg were determined in muscles of two freshwater fish species, namely perch (Perca flavescens) and walleye (Stizosedion vitreum). Samples of various sizes were collected from ten lakes of Northern Ontario with various chemical and limnological nature. The correlation analysis revealed a clear and strong antagonistic effect between Se and Hg in muscles for both perch (r= - 0.790) and walleye (r = - 0.973), with the concentrations of Hg decreasing exponentially as Se increase. Total Se concentrations of the lakes were linearly correlated to Se in fish muscles. Hg concentrations in fish from lakes near the Sudbury smelters were well below average values for fish in boreal shield lakes of this region.

INTRODUCTION

The phenomenon of biomagnification of the toxic element mercury through aquatic food chain is well known (Jackson, 1991). It has also been found recently that bioaccumulation of Hg can be affected by several environmental factors including pH, dissolved organic carbon (Wiener et al., 1990) and the presence of other pollutants such as PCB ( Cabana et al., 1994) Since one of the first study (Parizek & Ostadalova, 1967) reported an antagonistic effect of Se on Hg, many papers have been published on the subject (e.g. Koeman et al., 1973: Turner & Rudd, 1983; Barghigiani et al.,1991; Nuutien & Kukkonen, 1998). However there is not unanimity on the phenomenon until now. Two reasons can be used to explain that. First the chosen biological species in the trophic chain vary greatly from macroalgae to shark and their living habitats differ from freshwater ecosystems to marine environments. The second reason is that many studies are conducted under abnormally high concentrations of Se with short periods of exposure to the selected biological species in order to reduce experimental expenses. This could unrealistically change the normal processes of biological assimilation of trace metals as compared to natural aquatic conditions. Significant quantities of Se were introduced in water and sediments of lakes of the Sudbury area as a result of intense mining and smelting activities (Nriagu & Wong, 1983). However, a study on Hg assimilation by fish of the same area revealed that Hg concentrations in crayfish and perch increased with increasing distances between sampled lakes and smelters (Wren & Stokes, 1988). The objective of this study was to investigate the Se - Hg antagonistic effect using fish living in natural aquatic systems

METHODS

Lakes located between 3 to 300 km from the metal smelters of Sudbury and exhibiting a wide range of limnological and chemical characteristics were selected for the study. The detailed information for these lakes is given in Table 1. Yellow perch (Perca flavescens), a species that feeds mainly on zooplankton and benthic invertebrates and walleye (Stizosedion vitreum), a piscivore were selected in our investigation. Fish sampling was conducted in June-July of 1996 and April-October of 1997 when 9-20 fish of various size classes were collected from each lake. Samples of skeletal muscle tissues were dissected under clean room conditions, deeply frozen, then freeze-dried and ground into a fine powder. A precise 0.1 g fish muscle was weighed and digested with 1.0 mL 30% H₂O₂ (AG) and 2.5 mL of concentrated HNO₃ (Trace Metal Grade) at room temperature over night, then microwave-digested 10 times 1min at 720W. The quality of the digestion and analysis procedure was controlled by using the two Certified Reference Materials (CRM) DORM-2 (Dogfish Muscles) and TORT-2 (Lobster Hepatopancreas) from NRC-Canada at a frequency of one CRM digestion per 10 digested sample (rel error < ± 8 % for both Hg and Se). Triplicate digestion for each fish species from each lake was carried throughout all analyses (RSD < 5% and < 6% for Hg and Se respectively). Total Se and Hg in fish tissues were determined using graphite furnace atomic absorption spectrometry and cold vapour - atomic fluorescence spectrometry respectively. Instrumental variations were < 3 % and < 4% for Hg and Se respectively). Surface water samples were collected in carefully acid cleaned high density polyethylene bottles. The samples were filtered, subdivided, acidified, pre treated and stored in Teflon bottles in a freezer within less than 48 hours after collection. Determination and speciation of Se in lake water samples were done with hydride generation - atomic fluorescence spectrometry. The relative standard deviations were <5 %.

RESULTS AND DISCUSSION

The concentrations of total Se and Hg²⁺ in lakes are given in Table 1. The lakes are listed according to their decreasing distance from the Sudbury smelters.

Table 1. Chemical characteristics of the lakes.
	DOC mg/L	pH	Cu^2+ mM	Ni^2+ mM	Zn^2+ mM	Cd^2+ nM	Ca^2+ mM	Mg^2+ mM	SO4^2-mM	Cl^- mM	Tot Se nM	Hg^2+ pM
Hannah	3.3	7.2	0.44	2.39	0.18	1.78	0.26	0.15	0.22	1.88	9.21	20
Ramsey	3.0	7.7	0.22	1.57	0.06	0.89	0.38	0.19	0.22	1.69	7.74	Na
Laurentian	4.3	6.5	0.24	0.95	0.06	0.89	0.09	0.06	0.07	0.51	5.29	Na
Bethel	6.3	9.2	0.08	0.43	0.02	0.89	0.43	0.27	0.08	1.95	2.94	35
Vermilion	4.1	7.5	0.08	0.31	0.07	1.78	0.24	0.09	0.15	0.11	1.45	19
Whitson	3.4	6.7	0.27	2.39	0.17	0.89	0.18	0.08	0.19	0.75	5.97	26
Michiwakenda	8.4	7.2	0.04	0.02	0.05	nd	0.23	0.07	0.05	0.09	1.11	18
Geneva	3.3	6.7	0.02	0.03	0.03	0.89	0.06	0.03	0.06	0.07	1.22	18
Larder	6.2	8.0	0.24	0.74	0.08	na	0.34	0.18	0.19	0.14	1.67	180
Long	na	7.7	0.12	0.01	0.12	nd	0.51	0.34	0.04	0.13	1.20	9
na : not available.
nd : not detectable.

For each lake, the total concentration of Se and Hg (nmol/g dry wt) was separately plotted against the fish fork length. The mathematical expressions between Se or Hg concentrations and fish fork length and their corresponding coefficients of correlation were derived from the least squares method. Concentration values corresponding to fish fork lengths of 100 mm and 300 mm for perch and walleye respectively were selected. With this procedure, the age of fish could be approximately controlled and standardised. Total concentrations of Hg at the selected fish fork length were then plotted against those of Se for each species from the different lakes. Figure 1 represents the relationships of Se and Hg concentrations in perch and walleye fish muscles respectively. A mathematical treatment demonstrated that the best fitting for data of both perch and walleye could be obtained with the exponential function. Figure 1 clearly demonstrates that the concentrations of Hg decline exponentially with the increase of Se concentrations in fish muscles. The coefficients of correlation, r, between Hg and Se are - 0.6785 and - 0.9638 for perch and walleye respectively. The molar Se/Hg ratio varied from 14 (Long

lake) to 1744 (Hannah Lake) and 0.48 (Long Lake) to 163 (Ramsey Lake) in fish muscles for perch and walleye respectively. A remarkable antagonistic relationship between Se and Hg is confirmed for fish of different species living in largely different natural aquatic systems.

There is some controversy on the interaction between Hg and Se. Positive linear correlations between Hg and Se in the livers of marine mammals are reported (e.g. Koeman et al., 1973) whereas other studies find no significant correlation between Hg and Se in the muscles of some marine species (Barghigiani et al., 1991). Significantly negative correlations between concentrations of Hg in Oligochaeta worms and that of Se in sediments (Nuutien & Kukkonen, 1998) and between Hg and Se in perch muscles before undergoing a Se treatment of the lakes (Paulsson & Lundberg, 1991) are reported. However, in the study on perch, the inverse relationship between Hg and Se in muscles disappeared one and two years after the lakes were initially treated with sodium selenite.

Our studies also show that the concentration of total Se in fish tissues and that of dissolved total Se in lake waters are linearly correlated (Figure 2). Consequently the concentration of Hg in fish tissues also decreases exponentially with an increase of total Se in lake water (r= - 0.807 and – 0.951 for perch and walleye respectively). Recent studies have shown that total Hg concentrations in lakes of the Sudbury area vary widely from several tens to several hundred ng/L. The variation largely depends on the location of individual lakes, and on the analytical methods used to define total mercury. It is clear though that the concentrations in lakes close to the Sudbury smelters are significantly higher than many of the reported values (2 to 20 pM or 0.5 to 4 ng/L) for the lakes located in remote or boreal forest regions (Paulsson & Lundberg, 1991; Watras et al., 1994).

Studies on mercury assimilation indicate that accumulation of Hg is astonishingly high in fish tissues from pristine remote waters (Lindqvist, 1991; Watras & Bloom, 1992). In a study on eight Swedish headwater lake ecosystems of a boreal forest region (Lindqvist, 1991), concentrations of 0.9 - 9.9 nmol /g dry of Hg were measured in perch muscles while total Hg in those lake waters ranged from 4 - 120 pM (0.8 - 24 ng/L). Watras & Bloom (1992) found that even in water containing very low concentrations of Hg (5.5 pM or 1.1 ng/L), young yellow perch could still accumulate as high as 1.4 nmole Hg per g of dry muscle. These values are remarkably higher than those in perch collected in lakes close to Sudbury. Though the mechanisms of this antagonistic effect is not yet clear, our study has shown a strong evidence that the presence of Se in water is positively correlated to the concentration of Se in fish muscles and that the presence of Se in fish muscles has significantly reduced the accumulation capacity of Hg in fish muscles. Our study confirms that the presence of Se in aquatic systems is an important variable in the equation of bioaccumulation of mercury in aquatic biological species.

Even if this antagonistic effect between Se and Hg holds, one should be very careful in using Se an ameliorating ingredient against Hg, as Se possesses a complex nature and a serious potential toxicity. Teratogenic effects of Se are reported for freshwater fish living in a lake used as a coal power impoundment (Lemly, 1993). It is also noticed in this study that the concentrations of Se in apparently normal fish are very close to those measured in abnormal specimens. The complexity of the problem will require more research to be conducted in natural ecosystems with studies considering specifically the biochemical and physiological aspects of the various components of the food chain.

REFERENCES

Barghigiani G, Pellegrini D, D’Ulivo A, De Ranier S (1991), Mar. Pollut. Bull. 22:406-409.

Cabana G, Tremblay A, Kalff J, Rasmussen JB (1994), Can. J. Fish. Aquat. Sci. 51:381-389.

Jackson TA (1991), Can. J. Fish. Aquat. Sci. 48:2449–2470.

Koeman JH, Peeters WHM, Koudstaal-Hol CHM, Tjioe PS, De Goeil JJM (1973), Nature 245:385-386.

Lemly AD (1993), Ecotox. Environ. Safety 26:181-204.

Lindgvist O (1991), Water Air Soil Pollut. 55:143-177.

Nriagu JO, Wong HKT (1983), Nature 301:55-57.

Nuutien S, Kukkonen JVK (1998), Biogeochem. 40:267-278.

Parízek J, Ostádalová I (1967), Experientia 23: 142-143.

Paulsson K,.Lundbergh K (1991), Water Air Soil Pollut. 56:833-841.

Turner MA, Rudd JW (1983), Can. J. Fish. Aquat. Sci. 40:2228-2249.

Watras CJ, Bloom NS (1992), Limnol. Oceanogr. 37:1313-1318.

Watras CJ et al., (1994),. In: Mercury, Pollution: Integration and Synthesis (CJ Watras, JW Huckabee, Editors) Lewis Publ.

Wiener JG, Fitzgerald WF, Watras CJ, Rada RG (1990), Environ. Toxicol. Chem 9:909-918.

Wren CD, Stockes PM (1988), Ambio 17:28-30.