Environ. Sci. Technol. 2001, 35, 4830-4833
Equilibrium Partition Theory Applied to PCBs in Macrophytes CLAIRE VANIER AND DOLORS PLANAS* GEOTOP, Universite´ du Que´bec a` Montre´al, C.P. 8888 Succ. Centre-Ville, Montre´al, Que´bec H3C 3P8, Canada MICHEL SYLVESTRE
Introduction One of the most important goals of ecotoxicology is to determine the fate of anthropogenic contaminants in the environment. However, the physical-chemical and biological processes involved in the distribution of contaminants are extremely complex. The equilibrium partitioning theory (EPT) allows one to reduce this complexity. According to the EPT, the partitioning of a given compound between two phases is governed by the compound’s affinity for each of the phases. In a multiphase compartment system, such as an aquatic ecosystem, the affinity of hydrophobic contaminants for environmental compartments (i.e., water, sediments, and biota) can be conceptualized in terms of the partitioning of the compounds among the aqueous, the organic carbon, and the lipid phases of each compartment (1, 2). Thus, for a given dissolved concentration of a compound in water (Cw), the concentration in the sediments (Cs) is calculated by
(1)
where Koc is the organic carbon-water partition coefficient and foc is the percentage of organic carbon in sediments. Given that Koc is related to the 1-octanol-water partition coefficient (Koc ) xKow) of the compound (3-5), the concentration in the organic fraction of sediments (Cso ) Cs/foc) * Corresponding author e-mail:
[email protected]; telephone: (514)987-3000; fax: (514)987-3635. 9
(2)
where x is the constant of proportionality between Koc and Kow. In the biota, assuming that the partition coefficient between water and lipids is equal to Kow (4-6), the equilibrium concentration in lipids (Cli) is calculated by
(3)
From eqs 2 and 3, it follows that
The applicability of the equilibrium partition theory (EPT) for PCB accumulation in submerged macrophyte shoots was investigated using field data. The equilibrium state of PCBs between macrophyte shoots and sediments was verified by testing the direct proportionality (slope ) 1 in a log-log relationship) between congener concentrations in the lipids of macrophyte shoots (Cli) and in the organic fraction of sediments (Cso), using the slope-range method. A significant proportionality was found between Cli and Cso (slope of the log-log relationship: 0.978 ( 0.041; R2 ) 0.847; P < 0.001). The biota-sediment accumulation factor (BSAF) predicted from this relationship (3.74, SE: 1.14) was in the 1-4 range that is suggested in the literature. These results indicate that, in our study sites, EPT is applicable to PCBs in macrophyte shoots. This opens interesting perspectives in environmental monitoring.
4830
Cso ) xKowCw
Cli ) KowCw
Universite´ du Que´bec, Institut National de Recherche Scientifique-Sante´, 245 Boul. Hymus, Pointe-Claire, Que´bec H9R 1G6, Canada
Cs ) KocfocCw
is calculated by
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 24, 2001
Cli )x Cso
(4)
Cli ) BSAF Cso
(5)
or
where BSAF is the biota-sediment accumulation factor. Given that the Kow canceled out of the equation, the BSAF should be constant among different hydrophobic compounds (2, 5). The BSAF should also be relatively constant regardless of species or sites since variations in concentrations among species and sediments will be reflected in Cli and Cso, respectively. It follows that, at equilibrium and when no significant metabolic activity nor phase-transfer resistance occurs, Cli is directly related to Cso regardless of the characteristics of the compounds, sites, or species. This proportionality between Cli and Cso is the basis for the application of the EPT to the partitioning between biota and sediments. The great interest of EPT is its potential for predicting the hydrophobic contaminant concentration in any organism, using only the proportion of lipids in this organism and the concentration of the contaminant in the organic fraction of the sediments. The EPT has often been used as a theoretical basis for experimental (7-9) and field investigations (2, 1012). However, in several of these studies, there was a lack of adjustment between the observed concentrations of hydrophobic contaminants in organisms and those predicted by the EPT, when applied to the partitioning either between biota and water (7, 8, 13) or between biota and sediments (11, 12, 14, 15). Several explanations have been suggested to account for these observations, the most frequently cited being a selective accumulation of the compounds due to differences in assimilation/degradation/excretion patterns according to the compound’s characteristics (8). These differences could also be magnified along the food web (4, 15). Growth diluting effects, phase transfer resistance, and time to reach equilibrium were also suggested (2, 8, 13). The majority of previous studies have examined the accumulation of hydrophobic compounds in fish or aquatic invertebrates. Few of them have evaluated the potential of the EPT for the prediction of contamination in aquatic primary producers, particularly in natural conditions despite the fact that primary producers present several advantages for an EPT validation. The first advantage is that there is no interference caused by biomagnification in primary producers. Macrophytes in particular would be well-suited to the study since they have a lower growth rate as compared to algae and are sessile. Rooted macrophytes also have access to contaminants buried in sediments and may be involved 10.1021/es001519y CCC: $20.00
2001 American Chemical Society Published on Web 11/16/2001
in the mobilization of PCBs from the sediments. Experimental studies have shown uptake by roots, translocation between roots and shoots (22, 23), and uptake in plants from the water (9, 22) of organic contaminants. In the field, there has also been a growing interest with regards to the role of macrophytes in the redistribution of organic contaminants in aquatic ecosystems (17, 21, 24). It is often assumed that, in macrophytes, metabolism of hydrophobic contaminants is negligible and that contaminant accumulation is a passive process governed by the compound’s affinity for the lipids of plants (9, 16). However, controversy remains with regards to the importance of the lipid fraction as the accumulation site of hydrophobic contaminants in the biota (1, 17, 25, 26, 27). The objective of the present study was to evaluate the applicability of EPT to submerged macrophyte shoots using field PCB data. The applicability of the EPT was evaluated by testing the assumption of a direct proportionality between Cli and Cso. We chose this approach because, in addition to validating the use of BSAF for the prediction of PBC accumulation in macrophytes, it allowed for an evaluation of the correlation between Cli and Cso. The assumption of the independence of the PCB congener BSAF on their Kow values was also tested. Our hypotheses were tested using the data from a previous study (17) in which the results suggested that the PCB accumulation in macrophytes could have been a result of congener partitioning between sediments and macrophyte shoots.
Methods Details of the sampling procedure and PCB congener analysis were previously published (17) and will thereby be only briefly described here. The study was carried out in the mid-summer of 1992, in western part of Lake Saint-Franc¸ ois (74°40′ W, 45°00′ N), one of the most PCB-contaminated areas in the Saint Lawrence River (18). Samples of macrophyte shoots and of sediment cores (top 5-cm analyzed) were taken by scuba divers at three stations in the lake (stations were approximately 9 km from each other). At each station, samples were collected at five points along the perimeter of a circle (approximately 60 m). Macrophyte shoots of pooled species, mostly Myriophyllum sp. and Elodea canadensis, were cut at their base and cleaned of their epiphytes. The macrophyte and sediment samples were kept frozen (at -20 °C) in hexane-acetone-cleaned glass containers until analyzed. For PCB analysis, two aliquots (10 g wet weight) of sieved sediments (2-mm mesh) were used. For macrophytes, two aliquots (10 or 20 g wet weight) of crushed shoots were dried with sodium sulfate before extraction. The PCBs were extracted from each duplicate by sonication in hexane. The sediment extracts were treated with activated copper, and all extracts were purified in a Florisil column with hexane. PCB analyses were done separately on each duplicate, using a Perkin-Elmer Sigma-2 gas chromatograph equipped with an electron capture detector (GC/ECD). Individual quantification of 34 congeners (see ref 17 for the list of analyzed congeners) was made by comparing with the CLB-1 mixture standards (National Research Council of Canada, Halifax). A total of 10 g wet weight of sieved sediments was used to estimate the dry weight (70 °C) and the ash-free dry weight (500 °C, 1 h) of the samples. The loss on ignition was assumed to represent organic material in samples. The lipids of macrophytes (hexane-extractable lipids) were evaluated from shoot extracts. Expressed on the basis of lipid weight, the sum of the congener concentrations in macrophytes ranged from 9.6 to 405.0 mg kg-1. In sediments, the sum of the congener concentrations ranged from 4.4 to 256.0 mg kg-1, organic weight. Proportionality between Cli and Cso was tested by comparing the slope value found between log Cli and log Cso to
FIGURE 1. Relationship between PCB congener concentrations in macrophytes (lipid weight; log Cli) and in sediments (organic weight; log Cso) estimated from the slope-range method (SR). unity (Student’s t test, P ) 0.05; JMP.3, Statistical Analysis System Institute). Given that the usual method to estimate the slope of a relationship (ordinary least-squares: OLS) only minimizes the sum of the residual square values of the predicted variable and that, in natural conditions, regressor variables are also subject to random fluctuations, the slope between log Cli and log Cso was estimated using both the OLS and the slope-range method (SR). In the SR method, the parameters of the relationship are estimated using an “instrumental variable” to quantify the natural error variability in the regressor variable (for details on the method, see ref 20). The instrumental variable we used to estimate the variability of Cso was the wet weight congener concentrations that we had analyzed in a deeper sediment layer (10-15 cm) of the same cores (see ref 17). The assumption of the independence of the BSAF on the congener characteristics was tested by relating the congener concentration ratio (log BSAF ) log Cli/Cso) to the log Kow of congeners (ranging from 4.24 to 8.04; log Kow values taken from ref 19).
Results and Discussion The relationship between log Cli and log Cso that was estimated using the OLS method was highly significant (P < 0.0001; R2 ) 0.839), but the slope (0.903 ( 0.037; 95% confidence limit) of the relationship was
