Zebra mussels (Dreissena polymorpha) originating from a drinking water reservoir were exposed at seven sites in metal polluted watercourses in Flanders (Belgium), including to a cadmium and zinc pollution gradient. At each site one cage, containing twenty-five mussels, was exposed for six weeks. Mussels were collected after 10, 21 and 42 days. At each sampling, five mussels were taken from each cage and analysed for metal content and condition. Significant metal accumulation in zebra mussel was measured as a function of time at all sites for at least one metal. However, a steady state in accumulation was not always reached. Depending on the site, extremely high levels of cadmium, copper, nickel and zinc were measured in the mussels. Despite differences in metal accumulation, no significant differences in two condition indices were observed among the different sites. When data of all the sites were pooled there was no significant relationship between the condition and the metal load in the mussels, although a threshold could be distinguished. However, when this relationship was examined for the separate rivers, significant relationships were found. This indicates that besides metal pollution other environmental factors such as food availability affect the mussel condition.
Due to human activities, trace metals continuously enter the aquatic environment and as a consequence may become an environmental problem due to their persistent nature and toxicity to aquatic biota. In Flanders (Belgium), the quality of surface waters has improved over the last decade, as a result of increased municipal wastewater treatment (De Cooman et al., 2003). As a result, thanks to increased oxygen levels, partial restoration of fish communities was established in several watercourses (Bervoets et al., 2002). Nevertheless, in some cases metal pollution remains a serious problem, representing a significant risk to aquatic biota (Bervoets and Blust, 2003). To assess the risk of micropollutants (such as metals) present in the environment, it is necessary to know their bioavailability to aquatic organisms. The bioavailability of dissolved and sediment-bound metals is determined by the physical, chemical and biological characteristics of the overlying and interstitial water and the sediment (Luoma, 1983; van Griethuysen et al., 2003). Trace metal contaminants can be measured in the tissues of native organisms in combination with measurements of their condition. The feasibility of this so-called passive biological monitoring program depends on the presence of the same species at all sites in sufficient biomass or abundance. This problem can be overcome by using translocated organisms, called active biomonitoring. Instead of collecting indigenous organisms, animals can be collected at a clean reference site or cultured in the laboratory, and can subsequently be exposed at the sampling sites in cages (Salazar and Salazar, 1997). Zebra mussels (Dreissena polymorpha) fulfil the requirements of a good biomonitor for freshwater ecosystems. They have a relatively long life span, are sedentary and resistant to various types of pollution without suffering mortality. Their high filtration rate favours the bioaccumulation of contaminants, both from water and particulate material (Kraak et al., 1991; Hendriks et al., 1998). A previous study showed that levels of micropollutants in transplanted and indigenous zebra mussels were comparable when exposed for six weeks (Bervoets et al., 2004). This indicates that transplanted zebra mussels can give a reliable picture of the bioavailability of micropollutants present in aquatic ecosystems.
The aim of this study was to evaluate the bioavailability of trace metals to zebra mussels exposed at metal polluted watercourses. Bioaccumulation of trace metals was measured as a function of exposure time and related to the condition of the mussels.
Materials and methods
Test organism and field exposure
Zebra mussels (D. polymorpha) were collected at the drinking water reservoir of the Antwerp Drinking Water Company (Duffel, Belgium) in November 2001 and selected by length (17–20 mm). Random groups of 25 individuals were placed in polyethylene cages that allowed free circulation of water. Each cage consisted of two pond baskets tied together. The dimensions of the resulting cages are 11 × 11 × 22 cm with a mesh size of 2 × 4 mm. Cages were anchored by attaching a stone with a rope and secured to a peg in the bank with another 30 cm rope. All cages were free floating at a depth of 30 cm below the water surface.
Study area and sample processing
Eight sampling sites were selected in the river basin of the Nete, located in Flanders (Belgium), which is part of the Scheldt river basin (Figure 1; Table 1). All sites are known to have high to very high levels of at least one trace metal. Four of the selected sites (sites 1 to 4) were in the same river and are situated along a cadmium and zinc pollution gradient (Bervoets and Blust, 2003). At each site, a cage containing 25 mussels was deployed in situ at the end of November 2001. Five mussels were removed from each cage after 10, 21 and 42 d of exposure and survival was checked.
Two different indices were used to assess the effect of stress on the physiological integrity of the mussels. Before the start of the field exposure, 10 mussels from the reference site were dissected and their shell length (to 0.1 mm) and wet soft tissue weight (ww, to 0.1 mg) were measured. After drying at 60°C for 24 h, dry soft tissue weight (dw) and dry shell weight (sw) were determined on a Mettler H54 balance to the nearest 0.1 mg. The same measurements were taken on mussels after the field exposure. These measurements were used to calculate two condition indices: Tissue Condition Index (TCI = tissue dry wt./shell wt.) (Mersch et al., 1996; Soto et al., 2000) and the Shell Condition Index (SCI = tissue dry wt.*100/shell length; Payne et al., 1999).
In the laboratory, the soft body parts of the mussels were removed from the shell, byssus threads were removed and tissue was rinsed with deionised water. Samples for metal analysis were placed in pre-weighed acid washed polypropylene vials and dried for 24 h at 60°C. Subsequently, the biological material was digested in a microwave oven, by adding a mixture (5:1) of nitric acid (70%) and peroxic acid (30%) following the protocol as described by Blust et al. (1988). Digested samples were frozen at −20°C until further analysis. Silver, aluminum, arsenic, cadmium, cobalt, chromium, copper, mercury, manganese, nickel, lead and zinc were measured in the samples with either an Inductively Coupled Plasma-Atomic Emission Spectrometer (ICP-AES) or an Axial Inductively Coupled Plasma-Mass Spectrophotometer (ICP-MS, Varian Liberty Series II, Victoria, Australia) equipped with a microconcentric groove nebuliser (De Wit and Blust, 1998). Concentrations of the metals in the tissues were calculated on a dry weight basis and expressed as μ g g−1. Analytical accuracy was determined using certified reference material of the Community Bureau of Reference of the European Commission (Geel, Belgium); standard for trace elements in river sediment; Certified Reference Material 320 (CRM 320) and mussel tissue (CRM 278). Recoveries were within 10% of the certified values.
Relationship between condition indices and metal levels
To relate condition indices to tissue metal levels, the concentration of a single metal was divided by the mean concentration of that metal measured in the mussel tissue from clean sites in Flanders (Bervoets et al., 2004) and standardised values were then added. To this value the following formula was applied: MLj = [Σi(Cij/Cri)]/N, where MLj is the relative metal load in the mussel tissue at site j, Cij is the concentration of metal i at site j, Cri is the concentration of metal i at the reference site and N is the number of metals measured (modified after Meregalli et al., 2000). When Cij < Cri, the (Cij/Cri) was considered 1. Thus the relative ML is a measure of the enrichment of the mussel tissue with the measured metals, compared to those in zebra mussel from unpolluted sites. As a consequence, when no enrichment occurred, ML = 1.
Analysis of variance (ANOVA, with as post-hoc Duncan's multiple range test), linear and non-linear regressions were used to analyse the data. All data were tested for homogeneity of variance by the log-anova test and for normality by the Kolmogorov-Smirnov test for goodness of fit. If conditions for ANOVA were not fulfilled, the Kruskal-Wallis and Dunn's test were used. Significance levels of tests are indicated by asterisks according to the following probability ranges: * p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001. Statistical methods used are outlined in Sokal and Rohlf (1998). All statistical analysis were performed using the software package STATISTICA (StatSoft Inc., Tulsa).
Survival of the mussels
Table 1 gives the mortality in the cages as a function of time, along with the oxygen level and pH in the water. At site 3, the cage had disappeared after 6 weeks of exposure. At all sites with the exception of sites 5 and 6 the cumulative mortality was more than 50%. Mortality was always highest between 3 and 6 weeks of exposure. Only at three sites (sites 4, 5 and 6), was survival sufficiently high to take samples after 6 weeks of exposure. The high mortality could not be attributed to decreased oxygen levels since no significant relationship between oxygen levels and survival was found. At most sites however, where mortality was high, accumulation of sediment particles in the cages was observed between 3 and 6 weeks of exposure (Table 1).
Metal accumulation in zebra mussel
The mean metal concentrations in zebra mussels, at the different sites after 21 d of exposure are given in Table 2. At all sites levels of all metals were detectable in zebra mussel, with the exception of As at site 2. With the exception of Al, ANOVA indicated significant differences among sites for all metals measured. Highest metal levels were measured in mussels exposed at site 5 (Co, Cr, Hg, Pb), site 6 (Ag, As, Cu, Ni), and site 1 (Al, Cd, Zn). Lowest levels were measured in zebra mussels from the reference site and from site 4. Figure 2 shows the accumulation of some selected trace metals as a function of time; site 4 (Zn), site 5 (Ni) and site 6 (Cu). For site 5, metals that followed the same trend were Ag, Cd, Cu, Co and Hg and for site 6; Ag, As, Cd, Co and Ni. Figure 3 gives the accumulated concentration of Cd, Hg, and Zn along the pollution gradient in the Nete River after 3 weeks of exposure. Tissue concentrations of Cd and Zn followed the same concentration gradient of the river. An opposite gradient was observed for Hg with site 4 having highest concentration.
The mean condition indices of zebra mussels, at the different sites after 21 d of exposure are summarised in Table 3. No significant differences for the indices were found among the sites. Likewise, when only the pollution gradient was considered, no significant differences were found among the sites and the indices did not follow the pollution gradient. At individual sites however, the condition indices changed with time, lowest condition occurring at the longest period of exposure. At sites 4, 5 and 6, it was possible to examine the TCI for the maximum 6 weeks (Figure 4). Initially when deployed in the field the condition improved compared to the reference site. However, after 3 weeks of exposure, the TCI and SCI decreased to a level equal or lower than at the reference site. The same trend was observed at the other sites. When the data of all sites were pooled, significant correlations were found between TCI and SCI (r2 = 0.49; p < 0.001).
Relationship between metal load and condition indices
When individual condition indices (TCI) were related to the individual metal body burdens in the mussels (ML) using linear regression analysis, no significant relationships were found, although there appeared to be a threshold in metal load of 10 above which the condition indices were always low (Figure 5). However, when TCI was related to ML for the individual sites, in all cases except site 7, a significant negative relationship was found. This relationship for sites 5 and 6 is illustrated in Figure 6. A similar relationship was found for the SCI. In all cases the described variation was between 47 and 92%.
This study showed that transplanted zebra mussels could be used to assess metal bioavailability and effects at polluted sites in watercourses. Compared to published literature, high to very high levels of trace metals were measured in the mussels at all sites. Extremely high levels of Cd, Ni and Zn were found at sites 1, 5 and 6 exceeding by several times the highest levels reported in literature. As well, the accumulated levels of Co and Cu were very high at sites 5 and 6, respectively. The other metal levels were moderate compared to literature values (Mersch et al., 1996; Hendriks et al., 1998; Zimmerman et al., 2002; Camusso et al., 1994; de Lafontaine et al., 2000; Cope et al., 1999; Lowe and Day, 2002; Kwan et al., 2003). Both Cd and Zn clearly followed the concentration gradient between site 1 and 4. At sites 4, 5, and 6 metal accumulation was followed for 6 weeks. For most accumulated metals, equilibrium was reached between 3 and 6 weeks of exposure. This however was not the case for Cd at site 5 and for Ni at site 6 where accumulation further increased.
Bervoets et al. (2004) found a steady state for all metals was not reached before 14 weeks when zebra mussels were exposed in a metal contaminated lake for 18 weeks. However, after 6 weeks of exposure for most metals no significant differences were found in metal accumulation between resident and transplanted mussels at 12 sites. Mersch et al. (1993) exposed zebra mussels in the laboratory at 14°C to Cu and Cd and found after 27 d no steady state in accumulation at exposure concentrations of 50 and 44 μ g l−1 for Cu and Cd, respectively. Likewise, Camusso et al. (1994) observed no steady state in Cd and Pb accumulation in transplanted zebra mussels after 60 d of exposure in the river Po. Voets et al. (2004) found zebra mussels exposed to Cd during 30 d at 15°C did not reach a steady state. This together with the fact that exposure temperatures were low, means that the measured accumulated metal levels in our study after 3 weeks of exposure might be an underestimation of the levels under equilibrium.
The TCI found in our study were comparable to those found for zebra mussels in other studies (Mersch et al., 1996; Smolders et al., 2002). Condition factors based on the tissue weight/shell weight (TCI) or tissue weight/shell length (SCI) relationships are often used as measures for the well being of mussels (e.g., Mersch et al., 1996; Soto et al., 2000; Martel and Pihan, 2003). Some studies have investigated the effect of pollution on the TCI or SCI (Mersch and Pihan, 1993; Kilgour et al., 1994; Smolders et al., 2002; Roméo et al., 2003). Although significant differences in accumulation were found for all metals among the different sites, none of the condition indices were significantly different. The same was found by Martel et al. (2003) who exposed the freshwater mussel Elliptio complanata along a river in Québec and did not find significant differences in the TCI. In contrast, Mersch et al. (1996) found significant differences in TCI in transplanted zebra mussels along a pollution gradient. Also, Smolders et al. (2002) found significant differences in TCI values of zebra mussels exposed along a pollution gradient. Besides effects of pollution, the condition of zebra mussels is affected by differences in food availability at the different sites (Smolders et al., 2002; Roméo et al., 2003). In our study no significant relationship was found between the metal load of the individual zebra mussels and the individual condition indices for all pooled data,. However, it was possible to define threshold values for the ML above which TCI or SCI were always low. The high variability in the condition indices at low metal load could be attributed to differences in food availability. When the same relationship was investigated at individual exposure sites, in all but one case a significant relationship could be found. This indicates that in the absence of differences in food availability and other factors, the condition indices accurately reflect the effect of metal pollution on zebra mussels.
It can be concluded from this study that transplanted zebra mussel are suitable for the evaluation of metal bioavailability in watercourses. In order to be sure that a steady state has been reached, an exposure period of at least 6 weeks is required. Further, the results of our study show that the condition indices TCI and SCI are useful tools for the assessment of effects of metals on zebra mussels provided that the variation in other environmental factors such as food availability are limited or are considered in the monitoring program.
This project was supported by the University of Antwerp (project nr. BOF44704/UA and GOA-2001/1). L. Bervoets is a Postdoctoral Fellow of the Fund for Scientific Research - Flanders (FWO-Vlaanderen).