Challenges for the development of a biotic ligand model predicting copper toxicity in estuaries and seas



An effort is ongoing to develop a biotic ligand model (BLM) that predicts copper (Cu) toxicity in estuarine and marine environments. At present, the BLM accounts for the effects of water chemistry on Cu speciation, but it does not consider the influence of water chemistry on the physiology of the organisms. We discuss how chemistry affects Cu toxicity not only by controlling its speciation, but also by affecting the osmoregulatory physiology of the organism, which varies according to salinity. In an attempt to understand the mechanisms of Cu toxicity and predict its impacts, we explore the hypothesis that the common factor linking the main toxic effects of Cu is the enzyme carbonic anhydrase (CA), because it is a Cu target with multiple functions and salinity-dependent expression and activity. According to this hypothesis, the site of action of Cu in marine fish may be not only the gill, but also the intestine, because in this tissue CA plays an important role in ion transport and water adsorption. Therefore, the BLM of Cu toxicity to marine fish should also consider the intestine as a biotic ligand. Finally, we underline the need to incorporate the osmotic gradient into the BLM calculations to account for the influence of physiology on Cu toxicity. Environ. Toxicol. Chem. 2012;31:230–238. © 2011 SETAC


Historically, environmental quality criteria/standards (Environmental Quality Standards [EQSs] in the European Union and Water Quality Criteria [WQC] in the United States) for metals have been based on total or total dissolved metal concentrations, although many physiological and toxicological studies have demonstrated that water chemistry affects both bioavailability and toxicity of metals to aquatic organisms 1–4. The effects of water chemistry began to be incorporated into regulations from the 1970s onwards, when the European Agencies and the U.S. Environment Protection Agency (U.S. EPA) developed higher quality standards for metals in fresh waters than for those in marine waters 5. The threshold values in freshwaters were also adjusted for hardness to generate site-specific WQC/EQS 6. This represented a breakthrough because, for the first time, the regulatory authorities acknowledged the relevance of water chemistry to metal bioavailability and toxicity, even though other critical water chemistry variables, such as dissolved organic matter (DOM), were still overlooked. A further step in accounting for bioavailability was represented by the implementation in the United States of the water effects ratio (WER) procedure 7, which accounted for the difference between the toxicity of the metal in laboratory water and its toxicity in the water at the site. The WER is determined by dividing the endpoint (the concentration that is lethal to 50% of the test organisms [LC50]) obtained in the site water by the endpoint obtained in the laboratory-diluted water. This procedure, although effective, implies a number of site-specific tests and analyses, whereas the regulatory authorities have a need for easy-to-use tools to assess site-specific water quality standards for metals 8.

Speciation and toxicity models can respond to the need for ease of use because they incorporate the complexity of the environment into simplified systems based on specific assumptions. Their value does not depend on their capability to reproduce reality in the most exhaustive way, but actually on their usefulness and effectiveness. Their theory, along with the calibration data on which they are founded, must be robust enough to give realistic output data over a wide range of conditions. In this respect, one of the most relevant systems developed for the aquatic environment is the biotic ligand model (BLM) 9, originally conceived to predict metal availability and toxicity to freshwater fish and daphnids 10, 11. The model is essentially composed of chemistry-based and physiology-based parts (Fig. 1). The first is represented by the chemical speciation computations, both inorganic and organic. The inorganic speciation is relatively simple and is performed by the chemical equilibria in soils and solution (CHESS) model 12, whereas the organic speciation is computed by the Windermere humic acid model (WHAM), developed by Tipping and coworkers 13, 14. Modeling the complexation of metal cations by organic matter is challenging because of uncertainties regarding the characteristics of the ligands and their binding constants. The second part of the BLM accounts for the effect of the metal binding to the biotic ligand; it is based on the concept that toxicity effects occur when the metal-biotic ligand complex reaches a critical concentration.

Figure 1.

Scheme of the biotic ligand model. M-DOC = metal-dissolved organic carbon complex.

The BLM has gained acceptance in the scientific and regulatory communities over the last decade, when the U.S. EPA published its “Framework for Metal Risk Assessment” in 2007 15, which introduced the BLM as a tool to derive criteria for the protection of aquatic life. The BLM has successively been adopted by the Australian and New Zealand Environmental Protection Agencies as a tool for regulating metal concentrations in the aquatic environment 16. It is also in the process of being adopted by Environment Canada and the European Union. In 2008, the United Kingdom (UK) Environment Agency published a report on EQSs for trace metals in the aquatic environment 17, which recommended the use of BLM-type models to assess compliance with EQSs. In particular, they proposed a “field” BLM, the toxicity biotic model (TBM) that in the future may be optimized with field data and possibly include the effect of metal mixtures, which is considered an outstanding issue that needs to be addressed. They also developed and presented a tiered assessment approach 18 that takes into account background metal concentrations. Through this approach it would be possible to prioritize sites where the concentration of a metal exceeds the EQS expressed in terms of total dissolved concentration, thus focusing resources on sites at most risk.

In 2009, a European workshop discussed the incorporation of metal bioavailability into the regulatory framework ( and concluded that the application of the BLM in estuarine and marine environments should be evaluated carefully. In the marine environment, work should be focused on the DOM variable, given the different characteristics of riverine, marine, and estuarine organic matter, whereas in estuaries the complexity of the geochemical dynamics and the osmoregulatory responses of the estuarine organisms should be accounted for; otherwise, the effectiveness of the model might be compromised.

The BLM framework is based on both water chemistry and physiology. Water chemistry greatly influences Cu bioavailability and thus its toxicity: alkalinity, water hardness, pH, DOM, and salinity are all variables that affect Cu speciation 1–4. High salinity, in particular, has a protective effect against Cu toxicity because it reduces the availability of the metal as a consequence of both increased inorganic metal complexation and increased competition for binding sites by cations such as Na+, Mg2+, and Ca2+9, 19. In addition, salinity can also affect the physiology of aquatic organisms by influencing the mechanisms they adopt to maintain an osmotic balance and thus their responses to metal exposure 20, 21. Copper toxicity can vary depending on osmoregulatory strategies adopted by the organisms: osmoregulators and osmoconformers display different patterns of sensitivity to Cu, as well as differing physiological responses as salinity changes. The iso-osmotic point of the organism and, in general, the relative Na+ gradient between the internal fluids and the external environment, particularly the electrochemical gradient across the gill epithelium, can affect the osmoregulatory strategy of the organism and consequently its sensitivity to the metal 20, 22. The BLM does account for the effects of water chemistry on metal speciation, but it does not explicitly consider the influence of water chemistry on the physiology of the organisms. For this reason, although both physiology and water chemistry are likely to be critical in determining Cu toxicity, research on the BLM must shift from the chemistry-based side to the physiology-based side of the model, so that an understanding of the mechanism of toxicity may underpin predicting the effects of acute exposure. The aim of the present study is to discuss both these aspects, with particular regard to the effect of water chemistry on physiology, underlining the main issues that need to be addressed in future research, with a view to the development of a marine and estuarine version of the model.


Many estuarine and coastal zones are areas of special scientific, interest, and their habitats are protected. They are also among the ecosystems at greatest risk from anthropogenic impact, because they are affected by many human activities, including shipping, urbanization, and industrialization, which are leading to their degradation 23. One of the main concerns about these environments is the presence of metals, particularly Cu, because its usage as a biocide in antifouling painting coatings has increased considerably after the banning of tributyltin (TBT) 24. A recent survey of UK harbors, marinas, and estuaries observed that only between 10 and 30% of total dissolved Cu in these areas is present in the labile fractions (free Cu ion and inorganically bound Cu), which are believed to be the toxic forms, whereas most dissolved Cu is present in the nontoxic form (organically bound Cu) 25. These results suggest that the use of total Cu concentrations as an indicator of toxicity may lead to an overestimation of environmental risk. For this reason, efforts are ongoing to incorporate Cu speciation in EQS/WQC and to develop a BLM model more suitable to estuarine and marine environments 25.

Even though the BLM has shown some potential in predicting Cu toxicity to marine organisms, it still has some limitations. Considering only the chemistry-based side of the model, one of the main issues regards the metal complexation with organic matter, which is calculated using WHAM. This model has been developed and validated on data obtained at concentrations of metals and humic substance appropriate for freshwaters 13, 26. However, in estuarine and marine waters the concentrations are generally lower. Organic matter in these environments is also likely to have different characteristics and in particular different complexation capacities compared to fluvial material 27, 28. Thus, data on Cu speciation and DOM concentrations and complexation capacity are necessary for the development of a marine version of the BLM. Furthermore, some studies have shown that the source of DOM has a significant effect on acute Cu toxicity 28–32. This effect, although not specifically taken into account by the BLM at present, may be accounted for by refining either the percentage of active humic acid (%AHA) 33 or the DOM ligand density 29, which both affect organic matter complexation capacity in the model.

Correct modeling of Cu complexation by the organic matter appears to be critical in estuaries, where most of the metal is complexed by organic ligands 34; therefore, it may be worthwhile considering possible alternatives to WHAM, which has been shown to be the most robust in predicting organic complexation with metals in estuaries 34. Among the models that describe the interaction of organic matter with metals, two have been applied to estuaries with some success: the Stockholm humic model (SHM) 35 and the Nica-Donann model 36, which is part of the Visual MINTEQ model ( The main difference between them is in the modeling of the binding sites of the organic matter: WHAM and SHM consider discrete binding sites on the organic matter surface, whereas the Nica-Donann model uses a bimodal, continuous distribution of binding affinity values for protons and metal ions.

From a purely chemical perspective, an estuary represents a zone where seawater of high ionic strength is diluted by river water, resulting in a linear salinity gradient. The two main driving forces are the river and the tidal flows: the relative intensities of one or the other control the vertical stratification of the estuary and ultimately the mixing of river and seawater 37. Based on their vertical mixing, estuaries can be classified into three categories: partly mixed, well mixed, and salt-wedge. Even though this classification is often rather theoretical, it can still be useful in designing an experiment with the aim of predicting and modeling the fate and behavior of a chemical contaminant, such as Cu. In this context, dissolved metal species vary mainly as a function of salinity, and, to a lesser extent, to pH and redox conditions. In an estuary, pH generally shows significant variation, which can produce a critical change in Cu speciation. Model calculations have indicated that, in the absence of organic ligands, Cu speciation is dominated by carbonate complexation 38, which has a strong influence on the concentration of Cu2+, because an increase in pH from 7.6 to 8.2 results in a threefold reduction of the Cu2+ concentration 39.

In summary, the main issues that should be addressed in the development of a BLM for estuaries and sea are the interaction between Cu and organic matter, with particular regard to the different DOM sources, and the effect of pH on Cu speciation, considering the relevance of carbonate complexation at high salinity.


Osmoregulatory strategies and Cu toxicity in saltwater

All freshwater (FW) organisms osmoregulate to maintain the internal salt concentration above that of the external environment. They do this by an active uptake of Na+, which can be inhibited by Cu exposure 40 that may also affect the Na+ leakage pathway 41. Because this is true for both fish and invertebrates, it is one reason for the relative success of BLM in predicting acute Cu toxicity in both groups of organisms in freshwater, although the sensitivity to Cu displayed by FW organisms varies by more than three orders of magnitude among different taxa 20. However, saltwater (SW) organisms present wider intertaxa variability of osmoregulatory strategies. In particular, SW teleost fish (bony fish) are osmo- and ionoregulators, because they maintain extracellular ion concentrations well below that of the surrounding medium. For this reason, they need to intake seawater continuously and excrete the excess salts to compensate for water loss due to the osmotic gradient and the diffusion of salts into their tissues 20. Elasmobranch (cartilaginous) fish are ionoregulators and osmoconformers, because they regulate the ion concentration at approximately 50% of that in the SW, but they maintain the osmotic pressure of extracellular fluids similar to the surrounding medium by reabsorbing and retaining urea and other organic osmolytes in their tissues. This allows them to drink less seawater than teleosts, but they still face the problem of the diffusion of salts from the external SW, where the ion concentration is higher, into their body. They compensate for this by salt excretion in urine, by secretion from the rectal gland, and by salt transfer at the gill epithelium 42. Finally, most marine invertebrates are both iono- and osmoconformers, because their ionic composition and osmotic pressure are highly similar to those of the surrounding seawater 20.

Because it is known that acute Cu toxicity is mainly a consequence of its effect on ion transport 43–47, it is plausible that differences in osmoregulatory strategies may result in differences in Cu sensitivity among species and, as salinity changes, within the same species. If we assume that Cu exerts its toxic effect by disrupting the maintenance of a net Na+ gradient, we may hypothesize that osmoregulators would be more sensitive to Cu than are osmoconformers. Because the available data do not support this expectation (see Grosell et al. 20 for a review), the hypothesis that Cu acts as an osmoregulatory toxicant in seawater will be discussed.

In marine osmoconformers exposed to Cu, the principal cause of mortality appears to be either a disturbance in acid–base balance, related to impaired respiratory gas exchange, and/or an effect on ammonia excretion 48–51, although evidence also exists for Cu toxicity via oxidative stress 52, DNA damage 53, and metabolic inhibition as a secondary effect of oxygen uptake inhibition 54. In marine osmoregulators, the studies regarding the effects of Cu are equivocal. However, it is generally agreed that impaired ion regulation is the main effect of Cu exposure 44–47, 55 in both fish and invertebrates. Acute and chronic Cu exposure of the gulf toadfish (Opsanus beta) induced an increase in plasma Na+ and Cl concentrations, due to impaired osmoregulatory capacity both in the gill and in the intestine, resulting in fluid loss from muscle tissue 44. These results are in accordance with previous studies on the seawater-adapted flounder (Platichthys flesus) and the rainbow trout (Oncorhynchus mykiss) 56, 57. However, some exceptions exist: in other studies on the killifish (Fundulus heteroclitus) and the cod (Gadus morhua), disturbances in acid–base balance and ammonia excretion were the main observed effects 22, 58. In the experiments performed on F. heteroclitus, Cu toxicity was studied across a range of salinities, from freshwater to seawater 22. The hypothesis that Cu would act via the same general mechanism, regardless of salinity, was tested, but the results did not confirm this. Indeed, in FW the expected mechanism of toxicity was observed; in particular, a decrease in plasma Na+ concentration occurred, caused by the Cu-induced inhibition of the Na+/K+ ATPase activity. However, in SW no effect was observed either on Na+ homeostasis or on Na+/K+ ATPase activity. The only parameter affected by Cu exposure was ammonia excretion, which has also been reported to be affected by Cu in other studies 57, 58.

Taken together, these observations not only demonstrate the complex pattern of physiological responses to Cu, but also suggest that the main target of Cu toxicity might be a common factor, which controls ion transport along with acid–base balance and nitrogenous waste excretion. If we assume this, we can try to give an overall explanation for the main reported effects of Cu, both in osmoconformers and osmoregulators. In osmoconformers, acid–base and ammonia excretion disturbances seem to be the main causes of Cu-induced mortality, because maintaining an osmoregulatory gradient is not an issue for this group of animals. In osmoregulators, the disturbance of ion transport is the main effect, because maintaining a Na+ gradient is critical for them; however, acid–base balance and ammonia excretion are also affected, and in some cases are more sensitive endpoints (Table 1) 22.

Table 1. Summary of the effects of copper in osmoregulators and osmoconformers
Main effectsOsmoregulatory strategies
Acid-base balance disturbance caused by impaired gas exchange
Increase in plasma ammonia concentration caused by impaired ammonia excretion
Impaired intestinal and/or branchial ion transport 
Drinking rate disturbances

Carbonic anhydrase enzyme: A Cu target

A common denominator of the previously described mechanisms is the enzyme carbonic anhydrase (CA). In fact, this enzyme can bind Cu, which in turn can inhibit its activity, as has been shown in crustaceans exposed both in vivo and in vitro to this metal 59, 60. However, this mechanism of action in fish has been confirmed only by in vitro experiments 61, not by in vivo 22. This enzyme is present in a range of tissues, including branchial and intestinal epithelia, and is involved, directly or indirectly, in several physiological processes, including gas exchange, acid–base balance, Na+ and Cl transport, and ammonia/ammonium excretion 62, 63, which are all reported to be affected by Cu exposure. The main function of the enzyme is to facilitate the conversion of carbon dioxide and water to bicarbonates and protons 64. In the gill of FW fish, the protons produced by the hydration of carbon dioxide are involved in Na+ transport at the apical membrane, while the bicarbonates are exchanged for chloride. The inhibition of CA, besides influencing these mechanisms, can also affect ammonia excretion by reducing the diffusive trapping mechanism 65–67. Hence, the multiple functions of CA make it a candidate as the common factor that links together ion transport, acid–base balance, and nitrogenous waste excretion, thus offering a possible explanation of the complex pattern of physiological responses to Cu exposure. This hypothesis can be supported by giving further consideration to marine osmoregulators, with some distinctions between fish and crustaceans.

Marine fish have two sites of ion transport, the gill and the intestine, which are both relevant in the maintenance of the osmotic balance. In contrast to FW fish, ion transport in SW fish is not believed to be directly associated with CA in the gill 68, but it is in the intestine, which is responsible for taking up water to compensate for the diffusive loss of water to the concentrated external environment 68, 69. Water absorption is driven by the uptake of Na+ and Cl, which are then excreted at the gill. Ions are moved actively through the intestinal epithelium and water follows passively along the generated osmotic gradient, from the intestinal lumen into the blood. The presence of Cu (and also Ag 70–72, which has a similar toxic effect) in seawater can reduce drinking rates, depending on the exposure time, and interferes with the intestinal uptake of water 45 through the inhibition of the active ion uptake processes that drive water flux by osmosis. The driving force for the movement of ions across the intestinal epithelium is provided by the enzyme Na+/K+ ATPase; however, to date, the inhibition of its activity by metal exposure has not been demonstrated to be responsible for the impaired uptake processes in the intestine of marine fish. Therefore, it is plausible that one or several other proteins involved in ion transport are more sensitive targets. One of them is the enzyme CA, which has been shown to play a key role in osmoregulation 73. It has been demonstrated that both gene expression and enzymatic activity of CA can be modulated by salinity changes, similarly to the enzyme Na+/K+ ATPase, and that their response is tissue-specific, with significant differences between gills and intestine.

Considering the gill, the salinity dependence of CA expression and activity has been demonstrated in several studies, but this pattern still must be elucidated fully because some discrepancies exist among species. In killifish, an increase in CA gene expression was observed after transfer from intermediate salinity to freshwater 74; in the coho salmon (Oncorhynchus kisutch) and in the Mozambique tilapia (Oreochromis mossambicus), CA activity increased with increasing salinity 75, 76, while flounders kept at different salinities showed no significant differences in CA activity 77.

In the intestine, CA expression and activity increased two- to fourfold in killifish 22 after transfer to seawater. For rainbow trout, a salinity change induced a response that involved two isoforms of the enzyme CA: the cytosolic CA (CAc) and the extracellular isoform membrane-bound CA type IV (CAIV), localized at the apical region of the intestinal epithelium 78. The former usually displays the majority of the CA activity in the intestinal epithelium and provides the cellular substrate for the equation image anion exchanger on the apical membrane 79. The latter contributes to the deposition of CaCO3 in the intestinal lumen, which reduces the osmotic concentration of the intestinal fluid and thus facilitates ion transport and water absorption through the epithelium 80. In rainbow trout, an osmotic stress, such as caused by an abrupt transfer from freshwater to 65% saltwater, induced a transient increase of mRNA expression and of enzymatic activity, both of CAc and membrane-bound CAIV 78. A recent study on CA expression and activity, both in the intestine and in the gills of the gulf toadfish following transfer to salinity of 60 ppt, demonstrated that the CA was important for tolerance to hypersalinity 73. What emerges from these studies is that the enzyme CA plays a key role in osmoregulation and the modulation of its activity and expression appears to be a response to osmotic stress. If we hypothesize that CA is the main Cu target in SW, and that in fish it displays its osmoregulatory functions mainly in the intestine, then we might assume that the Cu target in SW is not (or not only) in the gill, but (also) in the intestine, and so consider the Cu speciation and bioavailability in this medium. It has been observed that Cu becomes less bioavailable as it moves along the intestine of the gulf toadfish: this may be because the absorption/excretion mechanisms at the intestinal epithelium progressively modify the chemical composition of the intestinal fluid and thus the speciation and availability of Cu 45.

Metabolically, marine crustaceans are less complex than fish, because the only tissue in which the enzyme CA may be involved in Cu-sensitive functions is the gill. However, although marine crustaceans are invertebrates, which are generally osmoconformers, several species can shift from osmoconformity to osmoregulation below a critical salinity, typically around 25 ppt 81, 82. The enzyme CA has been characterized as one of the central components of ion uptake in the branchial epithelium of crustaceans 82, because its expression is salinity sensitive and its inhibition has been linked to the disruption of ion transport and regulation 83. The increase in CA activity in response to low salinity appears to be a central feature of the transition from osmoconformity to osmoregulation. This is believed to be a common adaptive characteristic of all euryhaline marine crustaceans capable of osmotic- and ionoregulation 62, 84. If Cu exposure affects CA activity, it may interfere with the ability of these crustaceans to respond to osmotic stress and thus result in osmoregulatory imbalance.

Conceptually, from the modeling perspective, the validation of the hypothesis presented here would mean that the biotic ligand in marine fish is not only the gill, but also the intestine (as suggested also for Ag 71–73 and the relationship between toxic effect and metal accumulation at the binding site is modulated by salinity and in particular by the osmotic gradient.

Branchial permeability and mucus secretion

Another explanation of the toxic effects of Cu is a general alteration of the epithelial function of the gill. Most processes that have been reported to be affected by Cu toxicity are associated with ion regulation at the gill; therefore, a generalized mode of action has been hypothesized, possibly caused by mucus secretion 55. Mucus is thought to act as a buffer, preventing the metal from interacting with the site of toxicity; but, if a thick layer of mucus is produced, it may increase the diffusive distance at the gill surface and thus affect branchial processes by reducing branchial permeability. However, most studies on the effects of Cu exposure reported an increase, not a decrease, in branchial permeability, due to a displacement of Ca2+ at the gill surface, which controls the passive diffusion of Na+, Cl, and also Mg2+, all ions that demonstrated an increase in plasma concentrations after Cu exposure 44, 57.

Na+/K+ ATPase activity: Why not a target in saltwater?

An issue that may need to be elucidated is why the Na+/K+ ATPase activity is one of the targets of Cu in FW 85, whereas it is reported to be unaffected in SW 86. Because this enzyme is central to ion transport by the gills of both freshwater and marine fish, it may therefore be expected to respond similarly to Cu in the two environments. An explanation of these apparently puzzling observations can be found in a study conducted on the gill of the SW-adapted flounder 56, to which Cu was applied in both in vitro and in vivo experiments. The application of Cu to gill homogenates during the in vitro experiments caused a marked reduction in Na+/K+ ATPase activity, but the in vivo experiments showed that Cu exposure induced an increase in the number of Na+/K+ ATPase units in the gill. This increase counterbalances the reduced activity of each Na+/K+ ATPase site and thus resulted in an overall unchanged Na+/K+ ATPase activity in the gill tissue, as reported in most studies on marine fish gill physiology. Furthermore, the increase of Na+/K+ ATPase units is controlled by the production of cortisol 87, a hormone responsible for osmoregulatory balance in SW, the concentration of which is enhanced by Cu exposure. Another effect of higher levels of cortisol is the enhancement of protein catabolism 87, which increases ammonia production and thus plasma ammonia concentration, a reported effect of exposure to Cu. A further explanation of the apparent insensitivity of the Na+/K+ ATPase to Cu at high salinity is provided by the observation that rapid salinity change can induce the expression of different isoforms of the enzyme 88. It is plausible that different isoforms may display differential sensitivity to Cu 45.

Salinity changes: Relevance of the osmotic gradient

With the development of a BLM for marine and transitional waters in mind, the previous considerations imply that an estuarine/marine version of the BLM should give more relevance to the physiology, because it has been shown to be important in marine and transitional environments. At present, species differences in sensitivity are dealt with through adjustments of the stability constants characterizing the metal/gill interaction. This may be sufficient in FW, but in SW it cannot explain the variation in sensitivity to Cu shown by SW and estuarine organisms. The physiological mechanisms lying at the base of the osmotic regulation system are more various and variable in marine organisms than in freshwater ones and thus result in a large interspecies variation in Cu sensitivity and an intraspecies variation in Cu sensitivity when salinity changes 20. The second point is particularly relevant if we consider estuaries, transitional waters, or other environments characterized by fluctuating salinities.

Estuarine invertebrates display a wide range of osmoregulatory strategies. In general, at their iso-osmotic point they are osmoconformers. Above it, they are still generally osmoconformers, although a few (particularly decapod crustaceans) are able to osmoregulate 89. Below the iso-osmotic point they weakly osmoregulate up to a certain salinity level, after which they start strongly osmo- and ionoregulating 82, 90. A few estuarine invertebrates are osmo- and ionoregulators in the full tolerance range 91, 92. A good example of the transition from osmotic and ion conformity to regulation is given by the euryhaline green crab (Carcinus maenas), which is an osmoconformer in full-strength seawater, but at a critical salinity of 26 ppt, starts to actively uptake ions across the gill. The activation of the mechanisms of ion transport is correlated with an eightfold induction of the enzyme CA. The role of CA in ion transport and regulation in this species is confirmed by the observation that the ability of green crabs to regulate their hemolymph osmotic and ion concentrations is disrupted when branchial CA activity is inhibited 83. A similar mechanism was also observed in the blue crab (Callinectes sapidus) 93. If Cu exposure affects CA activity, which is a key feature of the acclimation of euryhaline organisms to salinity changes, it might be deduced that Cu exposure can theoretically affect the ability of euryhaline species to respond to osmotic stress.

Reviewing eight studies on the influence of salinity to acute Cu toxicity, Grosell et al. 20 pointed out that none of the studies showed a linear increase in tolerance with increasing salinity, as competitive interactions among cations would have suggested, and the displayed tolerance variations were not fully explained on the basis of Cu speciation. A nonlinear trend of LC50 values related to salinity gradient has also been reported in a recent study of Cu toxicity to the estuarine copepod Eurytemora affinis19. Here the lowest toxicity was observed around the iso-osmotic point of the organism, a fact also seen in the work by Grosell et al. 20. The trend of Cu toxicity related to salinity gradient displayed in this study (Fig. 2) suggests that the physiology of the organism was the driving factor influencing toxicity and thus supports the observation that euryhaline species are more tolerant to metal exposure at iso-osmotic salinities due to minimization of osmotic stress 94. By comparing measured and predicted values (calculated using the main equation of the BLM as modified by De Schampheleare et al. 95), Figure 2 also demonstrates that at present BLM predictions are inaccurate across such a wide salinity range (0–35 ppt); however, as indicated by the good agreement at high salinities, it may accurately predict toxicity in some limited situations 96. Because Cu disrupts the ability to sustain the osmotic gradient between internal fluids and the external medium, the greater the gradient, the higher the toxic effect. Including a parameter that accounts for the equilibrium potential (Ep) across the gill epithelium may thus reduce the bias between the predictions of the model and the observed toxicity, because the Ep undergoes a wide variation with salinity (Fig. 3).

Figure 2.

Acute Cu toxicity to juvenile killifish (Fundulus heteroclitus) as a function of ambient salinity. The triangles represent the 96 h LC50s values obtained by acute toxicity tests (data from Grosell et al. 20). The diamonds represent the values predicted by the basic biotic ligand model (BLM) equation as modified by De Schamphelaere et al. 95. Both the values are expressed as free metal ion activity.

Figure 3.

Absolute equilibrium potential (Ep) at different salinities (data from Grosell et al. 20).

Water and ions (including metals) are exchanged between the internal fluids and the external environment at different rates, in relation to the osmotic gradient and ultimately to the salinity 20, 97. This means that metal uptake and toxicity (assuming a direct relationship among metal accumulation at the biotic ligand, uptake rate, and toxic response) is modulated by the osmotic gradient. However, the mechanism is even more complex, because the mode of action of Cu (the biotic ligand in the BLM language) changes according to the osmoregulatory physiology, which varies with salinity. Thus, the relationship between Cu exposure and toxic response is influenced by salinity through a double interaction. On the one hand, Cu inhibits the ability of the organism to respond to an osmotic stress by interfering with its osmotic mechanisms (i.e., CA). Yet at the same time, a salinity variation produces an increase in Cu uptake due to a general increase in ion uptake rate, as well as, in some cases, a change in the Cu biotic ligand. At present, this complex relationship is clearly overlooked by the BLM, which assumes that the toxic response is directly related to the amount of Cu bioavailable and to its affinity for the binding sites, regardless of the water chemistry.

Conceptually, the challenge is to introduce the effect of water chemistry (salinity) on the physiology-based part of the model. Therefore, we hypothesize that at varying salinities (as in an estuary) the relationship between the [Cu]EC50 and the fraction of binding sites that need to be occupied by the metal to observe a toxic effect in 50% of the test organisms (equation image) is not linear and constant, but variable and modulated by the osmotic gradient.


In conclusion, considering the chemistry-based part of the model, it is necessary to improve the ability of the BLM to describe the interaction between Cu and organic matter: further investigations should be addressed at the different sources of DOM in estuaries and seas, with particular regard to their stability constants. Copper speciation studies in estuaries should also take into account the effect of pH, because it displays a significant change in this environment, and it influences the carbonates, which are thought to dominate inorganic Cu speciation in estuarine and sea waters.

As for the physiology-based part of the model, this presents the main issues of interest. We suggest that in estuaries and seas the water chemistry affects Cu toxicity not only by controlling its speciation, but also by affecting the osmoregulatory physiology of the organism, which in turn varies according to salinity. We present the hypothesis that the common factor that links together the main observed effects of Cu, both in osmoregulator and in osmoconformer organisms, is the CA enzyme, given its multiple functions and its salinity-dependent expression and activity. According to this hypothesis, we also suggest that the site of action of Cu in marine fish is not only the gill, but also the intestine, because this is where CA plays a role in ion transport and water adsorption. Thus, the BLM model of Cu toxicity to marine fish may also need to consider the intestine as a biotic ligand. It is therefore necessary to incorporate the osmotic gradient, and possibly the osmoregulatory strategy of organisms, into BLM calculations in order to use a mechanistic understanding to predict the impacts of acute Cu exposure in a wide range of salinities.


A. de Polo is grateful to Brunel University for the award of an Isambard Scholarship.