Chemicals used for maintenance of wood rafts in mussel farms : evaluation of their potential toxic risk to mussel culture

At the Galician coast, mussels are cultured on vertical ropes attached to floating rafts, which are cleaned in situ with tars or waterproof paints. The present study analyses the com position and the toxic risk of several compounds used during maintenance, in order to improve mussel-farming practices. The toxic risk was determined in relation to that of benzo(a)pyrene (BaP) by calculating the carcinogenic toxic equivalency (TEQ) and evaluating the mutagenic potential using a mutagenicity test based on Vibrio harveyi. The greatest toxic risk was determined for a random mixture of petroleum tars and diesel combusted oil (MTO). This mixture showed a concentration of polycyclic aromatic hydrocarbons 200-fold higher than that of coal tar (CT) and 256-fold higher than pine tar (PT) as well as significantly higher concentrations of Pb and Mn. The TEQ value of MTO was 581-fold and 486-fold higher than that of the CT and PT, respectively. In the waterproof paint (WP) analysed, hydrocarbons were not detected, but the Mn, Cr, Cu, Sn, and Ni content of the WP was significantly higher than that of the tars. Mutagenicity of the tars was dose-dependent and increased after metabolic activation. The MTO showed mutagenic effects that were significantly higher than those of CT and PT but still less than expected, suggesting that the mutagenic potential of all 3 mixtures depends on their concentration and composition, which determine their solubility and biodegradability. The WP did not show mutagenic effects. Our results suggest that the use of WP and PT is more suitable for the maintenance of rafts and could reduce the pollutant impact in the mussel farms.


INTRODUCTION
Galician Rías (NW Spain) are one of the largest productive areas of mussels in the world (Sánchez-Mata & Mora 2000, Caballero Miguez et al. 2009).The culture is carried out in floating rafts made with beams of eucalyptus wood supporting submerged cords, to which mussels attach and grow (Fig. 1).The raft maintenance and cleaning is usually performed in situ with tars or waterproof paints.Mussels, as filtering organisms, ingest particulate and dissolved forms of pollutants along with food, concentrating these pollutants to levels well above those in the surrounding seawater (Baumard et al. 1999).Therefore, the spillage of these substances around rafts entails a direct pollution risk on the immediate environment and the mussel culture itself, which can affect not only the quality of the product but also food safety if these chemicals are accumulated above the recommended values for human consumption (WHO 1991).Thus, it seems necessary to evaluate the toxicity of these compounds.
The compounds described are complex mixtures whose toxicological evaluation is difficult due to interactions of chemical, physical and physiological factors.Only their soluble or dispersed fraction in the water column is bioavailable and can be accumulated in marine organisms, exerting its toxic effect (McElroy et al. 1989).One of the most harmful effects of most toxic pollutants is mutagenicity, which has been related to carcinogenesis (Mortelmans & Zeiger 2000, Barton et al. 2005).Although toxicity is a broad concept that includes many other harmful effects, all compounds inducing mutagenesis or carcinogenesis are considered toxic.This consideration justifies that mutagenicity and carcinogenicity have been widely accepted for assessing the toxicity of chemicals.Based on these 2 effects, several authors have established toxic equivalency factors for each individual polycyclic aromatic hydrocarbon (PAH) relative to Benzo(a)pyrene, which is the most toxic PAH (Nisbet & La Goy 1992, US EPA 1993, Villeneuve et al. 2002).The toxic equivalency of a mixture can be calculated from these factors.However, this method does not consider physical or physiological interactions that may attenuate or enhance the mixture's toxicity.Therefore, in vivo assays, such as mutagenicity tests, would better reflect a mixture's actual toxicity.
The most commonly used mutagenicity test for the analysis of chemical compounds with legal recognition is the one described by Ames using genetically modified strains of Salmonella typhimurium (Ames 1971, Ames et al. 1975, Maron & Ames 1983, Mortelmans & Zeiger 2000).However, this bacterium has low survival in the presence of salts, so marine bacteria adapted to high osmotic pressures seem more suitable to evaluate chemical toxicity in marine environments (Czyz et al. 2002, Wegrzyn & Czyz 2003, Ohe et al. 2004, Podgórska & Wegrzyn 2007, Chec et al. 2008).For this reason, other mutagenicity assays using marine bacteria have been developed, such as the one based upon detection of neomycin-resistant mutants of Vibrio harveyi.The effectiveness of this assay has been compared to the Ames test, showing the highest sensitivity in samples with sea wa ter (Czyz et al. 2000, 2002, 2003, Słoczynska et al. 2010, Ruiz et al. 2013).
Biodegradation of toxic compounds by living organisms is conducted by microsomal enzymes that are NAD(P)H-de pendent (reductases and cytochromes P450), which produce compounds that are more soluble and easy to remove.However, these products are often more toxic than the parent compounds, resulting in a metabolic activation or bioactivation of toxicity (Lehr & Jerina 1977, Stegeman 1981, López-Barea & Pueyo 1998).Therefore, the mutagenic ability of the metabolic products of these compounds would indicate the real toxic potential of the compounds for organisms.A direct relationship among seawater mutagenicity from aquaculture farms, accumulation of high concentrations of pollutants in cultured mussels and development of gonadal neoplasias in these organisms has been shown by our group in previous papers focused in the Ría of Vigo (Ruiz et al. 2011(Ruiz et al. , 2013)).The present work is a study of the composition and toxic risk of 4 compounds generally used in maintenance and cleaning of floating rafts in order to identify their pollutant contribution in aquaculture areas and improve mussel farming practices, ensuring their sustainability as well as the product quality and consumer safety.

Composition of chemical compounds
Four compounds generally used in cleaning and maintenance of mussel rafts were provided by mussel farmers from the Ría of Vigo.These were a random mixture of petroleum tars and diesel combusted oil (MTO); a commercial coal tar (CT); tar of pine wood (PT); and a waterproof paint (WP).Their compositions in hydrocarbons and heavy metals were analysed.

Toxic equivalency
The toxic equivalency (TEQ) of the 4 compounds analysed was assessed based on the equivalency factors (TEF) of each individual hydrocarbon determined with respect to the carcinogenicity of BaP, which was given a value of 1 (Table 1) (Nisbet & Lagoy 1992, US EPA 1993, Villeneuve et al. 2002).The TEQ values were estimated as the sum of each PAH concentration multiplied by its specific TEF: TEQ = Σ (C i × TEF i ), where C i is the concentration of PAH i.

Samples preparation for mutagenicity assay
In a similar method as used by the mussel farmers, the tar samples were heated to make them more fluid.Then, the water-accommodated fractions of all 4 samples were prepared based on the recommendations of Singer et al. (2000).Each sample was dissolved in a minimum volume of dimethyl sulfoxide (DMSO) or acetone and diluted while stirred up to concentrations of 10 and 40 mg l −1 with artificial seawater, prepared according to MacLeod et al. (1954).Solutions were maintained in darkness at 20°C.Similarly, 6, 20, and 40 µg l −1 of BaP were prepared and taken as a mutagen reference.The BaP concentrations assayed were within the range used by the legally recognised mutagenicity tests (2 to 500 ng ml −1 ).The concentrations of tars and paint were determined basing on our previous papers about PAHs and metals accumulated in cultured mussels from the Ría of Vigo (Ruiz et al. 2011(Ruiz et al. , 2013)) (Nisbet & La Goy 1992, US EPA 1993) dilutions of DMSO and acetone were prepared in artificial seawater to achieve the same concentrations as in the samples.For the analyses of metabolic activation, BaP tars, WP and solvents were prepared at 50× concentration.
The reaction mixture contained 100 mM sodium phosphate buffer, pH 7.6, 33 mM KCl, 8 mM MgCl 2 , a NADPH-generating system (4 mM NADP, 5 mM glucose 6-phosphate, and 0.4 U ml −1 glucose 6-phosphate dehydrogenase), 10 µl S9, and 100 µl of the 50× concentration of the samples.The final volume of the reaction mixture was 0.5 ml.After incubation for 15 min at 30°C, the reaction mixtures were sterilized by filtration through 0.22 µm filters (Millipore) and diluted under stirring to a final volume of 5 ml with artificial seawater to reach the concentrations indicated above to test.

Mutagenicity assay
The Vibrio harveyi strains used in the mutagenicity assay were kindly provided by Dr. G. Wegrzyn (University of Gdansk, Poland): wild-type BB7 strain (Belas et al. 1982); BB7X, a strain that is very sensitive to UV irradiation and that bears the Tn5TpMSC insertion (Czyz et al. 2000, 2001, Sikora et al. 2006); and analogous strains called BB7M and BB7XM bearing plasmid pAB91273 that contains mucA and mucB genes, responsible for enhanced error-proneDNA repair, in addition to ampicillin and chloramphenicol resistance genes (Czyz et al. 2000).
V. harveyi strains were cultured at 30°C to mid-log phase (OD 575 approx.0.5).Then, 5 ml of each culture were centrifuged (5.000 × g, 20 min), and the bacterial pellet was resuspended in an equal volume of the samples (without and with metabolic activation) or of artificial marine water (control sample).After incubation over 1 generation (2 h at 30°C), bacterial suspensions were diluted in artificial seawater.Surviving cells were titrated on Marine BOSS plates, spreading 100 µl of a 1:10 6 dilution.Mutant cells were titrated on the same culture medium with neomycin (50 µg ml −1 ), spreading 100 µl of a 1:10 dilution.Following incubation at 30°C for 24 to 48 h, colony-forming units (CFU ml −1 ) were counted.All assays were performed in triplicate.
The mutation frequency is defined as the percentage of mutants with respect to survivors quantified in plates after exposure of the Vibrio culture to artificial marine water (spontaneous mutation frequency) or to the samples (induced mutation frequency).The ratio between the induced mutation frequency and the spontaneous mutation frequency provides the mutagenicity index (MI), which expresses the increase of the spontaneous mutation frequency, in absolute values, due to mutagen action.According to Ames et al. (1975), compounds that increase the spontaneous mutation frequency by 1.5-to 2-fold are considered potentially mutagenic (MI = 1.5 to 2), and those that increase it ≥2-fold are considered mutagenic compounds (MI ≥ 2).The bacterial mortality caused by the analysed samples was determined by calculating the percentage of survivors with respect to those in artificial seawater.

Statistical analysis
Statistical analysis was performed with SPSS 17.0 (SPSS).The distribution and homoscedasticity of all data were analysed by Kolmogorov-Smirnov and Levene tests, respectively.The data variations were assessed by Students's t-test or 1-way analysis of variance (ANOVA).In the latter, when ANOVA indicated significant differences, a Tukey's post-hoc test was used to determine significant differences be -tween groups.In both, the results were considered significant at p < 0.05.

Chemical composition
The maintenance and cleaning of mussel rafts is usually performed in situ with tars or waterproof paints, as described in the introduction.The most widely used of these compounds is an arbitrary mixture of petroleum tars and diesel combusted oil, followed in quantitative importance by coal and pine tars.The pine tar ist only initially used when the raft is installed due its higher cost.Recently, some fishermen have started using antifouling paint for these tasks.
The composition in hydrocarbons (13 parent PAHs) of these compounds is shown in Table 2.The WP is based on metallic oxides and artificial or vegetal resins, but no hydrocarbons were detected.The MTO contained PAHs concentrations (Σ13 parent PAHs) 200-and 256-fold greater than the CT and the PT, respectively.These latter showed similar hydrocarbon concentrations, although notable differences in their composition were observed: the coal tar presented a significantly greater percentage of PAHs with 3 benzene rings, and the pine tar contained more PAHs with 5 and 6 benzene rings (p < 0.01).Moreover, PAHs of 4 benzene rings were predominant in the arbitrary mixture of petroleum tars and diesel combusted oil (p < 0.01) (Fig. 2).PAHs containing 4 or more benzene rings are less water soluble and biodegradable, therefore tending to accumulate persistently (Jaward et al. 2004).
Pollutants accumulated in filtering organisms are indicative of a pollutant spill in the environment.For this reason, mussels and other bivalves have been established as sentinel organisms in monitoring programs (Goldberg et al. 1978, O'Connor 2002).Between 50 and 62% of the total accumulated PAHs in mussels cultured in the Ría of Vigo are hydrocarbons with 4 benzene rings (Ruiz et al. 2011).This percentage is similar to that found in the MTO (43%) (Fig. 2) and seems to agree with the preferential use of this mixture for raft cleaning.
The composition of a hydrocarbon mixture can identify its origin.In this regard, rates of different PAHs have been generally used as molecular indices to determine the process by which a hydrocarbon is generated: petrogenic origin (oil, diesel, or coal) and pyrolytic origin by petroleum combustion (vehicle and crude oil) or by biomass combustion (grass, wood, or coal) (Budzinski et al. 1997).In the present study, 6 of these indices were used to interpret the genesis of the tars analysed: Phe/An; An/An+Phe; Flt/Py; Flt/Flt+Py; BaA/BaA+Chry; IP/IP+Bper.Their range of values is summarized in Table 3A.Magi et al. (2002) also suggest the use of BaP as another molecular marker of PAH derivatives combustion since BaP is present in very low proportions in petroleum while it is abundantly formed during incomplete combustion.The values of these indices for the analysed tars (Table 3B) clearly evidenced the pyrolytic genesis of the MTO, while CT and PT seemed to be generated by mixed petrogenic and (%) PAHs by rings Fig. 2. Hydrocarbon families contained in different tar mixtures used to clean mussel rafts as a percentage of the total PAHs (Σ13 PAHs).Error bars = 1 SD.Significant variation (p < 0.01; Tukey's test).j 3-ring PAHs; j 4-ring PAHs; j 5ring PAHs; j 6-ring PAHs.MTO: mixture of petroleum tars and diesel combusted oil; CT: coal tar; PT: tar of pine wood pyrolytic processes.However, the predominant process for the CT and PT genesis seemed to differ: petrogenic for CT and pyrolytic for PT.These indices were also used by Ruiz et al. (2011) to determine the origin of PAHs accumulated in cultured mussels from the Ría of Vigo, demonstrating a chronic accumulation of pyrolytic hydrocarbons that increases in spring-summer.This is the season when the rafts are usually cleaned.
Heavy and trace metals are considered the most studied pollutant group in coastal environments because of their wide distribution, persistence, and toxicity.These metals are included in diverse industrial compounds, such as fuels and tars, lubricants, catalysts, paint drying agents, pigments, or pesticides (Laws 2000, Clark 2001).Some, such as Mn, Zn, and Cu, have physiological functions in living organisms but at high concentrations are also toxic (Goyer & Clarkson 2001).These metals are persistent because they are not biodegradable and they tend to accumulate in the food chain (Facchinelli et al. 2001).The trace metal content of the tested compounds is summarized in Table 4.The composition significantly differed between the tars and the waterproof paint.The WP showed significantly higher concentrations of Mn, Cr, Cu, Sn, and Ni (p < 0.01), while the tars were characterized by their significantly high concentration of Zn and Pb (p < 0.01).Some differences were also observed between the tars: the MTO contained higher concentrations of Pb and Mn (p < 0.01), while the CT showed higher concentrations of V and Ni (p < 0.01).
According to our previous results, the accumulation of heavy and trace metals in cultured mussels in the Ría of Vigo follows a seasonal pattern, with maximum concentrations in the spring and summer (Ruiz 2012, Ruiz et al. 2011, 2013).These works describe a significant accumulation of Zn (88%) and, in smaller proportions, of As (6.1%), Cu (3.3%) and Pb (1.4%), highly correlated with accumulation of IP.Among the compounds analysed in this work, only the MTO contained IP (Table 1), which again confirms the predominant use of this mixture in mussel farms.
Although the hydrocarbon and metal pollution can have a diverse source, the data shown seem to agree with the habitual use of the MTO in cleaning of mussel rafts and with the hydrocarbon class accumulated in mussels, confirming the contribution of these cleaning tasks to pollution of aquaculture areas.

Toxic equivalency
Not all PAHs and heavy and trace metals have the same toxicity.Among hydrocarbons, the PAHs with 3  (Colombo et al. 1989, Budzinski et al. 1997, Baumard et al. 1998a,b, Soclo et al. 2000, Yunker et al. 2002a,b).(B) Values of these molecular indices of 3 different tars used in cleaning of mussel rafts.MTO: mixture of petroleum tars and diesel combusted oil; CT: coal tar; PT: tar of pine wood; (j) petrogenic origin; (j) pyrolitic origin; (no shading) mixed origin.Other abbreviations as in Table 1 and 4 benzene rings are considered toxic but not carcinogenic.However, the PAHs containing 5 to 7 benzene rings are also highly mutagenic, carcinogenic, and teratogenic (Lehr & Jerina 1977, WHO 1989, Boström et al. 2002).The toxicological risk of the tars analysed was calculated as their toxic equivalency (TEQ) in relation to the BaP toxicity.
The toxic equivalency of the MTO was 581-and 486-fold higher than that of the CT and PT, respectively (Table 2).Although the PAH concentration in the PT was 1.3-fold less than that in the CT, its TEQ value was somewhat higher (Table 2) due to its larger percentage of PAHs with 5 and 6 benzene rings (Fig. 2), among which the most toxic compounds are present.Table 5 shows the toxic equivalency of the tars concentrations used in the mutagenicity test.The theoretical toxicity (TEQ value) of 10 and 40 mg l −1 of the MTO was 1.3-and 5.1-fold higher than that of 20 µg l −1 of BaP, respectively.However, the TEQ values calculated for the same concentrations of the CT and PT were 2 or 3 orders of magnitude lower than the BaP concentrations taken as a reference.
The toxicity of metals is mainly due to their affinity for sulphured and nitrogenised residues of biomolecules such as proteins (Nieboer & Richardson 1980, Rainbow 2006), which alters cation homeostasis (Ca +2 and Mg +2 ), affecting transport through the cell membrane, respiration rate, and metabolism and inducing oxidative stress (Khangarot & Rathore 2003, Muyssen et al. 2006).Among heavy and trace metals, Hg, Cd, and Pb are considered especially dangerous because of their toxicity, even at low concentrations, followed at some distance by Cu, Zn, and As, which also have mutagenic and carcinogenic effects (Florea & Büsselberg 2006, Prá et al. 2008, Obiakor et al. 2010).
As shown in Table 4, the tars analysed contained high concentrations of Zn (~38 mg kg −1 ) and Pb (2.7 to 9.7 mg kg −1 ).The Pb concentration of the MTO was 3.5-fold higher than that of the CT and PT, and the CT also contained significant concentrations of V, which has been related to development disorders, diseases of the nervous, hematologic, and immune systems, and cancer development (Mitchell 2007, ATSDR 2012).Moreover, the waterproof paint analysed contained a significant proportion of Cu and high Mn and Cr concentrations.These latter 2 metals, although they have a moderate toxicity, have been related to nervous, hepatic, and lung pathologies as well as to male reproductive diseases.Furthermore, Cr has carcinogenic effects (ATSDR 2004, 2012, Crossgrove & Zheng 2004).
The theoretical toxicity of the compounds used in the maintenance and cleaning of mussel rafts seems to agree with the observation of gonadal neoplastic disorders in mussels cultured in the Ría of Vigo, whose development is directly related to accumulation of hydrocarbons and metals, mainly PAHs with 4 to 6 benzene rings, Zn, and Cd (Alonso et al. 2001, Ruiz et al. 2013, 2014).

Mutagenicity
In previous sections, we identified the theoretical toxicity of the mixtures used to clean rafts in mussel farms based on their composition.However, conditions of the marine environment such as salinity and temperature may modify the solubility and bioavailability of chemicals and therefore their toxicity (McElroy et al. 1989, Singer et al. 2000, Schlautman et al. 2004).Furthermore, the biodegradation processes in living organisms can generate metabolic products more toxic than the parent compounds, thus increasing their harmful effect (Lehr & Jerina 1977, Stegeman 1981, López-Barea & Pueyo 1998).Therefore, the evaluation of marine pollutants toxicity should be performed using biological assays under similar conditions to those of the studied environment.
The most commonly used toxicity assay is the Ames mutagenicity test, based on the reversion detection of histidine-auxotrophic mutants of Salmonella typhimu rium strains, after their exposure to mutagens (Ames 1971, Maron & Ames 1983, Mortelmans & Zeiger 2000).However, the survival of S. typhimurium is dramatically reduced in marine water (Czyz et al. 2002, Ohe et al. 2004), so other mutagenicity assays using marine bacteria have been developed based on detection of reversion of dark mutants of Vibrio fischeri and V. harveyi or on detection of neomycin-resistant mutants of V. harveyi strains (Kwan et al. 1990, Czyz et al. 2000, 2002, 2003, Podgórska & Wegrzyn 2006, Słoczynska et al. 2010).In the present study, the latter assay was  selected for mutagenicity analysis of the cleaning mixtures used in mussel farms because of its greater sensitivity, simplicity, and low cost (Johnson 1992, Podgórska & Wegrzyn 2006, Chec et al. 2008, Ruiz et al. 2013).
Taking this into account, the mutagenic capacity of the water-accommodated fractions from the tars and the waterproof paint, before and after being metabolised by mouse liver microsomes, were tested in comparison to the mutagenic activity of BaP.The results are reported in Figs. 3 & 4.
The DMSO and acetone concentrations used as the solvents showed no mutagenic effects on any of the 4 strains of V. harveyi on which the test was based, giving mutation frequencies similar to the spontaneous mutation frequency.Survival of strains was > 95% in all the tests performed, demonstrating the suitability of the concentrations used in the mutagenicity test.
BaP is considered a premutagen whose epoxide, originated by microsomal metabolism, is a potent mutagen that interacts with DNA.The Ames test identifies this compound as a mutagen that increases its mutagenic capacity after metabolic activation (Ames et al. 1972, Maria et al. 2002, Jemnitz et al. 2004).In agreement with this, our results of the V. harveyi test showed a significant increase of the MI of BaP after its metabolic transformation (p < 0.1) (Fig. 3).Each strain used in this test showed a different sensitivity, as described by Czyz et al. (2000Czyz et al. ( , 2002)).However, the 4 presented a dose-dependent MI increment with respect to the concentrations of BaP (p < 0.05).Before metabolic activation, only the highest concentration of BaP (40 µg l −1 ) showed mutagenic or potentially mutagenic effects.But after metabolic activation, the lowest concentration (6 µg l −1 ) also showed potentially mutagenic effects, whereas the highest concentration increased the MI nearly 2-fold (Fig. 3).These results prove the validity of the V. harveyi test and its proportional response to the dose of the mutagen and metabolic activation.
Regarding the tars analysed, statistically significant proportional responses were observed according to dose and metabolic activation (p < 0.05) (Fig. 4).The MTO showed mutagenic effects clearly higher than those of the CT and the PT, but not proportional as would be expected by its much higher concentration of PAHs and toxic equivalency.The MI of MTO was also lower than expected with respect to the mutagenicity of BaP, so that concentrations of 10 and 40 mg l −1 of MTO showed similar values to 20 and 40 µg l −1 of BaP respectively, both before and after metabolic activation.The concentration of 10 mg l −1 of MTO showed mutagenic values (> 2) only after metabolic activation, but the concentration of 40 mg l −1 also showed mutagenic effects before metabolic activation.
Although the CT had a toxic equivalency somewhat lower than that of the PT (Table 5), it showed higher mutagenic effects (Fig. 4), probably due to the higher concentration of BaP and other less toxic PAHs such as Phe, An, and Py in its composition, which could have a synergistic effect enhancing the toxicity.Of these 2 tars, only the CT showed mutagenic effects after metabolic activation at the concentration of 40 mg l −1 (MI = 2.1 to 2.4).This same concentration of the PT after metabolic activation only showed potentially mutagenic values (MI = 1.6 to 1.8).The assay was performed with 4 different bacterial strains, before and after metabolic activation with the S9 fraction from the liver of male Wistar Han rat.V. harveyi strains: BB7 (j); BB7M (j); BB7X (j) and BB7XM (j).Limits of MI values (dashed lines): potentially mutagenic (1.5 < MI < 2) and mutagenic (MI > 2).: significant increase of MI with the BaP concentration (p < 0.05); Δ: significant increase of MI after metabolic activation (p < 0.05) (Tukey's test) The tars analysed are complex and variable mixtures whose bioavailability and toxic potential depend, as indicated above, on their solubility and partition coefficients, po tentially explaining the results ob served, which differed slightly from those expected.The solubility and par tition coefficient, in turn, vary depending on environmental conditions, such as salinity and temperature, but also on the concentration and composition of mixture itself, decreasing as the concentration of total hydrocarbons and the proportion of high molecular weight PAHs increase (Sin ger et al. 2000, Prak & Pritchard 2002, Schlautman et al. 2004).
Moreover, the waterproof paint showed no mutagenic effects, and its MI values were <1.5 at all concen trations assayed, before and after metabolic activation.Heavy and trace metals are persistent pollutants that are not biodegradable (Facchinelli et al. 2001), which justifies that this paint, composed of metallic oxides, does not display bioactivation.As in dicated above, acute and chronic exposure to some heavy metals and their bioacumulation has been related to carcinogenesis.However, their concentrations in the dilutions assayed of the waterproof paint seem very low and are unlikely to drive any mutagenic effect compared to the mutagenicity of PAHs.

CONCLUSION
Our results suggest the convenience of regulation and control of the maintenance tasks of mussel rafts, allowing only compounds with a low toxic risk.The data shown here reveal that the use of waterproof paints and pine tar would be suitable for the maintenance of rafts and could reduce the pollutant impact in the mussel farms.Nevertheless, if these compounds are used regularly, monitoring of the accumulation in mussels of the heavy metals and the PAHs of high molecular weight that the compounds contain should be performed.

Table 2 .
Hydrocarbon composition of different tars commonly used in the cleaning and maintenance of mussel rafts (mg kg −1 ).TEQ: toxic equivalency relative to BaP.MTO: mixture of petroleum tars and diesel combusted oil; CT: coal tar; PT: tar of pine wood.Other abbreviations as in Table1

Table 3 .
Molecular indices used as fingerprints to determinate the origin of hydrocarbons.(A) Characteristic values indicative of the pyrolytic or petrogenic origin of a hydrocarbon

Table 5 .
Toxic equivalency (TEQ) values of the different concentrations of BaP and tar mixtures used in the Vibrio harveyi mutagenicity test.(−):concentrationnot tested.Abbreviations as in Tables1 & 2