Baseline levels of biomarkers of oxidative damage in Odontesthes nigricans ( Pisces , Atherinopsidae ) from two coastal areas of the Beagle Channel , Argentina Niveles basales de biomarcadores de daño oxidativo en Odontesthes nigricans ( Pisces , Atherinopsidae ) en dos áreas costeras del Canal Beag

We assessed several biomarkers of oxidative damage on silversides Odontesthes nigricans from Varela and Golondrina bays, located on the North shore of the Beagle Channel (Tierra del Fuego, Argentina) and we tried to relate those responses to environmental data. Varela bay is visited mostly by local fi shermen and tourism during the spring-summer season; on the other hand Golondrina bay is a low urban impacted area, despite the population’s growth around it. We employed the condition factor as general biomarker of fi sh health and quantifi ed biomarkers of oxidative damage in gills and liver. Fishes from Golondrina bay showed a better condition as compared to those from Varela bay. The activity of enzymatic antioxidants and the content of non-enzymatic antioxidants were lower in gills than in the liver, independently of the sampling site. No signifi cant differences were found between sites in the gills. However, an increase in catalase and superoxide dismutase activities and a decrease in GSH/GSSG ratio were observed in the liver of fi shes from Varela bay, possibly being due to higher levels of environmental natural stresses, mainly metals in sediments. This study indicates an organ-dependent response and establishes different baseline levels of biochemical parameters for each bay that could be considered as values of normal homeostasis for wild silversides.


INTRODUCTION
Aquatic ecosystems are the main recipients of pollutants, which, over time, can cause serious consequences for biota that may not become apparent until changes occur at the population or ecosystem level, a point at which it may be too late to take ef fective countermeasures (Linde-Arias et al. 2008).The need to detect and assess the impact of pollutants, particularly at sublethal concentrations, on environmental quality has led to measure a range of biological responses in different species (Alves de Almeida et al. 2007, Viarengo et al. 2007, Vlahogianni & Valavanidis 2007, Ahmad et al. 2008).The use of biological responses to different environmental conditions, defi ned as biomarkers, has become a useful tool in environmental quality evaluation and risk assessment.Among the numerous ecotoxicological biomarkers, those based on responses at the molecular and cellular levels represent the earliest signals of environmental disturbance and they are commonly used for biomonitoring (Moore et al. 2004, Viarengo et al. 2007).Measuring the same biomarker in different locations simultaneously gives us information about the pollution status of the region and provides a better comprehension of the mechanisms of response of the organisms to pollutants (Giarratano et al. 2010).Although biomarkers can be helpful for gaining insight regarding the mechanisms causing obser ved ef fects of chemicals on whole-organism performance and may, in some cases, provide useful indicators of exposure, individual biomarker responses should not be expected to provide useful predictions of relevant ecological effects and probably not even predictions of whole-organism effects (Forbes et al. 2006).
Fishes are generally considered feasible organisms for pollution monitoring because they can be found virtually everywhere in the aquatic environment and they play a major ecological role in aquatic food-webs because of their function as carriers of energy from lower to higher trophic levels (Beyer 1996).The use of fi sh in environmental monitoring has become increasingly important in recent years in the investigation of natural variability, as well as anthropogenic substances, many of which function as prooxidants, accumulating in aquatic environments (Almroth et al. 2008).Prooxidants can induce morphological and physiological alterations in fi sh tissues (Bainy et al. 1996, Varanka et al. 2001), such as oxidative damage to lipids (lipid peroxidation).To counteract tissue damage, fi shes as well as other organisms have developed different enzymatic and non-enzymatic defensive mechanisms to protect themselves against reactive oxygen species (ROS) production (Winston & Di Giulio 1991).The enzymes superoxide dismutase (SOD) and catalase (CAT) are among the enzymatic antioxidants.Superoxide anions are dismutated by SOD to hydrogen peroxide (H 2 O 2 ), which are subsequently detoxified by CAT present in peroxisomes.Numerous non-enzymatic low-molecular-weight antioxidants, such as ascorbate, β-carotene, α-tocopherol and GSH, have also been described (Stegeman et al. 1992, Lopez-Torres et al. 1993).Ascorbate is an important water-soluble antioxidant which is known to directly scavenge ROS (Halliwell & Gutteridge 1999), acting also as a crucial micronutrient and per forming various vital physiological functions such as collagen synthesis and connective tissue formation (Wilson & Poe 1973).The α-tocopherol and β-carotene are lipid-soluble antioxidants which play a major role in protecting membranes from lipid peroxidation (van der Oost et al. 2003).Glutatione reduced (GSH) is the most abundant cellular thiol and it is involved in several metabolic processes (DeLeve & Kaplowitz 1991), playing a central role in the detoxification of ROS.As consequence of oxidizing conditions, two molecules of GSH are linked by a disulfide bond to comprise one molecule of oxidized glutathione (GSSG).Both total GSH and the ratio of GSH content/ GSSG content can be used as biomarkers for oxidative stress (Koivula & Eeva 2010).Swiergosz-Kowalewska et al. (2006) concluded that the GSH content/GSSG content ratio may be the best and the most sensitive indicator of metal toxicity for animals living in chronically contaminated environments.
The silverside Odontesthes nigricans (Richardson, 1848) is a native fi sh species in Tierra del Fuego (southernmost limit of its distribution) that can be found by following the tides in the coastal waters and estuaries of the Atlantic Ocean and Beagle Channel.It is a species of a regional socioeconomic relevance due to the sport and artisanal fi shing developed upon it.The oxidative metabolism of O. nigricans was previously characterized by Lattuca et al. (2009).O. nigricans was also chosen because another species of the same genus, Odontesthes bonariensis (Valenciennes, 1835), has been recently promoted as test species in a standardized acute toxicity test method in Argentina (IRAM 2007), due to its relatively high sensitivity to metals (Carriquiriborde & Ronco 2002, 2008).
The aim of this work was to determine baseline levels of a condition factor, as general biomarker of fish health, and of several biomarkers of oxidative damage, such as lipid radical content, enzymatic antioxidants (activities of superoxide dismutase and catalase), non-enzymatic antioxidants (α-tocopherol, β-carotene, ascorbate and glutathione contents) and general indexes of oxidative stress (ascorbyl/ascorbate and reduced glutathione/oxidized glutathione ratios) in gills and liver of O. nigricans from two coastal areas within Beagle Channel.Present study represents the first comprehensive investigation that set oxidative stress responses in fi shes from clean coastal waters of the Beagle Channel, essential information for a proper interpretation of future ecotoxicological data and environmental monitoring programs in this austral region.

METHODS
Authors wish to declare that this study w as conducted in accordance with Guidelines on the care and use of animals for scientifi c purposes established by the National Advisory Committee for Laboratory Animal Research (2004).

Sampling sites
Fishes were collected in spring (October) of 2007 at Golondrina and Varela bays, located on the North coast of the Beagle Channel (Tierra del Fuego, Argentina) (Fig. 1).Golondrina bay (54°50'S, 68°21'W) is located near Ushuaia city and has been recently considered by Duarte et al. (2011) as low urban impacted area, despite the population's growth around it.Varela bay (54°52'S, 67°16'W), located 70 km East from Golondrina bay, is considered a low impacted area, being mostly visited by local fi shermen and tourists during the spring-summer season.The nearest place where environmental data has been reported is Brown bay (54°52' S, 67°34'W) (Giarratano & Amin 2010), 20 km west of Varela bay.At that place, only a commercial production of mussels is being developed since 2001 (Hernando et al. 2008 1 ).

Fish collection and processing
Ten adult specimens (185 -210 mm total length, TL; 35 -55 g body weight, BW) were collected at both sampling sites by means of a seine net (60 m long, 1.5 m high and 12 mm stretched mesh) covering ca.200 m, from 1 m depth to the shallower littoral zone.Fishes were carried alive to the laboratory into 100 L tanks containing aerated seawater.In order to minimize stress generated during capture and transportation, fi shes were held in the laboratory during 24 h under similar fi eld conditions (8 ± 0.5 °C; light:dark photoperiod of 14:10 h).According to Cazenave et al. (2009) and Nahrgang et al. (2010), twenty-four hours after catch fi shes from both sampling sites were sacrifi ced by a blow to the head.TL of each individual was measured with a digital calliper (± 0.1 mm) and BW was recorded to the nearest 0.01 g.Condition of individual fi sh was indexed by Fulton's condition factor calculated as K = [BW x [TL 3 ] -1 ] x 100 (Bagenal & Tesch 1978).Considering that Lattuca et al. (2009) found no signifi cant differences between sexes in oxidative stress biomarkers for O. nigricans, sex differentiation was not made in this sudy.Gills and liver were dissected, weighed (± 10 -5 g) and stored at -80 ºC for 2 weeks, until analysis.

Environmental physicochemical parameters
At each sampling site, surface water temperature, pH, conductivity and dissolved oxygen were recorded once in situ by means of an HORIBA U-10 multiparameter device.Concentration of chlorophyll-a was measured according to Holm-Hansen & Riemann (1978) using a Sequoia Turner (model 450) fl uorometer.

Quantifi cation of lipid radicals by electron paramagnetic resonance (EPR)-spin trapping
Lipid radicals were detected by a spin trapping technique using N-t-butyl-α-phenyl nitrone (PBN).A 40 mM PBN stock solution was prepared in dimethyl sulfoxide (DMSO) immediately prior to use.The homogenates were prepared by adding DMSO-PBN (stock solution) in frozen tissue (10 to 20 mg).EPR spectra were obtained at room temperature using a Bruker spectrometer ECS 106, operating at 9.81 GHz with 50 kHz modulation frequency.EPR instrument settings for the spin trapping experiments were: microwave power, 20 mW; HOFFMEYER (2008).Informe Final: Monitoreo planctónico y ambiental para el desar rollo sustentable del cultivo comercial de mejillón (Mytilus edulis chilensis) en la zona de Almanza, Canal Beagle (Tierra del Fuego).1986).Quantifi cation was performed according to Kotake et al. (1996).

Content of lipid soluble antioxidants
The contents of α-tocopherol (α-TH) and β-carotene in homogenates from gills and liver were quantifi ed by reverse-phase HPLC with electrochemical detection using a Bioanalytical Systems LC-4C amperometric detector with a glassy carbon working electrode at an applied oxidation potential of 0.6 V. Samples were extracted with 1 mL of ethanol and 4 mL of hexane.
dismutase activity was determined by the epinephrine method (Misra & Fridovich 1972), based on the capacity of SOD to inhibit the autooxidation of epinephrine to adrenochrome at 480 nm.

Ascorbyl radical content (A • )
A • measurements were done using a Bruker ECS 106 spectrometer.Gills and liver homogenates were prepared by adding 100-150 mg of tissue to pure DMSO (1:3) (w/v).The spectra were scanned at room temperature under the following conditions: 50 kHz fi eld modulation, microwave power 20 mW, modulation amplitude 1G, time constant 655 ms, receiver gain 1 x 10 5 , microwave frequency 9.81 GHz, and scan rate 0.18 G/s (Giulivi & Cadenas 1993).The amount of spin adduct was calibrated using an aqueous solution of TEMPOL, introduced into the same cell used for spin trapping.EPR spectra of spin adduct and TEMPOL solutions were recorded at exactly the same spectrometer settings.The fi rst-derivative EPR spectra were doubly integrated to obtain the area intensity, and then the concentration of spin adduct was calculated using the ratio of these areas.Quantification was performed according to Kotake et al. (1996).

Ascorbate content (AH -)
The content of AH − was measured by reverse-phase HPLC with electrochemical detection using a LC-4C amperometric detector with a carbon working electrode at an applied oxidation potential of 0.6 V.The samples were homogenized in metaphosphoric acid (10 %, w/v) according to Kutnink et al. (1987).A Supelcosil LC-18 column was stabilized with metaphosphoric acid (0.8 %, w/v) and a freshly prepared solution of ascorbic acid in metaphosphoric acid (10 %, w/v) (1 µg mL -1 ) was used as standard.

Reduced glutathione (GSH) and oxidized glutathione (GSSG) contents
In order to assess GSH and GSSG contents, homogenates from gills and liver were prepared in 1.0 M HClO 4 and 2 mM EDTA (1:5) (w/v).After centrifugation at 29000 g and 4 °C for 20 min, the lipid layer was discarded.Meanwhile, the protein free supernatant was fi ltered with a 0.22 µm nylon membrane and immediately analyzed.The contents of GSH and GSSG were quantifi ed by reverse-phase HPLC analysis (isocratic modality) following the method described by Rodriguez-Ariza et al. (1994).The equipment employed was: a Perkin Elmer 250 LC bomb, a fixed phase: Supelcosil LC-18 column (5 µm size particle) 4.6 x 250 mm, a mobile phase: 20 mM NaH 2 PO 4 , pH 2.7, and an electrochemical detector: ESA Coulochem II with an analytical cell ESA 5011, at an applied potential of +0.45 V and 0.80 V and a 1.2 mL min -1 fl ow.Quantifi cation of GSH and GSSG was performed through a standard curve with a linear relationship between 5-200 ng GSH or GSSG, respectively.

Statistical analyses
Data in the text and tables are expressed as mean ± SE for three replicates of each of the fi ve individuals from each sampling site.Statistical test of Mann-Whitney was carried out using GraphPad InStat, version 3.01.When the two-tailed P value was < 0.05, it was considered signifi cant.

Enzymatic assays
Homogenates from gills and liver were prepared in 50 mM potassium phosphate 120 mM KCl (pH = 7.4).

RESULTS
Physicochemical parameters of seawater were similar in both bays (Table 1).
The condition of the whole body was used as an initial screening biomarker, measured through K condition factor.Dif ferences in K values were obser ved between sampling sites where fi shes from Golondrina bay had significant higher K factor than those from Varela bay (Fig. 3, Mann-Whitney test; U = 2.000; n = 10; P = 0.032).
in the present study as lipid radical content assessed by EPR.Lipid radicals in both organs combined with the spin trap PBN resulted in adducts that gave a characteristic EPR spectrum with hyperfi ne coupling constants of a N = 15.56G and a H = 2.79G, in concordance with computer spectral simulated signals obtained using the overall mentioned parameters.PBN itself was examined and no PBN spin adduct was obser ved.Spectra showed a signal from a second radical species with hyperfi ne coupling constants a N = 15.91 and a H = 3.19, corresponding to a carbon-centered endogenous radical of unknown origin.However, even though the signal was not fully identifi ed, it may refl ect the formation of a PBN-CH 3 adduct.Since hydroxyl radical formation in vivo may occur under these experimental conditions, its reaction with DMSO could result in the generation of methyl radical that could be trapped by PBN, forming the adduct PBN-CH 3 detected by EPR (Mason et al. 1994, Dikolova et al. 2001).Neither lipid radical content in gills (Golondrina bay: 3.7 ± 0.7 pmol mg -1 FW; Varela bay: 2.5 ± 0.9 pmol mg -1 FW; Mann-Whitney test; U = 7.000; n = 10; P = 0.309) nor in the liver of O. nigricans (Golondrina bay: 0.8 ± 0.2 pmol mg -1 FW; Varela bay: 0.99 ± 0.08 pmol mg -1 FW; Mann-Whitney test; U = 7.000; n = 10; P = 0.309) showed signifi cant differences between sampling sites.When comparing the lipid radical content between organs of fi shes from the same sampling site, only those from Golondrina bay showed signifi cant differences, with higher values in gills (Mann-Whitney test; U = 0.000; n = 10; P = 0.008).Chlorophyll-a (µg L -1 ) 0.29 ± 0.07 0.22 ± 0.07 Oxidative damage to lipid in the gills and the liver of silversides was estimated The content of non-enzymatic antioxidants and the activity of the antioxidant enzymes were also evaluated in both organs assayed.Independently of the sampling site, all of them were lower in gills than in the liver.R e g a r d i n g t h e n o n -e n z y m a t i c l i p i d antioxidants, the contents of α-tocopherol (Mann-Whitney test; U = 2.000; n = 10; P = 0.063) and β-carotene (Mann-Whitney test; U = 3.000; n = 10; P = 0.25) in the gills did not showed significant dif ferences between sampling sites (Table 2).In the liver (Table 2), the content of α-tocopherol was signifi cantly higher in fi shes from Varela bay than in those from Golondrina bay (Mann-Whitney test; U = 0.000; n = 10; P = 0.029), whereas the content of β-carotene did not show signifi cant differences between sampling sites (Mann-Whitney test; U = 8.000; n = 10; P = 0.421).The contents of both non-enzymatic lipid antioxidants were significantly higher in the liver than in the gills of fishes from the same sampling site (α-tocopherol for Varela bay: Mann-Whitney test; U = 0.000; n = 10; P = 0.016; α-tocopherol for Golondrina bay: Mann-Whitney test; U = 0.000; n = 10; P = 0.008; β-carotene for Varela bay: Mann-Whitney test; U = 0.000; n = 10; P = 0.036 and β-carotene for Golondrina bay: Mann-Whitney test; U = 1.000; n = 10; P = 0.015).
A high activity of CAT was also registered in the liver, being 53 % higher in fi shes from Varela bay than in those from Golondrina bay (Table 2, Mann-Whitney test, U = 27.000;n = 10; P = 0.050).The activity of CAT could not be determined in the gills due to the lack of sample.The activity of SOD in the gills did not differ signifi cantly between sampling sites (Table 2, Mann-Whitney test; U = 10.000;n = 10; P = 0.429), but it did in the liver, being 44 % higher in Varela bay than in Golondrina bay (Table 2, Mann-Whitney test; U = 0.000; n = 10; P = 0.009).No signifi cant differences were observed in the activity of SOD between organs at each sampling site (Table 2, Varela bay: Mann-Whitney test; U = 14.000; n = 10; P = 0.931; Golondrina bay: Mann-Whitney test; U = 10.000;n = 10; P = 0.240).
Typical EPR spectra of A • in the gills and in the liver of fi shes collected at the two sampling sites were observed with the characteristic two lines at g = 2.005 and a H = 1.8 G, in accordance to computer spectral simulated signals obtained with the parameters given in the Material and Methods section.Neither the content of A • (Mann-Whitney test; U = 11.000;n = 10; P = 0.841) and AH -(Mann-Whitney test; U = 9.000; n = 10; P = 0.548) nor the A • content/AH -content ratio in gills (Mann-Whitney test; U = 11.000;n = 10; P = 0.841) did differ signifi cantly between Varela and Golondrina bays (Fig. 4).Typical EPR spectra of A • in the liver are shown in Fig. 5A.The contents of A • (Mann-Whitney test; U = 0.000; n = 10; P = 0.050), AH -(Mann-Whitney test; U = 0.000; n = 10; P = 0.036) and the A • / AH -ratio (Mann-Whitney test; U = 0.000; n = 10; P = 0.050) assessed in the liver (Fig. 5B) were significantly dif ferent between both sampling sites.Signifi cant higher values of A • content in gills than in liver were measured at Varela (Mann-Whitney test; U = 1.000; n = 10; P = 0.016) and Golondrina (Mann-Whitney test; U = 1.000; n = 10; P = 0.016) bays.The A • /AH - ratio was also signifi cantly higher in the gills of fi shes from Golondrina bay than in those from Varela bay (Mann-Whitney test; U = 2.000; n = 10; P = 0.05; Figs. 4 and 5B).

DISCUSSION
In the present study, oxidative stress responses in fishes from Golondrina and Varela bays were employed to evaluate the usefulness of a multibiomarker approach in environmental monitoring programs in the Beagle Channel region.Golondrina bay is located near Ushuaia city, the southernmost location on Earth, and has been recently considered by Duarte et al. (2011) as low urban impacted area, despite the  population's growth around it.Varela bay does not receive anthropogenic inputs, been only visited by local fi shermen and tourists during a short part of the year.In order to assess the link between spatial and temporal changes in quality of coastal waters of the Beagle Channel and antioxidant responses, different works have been previously carried out in native (Duarte et al. 2011(Duarte et al. , 2012) ) and transplanted (Giarratano et al. 2010(Giarratano et al. , 2011) ) mussels Mytilus edulis chilensis (D'Orbigny, 1847) in the Beagle Channel.They found signifi cant differences in oxidative stress responses analyzed in relation to dissolved nutrients seawater and trace metals in the fi nest sediment fraction (< 62 µm).Sediments from Brown bay, which is only 20 km West of Varela bay, showed higher levels of Fe and Zn than those from Golondrina bay (Giarratano & Amin 2010, Duarte et al. 2012).Even though fi shermen's activities could be contributing to the trace metal enrichment of sediments from Brown bay, Giarratano & Amin (2010) believe that most of the concentrations measured derive from the Andean belt itself, since spurts of variable fl ow reach this coast associated to thawing and drift of materials contributed by Almanza river (Hernando et al. 2008).
The xenobiotics such as metals are known to induce ROS, which are responsible for oxidative stress (Viarengo et al. 1990, Almroth et al. 2005, Gonzalez & Puntar ulo 2011).Assuming that the higher content of Fe and Zn found in fi ne sediments from Brown bay could be similar in Varela bay due the its proximity, these metals could be resuspended in the water column by wave action and/or by physico-chemical changes.That fact could be responsible of the increase of the lipid radical content and the antioxidant activities in the liver, affecting O. nigricans from Varela bay through the decrease of the condition factor.However, measurements of trace metals in sediments from Varela bay should be carry out to test this hypothesis.
In several works the K condition factor has proven to be a useful biomarker to indicate dif ferent environmental stress conditions among sites with different characteristics (van der Oost et al. 2003, Linde-Arias et al. 2008).However, this parameter is not sensitive enough to give information of specifi c responses to the toxic substances in the environment.Presence of contaminants in the environment frequently leads to the depletion of energy reser ves as a compensator y mechanism to the higher demand of energy required by detoxifi cation processes (Guerlet et al. 2006).Fishes from Golondrina bay showed K values signifi cantly higher than those from Varela bay making evident that the former place is a site with less stressful environmental conditions.
The degree of oxidative damage on the fish O. nigricans was assessed through the estimation of lipid peroxidation products in the gills and in the liver.Gills were chosen for being the fi rst organ which gets in contact with the environment becoming a potential target for oxidative disruption.Liver was chosen because it possesses high potential for ROS generation, which seems to be effi ciently counterbalanced by powerful protective mechanisms to detoxify and repair damaged lipid and proteins (Oliveira et al. 2008, Lattuca et al. 2009, Nahrgang et al. 2010).Endogenous lipid radical products have extremely short half-lives and are present in low concentrations, making diffi cult their detection.Spin trapping analysis followed Reduced and oxidized glutathione and GSH / GSSG ratio in the gills and the liver of O. nigricans from the different sampling sites.Data are expressed as mean ± S.E. for three replicates of each of the fi ve individuals from each sampling site.
Glutatión reducido y oxidado y relación contenido de GSH / contenido de GSSG en las branquias y el hígado de O. nigricans de los distintos sitios de muestreo.Los datos están expresados como la media ± EE de tres réplicas para cada uno de los cinco individuos de cada sitio de muestreo.

Varela bay Golondrina bay
Gills Liver Gills Liver GSH (ng mg -1 FW) 10 ± 3 6 ± 1 6 ± 4 3 ± 1 GSSG (ng mg -1 FW) 3 ± 1 37 ± 9 3.3 ± 0.9 5 ± 2 GSH/GSSG 3.8 ± 0.9 0.2 ± 0.1 2.5 ± 0.8 1.3 ± 0.3 by EPR overcomes the limit of sensitivity of endogenous radicals in biological systems, and it has proved to be the least ambiguous method to detect shor t-lived and reactive free radicals generated in low concentrations in biological systems (Luo et al. 2006) Fishes tend to adapt to oxidative conditions by mobilizing enzymatic as well as nonenzymatic antioxidant defences (Ahmad et al. 2008, van der Oost et al. 2003).SOD and CAT activities, the α-tocopherol content and the high A • /AH -ratio found at Varela bay suggest that the natural environmental stress at this site is higher than at Golondrina bay.The enzymatic antioxidants SOD and CAT provide the fi rst defence against oxygen toxicity.A simultaneous induction of the activities of SOD and CAT has been obser ved in organisms exposed to pollutants (Dimitrova et al. 1994).In that sense, such relationship found in liver in this study could be attributable to higher natural levels of some trace metals in Varela bay.Jena et al. (2009) proposed the use of ascorbic acid as a biomarker for environmental monitoring.The content of AH -in Golondrina bay was higher than in Varela bay.This may be either due to dietary defi ciency or the utilization of ascorbic acid in response to elevated ROS generation (Oakes et al. 2004).The pattern of variation in ascorbic acid content in O.nigricans is comparable to that repor ted by Jena et al. (2009) for Perna viridis (Linnaeus, 1758) from the coast of India.Further studies would fully describe the use of ascorbic acid as a biomarker for environmental monitoring in the studied areas.
GSH is a co-factor for GPx and GST enzymes and is also an effective protection capable of quenching oxyradicals (Ross 1988).It may be either a result of the novo synthesis as previously demonstrated in fi sh or a transfer from other organs (Deneke & Fanburg 1989, Oliveira et al. 2009).An increment of GSH was observed in the liver of fi shes from Varela bay, which could be due to the fact that most GSH is synthesized in liver and transported to the bloodstream for supply to other tissues.The same explanation was given by Almroth et al. (2008) for rainbow trout Oncorhynchus mykiss (Walbaum, 1792).Similarly, DiGiulio et al. (1993) observed high levels of GSH in catfish exposed to polluted sediment.The ratio between the reduced GSH and oxidized forms of GSH (GSSG) plays an important role in intercellular signalling and gene regulation (Almroth et al. 2008) and may be a potential biomarker for oxidative stress (van der Oost et al., 2003;Jena et al., 2009).In the current study, the lowest hepatic GSH/GSSG ratio was found at Varela bay indicating a more oxidized redox status.Similar responses of decreased GSH/GSSG ratio have been shown in Perna perna (Linnaeus, 1758) exposed to lead and paraquat (Dafre et al. 2004) and in Mytilus galloprovincialis (Lamarck, 1819) in response to an oil spill in Spain (Sureda et al. 2011).
Taken as a whole, the antioxidant responses were higher in the liver making it an ideal tissue for biomarker studies in O. nigricans.These results are in concordance with those found by Lattuca et al. (2009) who showed that the liver of O. nigricans exhibited a better control of the oxidative damage than the gills, allowing minimization of intracellular damage when it is exposed to environmental stressful conditions.The current study indicates that this species also displays an antioxidant response spatially differentiated, in order to counteract pro-oxidant natural factors present in these marine habitats.
We studied silversides from two coastal environments from Beagle Channel and

TABLE 1
Physicochemical parameters of water from the different sampling sites.Data are expressed as mean ± S.E. for three replicates of each of the fi ve individuals from each sampling site.Parámetros fi sicoquímicos del agua de los distintos sitios de muestreo.Los datos están expresados como la media ± EE de tres réplicas para cada uno de los cinco individuos de cada sitio de muestreo.

TABLE 2
Antioxidant capacity of O. nigricans f rom the different sampling sites.Data are expressed as mean ± S.E. for three replicates of each of the fi ve individuals from each sampling site.ND: not determined.Capacidad antioxidante de O. nigricans de los distintos sitios de muestreo.Los datos están expresados como la media ± EE de tres réplicas para cada uno de los cinco individuos de cada sitio de muestreo.ND: no determinado.