Aquatic Toxicology 105 (2011) 508517

Contents lists available at ScienceDirect

Aquatic Toxicology
journal homepage: www.elsevier.com/locate/aquatox

Risks of seawater ozonation in recirculation aquaculture Effects of oxidative stress on animal welfare of juvenile turbot (Psetta maxima, L.)
S. Reiser a,b, , S. Wuertz b,c , J.P. Schroeder a,c , W. Kloas b,d , R. Hanel e,a
a

Leibniz-Institute of Marine Sciences, IFM-Geomar, Duesternbrooker Weg 20, 24105 Kiel, Germany Leibniz-Institute of Freshwater Ecology and Inland Fisheries, Mueggelseedamm 310, 12587 Berlin, Germany Gesellschaft fuer Marine Aquakultur, Hafentoern 3, 25761 Buesum, Germany d Institute of Biology, Department of Endocrinology, Humboldt University, Invalidenstrasse 43, 10115 Berlin, Germany e Institute of Fisheries Ecology, Johann Heinrich von Thuenen-Institute (vTi), Federal Institute for Rural Areas, Forestry and Fisheries, Palmaille 9, 22767 Hamburg, Germany
b c

a r t i c l e

i n f o

a b s t r a c t
Ozone is frequently used for water treatment and disinfection in recirculating aquaculture systems. However, due to the fragmentary data on chronic toxicity of ozone produced oxidants (OPO) and its safe concentrations, the daily application of ozone in aquaculture is challenging. To evaluate the chronic effects of sublethal OPO concentrations, juvenile turbot (Psetta maxima, L.) were exposed to OPO concentrations of 0.06, 0.10 and 0.15 mg/l for 21 days. Gills were analysed for histopathological alterations and mRNA expression of heat shock protein 70 (hsp70), hsp90 as well as glutathione-S-transferase (gst) were determined in the gills and the liver after 1 d, 7 d and 21 d. Histopathologic ndings conrmed adverse effects at 0.100.15 mg/l, but these (necrosis, lamellar clubbing, hypertrophy, hyperplasia) could only be observed after an extended exposure (mostly 21 d), and were considered as irreversible tissue damage. Hsp70 expression in the gills was only signicantly increased at the highest OPO concentration (0.15 mg/l) on 1 d and 7 d, and returned to basic levels until day 21. Hsp90 mRNA was already increased at 0.10 mg/l after 1 and 7 days of exposure, and again was comparable to the control group on day 21. In contrast, elevated gst mRNA expression was only observed on day 7 at 0.10 mg and 0.15 mg/l. Although similar trends were observed in the liver for all markers, differences were only signicant in exceptional cases due to the high individual variation observed. Thus, mRNA expression in the gills rather than in the liver is recommended as a marker to characterize OPO-induced oxidative stress in turbot. It has to be noted that mRNA expression returned to basic levels on day 21 regardless the actual OPO concentration, suggesting a collapse of adaptive mechanisms as a possible explanation for the observed tissue damage. 2011 Elsevier B.V. All rights reserved.

Article history: Received 30 May 2011 Received in revised form 3 August 2011 Accepted 8 August 2011 Keywords: Turbot RAS GST HSP70 Heat shock protein HSP90 Histology Oxidative stress Ozone

1. Introduction Governed by the perception that many wild stocks are rapidly being depleted and ow-through aquaculture pollutes local water resources, recirculating aquaculture systems (RAS) are increasingly being used to meet the requirements of a sustainable aquaculture production (Waller, 2000; Martins et al., 2010). An effective reduction of pathogens such as bacteria, viruses and fungi is of particular importance in intensive RAS, which is mostly accomplished by the use of oxidizing disinfectants (Liltved et al., 1995). The most frequently applied disinfectant in marine RAS is ozone, comprising additional benets like the removal of metabolic waste products, toxicants or dissolved organic matter (Rosenthal and

Corresponding author at: Institute for Hydrobiology and Fisheries Science, University of Hamburg, Olbersweg 24, 22767 Hamburg, Germany. Tel.: +49 40 428387874; fax: +49 40 42838 6618. E-mail address: [email protected] (S. Reiser). 0166-445X/$ see front matter 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.aquatox.2011.08.004

Otte, 1979; Summerfelt and Hochheimer, 1997). However, excessive ozonation may severely impact farmed sh by causing adverse effects including histopathologic tissue damage (Richardson et al., 1983; Reiser et al., 2010), alterations in feeding behaviour (Reiser et al., 2010) as well as oxidative stress (Ritola et al., 2000, 2002b; Livingstone, 2003). Although ozone is intensively used in marine aquaculture systems, knowledge on the toxic impact of ozone on marine aquaculture organisms is limited. Findings from freshwater studies cannot be projected to marine conditions, since emerging ozone-induced oxidants differ substantially in marine and freshwater systems (Hoign et al., 1985; Oemcke and van Leeuwen, 1998). Ozone has a very short half life in seawater of only a few seconds (Haag and Hoign, 1984). It instantaneously reacts with different chemical compounds resulting in the formation of several reactive species, generally summarized as ozone-produced oxidants (OPO). Particularly halogen ions are oxidized to halogenated oxidants by ozone (Haag and Hoign, 1984; von Gunten, 2003; Perrins et al., 2006). Low concentrations of iodide ions in typical seawater as well as the slow rst order rate constant for

S. Reiser et al. / Aquatic Toxicology 105 (2011) 508517

509

ozones reaction with chloride ions (k = 3.0 103 M/s) (Hoign et al., 1985) make the respective oxidation products less important. However, with relatively high bromide-ion concentrations of 6070 mg/l and an ozone reaction rate constant of approximately 160 M/s, there is a high potential for the formation of brominated oxidants in ozonated seawater (Hoign et al., 1985). Hence, the brominated oxidants free bromine (HOBr/OBr ) and bromamines (NH2 Br, NHBr2 ) represent the majority of OPO in marine RAS. As these OPO are more stable compared to ozone itself (Blogoslawski et al., 1976; Crecelius, 1979), they have a high potential to accumulate within RAS (Oemcke and van Leeuwen, 1998; Herwig et al., 2006; Perrins et al., 2006) and become toxic to the cultivated organisms. OPOs were reported to oxidize membrane lipids of epithelia cells, thereby affecting respiratory function and osmoregulation (Giese & Christensen, 1954; Block et al., 1978; Richardson et al., 1983). Furthermore some OPOs are suggested to enter the cell. Since cellular stress response represents a defence reaction of cells to damage that environmental forces inict on macromolecules, it is not stressor specic (Iwama et al., 1998; Basu et al., 2002). Among several molecules involved, heat shock proteins (hsp) comprise a family of highly conserved cellular proteins upregulated in response to cellular stress, i.e. damaged macromolecules (Iwama et al., 1998; Lewis et al., 1999) which are further specied according to their molecular weight (kDa). In unstressed cells, there is a basic production of hsp. Hsp are involved in numerous aspects of protein metabolism maintaining cellular integrity and homeostasis (Fink and Goto, 1998) with hsp70 and hsp90 being the most frequently studied heat shock proteins. Hsp70 assists the folding of nascent polypeptide chains, acts as a molecular chaperone and mediates the repair and degradation of altered or denatured proteins. Hsp90 supports numerous components of the cytoskeleton, enzymes and steroid hormone receptors (Fink and Goto, 1998). Hsp70 and hsp90 have been used intensively as biomarkers for the evaluation of stressors in sh including toxicants (Hassanein et al., 1999; Boone and Vijayan, 2002), stocking density (Gornati et al., 2004, 2005) and poor water quality such as high nitrite loads (Deane and Woo, 2007). In stressed cells, hsps act as an alarm response, maintaining cellular functioning by increased biosynthesis. Contrary to hsps, detoxifying enzymes are specic to a stressor, i.e. substrate specic, ameliorating stressful states by detoxication and re-establishment of homeostasis. In general, detoxication is divided in two phases, i.e. phase-I enzymes that metabolize the stressor/toxicant in a rst step, while phase-II enzymes act on secondary metabolites induced by the primary stressor, e.g. oxidized lipids, etc. (van der Oost et al., 2003). Glutathione S-transferases (GSTs) represent a multigene family of primarily soluble enzymes that are involved in detoxication of electrophilic and genotoxic xenobiotics, as well as endogenous reactive intermediates produced during cellular oxidative stress like peroxidative products of DNA and lipids (Hayes and Pulford, 1995; van der Oost et al., 2003). Hayes and Pulford (1995) suggested that the level of gst expression determines the sensitivity of cells towards a broad spectrum of toxicants. As an enzyme contributing to phase-II metabolism, Gsts conjugate electrophilic xenobiotics and phase-I metabolites, i.e. organic peroxides, with the nucleophilic tripeptide glutathione (GSH) from the intracellular pool (Halliwell and Gutteridge, 1999; van der Oost et al., 2003). Thus, Gst inactivates lipoperoxidation products, lipid hydroperoxides and their derivates (Salinas and Wong, 1999). So far, alterations in Gst activity have been observed in sh exposed to different toxicants (Ferrari et al., 2007; Bouraoui et al., 2008; Pandey et al., 2008) as well as ozone (Ritola et al., 2000). To safeguard good welfare within aquaculture production systems, potential stressors like OPO have to be evaluated, in particular

with regard to chronic effects. In the present study, we monitored chronic adverse effects on juvenile turbot (Psetta maxima) caused by ozonation and evaluated the suitability of hsp70, hsp90 and gst as biomarkers for oxidative stress. Upon exposure of juvenile turbot to three different sublethal OPO concentrations (0.06, 0.10 and 0.15 mg/l) for 21 days, hsp70, hsp90 and gst mRNA expression in the liver and gills were determined by RT-qPCR. The results were discussed with regard to histopathological alterations and abnormalities of the gills. Major aim of the present study was to ultimately dene a threshold for a safe, non-hazardous OPO concentration for juvenile turbot. We used the nomenclature ozone-produced oxidants (OPO) to refer to the total residual oxidant concentration, measured spectrophotometrically as total residual chlorine.

2. Materials and methods 2.1. Experimental setup Juvenile turbot (12.2 1.0 cm TL, 36.4 8.2 g wet weight) were purchased from a commercial turbot hatchery (Maximus Fry, Denmark) and randomly stocked to 12 identical recirculation systems providing triplicates of 23 sh for each OPO treatment and the control group. Recirculation systems were set up in a temperature controlled lab in order to ensure constant environmental conditions as recommended by Hurlbert (1984). Each recirculation system comprised of a 200 l berglass tank (Chemowerk) lled with approximately 150 l of ltered natural seawater and an individual biolter and foam fractionator (by-pass operated). Following an acclimation period of 10 days, juvenile turbot were exposed to OPO concentrations (SD) of 0.00, 0.06 0.01, 0.10 0.01 and 0.15 0.02 mg/l for 21 days. Fish were kept at a 12:12 L:D photoperiod (Pichavant et al., 1998) and fed with a commercial pellet feed (DAN-EX 1562, 1.5 mm; Dana Feed) to apparent satiation (accounting for approximately 2% of body weight) in 48 h intervals to minimize uctuations in OPO concentrations. Ozone gas was produced by electrical discharge ozone generators (ModelC 200, Erwin Sander Elektroapparatebau GmbH) using compressed air. Ozone-enriched air was injected into the seawater via foam fractionators (Model 1 AH 1100, Erwin Sander Elektroapparatebau GmbH) operating in counter-ow mode to ensure maximum diffusion of ozone. The retention time in the foam fractionator was set to approximately 1 min. Ozonated water was introduced at high ow rates (600 l/h) via vertical spray bar and supply of ozone was controlled and regulated by a redox potential controller (Erwin Sander Elektroapparatebau GmbH) using the redox potential as proxy for the total oxidant concentration (Buchan et al., 2005). OPO concentrations were determined spectrophotometrically in 2 h intervals over the entire experimental period and adjusted if necessary, using the colorimetric N,Ndiethyl-p-phenylenediamine (DPD) method (DPD Total Chlorine powder pillows, Hach) at 530 nm with a DR/2800 Spectrophotometer (Hach Lange), as recommended for the quantication of OPO in seawater (American Public Health Association, 1989; Buchan et al., 2005). OPO concentrations were expressed as chlorine equivalents. Physical and chemical water parameters (temperature, salinity, pH, dissolved oxygen, total ammonia-N, nitrite-N) were measured at 2 day intervals as described (Reiser et al., 2010). Mean water quality (SD) was: temperature: 14.29 0.45 C, salinity: 18.10 0.25, pH: 7.57 0.07, dissolved oxygen concentration: 10.48 0.16 mg/l, total ammonia-N (TAN): 0.93 0.53 mg/l and nitrite-N: 0.41 0.84 mg/l.

510

S. Reiser et al. / Aquatic Toxicology 105 (2011) 508517

Table 1 Primers used for real-time RT-qPCR providing NCBI accession code, nucleotide sequence, position and amplicon size. Act (-actin), 2m (-2-mikroglobuline), pol (RNApolymerases II), hsp70 (heat shock protein 70), hsp90 (heat shock protein 90) and gst (glutathione-S-transferase) genes. Target act 2 m pol hsp70 hsp90 gst Primer actF actR 2mF 2mR polF polR hsp70F hsp70R hsp90F hsp90R gstF gstR Sequence 5 -AgCACggTATTgTgACCAAC-3 3 -AgCTTggATggCAACgTAC-5 5 -gATCTgCCACgTgAgCg-3 3 -AggAACggCACgCTCTT-5 3 -AgACgCACAgACCTCACg-5 3 -ggTgCCATgTCgACTAgAgAC-5 5 -CTgTCCCTgggTATTgAgAC-3 3 -gAACACCACgAggAgCA-5 5 -CCgCCTACCTCgTTgC-3 3 -gTAgCCgATgAACTgCgAgT-5 5 -gggTTCgCATCgCTTTT-3 3 -ggCCTggTCTCgTCTATgTACT-5 Position 142 327 56 183 96 285 1162 1366 401 610 9 183 Accession nr EU686692 EU878757 EU925140 EF191027 EU099575 DQ848966 Amplication size 185 127 189 204 209 174

2.2. Sampling Following 1, 7 and 21 days of OPO exposure, sh were randomly sampled by dip-netting and immediately killed in icewater as recommended (FAO, 2009). After dissecting the gill apparatus, rst and second left gill arches were sampled for gene expression analysis and stored in RNAlater. For histological analysis the second right gill arch was stored in 4% sodiumtetraborate-buffered formalin. A liver sample was extracted through an incision in the ventral cavity. For gene expression analysis, approximately 100 mg of liver and gill tissue were immediately incubated in 1 ml RNAlater (100 mg/ml) for 24 h at 4 C and stored at 20 C. 2.3. Histology Gill arches were dehydrated in a graded series of ethanol, cleared with xylene and inltrated and embedded in parafn according to Wuertz (2005). Cross-sections of 2 m were haematoxylineosin stained and analysed using a Leica DMRBE microscope equipped with a digital camera (DFC320, Leica). Five primary laments from the second right gill arch per individual were randomly chosen and 20 lamellae per lament were blindly selected and analysed. The histopathological alterations were recorded as percentage of affected secondary lamellae and assigned to the categories not affected, necrotic, lamellar fusion, epithelial lifting, clubbed lamellar tips, hypertrophy and hyperplasia according to Mallatt (1985), Crespo et al. (1987), Frances et al. (1998) and Gisbert et al. (2004). 2.4. RNA extraction Extraction of total RNA from stored tissue samples was performed using TRIZOL (Invitrogen) according to Kroupova et al. (2011). In brief, tissue samples were homogenized twice in 700 l TRIZOL for 1.5 min at 30.0 l/s using a tissue lyser (Qiagen), 600 l of TRIZOL were additionally added, followed by centrifugation for 10 min (12000 g, 4 C). To assure a high RNA quality, a subsample of the supernatant (300 l liver, 700 l gill) was used for the subsequent extraction. Adding further TRIZOL to 1.3 ml and 250 l of chloroform, samples were vortexed (15 s), and incubated for 10 min at room temperature (RT). Upon centrifugation (15 min, 12000 g, 4 C), 270 l of the aqueous phase were transferred to a new vial, 270 l isopropanol (Roth) were added and the sample was incubated at room temperature for 10 min. After centrifugation (15 min, 12000 g, 4 C), isopropanol was removed, the pellet was washed with 70% ethanol (300 l) and allowed to dry after the removal of ethanol. Pellets were redissolved in DEPC-treated water (0.01%). A DNAse I treatment was used to assure DNA-free RNA for transcription (Qiagen). Total RNA content was measured by Ribogreen RNA-quantication kit (Invitrogen) in a Spectrauor

plus microplate reader (Tecan) and RNA integrity was controlled on a RNA 6000 Nano LabChip (Agilent Technologies) as described (Trubiroha et al., 2009). RNA (1 g) was transcribed using a poly(dT) primer [CCTGAATTCTAGAGCTCA(T)17] in a two step procedure, (1) annealing of poly(dT) primer at 70 C for 2 min [reaction mix: 0.625 mol poly(dT)], chilled on ice for 3 min followed by (2) synthesis at 36 C for 60 min [reaction mix: AMV buffer, 10 U AMV-RT, 0.5 mM dNTP] and denaturation at 94 C for 3 min. Genomic DNA contamination was excluded by an -RT control where the AMV-RT had been replaced by DEPC treated water. 2.5. Real-time quantication Using gene-specic primers (Table 1), real-time PCR was carried out with a Mx3005p qPCR Cycler (Stratagene) using hot start polymerase (Platinum, Invitrogen) and SYBR Green (25 l reaction volume: 4 l diluted cDNA, 400 nM of each primer, 1* Taq buffer, 5 mM MgCl2 , 10 nM of each dNTP, 1*SYBR-Green I solution, 1 U Taq). Specic amplication of the assays was conrmed by direct sequencing and multiple sequence alignment analysis as previously described (Wuertz et al., 2007). PCR conditions were: 95 C initial denaturation for 8 min, followed by 40 cycles of 95 C denaturation for 30 s, 56 C primer annealing for 30 s and 72 C polymerisation for 40 s. A melting curve was included at the end of the PCR to conrm specic amplication. All PCR reactions were run in duplicate, and only considered if Ct < 0.5. For normalisation, three commonly used housekeeping genes (-actin (act), -2-mikroglobuline (2m), RNApolymerases II (pol)) were evaluated comparing CT values of a subset of samples over time, between and within treatments. Evaluation revealed 2m as most constantly expressed housekeeping gene. It was consequently used for normalisation by the comparative CT method (Pfaf, 2001). For each run, quantities were calculated from a standard dilution series of a pooled liver and a pooled gill sample compensating for lot-to-lot and run-to-run variations. The target transcript expression of genes is presented as fold increase of the control at each day. 2.6. Statistical analysis Data are presented as mean standard deviation and were analysed using STATISTICA 6.1 (StatSoft). Data were calculated as mean for each single tank and compared between replicates and treatments (Hurlbert, 1984). When conformed, factorial ANOVAs were conducted and signicance was analysed by Tukey HSD test for equal sample sizes or unequal N HSD test, respectively. If a signicant interaction between the two factors exposure time and OPO concentration was observed, the effect of each single factor was investigated by one-way ANOVA. If normality or homogeneity

S. Reiser et al. / Aquatic Toxicology 105 (2011) 508517

511

Fig. 1. Histopathological alterations of the secondary gill laments of juvenile turbot (Psetta maxima) upon exposure to OPO concentrations of 0.06, 0.10 and 0.15 mg/l compared to the control at day 1, 7 and 21. The major histopathological categories analysed are not affected, necrosis, lamellar clubbing, hypertrophy and hyperplasia as described in the text.

of variances were not conrmed, even after transformation, data were analysed with the Scheirer-Ray-Hare test (Sokal and Rohlf, 1995). Here, subsequent analyses within levels of each factor were carried out using Kruskal-Wallis test. 3. Results 3.1. Gill histology Several histopathological deteriorations were detected and categorised as not affected, necrosis, lamellar fusion, epithelial lifting, lamellar clubbing, hypertrophy and hyperplasia (Fig. 1). Lamellar fusion and epithelial lifting were detected in just few specimens analysed, which were not correlated to the OPO concentration and therefore excluded from further analysis. Lamellar clubbing, hypertrophy and hyperplasia occurred in all specimens analysed. Necrotic abnormalities were observed in 69.4% of the investigated specimens and were also present in controls. The total amount of affected lamellae remained almost stable in control animals (Fig. 1), showing a slight increase in hyperplasia at day 21. In OPO-exposed turbot the amount of gill laments showing no histopathological damage decreased steadily over the experiment (Fig. 1) with the 0.10 and the 0.15 mg/l treatment showing the most pronounced changes in gill histology. Fish exposed to OPO concentrations of

0.10 and 0.15 mg/l exhibited the most pronounced alterations in gill histology. 3.2. Evaluation of housekeeping genes Relative mRNA expression of the commonly used housekeeping genes act, 2m and pol are presented in Figs. 2 and 3. In the liver, mRNA expression of act and 2m was not inuenced by OPO concentration (act: two-way ANOVA, F(1, 16) = 0.24, p > 0.05; 2m: two-way ANOVA, F(1, 14) = 0.06, p > 0.05) nor by the time of exposure (act: two-way ANOVA, F(1, 16) = 0.02, p > 0.05; 2m: two-way ANOVA, F(1, 14) = 0.22, p > 0.05) (Fig. 2). No signicant interaction was observed for both housekeeping genes (p > 0.05). In contrast, pol mRNA expression in liver was signicantly affected by time of exposure (two-way ANOVA, F(1, 16) = 14.58, p < 0.01) but not by OPO concentration (two-way ANOVA, F(1, 16) = 0.98, p > 0.05). Upon exposure to 0.15 mg/l OPO, pol mRNA levels were significantly higher after 1 d compared to 21 d (two-way ANOVA, F(1, 8) = 12.67, p < 0.01). Similarly, act and 2m in the gills were affected neither by OPO concentration (act: two-way ANOVA, F(1, 16) = 3.33, p > 0.05; 2m two-way ANOVA, F(1, 16) = 0.57, p > 0.05) nor by the duration of the experiment (act: two-way ANOVA, F(1, 16) = 0.001, p > 0.05; 2m: two-way ANOVA, F(1, 16) = 1.24, p > 0.05) (Fig. 3). No signicant

512

S. Reiser et al. / Aquatic Toxicology 105 (2011) 508517

Fig. 2. mRNA transcription of the housekeeping genes actin (act), 2 microgobulin (2m) and RNA-Polymerases II (pol) in the liver of control specimens and sh exposed to the highest OPO concentration (0.15 mg/l) after 1 and 21 days of exposure. Different lower case letters indicate signicant differences between exposure times.

interaction was observed (p > 0.05). In contrast, mRNA expression of pol was signicantly affected by OPO concentration (two-way ANOVA, F(1, 16) = 13.37, p < 0.01) but not by time of exposure (twoway ANOVA, F(1, 16) = 2.37, p > 0.05). In both tissues, actin and 2m were thus regarded as suitable housekeeping genes. Here, 2m was chosen for further analyses as actin mRNA expression in the target tissues was slightly more variable than that of 2m. 3.3. mRNA expression in liver All biomarker mRNAs, those of hsp70, hsp90 and gst in the liver increased upon OPO exposure compared to control specimens until day 7 and returned to basal levels by day 21 (Fig. 4). For hsp70, a 1.93 fold increase was observed at 0.15 mg/l after 1 d and a 1.68 fold increase after 7 d of exposure when compared to control specimens. Hsp70 was signicantly affected by exposure time (two-way ANOVA, F(2, 24) = 5.61, p < 0.01) but not by OPO concentration (two-way ANOVA, F(3, 24) = 1.37, p > 0.05). However, subsequent Tukey HSD did not

Fig. 4. Mean (SD, n = 3) normalized mRNA expression of (a) hsp70, (b) hsp90 and (c) gst genes in the liver of juvenile turbot P. maxima upon exposure to OPO concentrations of 0.06 mg/l, 0.10 mg/l and 0.15 mg/l compared to the control at day 1, 7 and 21. Expression was normalised to 2m and set as fold increase of control at each sampling day. Error bars denote standard deviations.

Fig. 3. RNA transcription of the housekeeping genes actin (act), 2 microgobulin (2m) and RNA-Polymerases II (pol) in the gills of control specimens and sh exposed to the highest OPO concentration (0.15 mg/l) after 1 and 21 days of exposure. Different upper case letters indicate signicant differences between control and the OPO treatment.

show any differences between the different OPO treatments (p > 0.05). Exposure to an OPO concentration of 0.15 mg/l revealed a 1.72 fold increase in hsp90 after 1 d and a 1.38 fold increase after 7 d. Again, expression returned to basic levels by day 21 and no signicant differences between treatment groups and exposure time were observed (OPO concentration: twoway ANOVA, F(3, 24) = 0.56, p > 0.05; exposure time: two-way ANOVA, F(2, 24) = 0.84, p > 0,05). At an OPO concentration of 0.15 mg/l, gst exhibited a 2.06 fold elevation at day 1 and a 2.07 fold increase at day 7. In the 0.06 mg/l and 0.10 mg/l OPO treatment groups, gst mRNA increased to 1.62 and 1.66 at day 1, respectively. As observed for hsp, the mRNA expression of gst returned to basic levels on day 21 and did not show any

S. Reiser et al. / Aquatic Toxicology 105 (2011) 508517

513

statistical differences between the groups (OPO concentration: Scheirer-Ray-Hare two-way ANOVA, H(3) = 2.02, p > 0.05; exposure time: Scheirer-Ray-Hare two-way ANOVA, H(2) = 2.85, p > 0.05). 3.4. mRNA expression in the gills mRNA expression of the investigated biomarker genes in the gills revealed similar patterns as in liver. Still, individual variation as indicated by the standard deviation was lower (Fig. 5). Expression of hsp70, hsp90 and gst in the gills were all affected by OPO concentration (hsp70: F(3, 24) = 15.97, p < 0.001; hsp90: F(3, 24) = 9.06, p < 0.001; gst: F(3, 24) = 6.37, p < 0.01) as well as by exposure time (hsp70: F(2, 24) = 15.77, p < 0.001; hsp90: F(2, 24) = 26.24, p < 0.001; gst: F(2, 24) = 21.06, p < 0.001). For all three genes, a signicant interaction occurred (hsp70: F(6, 24) = 2.64; p < 0.05; hsp90: F(6, 24) = 6.17, p < 0.001; gst: F(6, 24) = 8.58, p < 0.001). Therefore, subsequent one-way-ANOVAs were conducted to separately investigate the effects of OPO concentration and time of exposure. Turbot exposed to OPO showed strong alterations in hsp70 compared to control specimens at day 1 (one-way ANOVA, F(3, 8) = 5.32, p < 0.05) and day 7 (one-way ANOVA, F(3, 8) = 16.05, p < 0.001) (Fig. 5 a). Fish exposed to an OPO concentration of 0.15 mg/l exhibited around 1.75 fold more hsp70 compared to control specimens (p < 0.05) at day 1 and around 1.88 fold more at day 7. The increase in hsp70 mRNA in sh exposed to 0.15 mg/l at day 7 was signicant compared to control specimens (p < 0.01), to sh exposed to 0.06 mg/l (p < 0.01), where hsp70 was elevated around 1.27 fold, as well as to sh of the 0.10 mg/l treatment (p < 0.05). After 21 days of OPO exposure, no signicant difference between controls and OPO treatments was observed (one-way ANOVA, F(3, 8) = 3.14, p > 0.05). As for hsp70, a signicant effect of OPO concentration on hsp90 mRNA level was detected at day 1 (one-way ANOVA, F(3, 8) = 7.55, p < 0.05) and day 7 (one-way ANOVA, F(3, 8) = 14.71, p < 0.01) (Fig. 5b). Following 1 d of exposure, sh in the 0.15 mg/l OPO treatment exhibited around 2.46 fold more hsp90 than control specimens (p < 0.01) and 2.16 fold more at day 7. Hsp90 in sh exposed to an OPO concentration of 0.15 mg/l for 7 d was significantly elevated compared to that in control sh (p < 0.01) as well as in sh exposed to 0.06 mg/l (p < 0.05). Hsp90 mRNA levels of sh exposed to 0.10 mg/l were signicantly elevated compared to those from control sh (p < 0.05) but not compared to sh exposed to 0.06 mg/l (p > 0.05). At day 21, no signicant effect of OPO concentration on hsp90 levels was detected (F(3, 8) = 1.40, p > 0.05). Here, hsp90 decreased beyond control levels within all OPO treatments. Gst expression (Fig. 5c) was not found to be altered by OPO at day 1 (one-way ANOVA, F(3, 8) = 3.33, p > 0.05). However, a gradual increase in the amount of gst mRNA with increasing OPO concentration was observed at day 7 (one-way ANOVA, F(3, 8) = 21.45, p < 0.001). Turbot exposed to 0.15 mg/l OPO showed an up to 2.03 fold increase in gst expression compared to control specimens (p < 0.001), whereas sh exposed to 0.10 mg/l (p < 0.01) showed an 1.49 fold increase. In addition, the gst mRNA level in sh exposed to an OPO concentration of 0.10 mg/l was signicantly elevated compared to control specimens (p < 0.05). Gst mRNA levels in OPOexposed specimens returned to control levels at day 21, showing no signicant differences among treatments (one-way ANOVA, F(3, 8) = 1.28, p > 0.05). 4. Discussion In the past 20 years, the technical development of RAS for a sustainable marine nsh production has become an emerging technology in aquaculture. Ozonation, due to its high oxidative capacity has been established as an effective water treatment with proven efcacy of disinfection in routine application (Rosenthal and

Fig. 5. Mean (SD, n = 3) normalized mRNA expression of (a) hsp70, (b) hsp90 and (c) gst in the gills of juvenile turbot P. maxima upon exposure to OPO concentrations of 0.06 mg/l, 0.10 mg/l and 0.15 mg/l compared to the control at day 1, 7 and 21. Expression was normalised to 2m and set as fold increase of control at each sampling day. Error bars denote standard deviations. At the same exposure time, columns with different upper case letters indicate signicant differences among concentrations, within the same concentration, different lower case letters indicate signicant differences between exposure times, means not sharing a common letter are signicantly different. Letters are not shown if there was no signicant difference within one sampling day or between sampling days.

Otte, 1979; Liltved et al., 1995; Summerfelt and Hochheimer, 1997). However, with regard to the high oxidative potential, even low OPO concentrations can adversely affect the physiology of shes (Richardson et al., 1983; Reiser et al., 2010). Up to now, numerous studies focused on acute ozone toxicity addressing freshwater shes (e.g. Wedemeyer et al., 1979; Paller and Heidinger, 1980; Fukunaga et al., 1992; Morita et al., 1995; Ritola et al., 2000, 2002b),

514

S. Reiser et al. / Aquatic Toxicology 105 (2011) 508517

but just a few have considered estuarine and marine nsh species (Toner and Brooks, 1975; Hall et al., 1981; Richardson and Burton, 1981; Richardson et al., 1983; Jiang et al., 2001; Jones et al., 2006). Because of the different ozone chemistry in both media, results obtained from freshwater cannot be projected to sh kept under marine conditions. Furthermore, studies focusing on the acute toxicity are of minor relevance for an ozone application in aquaculture production, since ozone concentrations remain far below acutely toxic levels. For daily aquaculture practice, thresholds for safe levels have to consider chronic toxicity at sublethal ozone and OPO concentrations. In this context, safe represents a level which does not cause an adverse, sublethal effect upon long-term exposure (Sprague, 1971). Upon chronic exposure of turbot to sublethal OPO concentrations we observed adverse effects in gill histology, increasing in a time- and concentration-dependent manner. Among these abnormalities, necroses increased at higher concentrations. Although necrotic cells may be replaced (Karan et al., 1998; Cerqueira and Fernandes, 2002), these irreversible damages indicate an exhaustion of adaptive countermeasures at OPO concentrations 0.10 mg/l. Clubbing, hypotrophy and hyperplasia may be regarded as adaptive defence responses towards irritants such as OPO in the process water (Mallatt, 1985). Irritants in the water usually affect the gills rst as they are just covered with a thin epithelium, providing only a slight barrier between blood and the aquatic environment (Evans, 1987). The categories of histological abnormalities identied here have been observed in numerous studies on different irritants (reviewed by Mallatt, 1985). Most prominent, hyperplasia of epithelia as well as chloride cells have been repeatedly reported indicative for chronic inammation and distress at the cellular level. This has widely been reported in sh, e.g. encapsulation of holdfasts of parasites (Lovy et al., 2007) or upon exposure to toxicants (Mallatt, 1985). Hyperplasia, cell proliferation at the tips of the secondary laments, i.e. lamellar clubbing, as well as hypertrophy are considered to be not pathognomonic as they have been described to arise regardless of the irritant (Frances et al., 1998; Richardson et al., 1983; Mallatt, 1985). In general, they have often been interpreted as a way to block the entry of toxicants by increasing the toxicantblood diffusion distance, thereby reducing further dispersion via the circulatory system (Morgan and Tovell, 1973; Mallatt, 1985). Additionally, detoxication should be enhanced as the amount of contributing cells is increased. Massive proliferation of chloride cells might indicate ion deciency and therefore represent an attempt to re-establish cellular homeostasis (Laurent & Perry, 1991). Chloride cells may also affect toxicant extrusion or neutralization (Mallatt, 1985). However, excessive hyperplasia and hypertrophy could also affect respiration and ultimately gill functionality as diffusion processes like gas-exchange, excretion and acidbase regulation are exacerbated through the thickened gill epithelium (Evans et al., 2005). In contrast, necroses represent a deterioration of the gill epithelium, which disrupt cellular integrity of the tissue and ultimately constrains gill functionality (Mallatt, 1985). Respiration is retained at deteriorated gill epithelia as, usually, irritation of the gill epithelium and cell ruptures entail mucus proliferation and therefore restrict oxygen permeability as shown for white perch exposed to OPO (Richardson et al., 1983). Here excessive mucus production was associated with signicantly increased hematocrit at OPO concentrations of 0.10 mg/l to compensate for reduced oxygen permeability. However, Reiser et al. (2010) reported slightly but not signicantly increased haemoglobin and hematocrit levels in turbot upon 1 day of OPO exposure, suggesting a compensation of reduced oxygen uptake just after immediate contact with OPO. Following chronic exposure, no difference between controls and specimens subjected to 0.15 mg OPO/l was detected, possibly accounting for increased mucous secretion upon OPO exposure only at the rst

day. Nevertheless, proliferation of neither mucous cells nor mucus itself has been observed (Reiser et al., 2010). Insufciency and collapse of the aforementioned protective mechanisms in the gills causes rupture of the surrounding epithelium and enables inltration and spread of the irritants through the dehisced cells. Here, necroses were just observed to a minor extent accumulating in a non-signicant manner upon exposure to high concentrations. The observed histopathological abnormalities were less severe compared to those found by Richardson et al. (1983) in white perch at comparable OPO levels: massive deteriorations were reported at OPO levels of 0.10 mg/l and to 0.15 mg/l. This further conrms the robust nature of turbot (van Ham, 2003), which can be explained by the passive and benthic mode of life compared to the highly mobile white perch, having higher oxygen requirements. Additionally, the turbot used for this study were already adapted to aquaculture conditions potentially making them more robust towards higher OPO concentrations. Expression proles of hsp70, hsp90 and gst mRNA congruent to the histological ndings were observed, revealing a concentration-dependent response but not a continuous timedependent increase. Upon exposure, transcription of hsp70 and hsp90 genes increased within 1 day, succeeded by gst, reaching its climax on day 7. Still, all biomarkers returned to normal on day 21. This induction of gene expression upon exposure clearly indicates an OPO-induced oxidative stress. In general, there is a baseline production of reactive oxygen species (ROS) in any cell as ROS arise from numerous cellular processes. Implying disturbances in the normal redox state of tissues, oxidative stress due to enhanced formation of ROS involves the production of peroxides and free radicals at the cellular level and results in progressing denaturation of proteins and oxidation of lipids (Livingstone, 2003; van der Oost et al., 2003). Additionally, OPO might inltrate the cells via the irrigated gills, get dispersed via the circulatory system and locally contribute to oxidative processes. The increase of both hsp mRNAs towards signicant levels illustrates that OPO concentrations as low as 0.10 mg/l can already modulate cellular homeostasis in turbot. Heat shock proteins have been identied as components of the secondary stress response cascade (Barton et al., 2002) with hsp70 and hsp90 being considered as the most common markers of cellular stress (Lewis et al., 1999; Iwama et al., 1998). Alterations in hsp translation across baseline expression represent a rst alarm response towards cellular stress, indicating protein damage and onset of protective mechanisms (Iwama et al., 2004). Hsp70 and hsp90 mainly function as molecular chaperones mediating the repair of altered or denatured proteins, play a crucial role in protein folding and support numerous components of the cytoskeleton that might be affected by ROS (Fink and Goto, 1998). Due to the lack of introns in hsp genes, the hsp response is immediately established after proteins are damaged (Iwama et al., 2004), explaining their high mRNA expression upon short term exposure of just one day in the present study. Among the hsp family, hsp70 has been reported most sensitive towards stressors compared to hsp90, which was attributed to the major importance of hsp70 as a molecular chaperone (Iwama et al., 2004). However, in the present study mRNAs of hsp90 and hsp70 were similarly increased, revealing a more differentiated expression between OPO concentrations. The hsp response decreased at day 7 and returned to basal levels by 21 days of exposure. Unfortunately, data concerning the duration and dynamics of the hsp response towards persistent stressors is limited (Iwama et al., 2004). In the present study, the decrease in expression levels at day 7 potentially indicates an exhaustion of the hsp response but this remains speculative. Gornati et al. (2004) observed increased hsp70 and hsp90 with a persistent stressor still after 3 months in sea bass reared at high densities contradicting hsp exhaustion after 3 weeks of exposition.

S. Reiser et al. / Aquatic Toxicology 105 (2011) 508517

515

In contrast to an exhaustion of the hsp response, cellular adaptation involving an onset of detoxication processes could reduce oxidative stress and decreases the expression of hsp. Enzymes like catalases, superoxide dismutases, glutathione peroxidases, transferases and reductases, i.e. phase I and phase II metabolisation (Livingstone, 2003; van der Oost et al., 2003) have been reported to increase after ozone exposure re-establishing cellular homeostasis (Ritola et al., 2002a, 2002b). These enzymes catalyse the scavenging of oxidized radicals and prevent cellular damage by ROS (van der Oost et al., 2003). Consequently, future studies need to involve expression analyses of phase I metabolizing enzymes. In general, an upregulation of transcripts with products involved in detoxifying mechanisms has been observed in the present study, as indicated by gst mRNA induction on day 7. As a biotransformation phase II enzyme, gst is activated after cellular damage, e.g. lipid peroxidation derived from phase I metabolisation (van der Oost et al., 2003) and inactivates secondary oxidative metabolites like oxidants that arise from lipoperoxidation, lipid hydroperoxides and their derivates. Additionally, gst detoxies ROS by conjugation to glutathione (Salinas and Wong, 1999). Altered gst activity has been previously reported in salmonids following ozone (Ritola et al., 2000, 2002a,b), metal (Cunha et al., 2007) and organic trace pollutant (reviewed by van der Oost et al., 2003) exposures. In the present study, gst mRNA response was delayed compared to hsp70 and hsp90 responses, showing highest transcription levels upon 7 days of OPO exposure. The lack of a gst response at day 1 might be due to the absence of phase I metabolites or a delayed induction of mRNA expression. Since gst mRNA induction peaks on day 7 accompanied by a sharp decrease in hsp mRNA expression, effective detoxication is suggested here. Still, the return of gst mRNA to basal levels is incongruent, but may be interpreted as a further indication of an exhaustion of compensatory mechanisms. Hsp as well as gst transcription levels in the gills were found to be more affected by OPO compared to those in the liver, coinciding with previous studies that revealed tissue-specic gene expression (Smith et al., 1999; Rabergh et al., 2000; Deane and Woo, 2007). However, in contrast to the gills the liver might be less exposed to OPO. A potential dispersion via the circulation might not trigger responses comparable to the gills in other tissues. Nevertheless, a differential susceptibility of gills and liver towards OPO has to be considered although it seems more likely that reduced OPO concentrations in the liver were encountered. Still, oxidative stress is locally restricted and the gills might be affected in particular due to the instantaneous contact with the irritant in the medium (Mallatt, 1985; Evans et al., 2005). However, despite the pronounced effects on the biomarkers in the gills, the liver has been clearly affected by OPO. Even though alterations were not signicant similar patterns as in the gills were observed, suggesting a substantial risk of OPO toxication for most tissues and organs. Interestingly, the results gained from the histological analysis are in good accordance with the molecular oxidative stress markers with regard to the concentration response of OPO and resulted in similar conclusions on hazardous OPO levels. Gill histopathology served as a delicate marker for OPO induced irritations and complemented the results obtained by mRNA expression analysis of genes. Furthermore, predicted exhaustion of the cellular defence (hsp) and detoxication mechanisms (gst) illustrate the severity of the observed adverse effects. However, histological alterations have been found to accumulate until the last day of OPO exposure, whereas mRNA expression levels returned to resting levels by the end of the experiment. Thus, for monitoring purposes, the here investigated molecular biomarkers may not be suited for the determination of safe ozone concentrations. In conclusion, evaluation of mRNA levels in liver and gills clearly demonstrate that sublethal OPO concentrations of 0.15 mg/l increase hsp70 and hsp90 mRNAs, whereas gst mRNA expression

become already modulated at OPO concentrations 0.10 mg/l in the gills of juvenile turbot. The results obtained in this study clearly extend and complement the ndings of an earlier study (Reiser et al., 2010), that reported adverse effects to sublethal OPO concentrations 0.10 mg/l, affecting behaviour, gill functionality and cortisol levels in juvenile turbot. Congruently, at an OPO concentration of 0.06 mg/l no signicant alteration in the physiological and histological parameters nor mRNA expression were observed. Hence, summarizing the ndings of both studies, an OPO concentration of 0.06 mg/l can be suggested as a safe level for the rearing of juvenile turbot. Since the use of residual OPO concentrations as low as 0.06 mg/l was proven to provide a satisfactory reduction in bacterial loads (Sugita et al., 1992; Schroeder et al., 2010), adequate microbiological control is still feasible under safe OPO levels. These ndings are highly valuable for ozone application in aquaculture routine, although it would be desirable to validate these results by extending the exposure time to an even longer period. However, the results provide a rst insight concerning the chronic effect of OPO on sh wellbeing in aquaculture which is becoming increasingly important in terms of marketing issues and consumer perception towards farmed sh. Acknowledgements This study was nancially supported by the German Federal State of Schleswig-Holstein, the European Regional Development Fund (ERDF) through the NEMO project and the Financial Instrument for Fisheries Guidance (FIFG) programme of the European Union through the project The toxicity of ozone in marine recirculation systems. References
American Public Health Association, 1989. Standard Methods for the Examination of Water and Wastewater , 17th ed. American Public Health Association, Washington, DC, USA, 1624pp. Barton, B.A., Morgan, J.D., Vijayan, M.M., 2002. Physiological and condition-related indicators of environmental stress in sh. In: Marshall, A.S. (Ed.), Biological Indicators of Aquatic Ecosystem Stress. American Fisheries Society, Bethesda, USA, pp. 111148. Basu, N., Todgham, A.E., Ackerman, P.A., Bibeau, M.R., Nakano, K., Schulte, P.M., Iwama, G.K., 2002. Heat shock protein genes and their functional signicance in sh. Gene 295, 173183. Block, R.M., Burton, D.T., Gullans, S.R., Richardson, L.B., 1978. Respiratory and osmoregulatory responses of white perch (Morone americana) exposed to chlo and ozone in estuarine waters. In: in: Jolley, L.R., Gorchev, H., Hamilton, D.H. (Eds.), Water Chlorination: Environmental Impact and Health Effects, 2. Arbor Science Publisher, Michigan, USA, pp. 351360. Blogoslawski, W.J., Farrell, L., Garceau, R., Derrig, P., 1976. Production of oxidants in ozonized sea water. In: Rice, R.G., Pichet, P., Vincent, M.A. (Eds.), Proceedings of the Second International Symposium on Ozone Technology. Ozone Press International, Jamesville, NY, pp. 671681. Boone, A.N., Vijayan, M.M., 2002. Constitutive heat shock protein 70 (HSC70) expression in rainbow trout hepatocytes: effect of heat shock and heavy metal exposure. Comp. Biochem. Physiol. C 132, 223233. Bouraoui, Z., Banni, M., Ghedira, J., Clerandeau, C., Guerbej, H., Narbonne, J.F., Boussetta, H., 2008. Acute effects of cadmium on liver phase I and phase II enzymes and methallothionein accumulation on sea bream Sparus aurata. Fish Physiol. Biochem. 34, 201207. Buchan, K.A.H., Martin-Robichaud, D.J., Benfey, T.J., 2005. Measurement of dissolved ozone in seawater: a comparison of methods. Aquacult. Eng. 33, 225231. Cerqueira, C.C.C., Fernandes, M.N., 2002. Gill tissue recovery after copper exposure and blood parameter responses in the tropical sh Prochilodus scrofa. Ecotoxicol. Environ. Safe. 52, 8391. Crecelius, E.A., 1979. Measurements of oxidants in ozonized seawater and some biological reactions. J. Fish. Res. Board Can. 36, 10061008. Crespo, S., Sala, R., Marlasca, M.J., 1987. Different cell types in the gill epithelium of juvenile turbot Scophthalmus maximus. Aquaculture 67, 216217. Cunha, I., Neuparth, T., Caeiro, S., Costa, M.H., Guilhermino, L., 2007. Toxicity ranking of estuarine sediments on the basis of Sparus aurata biomarkers. Environ. Toxicol. Chem. 26, 444453. Deane, E.E., Woo, N.Y.S., 2007. Impact of nitrite exposure in endocrine, osmoregulatory and cytoprotective functions in the marine teleost Sparus sarba. Aquat. Toxicol. 82, 8593. Evans, D.H., 1987. The sh gill: site of action and model for toxic effects of environmental pollutants. Environ. Health Perspect. 71, 4758.

516

S. Reiser et al. / Aquatic Toxicology 105 (2011) 508517 Martins, C.I.M., Eding, E.H., Verdegem, M.C.J., Heinsbroek, L.T.N., Schneider, O., Blancheton, J.P., Roque dOrbcastel, E., Verreth, J.A.J., 2010. New developments in recirculating aquaculture systems in Europe: a perspective on environmental sustainability. Aquacult. Eng. 43, 8393. Morgan, M., Tovell, P.W.A., 1973. The structure of the gill of the trout, Salmo gairdneri (Richardson). Z. Zellforsch. Mikrosk. Anat. 142, 147162. Morita, J., Suzuki, T., Kimura, S., Hara, A., Takama, K., 1995. Effect of low level ozone exposure on the serum TBA-reactive substance (TBA-RS) level and superoxide dismutase and catalse activities in rainbow trout. Fish. Sci. 61, 890891. Oemcke, D.J., van Leeuwen, J., 1998. Potential of Ozone for Ballast Water Treatment. CRC Reef Research Centre, James Cook University, Australia, 40 pp. Paller, M.H., Heidinger, R.C., 1980. Mechanisms of delayed ozone toxicity to bluegill Lepomis macrochirus (Ranesque). Environ. Pollut. A 22, 229239. Pandey, S., Parvez, S., Ansari, R.A., Ali, M., Kaur, M., Hayat, F., Ahmad, F., Raisuddin, S., 2008. Effects of exposure to multiple trace metals on biochemical, histological and ultrastructural features of gills of a freshwater sh, Channa punctata Bloch. Chem. Biol. Interact. 174, 183192. Perrins, J.C., Cooper, W.J., van Leeuwen, J.H., Herwig, R.P., 2006. Ozonation of seawater from different locations: formation and decay of total residual oxidants-implication for ballast water treatment. Mar. Pollut. Bull. 52, 10231033. Pfaf, M.W., 2001. A new mathematical model for relative quantication in real-time RT-PCR. Nucleic Acids Res. 29, e36. Pichavant, K., Person-Le Ruyet, J., Severe, A., Le Roux, A., Quemener, L., Boeuf, G., 1998. Effects of photopertiod on juvenile turbot (Psetta maxima): physiological status and growth. B. Fr. Peche. Piscic. 350351, 265277. Rabergh, C.M.I., Airaksinen, S., Soitamo, A., Bjorklund, H.V., Johansson, T., Nikinmaa, M., Sistonen, J., 2000. Tissue-specic expression of zebrash (Danio rerio) heat shock factor 1 mRNAS in response to heat stress. J. Exp. Biol. 203, 18171824. Reiser, S., Schroeder, J.P., Wuertz, S., Kloas, W., Hanel, R., 2010. Histological and physiological alterations in juvenile turbot (Psetta maxima, L.) exposed to sublethal concentrations of ozone-produced oxidants in ozonated seawater. Aquaculture 307, 157164. Richardson, L.B., Burton, D.T., 1981. Toxicity of ozonated estuarine water to juvenile blue crabs (Callinectes sapidus) and juvenile Atlantic menhaden (Brevoortia tyrannus). Bull. Environ. Contam. Toxicol. 26, 171178. Richardson, L.B., Burton, D.T., Block, R.M., Stavola, A.M., 1983. Lethal and sublethal exposure and recovery effects of ozone-produced oxidants on adult white perch (Morone americana Gmelin). Water Res. 17, 205213. Ritola, O., Lyytikainen, T., Pylkko, P., Molsa, H., Lindstrom-Seppa, P., 2000. Glutathione-dependant defence system and monooxygenase enzyme activities in Arctic charr Salvelinus alpinus (L.) exposed to ozone. Aquaculture 185, 219233. Ritola, O., Peters, L.D., Livingstone, D.R., Lindstrom-Seppa, P., 2002a. Effects of in vitro exposure to ozone and/or hyperoxia on superoxide dismutase, catalase, glutathione and lipid peroxidation in red blood cells and plasma of rainbow trout, Oncorhynchus mykiss (Walbaum). Aquacult. Res. 33, 165175. Ritola, O., Livingstone, D.R., Peters, L.D., Lindstrom-Seppa, P., 2002b. Antioxidant processes are affected in juvenile rainbow trout (Oncorhynchus mykiss) exposed to ozone and oxygen-supersaturated water. Aquaculture 210, 19. Rosenthal, H., Otte, G., 1979. Ozonation in an intensive sh culture recycling system. Ozone-Sci. Eng. 1, 319327. Salinas, A.E., Wong, M.G., 1999. Glutathione S-transferases a review. Curr. Med. Chem. 6, 279309. Schroeder, J.P., Grtner, A., Waller, U., Hanel, R., 2010. The toxicity of ozone-produced oxidants to the Pacic white shrimp Litopenaeus vannamei. Aquaculture 305, 611. Smith, T.R., Tremblay, G.C., Bradley, T.M., 1999. Characterization of the heat shock protein response of Atlantic salmon (Salmo salar). Fish Physiol. Biochem. 20, 279292. Sokal, R.R., Rohlf, F.J., 1995. Biometry. In: The Principles and Practice of Statistics in Biological Research, second ed. WH Freeman & Co, New York, USA, 859 pp. Sprague, J.B., 1971. Measurement of pollutant toxicity to sh III. Sublethal effects and safe concentrations. Water Res. 5, 245266. Sugita, H., Asai, T., Hayashi, K., Mitsuya, T., Amanuma, K., Maruyama, C., Deguchi, Y., 1992. Application of ozone disinfection to remove Enterococcus seriolicida, Pasteurella piscicida, and Vibrio anguillarum from seawater. Appl. Environ. Microbiol. 58, 40724075. Summerfelt, S.T., Hochheimer, J.N., 1997. Review of ozone processes and applications as an oxidizing agent in aquaculture. Prog. Fish-Cult. 59, 94105. Toner, R.C., Brooks, B., 1975. The effect of ozone on four species of phytoplankton, crab zoea and megalops and the Atlantic silverside Menidia menidia. In: Blogoslawski, W.J., Rice, R.G. (Eds.), Aquatic Applications of Ozone. International Ocean Institute, Stamford, Connecticut, USA, pp. 154160. Trubiroha, A., Wuertz, S., Frank, S.N., Sures, B., Kloas, W., 2009. Expression of gonadotropin subunits in roach (Rutilus rutilus, Cyprinidae) infected with plerocercoids of the tapeworm Ligula intestinalis (Cestoda). Int. J. Parasitol. 39, 14651473. van der Oost, R., Beyer, J., Vermeulen, N.P.E., 2003. Fish bioaccumulation and biomarkers in environmental risk assessment: a review. Environ. Toxicol. Pharmacol. 13, 57149.

Evans, D.H., Piermarini, P.M., Choe, K.P., 2005. The multifunctional sh gill: dominant site of gas exchange, osmoregulation, acidbase regulation and excretion of nitrogenous waste. Physiol. Rev. 85, 97117. FAO, 2009. Psetta maxima (Linnaeus, 1758). Cultured aquatic species information programme. Fisheries and Aquaculture Department, Rome, Italy; www.fao.org/shery/culturedspecies/Psetta maxima/en. Ferrari, A., Venturino, A., de Dangelo, A.M.P., 2007. Effects of carbaryl and azinphos methyl on juvenile rainbow trout (Oncorhynchus mykiss) detoxifying enzymes. Pestic. Biochem. Physiol. 88, 134142. Fink, A.L., Goto, Y., 1998. Molecular Chaperones in the Life Cycle of Proteins: Structure, Function, and Mode of Action. Marcel Dekker, New York, USA, 626 pp. Frances, J., Allan, G.L., Nowak, B.F., 1998. The effects of nitrite on the short-term growth of silver perch (Bidyanus bidyanus). Aquaculture 163, 6372. Fukunaga, K., Suzuki, T., Arita, M., Suzuki, S., Hara, A., Yamauchi, K., Shinriki, N., Ishizaki, K., Takama, K., 1992. Acute toxicity of ozone against morphology of gill and erythrocytes of Japanese charr (Salvelinus leucomaenis). Comp. Biochem. Physiol. C 101, 331336. Giese, A.C., Christensen, E., 1954. Effects of ozone on organisms. Physiol. Zool. 27, 101115. Gisbert, E., Rodriguez, A., Cardona, L., Huertas, M., Gallardo, M.A., Sarasquete, C., SalaRabanal, M., Ibarz, A., Snchez, J., Castell-Orvay, F., 2004. Recovery of siberian sturgeon yearlings after an acute exposure to environmental nitrite: changes in the plasmatic ionic balance, Na+ K+ ATPase activity and gill histology. Aquaculture 239, 141154. Gornati, R., Papis, E., Rimoldi, S., Terova, G., Saroglia, M., Bernardini, G., 2004. Rearing density inuences the expression of stress-related genes in sea bass (Dicentrarchus labrax, L.). Gene 341, 111118. Gornati, R., Gualdoni, S., Cavaliere, R., Terova, G., Saroglia, M., Bernardini, G., 2005. Molecular biology and sh welfare: a winning combination. Aquacult. Int. 13, 5155. Haag, W.R., Hoign, J., 1984. Kinetics and products of the reactions of ozone with various forms of chlorine and bromine in water. Ozone-Sci. Eng. 6, 103114. Hall, L.W., Burton, D.T., Richardson, L.D., 1981. Comparison of ozone and chlorine toxicity to the developmental stages of striped bass, Morone saxatilis. Can. J. Fish. Aquat. Sci. 38, 752757. Halliwell, B., Gutteridge, J.M.C., 1999. Free Radicals in Biology and Medicine, third ed. Oxford University Press, Oxford, UK, p. 704. Hassanein, H.M.A., Banhawy, M.A., Soliman, F.M., Abdel-Rehim, S.A., Muller, W.E.G., Schroder, H.C., 1999. Induction of hsp70 by the herbicide oxyuoren (goal) in the Egyptian Nile sh, Oreochromis niloticus. Arch. Environ. Contam. Toxicol. 37, 7884. Hayes, J.D., Pulford, D.J., 1995. The glutathione S-transferase supergene family: regulation of GST and the contribution of the isoenzymes to cancer chemoprotection and drug resistance. Crit. Rev. Biochem. Mol. Biol. 30, 445600. Herwig, R.P., Cordell, J.R., Perrins, J.C., Dinnel, P.A., Gensemer, R.W., Stubbleeld, W.A., Ruiz, G.M., Kopp, J.A., House, M.L., Cooper, W.J., 2006. Ozone treatment of ballast water on the oil tanker S/T Tonsina: chemistry, biology and toxicity. Mar. Ecol. Prog. Ser. 324, 3755. Hoign, J., Bader, H., Haag, W.R., Staehelin, J., 1985. Rate constant of reactions of ozone with organic and inorganic compounds in water-III. Water Res. 19, 9931004. Hurlbert, H.H., 1984. Pseudoreplication and the design of ecological eld experiments. Ecol. Monogr. 54, 187211. Iwama, G.K., Thomas, P.T., Forsyth, R.B., Vijayan, M.M., 1998. Heat shock protein expression in sh. Rev. Fish Biol. Fisher. 8, 3556. Iwama, G.K., Afonso, L.O.B., Todgham, A., Ackerman, P., Nakano, K., 2004. Are hsps suitable for indicating stressed states in sh? J. Exp. Biol. 207, 1519. Jiang, G., Liu, Y., Yang, D., Lue, Y., 2001. The toxicity of ozonated seawater to the Penaeus chinensis and Paralichthys olivaceus. Mar. Sci. 25, 1113. Jones, A.C., Gensemer, R.W., Stubbleeld, W.A., van Genderen, E., Dethloff, G.M., Cooper, W.J., 2006. Toxicity of ozonated seawater to marine organisms. Environ. Toxicol. Chem. 25, 26832691. Karan, V., Vitorovic, S., Tutundzic, V., Poleskic, V., 1998. Functional enzymes activity and gill histology of carp after copper sulphate exposure and recovery. Ecotoxicol. Environ. Safe. 40, 4955. Kroupova, H., Trubiroha, A., Wuertz, S., Kloas, W., 2011. Stage dependent differences in RNA composition and content affect the outcome of expression proling in roach (Rutilus rutilus) ovary. Comp. Biochem. Physiol. A 159, 141149. Laurent, P., Perry, S.F., 1991. Environmental effects on sh gill morphology. Physiol. Zool. 64, 425. Lewis, S., Handy, R.D., Cordi, B., Billinghurst, Z., Depledge, M.H., 1999. Stress proteins (HSPs): methods of detection and their use as an environmental biomarker. Ecotoxicology 8, 351368. Liltved, H., Hektoen, H., Efraimsen, H., 1995. Inactivation of bacterial and viral sh pathogens by ozonation or UV irradiation in water of different salinity. Aquacult. Eng. 14, 107122. Livingstone, D.R., 2003. Oxidative stress in aquatic organisms in relation to pollution and aquaculture. Rev. Med. Vet.-Toulouse 154, 427430. Lovy, J., Wright, G.M., Speare, D.J., 2007. Ultrastructural examination of the host inammatory response within gills of netpen reared Chinook salmon (Oncorhynchus tshawytscha) with microsporidial gill disease. Fish Shellsh Immunol. 22, 131149. Mallatt, J., 1985. Fish gill structural changes induced by toxicants and other irritants: a statistical review. Can. J. Fish. Aquat. Sci. 42, 630648.

S. Reiser et al. / Aquatic Toxicology 105 (2011) 508517 von Gunten, U., 2003. Ozonation of drinking water: part II. Disinfection and byproduct formation in presence of bromide, iodide or chlorine. Water Res. 37, 14691487. van Ham, E.H., 2003. The physiology of juvenile turbot (Scopthalmus maximus, Ranesque) and Atlantic halibut (Hippoglossus hippoglossus, L.) under different stress conditions. PhD Thesis, University of Nijmegen, The Netherlands. Waller, U., 2000. Tank culture and recirculating systems. In: Black, K.D. (Ed.), Environmental Impacts of Aquaculture. Shefeld Academic Press, Shefeld, UK, pp. 99127.

517

Wedemeyer, G.A., Nelson, N.C., Yasutake, W.I., 1979. Physiological and biochemical aspects of ozone toxicity to rainbow-trout (Salmo gairdneri). J. Fish. Res. Board Can. 36, 605614. Wuertz, S., 2005. The paracrine IGF-I system and its role in the regulation of gonad maturation in sterlet, Acipenser ruthenus Linnaeus, 1758. PhD Thesis, dissertation.de, ISBN 3-86624-119-4. Wuertz, S., Gessner, J., Kirschbaum, F., Kloas, W., 2007. Expression of IGF-I and IGF-I receptor in male and female sterlet, Acipenser ruthenus Evidence for an important role in gonad maturation. Comp. Biochem. Physiol. A 147, 223230.