Aquatic Toxicology 74 (2005) 242–253

Impact of cadmium contamination and oxygenation levels onbiochemical responses in the Asiatic clamCorbicula fluminea

Alexia Legeay∗, Maud Achard-Joris, Magalie Baudrimont,Jean-Charles Massabuau, Jean-Paul Bourdineaud

Laboratoire d’Ecophysiologie et Ecotoxicologie des Syst`emes Aquatiques (LEESA), UMR 5805-OASU,Universite Bordeaux 1 and CNRS, Place Peyneau, 33120 Arcachon, France

Received 6 December 2004; received in revised form 11 May 2005; accepted 11 May 2005

Abstract

The aim of this work was to evaluate the potential utility of several biochemical parameters as indicators of the toxic effects ofcadmium in the freshwater clamCorbicula flumineaunder two levels of oxygenation (normoxia 21 kPa and hypoxia 4 kPa). Thesevariations in oxygenation are representative of the natural environments of bivalves living at the bottom of the water column,where hypoxic episodes may occur regularly. Cadmium accumulation, metallothionein synthesis, MXR protein induction,lipoperoxidation and antioxidant enzyme activities (catalase, glutathione reductase and total and selenium-dependent glutathioneperoxidases) were assessed in the gills ofC.flumineain four experimental conditions: normoxia, hypoxia, normoxia with cadmiumand hypoxia with cadmium ([Cd] = 30�g l−1) over a 14-day period. Behavioural reactions were also followed for the duration

s underdthis washypoxicimpact

decisivecadmium-mination,portance

ol-es.

of the experiment by monitoring clam activity and valve movements. This study is a first report on biochemical responsecadmium contamination and hypoxia and will enable us to determine better biomarkers forC. flumineaas they were measuresimultaneously. In metal-exposed animals, we found an increasing accumulation of cadmium in the gills with time, andmore severe in hypoxic conditions. Metallothionein synthesis occurred in contaminated clams and was precocious inconditions. MXR protein induction appeared promising due to its quick and significant response to metal with a strongfrom hypoxic contamination. On the other hand, in our experimental conditions, antioxidant parameters did not showresponses to contamination and hypoxia, except glutathione peroxidases which decreased systematically with time in aindependent manner. Lipid peroxidation, expressed as malondialdehyde content, was not stimulated by normoxic contaas has been shown in other studies, but was stimulated under hypoxic cadmium contamination. Our study confirms the imof a multi-biomarker approach in environmental studies as some are not appropriate to all organisms.© 2005 Elsevier B.V. All rights reserved.

Keywords:Antioxidant enzymes; Lipid peroxidation; Metallothionein; MXR; Cadmium; Hypoxia;Corbicula fluminea

∗ Corresponding author. Tel.: +33 556 223 920;fax: +33 556 549 383.

E-mail address:a.legeay@epoc.u-bordeaux1.fr (A. Legeay).

1. Introduction

Aquatic systems are contaminated by different plutants, including metals, as a result of man’s activiti

0166-445X/$ – see front matter © 2005 Elsevier B.V. All rights reserved.doi:10.1016/j.aquatox.2005.05.015

A. Legeay et al. / Aquatic Toxicology 74 (2005) 242–253 243

Bivalve molluscs are known to accumulate high con-centrations of heavy metals in their tissue and arewidely used as bioindicators for pollution in marine(Regoli and Orlando, 1994; Geret et al., 2003) andfreshwater (Cossu et al., 1997; Doyotte et al., 1997;Vidal et al., 2001) environments. They usefully com-plement chemical analyses as they provide a time-integrated indication of environmental contamination.A panel of different parameters is in current use infield studies including behavioural analyses or mea-surements of biochemical biomarkers. Among thesebiomarkers, the evaluation of oxygen reactive species(ROS) has been well developed in the past 2 decades.In fact, a consequence of aerobic life is the produc-tion of ROS as a result of normal oxygen metabolism.To protect themselves against the possible ravages ofoxyradical production, aerobic organisms have evolveda complex system of defences including antioxidantenzyme catalase (CAT), glutathione reductase (GRD)and glutathione peroxidases (GPX). In the environ-ment, numerous anthropogenic pollutants, includingmetals, provoke oxidative stress in aquatic organ-isms resulting, among other things, in lipid peroxida-tion (LPO), through oxidation of polyunsaturated fattyacids, DNA damage and altered calcium homeosta-sis (Brennan and Schiestl, 1996; Stohs and Bagchi,1995; Stohs et al., 2000). Cadmium (Cd) is one ofthe metals which is known to increase ROS forma-tion and promote oxidative stress in organisms rang-ing from yeast (Brennan and Schiestl, 1996) to ver-t alw con-t b;S al.,2 )t nceda ROSf ofR

a-s hesel wnf eiri csi 1;B 998;G ess st an

excess of Cd in the cells, while MTs intervene later(Geret et al., 2002b).

A new biochemical marker has recently been studiedonCorbicula flumineain relation to metal contamina-tion. Specifically, the induction of a multi-xenobioticresistance (MXR) protein was measured to establish ifa dose-dependent relationship existed between MXRexpression and metallic stress (Achard et al., 2004).

The aim of the present work was to study sev-eral biochemical responses to cadmium contaminationin C. flumineaand assess and compare their poten-tial use as biomarkers. This freshwater bivalve is aburrowing species living in the upper layers of sedi-ments (Briton and Morton, 1982). Due to its positionat the bottom of the water column, this species canbe naturally submitted to hypoxic episodes. Under twooxygenation levels in the water (normoxia 21 kPa andhypoxia 4 kPa), we measured several parameters, suchas cadmium accumulation, metallothionein synthesis,MXR protein induction, lipoperoxidation and antioxi-dant enzyme activities (catalase, glutathione reductaseand glutathione peroxidases), inC. flumineagills. Fur-thermore, during the whole experiment, one batch ofclams was equipped for valvometric studies to esti-mate behavioural response to cadmium contaminationin these oxygenation conditions. This study, the firston biochemical responses under cadmium contamina-tion and hypoxia, will allow us to improve biomarkerchoice forC. flumineaas both sets of responses weremeasured simultaneously.

sena etalu rgee iron-m rgea

2

2

u-a rin( ra-t shelll r-r

ebrates (Stohs et al., 2000) and several studies deith antioxidant systems in response to cadmium

amination (Almeida et al., 2002; Geret et al., 2002iraj Basha and Usha Rani, 2003; Company et004). As described inLivingstone et al. (1990,

he enzymatic antioxidant system can be enhas an adaptation or a compensatory reaction to

ormation, or impaired indicating a toxic effectOS.Metallothioneins (MTs) are also currently me

ured in response to cadmium contamination. Tow molecular weight, cystein-rich proteins are knoor their protective action against toxic metals. Thnduction following cadmium exposure in molluss well documented (Bebianno and Langston, 199audrimont et al., 1997; Bebianno and Serafim, 1eret et al., 2002b). Otherwise, antioxidant enzym

eem to act as the first defence mechanism again

From various different tissues, the gills were chos they represent the main barrier involved in mptake via the direct route of exposure, with a laxchange area in contact with the surrounding envent. Moreover, they are known to accumulate lamounts of cadmium (Baudrimont et al., 1997).

. Materials and methods

.1. Animals

Adult C. flumineabivalves were collected manlly in June 2003 from the L’Isle River at Saint-SeuAquitaine region, France). On arrival at the laboory, the clams were calibrated and the averageength was 27.5± 0.1 mm (mean± standard error) coesponding to a mean wet weight of 0.51± 0.01 g for

244 A. Legeay et al. / Aquatic Toxicology 74 (2005) 242–253

the whole soft bodies. Before experiments, clams werekept in large tanks (volume: 200 l) filled with aeratedtap water over a 10 cm quartz sand layer (SILAQ, Mios,France) where they could reproduce natural behaviour(burying). They were fed twice a week with a suspen-sion of unicellular algae (Scenedesmus subspicatus)over a 1-month maintenance period allowing acclima-tion.

2.2. Experimental design

Experiments were performed on acclimated fresh-water bivalves in four experimental units (EU) con-sisting of polyvinyl chloride tanks (80 cm× 60 cm×30 cm) lined with plastic bags (Plastiluz, alimentarystandard). The bottom of the tanks was covered witha 5 cm layer of quartz sand, rather than natural sub-strate which could introduce indirect contamination,and they were filled with 120 l of dechlorinated tapwater. The water column was permanently controlledfor temperature (15± 0.5◦C), pH (7.00± 0.05, pH-CO2-stat) and oxygenation levels (either normoxia,partial pressure of oxygen,PO2 = 21 kPa, or hypoxia,PO2 = 4 kPa; corresponding O2 concentrations, 10 and2 mg l−1) and some water was added to compen-sate for evaporation in these closed systems. Exper-iments occurred under natural light conditions andanimals were fed twice a day with 500 ml of analgae suspension ofS. subspicatus(2× 105 cells l−1).Each tank contained 127 bivalves and four experi-m ento s (0o aN an,M tedc on,d en-s eachE tra-t etala( msw thes eri-m dis-sf heref ly

selected animals per tank were equipped for valvome-try and their valve activity monitored during the entireexperiment.

2.3. Valvometry

To evaluate the impact of hypoxic stress on clambehaviour, with or without cadmium contamination,we recorded their body activity and valve move-ments throughout the experiment with an impedancevalvometer (Tran et al., 2003). This was performed onclams equipped with two light platinum electrodes (Ø3 mm; weight 20 mg) fixed on each valve. This allowedthe animal to position itself in the sand and movefreely without experimental constraint. The measure-ment principle was based on the application of Ohm’sLaw, U=RI, since the apparatus measured a currentwhich varied according to the distance between theelectrodes. A total of 317 h (13.2 days) of signal wasanalyzed per condition.

2.4. Cadmium determination

At days 1, 3, 7 and 14, gills from five individu-als in each tank were dissected, dried on absorbentpaper before being weighed (fw). Tissue samples weredigested with nitric acid in a pressurized medium at100◦C for 3 h. After dilution of the digestates in 20 mlultrapure water (MilliQ plus), cadmium concentrationswere determined by atomic absorption spectropho-t ubea tionl ree tyo ea-s rials( fishl

2

2e to

cw 14o( sam-p .2,

ental conditions were applied, combining differxygenation levels and cadmium concentrationr 30�g l−1). Hypoxia was obtained by bubbling2/air gas mixture via mass flow controllers (Tylodel FC-260) driven by a laboratory-construc

ontrol unit. To maintain cadmium contaminatiaily additions of aqueous metal solution compated for the decrease in Cd concentration inU, determined from measuring metal concen

ions in water samples collected each morning. Mdditions were made from a CdCl2 stock solution100 mg l−1 Cd). At days 1, 3, 7 and 14, some claere removed and replaced by others to maintainame density over the 14-day period of the expent. Sampled clams were either immediately

ected for MXR determination or stored at−80◦C forurther measurements and under nitrogen atmospor MT determination. In addition, five random

ometry with Zeeman correction, using graphite ttomizer (AAS Thermoptec M6Solaar). The detec

imit was 0.1�g l−1 Cd. All metal concentrations wexpressed as ng g−1 wet weight tissue. The validif the method was checked periodically against murements of certified biological reference mateTort-2: lobster hepatopancreas and Dolt-2: dogiver from NRCC-CNRC, Ottawa, Canada).

.5. Biochemical analyses

.5.1. Metallothionein determinationAs MTs are known to occur belatedly in respons

admium contamination (Rosijadi and Klerks, 1989),e determined MT concentrations at days 7 andnly, in gills from nineC. flumineaper conditionthree replicates of three pooled samples). Tissueles were homogenized in Tris–HCl 25 mM, pH 7

A. Legeay et al. / Aquatic Toxicology 74 (2005) 242–253 245

with a blender (Ultra-Turrax T-25), centrifuged at20,000×g for 1 h at 4◦C and quantified with themercury-saturation assay (MSA,Baudrimont et al.,1997). This technique, using cold inorganic mercury,is based on the quantification of mercury (Hg) boundto the saturated-MT after protein denaturation withtrichloroacetic acid (MTs are not denatured) and thescavenging of excess Hg not bound to the MT withlyophilized beef hemoglobin (25 g l−1, Sigma H-2500)prepared in Tris–HCl 30 mM, pH 8.2. After 20,000×gcentrifugation for 20 min, the final supernatant wasthen quantitatively recovered and analyzed for Hgdetermination by flameless atomic absorption spec-trometry (AMA 254, Altec, Prague, Czech Republic).The detection limit was estimated at 1 ng Hg.

Reference samples or “blanks” (200�l Tris–HClhomogenization buffer + 200�l trichloracetic 10%HgCl2 50 mg l−1 solution + 200�l beef hemoglobinsolution) were prepared to monitor the Hg complex-ation efficiency of the hemoglobin. Under our experi-mental conditions, an average burden of 1.2 ng Hg wasmeasured in these reference samples, compared withthe 10,000 ng initially added (0.01%). The mean of thethree blank values measured in each analytical run wasdeducted from the Hg burdens measured in each tissuesample replicate.

Simultaneously, a standard of purified rabbit liverMT (Sigma M-7641) was used to check the validity ofthe saturation assay by determining the ratio betweenthe binding sites measured after Hg-saturation and thep andp i-nt ndi-t was1 llsw perg hee isu notb

2ved

f ot-t then tis-s rrax

T-25) in 1 ml of fresh lysis buffer (10 mM KCl, 1.5 mMMgCl2, 10 mM Tris–HCl pH 7.4 and 2 mM phenyl-methylsulfonyl fluoride, PMSF). Sodium dodecyl sul-fate (2% final concentration) was added, samples werethen sonicated in a water bath for 10 min and cen-trifuged at 1000×g for 10 min. Protein concentrationsfrom the supernatants, kept at−20◦C, were determinedwith the bicinchoninic acid assay (Pierce, BCA Pro-tein Assay kit). Thirty micrograms of proteins fromeach protein extract were resolved on 7.5% polyacry-lamide gels and transferred to Immun-blot polyvinyli-dene difluoride (PVDF) membrane (Bio-Rad) with asemi-dry transfer cell (Bio-Rad), according to man-ufacturer’s instructions. Immunolabelling was per-formed using the P-glycoprotein-specific monoclonalantibody C219 (dilution: 1/500; Signet, Dedham, MA,USA), which recognizes a highly conserved aminoacid sequence common to all MDR-protein (Endicottand Ling, 1989), and goat anti-mouse peroxidase-conjugated secondary antibody (dilution: 1/10,000,Jackson Immunoresearch, local supplier: Perbio Sci-ence, Brebieres, France). Bands were detected bychemiluminescence according to Amersham protocols.Digitalised photos of the gels were used for subsequentimage analysis with the Scion Image program (Scioncorporation).

2.5.3. Antioxidant enzyme assaysThe study of enzymatic activities and malondialde-

hyde (MDA) concentration was carried out on threep s.G der( ferc itor.Tf g as inu-o

ea-s ei( lc d0H as�

vityw ance

otential binding sites indicated by the supplierreviously verified by Hg, Cd, Zn and Cu determations on purified MT solution (10�g MT ml−1 in

he homogenization buffer). In our experimental coions, the average value of the recovery percentage00.17± 4.6%. MT concentrations in the clams’ giere expressed in nanomoles of Hg-binding sitesram± S.E. (wet weight). Owing to the fact that txact quantity of Hg binding sites per MT moleculenknown for this species, MT concentrations cane expressed in nmol MT g−1 (wet weight).

.5.2. Western blotting for MXRAfter each exposure time, nine clams were remo

rom each of the four conditions for Western bling. Animals were dissected on ice, gills fromine individuals were pooled and 400–600 mg ofue was homogenized with a blender (Ultra-Tu

ools of four clams’ gills from the four conditionills were weighed and homogenized with a blen

Ultra-Turrax T-25) in 2 ml of 50 mM phosphate bufontaining 1 mM PMSF as serine protease inhibhe homogenates were centrifuged at 100,000×g

or 80 min. Enzyme activities were measured usinpectrophotometer (Shimadzu UV-1601) by contus registration at 25◦C.

Catalase (EC 1.11.1.6.). CAT activity was mured according toGreenwald (1985)by the decreasn absorbance at 240 nm due to H2O2 consumptionε =−0.0436 mM−1 cm−1). A final volume of 3 monsisted of 2900�l 50 mM K-phosphate buffer an.1 mM EDTA, 70�l protein sample and 30�l 1 M2O2. CAT activity in the gills was expressedmol min−1 mg−1 total protein.Glutathione reductase (EC 1.6.4.2.). GRD acti

as measured by following the decrease in absorb

246 A. Legeay et al. / Aquatic Toxicology 74 (2005) 242–253

at 340 nm (ε =−6.22 mM−1 cm−1) due to the oxida-tion of NADPH (Ramos-Martinez et al., 1983). Theassay mixture contained in a final volume of 1 ml930�l 50 mM K-phosphate buffer and 0.1 mM EDTA,10�l 12 mM NADPH, 50�l protein sample and 10�l20 mM oxidized glutathion. Activities were expressedas nmol min−1 mg−1.

Glutathione peroxidases (EC 1.11.1.9. and EC2.5.1.18.). Glutathione peroxidase activities weremeasured following the formation of GSSG in anenzyme system coupled with glutathione reduc-tase (Lawrence and Burk, 1976). The decrease inNADPH due to its oxidation was followed at 340 nm(ε =−6.22 mM−1 cm−1). Se-dependent (SeGPX) andtotal peroxidase (total GPX) activities were mea-sured using H2O2 and cumene hydroperoxide as sub-strates, respectively. A final volume of 1 ml con-tained 930�l 50 mM K-phosphate buffer and 0.1 mMEDTA, 33�l 150 mM reduced glutathione, one unitglutathione reductase, 10�l 12 mM NADPH, 3.3�l1.125 M sodium azide, 50�l protein sample or blank,and 10�l 20 mM H2O2 or 200 mM cumene hydroper-oxide.

2.5.4. Lipid peroxidation determinationPolyunsaturated fatty acid peroxides generate MDA

on decomposition. The determination method usedwas designed to measure thiobarbituric acid reac-tives (TBARs), including MDA, by spectrophotome-t rat st tedbo umeo thf ti-m UV-

1601) against a blank consisting of 1.5 ml 5% coldtrichloroacetic acid mixed with 1.5 ml 1% thiobar-bituric acid, and using malondialdehyde bis acetal(tetramethoxypropan, 10838-3, Sigma) as standard.Concentrations of lipid peroxidation compounds in thegills were expressed as�mol mg−1 protein.

2.5.5. Total protein determinationFor each homogenate, total protein concentrations

were measured by the Lowry method (Lowry et al.,1951) using bovine serum albumin as reference stan-dard material. Protein concentrations were expressedas mg ml−1 or mg g−1 wet weight.

2.6. Statistics

The data were analyzed by non-parametricKruskal–Wallis ANOVA and Mann–Whitney test withthe Statistica 5.0 and Sigma Stat 2.0 programs. Thelevel of significance was set at 0.05.

3. Results

3.1. Animal behaviour and valve activity

No difference in body movements, burrowingactivity or displacement was noticed as a functionof experimental condition. In bivalves, valve openingt tionsi ev d ine oure sig-n n-t rom4 rvet

TV xposu )

[Cd

O 127O 38.8O 0.8

V s (317 dies.

ry using the thiobarbituric acid method (Esterbauend Cheeseman, 1990). Briefly, 450�l of 5% cold

richloroacetic acid was added to 150�l homogenateo precipitate proteins. The precipitate was pelley centrifugation (10,000×g for 15 min) and 550�lf supernatant was reacted with an equal volf 1% thiobarbituric acid in a boiling water ba

or 10 min. After cooling, TBAR levels were esated at 532 nm (spectrophotometer Shimadzu

able 1alve opening behaviour ofC. flumineain response to cadmium e

Normoxia

[Cd] = 0�g l−1

pening time (ind−1) 153 h 49 minpening time (ind−1 day−1) (%) 48.5± 6.2pening events (ind−1 day−1) 0.86± 0.05

alve closure activity was monitored and analyzed for 13.2 day

ime represents a behavioural response to varian environmental parameters.Table 1focuses on thalve closure responses during the 14-day perioach experimental condition. We show that, inxperimental conditions, opening time decreasedificantly in hypoxia. In normoxia, cadmium co

amination induced a decrease in opening time (f8.5± 6.2 to 38.8± 5.4%), whereas we did not obse

his effect under hypoxia (27.8± 5.1 and 31.0± 3.7%

re and oxygenation level (normoxia 21 kPa and hypoxia 4 kPa

Hypoxia

] = 30�g l−1 [Cd] = 0�g l−1 [Cd] = 30�g l−1

h 19 min 88 h 21 min 88 h 14 min± 5.4 27.8± 5.1 31.0± 3.7

3± 0.19 0.65± 0.07 1.10± 0.15

h). Five clams per condition were equipped for valvometric stu

A. Legeay et al. / Aquatic Toxicology 74 (2005) 242–253 247

opening time without or with cadmium). These resultssuggest that animals are able to detect Cd in waterand modify their behaviour as a protective responsein normoxia. In turn, this adaptative response wasnot observed under hypoxia presumably due to theconstraint imposed by the O2-supply requirement. Infact, when cadmium contamination was performedin hypoxia, the number of valve movements doubled(1.1± 0.15 instead of 0.65± 0.07 ind−1 day−1) but thedaily opening time remained constant in comparisonwith the uncontaminated hypoxic condition. AsC.fluminea is known to hyperventilate in hypoxia, wethought that different Cd accumulation rates in theanimals and thus different cellular responses betweennormoxia and hypoxia may occur. Consequently, wefirst characterized the metal tissue contamination lev-els in each condition.

3.2. Cadmium accumulation

Cadmium concentrations in the gills of clamsexposed to different oxygenation and cadmium levelsfor 14 days are reported inFig. 1. Cd levels in thegills of normoxic and hypoxic uncontaminated clamsdid not change during the course of the experimentand remained around 118± 11 ng g−1 (normoxic) and132± 14 ng g−1 (hypoxic) wet weight. In Cd-exposedclams, Cd levels increased regularly with time of expo-sure. At day 14, significantly, the hypoxic contami-nated clams accumulated around 30% more Cd than

Ft r-m

Fig. 2. Metallothionein concentrations (mean± S.E.) in gills of Cd-exposed and non-exposed clamsC. fluminea, under two levels ofoxygenation, after 7 and 14 days of treatment.

normoxic contaminated animals (8001± 827 ng g−1

against 5970± 469 ng g−1).

3.3. Metallothionein concentrations

Due to the late response of MT, this parameterwas only measured after 7 and 14 days of exposure(Fig. 2A) together with the cadmium concentrationin the soluble fraction of the gills (just before theMT Hg-saturation,Fig. 2B). After 7 days of metalexposure, a significant induction of MT synthesis inthe gills was observed compared with normoxic andhypoxic controls (+37 and +173% in normoxic- andhypoxic-exposed clams, respectively). This synthe-sis was significantly higher in hypoxic contaminatedbivalves than in normoxic contaminated ones. Notethat the MT induction appeared well correlated to Cdconcentration in the soluble fraction but not in thewhole gill (compareFig. 1 and Fig. 2B). Indeed, atthis time, cadmium concentration in the soluble frac-tion of gills under hypoxic conditions was significantlyhigher than under normoxia (6.16± 0.8�g Cd g−1

against 4.28± 0.6�g Cd g−1 ww). No difference

ig. 1. Evolution of the cadmium concentration (mean± S.E.) withime in gills of C. flumineaunder two levels of oxygenation (nooxia 21 kPa and hypoxia 4 kPa).

248 A. Legeay et al. / Aquatic Toxicology 74 (2005) 242–253

Fig. 3. Western blot showing MXR protein levels inC. flumineagills after cadmium exposure under two levels of oxygenation (panels A andB). This pattern is representative of five Western blots. Panel C shows the evolution of MXR protein level in the gills as a function of exposuretime. MXR protein levels correspond to the integrated band intensities taken from the blot, as determined with Scion Image program (in arbitraryunits). This digital treatment of the blot’s band intensities does not pretend to be quantitative but helps readers see the evolution of MXR proteinpattern.

was observed between days 7 and 14 in hypoxicCd-exposed clams (4.18± 0.22 nmol Hg g−1 against4.31± 0.95 nmol Hg g−1) whereas a slight increaseoccurred in normoxic ones (2.68± 0.17 to 3.5± 0.41).We suggest that a plateau value could have beenreached at day 7 in hypoxic conditions.

3.4. MXR induction

Fig. 3shows Western blots of clams’ gills after 1, 3,7 and 14 days of experiment. A band at 135 kDa, cor-responding to MXR, consistently shows the strongeststaining in the lanes loaded with protein from bivalvesexposed to cadmium (panel B). Semi-quantitative anal-ysis of the bands (panel C) shows differential proteinlevels between (i) normoxic and hypoxic clams inexposed and non-exposed clams and (ii) uncontami-nated and contaminated clams (compare panels A andB) suggesting that MXR protein levels are induced bothby cadmium and oxygenation. Surprisingly, MXR pro-

tein levels were higher in normoxic than in hypoxicclams. As shown in a previous study carried out in sim-ilar experimental conditions (Achard et al., 2004), inuncontaminated bivalves, we observed an increase withtime in normoxia whereas hypoxic levels remainedconstant (Fig. 3C). Under Cd exposure, a pronouncedMXR induction was observed (Fig. 3B and C). MXRlevels increased with time in normoxic and hypoxicmetal-exposed clams and highest levels were reachedunder normoxic conditions. Despite these lower MXRlevels under hypoxia, note that normoxic induction wastwice that observed at day 14 compared with the con-trol animals, whereas a nine-fold induction was reachedunder hypoxia.

3.5. Antioxidant enzyme activities and MDAcontents

The antioxidant enzyme activities in gills of the fourgroups of clams are presented inFig. 4. No difference

A. Legeay et al. / Aquatic Toxicology 74 (2005) 242–253 249

Fig. 4. Responses of catalase (CAT), glutathione reductase (GRD), total glutathione peroxidase (totGPX) and selenium-dependent peroxidase(SeGPX) in gills of Cd-exposed and non-exposedC. flumineaclams, under two levels of oxygenation, after 3, 7 and 14 days of treatment.Enzymatic activities are expressed as mean± S.E.

was observed for catalase activities in the four condi-tions. In this experiment, neither hypoxia nor cadmiuminduced an effect on this enzyme inC. fluminea.

For GRD activity we observed a significant decreasebetween days 3 and 14 in normoxic conditions. Thisdecrease reached 38% without cadmium and 31% withcadmium. In hypoxic conditions, we noted a relativestability of the activity from 3 to 14 days withoutcadmium, whereas a decrease occurred at day 7 in Cd-hypoxic conditions before a recovery at day 14.

A significant decrease in total GPX activities wasobserved in the two normoxic groups. This decreasewas continuous from 3 to 14 days and more severe inthe presence of Cd (−76% of activity between days 3and 14 instead of−62% without metal). Under hypoxicconditions, total GPX activity decreased slightly in theabsence of Cd (−32% between days 3 and 14) whereasit fell between days 7 and 14 in the presence of Cd(−60%). In brief, total GPX activities decreased withtime in all conditions, with an acute effect under metalexposure. The greatest decrease took place in normoxiccontaminated clams.

The kinetics of SeGPX activities were quite sim-ilar to total GPX ones. A decrease with time was

also observed for all groups except the uncontami-nated hypoxic one. This decrease was more markedin normoxic contaminated clams:−53% in normoxicuncontaminated clams;−60 and−53% with Cd undernormoxia and hypoxia.

Cadmium effects on lipid peroxidation are illus-trated inFig. 5. Apart from the control group, where

Fig. 5. Malondialdehyde concentrations (MDA, mean± S.E.) ingills of Cd-exposed and non-exposedC. flumineaclams, under twolevels of oxygenation, after 3, 7 and 14 days of treatment.

250 A. Legeay et al. / Aquatic Toxicology 74 (2005) 242–253

the increase was not significant, MDA concentrationsincreased significantly from days 3 to 14 in the otherthree conditions. However, note that under normoxicconditions the presence of Cd exerted no influence onMDA generation at the three exposure times. However,the greatest quantities of Cd accumulated in hypoxicconditions stimulated lipid peroxidation (×2.7 and 2.2at days 7 and 14, respectively).

4. Discussion

Laboratory studies are useful for evaluating theresponse to metal contamination and for a better under-standing of cellular impact in order to choose the mostuseful parameters to follow up in field experimenta-tions. The aim of this experiment was to investigate theresponses of several biomarkers to cadmium exposure(30�g l−1) under normoxic and hypoxic conditions (21and 4 kPa in the water) in gills of the freshwater bivalveC. fluminea, and to determine which are the most effi-cient.

In metal-exposed animals, we showed an increasingaccumulation of Cd in gills with time. This accumula-tion was more severe under hypoxic conditions. In fact,a decrease in water oxygenation induced ventilatorystimulation which enhanced the biological availabil-ity of Cd in C. fluminea(Tran et al., 2001). Despitea lower total opening time in hypoxic contaminatedclams in our experimental conditions (31.0± 3.7%at uxo ani , ast te Tsw lamR 8m s(t aysa ,1 on-c dern by2 Tsa

they are in many other molluscs. Indeed, MT con-centrations showed a significant correlation with Cdbioaccumulation in the soluble fraction taking intoaccount all our experimental data (Rspearman= 0.87).Moreover, we showed in this study that hypoxic Cdcontamination stimulates their synthesis. We shouldalso bear in mind that MTs may have an antioxidantrole in molluscs as the highly toxic hydroxyl radi-cal can be scavenged by MTs (Viarengo et al., 1999,2000).

As shown in previous studies, cadmium inducesMXR in molluscs (Eufemia and Epel, 2000; Achardet al., 2004). Our experiment demonstrated that thisresponse is very precocious since after one day’s expo-sure we observed a significant induction in three groupsof organisms out of four (no induction in the hypoxicuncontaminated group). Surprisingly, the MXR basallevel increased with time in the control normoxic groupalthough it remained lower than in the two groups ofCd-exposed clams. In accordance withAchard et al.’s(2004) findings, we suggest that the accumulation oforganic waste resulting from the 14 days experimentalperiod without water renewal, could contribute to thisMXR induction. Interestingly, the basal level of MXRstayed constant under hypoxic conditions in uncon-taminated clams, suggesting that the gene expressionencoding for the MXR protein should be triggered toambient oxygen level. The underlying mechanisms arepresently unclear and obviously require further exper-iments. Note finally that, following metal exposure,t tedu owst fore weo per-tw

sys-te andUe ium( ani utaseas ds.Tm

gainst 38.8± 5.4% in normoxic conditions,Table 1),his increase in ventilation allowed a greater inflf metal into the gills. This rise in Cd induced

ncrease in MT synthesis in contaminated clamshey play a key role in metal detoxication (Baudrimont al., 2003). Indeed, cadmium is known to induce Mynthesis inC. fluminea(Baudrimont et al., 1997) asell as in other molluscs, such as the marine cuditapes decussatus(Bebianno and Serafim, 199),ussels (Bebianno and Langston, 1991) and oyster

Rosijadi and Klerks, 1989; Geret et al., 2002a), wherehe induction of MT synthesis started only 4 dfter Cd exposure (200�g l−1, Rosijadi and Klerks989). In our study, after 7 days exposure, MT centrations were multiplied by a factor of 1.4 unormoxia and 2.7 under hypoxia and at day 14.5 under normoxia and 3.3 under hypoxia. Mppeared to be a good biomarker inC. fluminea, as

he ratio of MXR induction was widely exacerbander hypoxia as compared with normoxia. This sh

hat monitoring MXR induction is of great interestcotoxicological studies. This is especially true asbserved the highest inductions under hypoxia, a

inent environmental parameter forC. flumineain theild.Several authors have dealt with the antioxidant

em in response to cadmium contamination (Almeidat al., 2002; Geret et al., 2002b; Siraj Bashasha Rani, 2003; Company et al., 2004). Almeidat al. (2002)showed that acute exposure to cadm0.35–3.0 mg l−1) over a 60-day period inducedncrease in antioxidant defenses (superoxide dismnd GPX) in survival fishOreochromis niloticusand noignificant alteration in lipid peroxidation compounhe same results were observed inOreochromisossambicusafter 15 days Cd-exposure (5 mg l−1) but

A. Legeay et al. / Aquatic Toxicology 74 (2005) 242–253 251

this was followed by a decrease in antioxidant enzymeactivities after 30 days of exposure (Siraj Basha andUsha Rani, 2003). On the other hand, Cd produced aninhibitory effect in the enzymatic defence system in thegills of the hydrothermal vent musselBathymodiolusazoricuswith a significant decrease in CAT and totalGPX activities and an increase in LPO levels (Companyet al., 2004). Inhibition of GPX activities has alsobeen reported in the gills of Cd-contaminated marineclamsR. decussatusconcomitant with an increase inMDA levels (Geret et al., 2002b). Such effects havealso been seen under various toxicant situations, as inthe freshwater bivalveUnio tumiduswhen exposed tothiram (a dithiocarbamate known to induce oxidativestress in mammals) where SeGPX activities were alsomarkedly lower compared with controls in the gills oftreated mussels (−66%). The same pattern was foundfor GRD activity with a loss of 63% (Doyotte et al.,1997). In the same study, some field experiments wereconducted in a complex industrial effluent where poly-cyclic aromatic hydrocarbons, polychlorobiphenylsand metals were mixed. Once again lower SeGPX andGRD activities were observed, whereas CAT was notparticularly modified under these stressful conditions(Doyotte et al., 1997). These results show that Cd con-tamination does not necessarily lead to the inductionof antioxidant enzymes but rather to their depletion.In keeping with this, we here report a lack of influenceof oxygenation levels and Cd on CAT, a normoxicCd-independent decrease in GRD, and an oxygen-a s ofG ceo inC iv-i to8t orec ved,t opew oft indti erP

s e inh

ssb lex

environmental conditions when pollutant concentra-tions are low. Clear-cut inductible responses are onlyobtained with very high concentrations of metal, butwhich are environmentally irrelevant (0.35–3 mg l−1

Cd inAlmeida et al., 2002; 5 mg l−1 Cd inSiraj Bashaand Usha Rani, 2003).

The MDA results showed a small increase with timein the control group, probably related to waste accu-mulation. Although several studies have shown a LPOincrease under normoxic cadmium contamination inmolluscs (Geret et al., 2002b; Company et al., 2004),no significant change in TBARs compared with con-trol appeared at any of our experimental times exceptat days 7 and 14 in hypoxic contaminated clams. Atthese times, no significant differences were observedbetween normoxic and hypoxic antioxidant enzymeactivities (CAT, GRD and GPX) in contaminated ani-mals. Thus, the increase in TBARs in hypoxic bivalvesis not due to a lower antioxidant enzyme levels. Wesuggest that in this group, where the MXR levels arelower than in normoxic contaminated clams, xenobi-otics, such as MDA, accumulated rapidly. Otherwise, astudy onC. flumineashowed that during valve closureclams depressed their energy metabolism to less than10% of the standard metabolic rate measured when thevalves are open (Ortmann and Grieshaber, 2003). Wesuggest that under our hypoxic conditions less ATPwould be produced, and less would therefore be avail-able for the various ATPase which could be implied indetoxicant mechanisms.

kerf nalF ual-i mst nvi-r to allo ainia antr oxi-d s noc ions.

R

A 004.the

nd Cd-independent time decrease in both typePX activities. In contrast, a study of the influenf abiotic factors on oxidative stress indicators. flumineashowed a slight increase in CAT act

ty in the entire soft bodies under hypoxia from 6�mol min−1 mg−1 prot (Vidal et al., 2002). Although

he authors recognize that it would have been moherent if a decrease in CAT activity were obserhey suggest that an increase in ventilation could cith the low oxygen concentration within the 5 days

he experiment. Nevertheless, we have to keep in mhat, like many water breathers,C. flumineamaintainsts arterial bloodPO2 at 3 kPa whatever the watO2 within the range 4–40 kPa (Tran et al., 2000)o the pool of free radicals should not increasypoxia.

In conclusion, the induction of oxidative streiomarkers is hardly predictable, especially in comp

C. flumineais a tool currently used as a biomaror water contamination, for instance in the natiorench program ECODYN which assesses water q

ty in the Lot-Garonne continuum. Our study confirhe importance of a multi-biomarker approach in eonmental studies as some are not appropriaterganisms. Cadmium accumulation and MT rem

ndispensable biomarkers inC. flumineaand MXRppears to be promising with a quick and significesponse under contamination and hypoxia. Antiant enzymes did not live up to our expectations alear results were seen in our experimental condit

eferences

chard, M., Baudrimont, M., Boudou, A., Bourdineaud, J.P., 2Induction of a multixenobiotic resistance protein (MXR) in

252 A. Legeay et al. / Aquatic Toxicology 74 (2005) 242–253

Asiatic clamCorbicula flumineaafter heavy metals exposure.Aquat. Toxicol. 67, 347–357.

Almeida, J.A., Diniz, Y.S., Marques, S.F.G., Faine, L.A., Ribas, B.O.,Burneiko, R.C., Novelli, E.L.B., 2002. The use of the oxida-tive stress responses as biomarkers in Nile tilapia (Oreochromisniloticus) exposed to in vivo cadmium contamination. Environ.Int. 27, 673–679.

Baudrimont, M., Andres, S., Durrieu, G., Boudou, A., 2003. The keyrole of metallothioneins in the bivalveCorbicula flumineaduringthe depuration phase, after in situ exposure to Cd and Zn. Aquat.Toxicol. 63, 89–102.

Baudrimont, M., Metivaud, J., Maury-Brachet, R., Ribeyre,F., Boudou, A., 1997. Bioaccumulation and metallothioneinresponse in the Asiatic clam (Corbicula fluminea) after exper-imental exposure to cadmium and inorganic mercury. Environ.Toxicol. Chem. 16, 2096–2105.

Bebianno, M.J., Serafim, M.A., 1998. Comparison of metalloth-ionein induction in response to cadmium in the gills of the bivalvemolluscsMytilus galloprovincialisandRuditapes decussatus.Sci. Total Environ. 214, 123–131.

Bebianno, M.J., Langston, W.J., 1991. Metallothionein inductionin Mytilus edulisexposed to cadmium. Mar. Biol. 108, 91–96.

Brennan, R.J., Schiestl, R.H., 1996. Cadmium is an inducer of oxida-tive stress in yeast. Mutat. Res. 356, 171–178.

Briton, J.C., Morton, B., 1982. A dissection guide, field and lab-oratory manual for the introduced bivalveCorbicula fluminea.Malacol. Rev. 3 (Suppl.), 1–82.

Company, R., Serafim, A., Bebianno, M.J., Cosson, R., Shillito, B.,Fiala-Medioni, A., 2004. Effect of cadmium, copper and mercuryon antioxidant enzyme activities and lipid peroxidation in thegills of the hydrothermal vent musselBathymodiolus azoricus.Mar. Environ. Res. 58, 377–381.

Cossu, C., Doyotte, A., Jacquin, M.C., Babut, M., Exinger, A.,Vasseur, P., 1997. Glutathione reductase, selenium-dependent

ida-f. 38,

D 997.rel-gills

E in-58,

E alde-4-

E bi-sselon-

G 002a.and

sostrea gigas) and the mussel (Mytilus edulis). Aquat. LivingResour. 15, 61–66.

Geret, F., Serafim, A., Barreira, L., Bebianno, M.J., 2002b. Effectof cadmium on antioxidant enzyme activities and lipid peroxida-tion in the gills of the clamRuditapes decussatus. Biomarkers 7,242–256.

Geret, F., Serafim, A., Bebianno, M.J., 2003. Antioxidant enzymeactivities, metallothioneins and lipid peroxidation as biomarkersin Ruditapes decussatus? Ecotoxicol 12, 417–426.

Greenwald, R.A., 1985. Handbook of Methods for Oxygen RadicalResearch. CRC Press, Boca Raton, FL.

Lawrence, R.A., Burk, F.B., 1976. Glutathione peroxidase activity inselenium-deficient rat liver. Biochem. Biophys. Res. Commun.71, 952–958.

Livingstone, D.R., Garcia Martinez, P., Michel, X., Narbonne, J.F.,O’Hara, S., Ribera, C., Winston, G.W., 1990. Oxyradical produc-tion as pollution-mediated mechanism of toxicity in the commonmussel,Mytilus edulisL., and other molluscs. Funct. Ecol. 4,415–424.

Lowry, O.H., Rosenbrough, N.J., Farr, A.L., Randall, R.J., 1951. Pro-tein measurement with the folin phenol reagent. J. Biol. Chem.193, 265–275.

Ortmann, C., Grieshaber, M.K., 2003. Energy metabolism and valveclosure behaviour in the Asian clamCorbicula fluminea. J. Exp.Biol. 206, 4167–4178.

Ramos-Martinez, J.I., Bartolome, T.R., Pernas, R.V., 1983. Purifi-cation and properties of glutathione reductase from hepatopan-creas ofMytilus edulis. Comp. Biochem. Physiol. 75 B, 689–692.

Regoli, F., Orlando, E., 1994. Accumulation and subcellular distri-bution of metals (Cu, Fe, Mn, Pb and Zn) in the mediterraneanmusselMytilus galloprovincialisduring a field transplant exper-iment. Mar. Pollut. Bull. 28, 592–600.

Rosijadi, G., Klerks, P.L., 1989. Kinetic analysis of cadmium bindingto metallothionein and other intracellular ligands in oyster gills.

S idant-

S the1–

S atives. J.

T ain-vels,

T ena-iaticn.

T 003.e toicol.

glutathione peroxidase, glutathione levels, and lipid peroxtion in freshwater bivalves,Unio tumidus, as biomarkers oaquatic contamination in field studies. Ecotox. Environ. Saf122–131.

oyotte, A., Cossu, C., Jacquin, M.C., Babut, M., Vasseur, P., 1Antioxidant enzymes, glutathione and lipid peroxidation asevant biomarkers of experimental or field exposure in theand the digestive gland of the freshwater bivalveUnio tumidus.Aquat. Toxicol. 39, 93–110.

ndicott, J.A., Ling, V., 1989. The biochemistry of P-glycoprotemediated multidrug resistance. Annu. Rev. Biochem.137–171.

sterbauer, H., Cheeseman, K.H., 1990. Determination ofhydic lipid peroxidation products: malonaldehyde andhydroxynonenal. Methods Enzymol. 186, 407–421.

ufemia, N.A., Epel, D., 2000. Induction of the multixenootic defense mechanism (MXR), P-glycoprotein, in the muMytilus californianusas a general cellular response to envirmental stresses. Aquat. Toxicol. 49, 89–100.

eret, F., Jouan, A., Turpin, V., Bebianno, M.J., Cosson, R.P., 2Influence of metal exposure on metallothionein synthesislipid peroxidation in two bivalve mollusks: the oyster (Cras-

J. Exp. Zool. 251, 1–12.iraj Basha, P., Usha Rani, A., 2003. Cadmium-induced antiox

defense mechanism in freshwater teleostOreochromis mossambicus(Tilapia). Ecotox. Environ. Saf. 56, 218–221.

tohs, S.J., Bagchi, D., 1995. Oxidative mechanisms intoxicity of metal ions. Free. Radic. Biol. Med. 18, 32336.

tohs, S.J., Bagchi, D., Hassoun, E., Bagchi, M., 2000. Oxidmechanisms in the toxicity of chromium and cadmium ionEnviron. Pathol. Toxicol. Oncol. 19, 201–213.

ran, D., Boudou, A., Massabuau, J.C., 2000. Mechanism for mtaining oxygen consumption under varying oxygenation lein the freshwater clamCorbicula fluminea. Can. J. Zool. 782027–2036.

ran, D., Boudou, A., Massabuau, J.C., 2001. How water oxygtion level influences cadmium accumulation pattern in the AsclamCorbicula fluminea: a laboratory and field study. EnviroToxicol. Chem. 20, 2073–2080.

ran, D., Ciret, P., Ciutat, A., Durrieu, G., Massabuau, J.-C., 2Estimation of potential and limits of bivalve closure responsdetect contaminants: application to cadmium. Environ. ToxChem. 22, 116–122.

A. Legeay et al. / Aquatic Toxicology 74 (2005) 242–253 253

Viarengo, A., Burlando, B., Cavaletto, M., Marchi, B., Ponzano, E.,Blasco, J., 1999. Role of metallothionein against oxidative stressin the musselMytilus galloprovincialis. Am. J. Physiol. (Regul.Integr. Comp. Physiol. 46) 277, R1612–R1619.

Viarengo, A., Burlando, B., Ceratto, N., Panfoli, I., 2000. Antioxidantrole of metallothioneins: a comparative overview. Cell. Mol. Biol.46, 407–417.

Vidal, M.L., Basseres, A., Narbonne, J.F., 2001. Potential biomarkersof trichloroethylene and toluene exposure inCorbicula fluminea.Environ. Toxicol. Pharm. 9, 87–97.

Vidal, M.L., Basseres, A., Narbonne, J.F., 2002. Influence of temper-ature, pH, oxygenation, water-type and substrate on biomarkerresponses in the freshwater clamCorbicula fluminea. Comp.Biochem. Physiol. 132C, 93–104.