The use of antioxidant enzymes in freshwater biofilms: Temporal variability vs. toxicological responses

Aquatic Toxicology 136–137 (2013) 60–71

Contents lists available at SciVerse ScienceDirect

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

The use of antioxidant enzymes in freshwater biofilms: Temporal variability vs. toxicological responses Chloé Bonnineau a,∗ , Ahmed Tlili b,c , Leslie Faggiano a , Bernard Montuelle b,d , Helena Guasch a a

Institute of Aquatic Ecology, Campus Montilivi, 17071 Girona, Spain Irstea, UR MALY, 5 rue de la Doua, 69626 Villeurbanne Cedex, France c Institute of Freshwater Ecology and Inland Fisheries, Alte Fischerhütte 2, Neuglobsow, Germany d INRA – UMR Carrtel, 75 avenue de Corzent, BP 511, 74203 Thonon les Bains, France b

a r t i c l e

i n f o

Article history: Received 8 December 2011 Received in revised form 7 March 2013 Accepted 12 March 2013 Keywords: Catalase Periphyton Oxyfluorfen Tolerance acquisition

a b s t r a c t This study aims to investigate the potential of antioxidant enzyme activities (AEA) as biomarkers of oxidative stress in freshwater biofilms. Therefore, biofilms were grown in channels for 38 days and then exposed to different concentrations (0–150 ␮g L−1 ) of the herbicide oxyfluorfen for 5 more weeks. Under control conditions, the AEA of biofilms were found to change throughout time with a significant increase in ascorbate peroxidase (APX) activity during the exponential growth and a more important role of catalase (CAT) and glutathione reductase (GR) activities during the slow growth phase. Chronic exposure to oxyfluorfen led to slight variations in AEA, however, the ranges of variability of AEA in controls and exposed communities were similar, highlighting the difficulty of a direct interpretation of AEA values. After 5 weeks of exposure to oxyfluorfen, no clear effects were observed on chl-a concentration or on the composition of other pigments suggesting that algal group composition was not affected. Eukaryotic communities were structured clearly by toxicant concentration and both eukaryotic and bacterial richness were reduced in communities exposed to the highest concentration. In addition, during acute exposure tests performed at the end of the chronic exposure, biofilms chronically exposed to 75 and 150 ␮g L−1 oxyfluorfen showed a higher CAT activity than controls. Chronic exposure to oxyfluorfen provoked then structural changes but also functional changes in the capacity of biofilm CAT activity to respond to a sudden increase in concentration, suggesting a selection of species with higher antioxidant capacity. This study highlighted the difficulty of interpretation of AEA values due to their temporal variation and to the absence of absolute threshold value indicative of oxidative stress induced by contaminants. Nevertheless, the determination of AEA pattern throughout acute exposure test is of high interest to compare oxidative stress levels undergone by different biofilm communities and thus determine their antioxidant capacity. © 2013 Elsevier B.V. All rights reserved.

1. Introduction In freshwater ecosystems, biofilm communities are now recognized as pertinent indicators of perturbations (Sabater and Admiraal, 2005). These complex communities, composed of algae, bacteria, fungi, and protozoa, are embedded in a matrix constituted by extra-polymeric substances (EPS). They live attached to different types of substrates (cobbles, wood, sand, etc.) and are the

Abbreviations: EPS, extra-polymeric substances; AEA, antioxidant enzyme activities; ROS, reactive oxygen species; CAT, catalase; APX, ascorbate peroxidase; GR, glutathione reductase. ∗ Corresponding author at: Universitat de Girona, Facultat de Ciències, Institute of Aquatic Ecology, Avinguda Montilivi s/n. 17071 Girona, Spain. Tel.: +34 479818036. E-mail address: [email protected] (C. Bonnineau). 0166-445X/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.aquatox.2013.03.009

main primary producers in open streams (Romaní, 2010; Stevenson et al., 1996). To assess biofilm status, different structural and functional variables are usually determined. They include community composition (mostly of diatoms), photosynthesis, biomass and heterotrophic activity (Sabater et al., 2007; Weitzel, 1979). To complete the information given by these indicators, we propose the use of antioxidant enzyme activities (AEA) in biofilms as indicators of oxidative stress induced by toxicants. In fact antioxidant enzymes participate in the regulation of reactive oxygen species (ROS) levels to avoid their accumulation and the resulting oxidative stress (Mittler, 2002). Previous studies highlighted the interest of AEA as sensitive markers of stress induced by organic and inorganic toxicants. Dewez et al. showed that the catalase (CAT) activity was a more sensitive biomarker of fludioxionil toxicity than photosynthetic parameters in Scenedesmus obliquus (Dewez et al., 2005). In freshwater biofilms, AEA were found to be more sensitive to

C. Bonnineau et al. / Aquatic Toxicology 136–137 (2013) 60–71

copper toxicity than photosynthetic parameters (Guasch et al., 2010). The present study focused on three important antioxidant enzymes: CAT, ascorbate peroxidase (APX) and glutathione reductase (GR). CAT (EC 1.11.1.6) and APX (EC 1.11.1.11) catalyse the transformation of hydrogen peroxide in water and oxygen mainly in peroxisomes and chloroplast, respectively (Chelikani et al., 2004; Lesser, 2006). GR (1.8.1.7) participates in this reaction by regenerating the cofactor needed by APX (ascorbate-glutathione cycle, Mittler, 2002). Since temporal variations affect function and structure of communities strongly, functional biomarkers chosen to reflect perturbations, such as AEA, may also change due to temporal variability. Indeed, biofilms are very dynamic communities in which changes in biomass due to processes of attachment, colonization, exponential growth, senescence and sloughing (Biggs, 1996) are concomitant to species succession (Hudon and Bourget, 1981; Peterson and Stevenson, 1990). These processes are linked to changes in community functioning. For instance, Sabater and Romaní (1996) found a higher respiratory activity in younger rather than in mature biofilms from an undisturbed Mediterranean stream. Romaní et al. (2008) also observed that the release of extracellular bacterial enzymes allowing organic matter compound degradation in the EPS matrix was higher at the beginning of the biofilm formation than at the end of colonization. Though, in ecotoxicology, the effect of these temporal variations is reduced by using same age communities (Clements and Newman, 2002), the temporal variation represents an estimate of the “natural” range of variation and thus may still be a pertinent scale to interpret the importance of further variations related to disturbances. Since enzymes are sensitive to different factors (e.g. pH, temperature), changes in biofilm environmental conditions due to growth are expected to provoke variations in AEA. To our knowledge, patterns of AEA in freshwater biofilms throughout time are unknown. Thus, the first aim of this study was to determine the “temporal” range of variation of AEA of the biofilm. This background information is essential to interpret the importance of AEA’s variations in response to chemical exposure and thus to use AEAs as biomarkers of oxidative stress. In the present study, temporal changes in AEA of non-exposed biofilms were investigated on mature biofilm communities established after several weeks of colonization, during their transition from an exponential to a slow growth phase. Mature communities were used because ecotoxicological tests are better performed on those communities to test toxicant impact on a fully active community (Clements and Newman, 2002). The study of the temporal variability of AEA in non-exposed communities was complemented with more traditional biofilm metrics, such as biomass variables, photosynthetic parameters (Sabater et al., 2007), relative abundance of the different algal groups (based on marker pigments; Jeffrey et al., 1997) and some antioxidant pigments (e.g. carotenoids; Pinto et al., 2003). In the second part of this study, the temporal variability of AEA observed in a control situation was compared to the variations of AEA in response to contamination during acute and chronic ecotoxicological tests. For these tests, the herbicide oxyfluorfen was selected since it is representative of compounds likely to be tested in ecotoxicological tests and it is expected to provoke oxidative stress. Indeed, this diphenyl-ether herbicide inhibits chlorophyll-a biosynthesis and provokes the accumulation in the cytoplasm of protoporphyrin IX, a potent photo-sensitizer that generates high levels of singlet oxygen and so oxidative stress (Aizawa and Brown, 1999; Duke et al., 1991). Though its use has been recently reapproved by the UE, oxyfluorfen exposure has been shown to provoke oxidative stress in algae (Geoffroy et al., 2003; Kunert et al., 1985; Sandmann and Böger, 1983) and cyanobacteria (Sheeba et al., 2011). Consequently, the European Food Safety Authority (EFSA)

61

pointed out the high risk for algae by this compound as well as the need for further studies on its potential impact on aquatic organisms (EFSA, 2010). Based on these previous studies, oxyfluorfen is expected to provoke oxidative stress and so changes in AEA also in biofilms, but has never been tested. In the present study, changes in AEA after acute and chronic exposure to oxyfluorfen were compared to changes in more traditional metrics as described earlier. In addition, after five weeks of exposure, the structure (bacterial/eukaryotic diversity, algal composition) and the function (AEA response in short-term toxicity tests) of the exposed and control communities were compared to assess whether AEA plays a role in the selection of more resistant species expected to occur under chronic exposure of a community to a critical level of contaminant. The objectives of the present study were then: 1. to characterize the pattern of temporal variation of AEA in mature biofilms. 2. to compare the temporal variation of AEA and the toxicological variation provoked by the exposure to oxyfluorfen, a toxicant inducing oxidative stress in mature biofilms. 3. to determine the influence of chronic exposure on the capacity of biofilms to respond to a sudden increase in oxidative stress (induced by oxyfluorfen). 2. Materials and methods 2.1. Microcosm setup Colonization and exposure were performed in an indoor microcosm system consisting of 7 recirculating channels previously described by Serra et al. (2009a). Biofilms were allowed to colonize sandblasted glass substrata of 1.4 and 17 cm2 installed in the bottom of each channel. In each channel, 10 L of dechlorinated tap water was used as a culture medium and changed 3 times a week; aquarium pumps allowed water recirculation. At each water renewal, phosphate was added to a final nominal concentration of 30 ␮g L−1 to avoid nutrient depletion and P limitation. A cooling bath maintained the water temperature at 20 ◦ C. Once a week during the first 5 weeks of colonization, an original inoculum of biofilm collected, during the Spring season, from the river Llémana (NE Spain, Serra et al., 2009a) was added to each channel. Light was provided by halogen lamps (80–120 ␮mol photons m−2 s−1 ) with a light regime of 12 h:12 h light:dark. After 5 weeks of colonization, on day 38, biofilms were exposed to increasing concentrations of oxyfluorfen (CAS: 42874-03-3) following an exponential design (Ricart et al., 2009). Three channels were used as controls and the remaining 5 channels were exposed to 3, 7.5, 15, 75 or 150 ␮g L−1 of oxyfluorfen. Oxyfluorfen was added in each channel from a stock solution at 15 mg L−1 in 2.5% acetone to obtain 0.025% acetone in each channel, acetone was also added in a similar amount in control channels. At each water renewal, toxicant and/or acetone (when appropriate) were added to compensate for potential degradation of the toxicant and to ensure a maximal exposure. To characterize the temporal pattern of AEA and to link it with changes in other biological variables, the 3 control channels were sampled on days 33, 36, 38, 39, 41, 59, 66 and 73. At each sampling and from each control channel, 3 samples (each consisting of three 1.4 cm2 glass substrata) were collected randomly for AEA measurements, 5 samples (1.4 cm2 glass substrata each) for the measurement of photosynthetic efficiency and 3 samples (1.4 cm2 glass substrata each) for pigment analyses. To determine the variability of AEA due to contaminant exposure, AEA but also photosynthetic efficiency, protein content and wet weight of biofilms from all channels were measured after

62

C. Bonnineau et al. / Aquatic Toxicology 136–137 (2013) 60–71

exposure to oxyfluorfen during 6 h (day 38), 24 h (day 39), 48 h (day 41), 3 weeks (day 59), 4 weeks (day 66) and 5 weeks (day 73). To determine the structural changes in biofilm communities after 5 weeks of exposure to oxyfluorfen, samples were also collected for pigment (3 samples of 1.4 cm2 glass substrata each per channel) and DGGE (1 sample of 17 cm2 glass substrata per channel) analyses, just before the start of the exposure (day 38) and 5 weeks after the exposure (day 73). 2.2. Short-term toxicity tests To test the capacity of antioxidant response of biofilms acquired after 5 weeks of exposure to oxyfluorfen, short-term toxicity tests, based on AEA, were performed. Biofilms from the different channels were exposed to increasing concentrations of oxyfluorfen (0, 1.5, 15, 75, 150 and 1000 ␮g L−1 ) in a microcosm set-up previously described (Bonnineau et al., 2010). Briefly, 9 glass substrata of 1.4 cm2 were used for each concentration. Each glass substratum was incubated in a vial containing 10 ml of colonization medium and the corresponding toxicant concentration. Samples were incubated under the same conditions as the colonization, using a single-speed orbital mixer (KS260 Basic, IKA® ) to maintain constant agitation. After 6 h of exposure, for each concentration and each channel, 3 samples, each consisting of 3 glass substrata, were collected for AEA measurements. 2.3. Biofilm parameters 2.3.1. AEA Sampling, protein extraction (by homogenisation followed by glass beads disruption) and AEA measurements were performed as described previously (Bonnineau et al., 2011). The protein concentration was measured in triplicates for each sample by the method of Bradford (1976) using dye reagent concentrate from Bio-Rad (Bio-Rad Laboratories GmbH, Germany) and bovine serum albumin as a standard. The final concentration of protein was then expressed in ␮g mg−1 of biofilm wet weight. AEA measurements were performed as previously described (Bonnineau et al., 2012) in microtiter plates (UV-Star 96 well plate, Greiner® ), changes in absorbance were followed using a microtiter plate reader Synergy4 (BioTek® ). For all assays, the optimal protein concentration was determined using protein amounts between 0.5 and 6.5 ␮g. CAT activity was measured spectrophotometrically by following the decomposition of H2 O2 at 240 nm and 25 ◦ C during 2 min (Aebi, 1984). After determination of the optimal substrate concentration, the 250 ␮L reaction mixture contained in final concentration 80 mM of potassium phosphate buffer (pH 7.0) and 2 ␮g of proteins. The reaction was started by adding 35 mM of H2 O2 . CAT activity was calculated as ␮mol H2 O2 mg prot.−1 min−1 (extinction coefficient, ε: 0.039 cm2 ␮mol−1 ). Oxidation of sodium ascorbate by APX was measured at 290 nm and 25 ◦ C for 2 min according to Nakano and Asada (1981). After determination of the optimal substrate concentration, the 250 ␮L reaction mixture contained in final concentration: 80 mM of potassium phosphate buffer (pH 7.0), 150 ␮M of sodium ascorbate and 2 ␮g of proteins. The reaction was started by adding 4 mM of H2 O2 . APX activity was calculated as ␮mol ascorbate mg prot−1 min−1 (ε: 2.8 cm2 ␮mol−1 ). The oxidation of NADPH by GR was determined by measuring the decrease in absorbance at 340 nm and 25 ◦ C for 2 min (Schaedle and Bassham, 1977). After determination of the optimal cofactor (NADPH) concentration, the 200 ␮L reaction mixture contained in final concentration: 100 mM Tris–HCl (pH 7.5), 1 mM EDTA, 1 mM oxidized glutathione and 4 ␮g of proteins. The reaction was started

by adding 0.25 mM NADPH. GR activity was calculated as ␮mol NADPH mg prot−1 min−1 (ε: 6.22 cm2 ␮mol−1 ). 2.3.2. Photosynthetic parameters For each sample, in vivo photosynthetic efficiency determination was performed using a PAM (Pulse Amplitude Modulated) fluorometer. For technical reasons, measurements from day 33 to 41 were performed using a PhytoPAM (Heinz Walz, Effeltrich, Germany) while a MiniPAM (Heinz Walz, Effeltrich, Germany) was used for measurements from day 59 to day 73. The distance between the fibre optics and the sample surface was set at 2 mm. The fluorescence signal was determined by the emitter-detector unit (PHYTO-EDF). After light acclimation, 5 strong saturating pulses of light (8000 ␮mol photons m−2 s−1 ) were applied to the samples to obtain the fluorescence signal at the steady-state (F), the maximal fluorescence yield (Fm ) of an actinic-adapted sample and the minimal fluorescence yield (Fo ). These parameters were used to calculate the photosynthetic efficiency (Ph. eff. = Fv /Fm with Fv = Fm − F) following Genty et al. (1989). All calculations were done using the fluorescence signal recorded at 665 nm and are given as relative units of fluorescence. 2.3.3. Pigment analysis by high pressure liquid chromatography (HPLC) Samples were stored in 15 mL tubes at −80 ◦ C until further analysis. Pigment extraction was performed by ultrasonication as described by Dorigo et al. (2007). Determination of lipophilic pigment composition of biofilm was performed by HPLC as described by Tlili et al. (2008). The injection volume was 100 ␮L of purified biofilm extract and pigments were separated on a 4.6 mm × 250 mm column (Waters Spherisorb ODS5 25 ␮m). Pigment identification was done based on their retention time and absorption spectrum according to the Scientific Committee for Oceanic Research (SCOR, Jeffrey et al., 1997). For each sample, the relative abundance (expressed as the percentage of the sum of the areas for all the pigments in the sample) of each pigment was calculated. In addition, standard o chlorophyll-a was used to quantify its concentration in each sample, final concentrations are given in ␮g cm−2 . 2.3.4. DNA extraction – amplification and denaturing gradient gel electrophoresis (DGGE) analysis For each sample, biofilm was removed from the glass substrata with a cell scraper (Nunc, Wiesbaden, Germany) and put into a 15 mL tube. Samples were then centrifuged for 30 min at 10,000 × g and 4 ◦ C to remove the excess of water and stored at −80 ◦ C. Nucleic extraction, PCR amplification of eukaryotic 18S rRNA gene fragments and bacterial 16S rRNA gene fragments and their DGGE analysis were performed as described by Tlili et al. (2008). Samples collected on day 38 and after 5 weeks of exposure were loaded on a same gel to allow comparison between samples. 2.4. Statistical analyses The R software (R Development Core Team, 2008; Ihaka and Gentleman, 1996), the ‘ade4’ (Dray and Dufour, 2007), the ‘proxy’ (Meyer and Butcha, 2010) and the ‘vegan’ (Oksanen et al., 2010) packages were used to perform statistical analyses. The statistical significance for all the analyses was set at p < 0.05. For each sampling time and each biological variable, a mean was calculated from the different samples collected from one channel. These mean values (one per channel and per time) were then used in further analyses as independent replicates.

C. Bonnineau et al. / Aquatic Toxicology 136–137 (2013) 60–71

2.4.1. Temporal variation The temporal variation of biofilms was studied using biofilms from the 3 unexposed channels sampled at day 33, 36, 38 (t = 0 h only), 39, 41, 59, 66 and 73. To determine the different growth phases of biofilms, chlorophyll-a concentration was adjusted to the sigmoidal model previously proposed by Romaní (2010): chla = K/(1 + exp(−r × (day − d0 )) where K is the carrying capacity (maximal chlorophyll-a concentration reached), r the growth rate and d0 the time when maximal growth rate is achieved. Differences between colonization time in terms of AEA were estimated by analysis of variance (ANOVA) and post hoc analysed by a Tukey test. To understand the temporal variation of all the biological variables and to determine the importance of differences between the replicated channels, a multivariate approach was used. Two matrices were constructed (with samples as rows and biological variables as columns). The pigment matrix contained the relative abundance of each pigment (normalized using an arcsine square root transformation) and the function-biomass matrix contained different log-transformed variables: AEA, photosynthetic efficiency and biomass variables (chlorophyll-a in ␮g cm−2 , protein in ␮g mg−1 , wet weight in mg). Then, for each matrix, two between-PCAs (Principal Component Analysis) were carried out in which, for each sample, an information factor was added (i.e. the sampling day for the factor time or the channel number for the factor channel). The percentage of variance explained by one factor was calculated as the ratio between the sum of the eigenvalues of the between-PCA and the sum of the eigenvalues of the PCA (Dolédec and Chessel, 1987; Dray and Dufour, 2007). To measure the concordance between the two matrices (pigment and function-biomass), a co-inertia analysis was performed. This multivariate technique analyses co-structure by maximizing covariance between two matrices (Dray et al., 2003a; Dolédec and Chessel, 1994) and the calculation of the RV-coefficient allowed estimating the degree of concordance between the matrices (Robert and Escoufier, 1976). A Monte-Carlo permutation test on the sum of the eigenvalues of the co-inertia analysis was also performed to assess the significance of the RV-coefficient (Heo and Gabriel, 1998). 2.4.2. Oxyfluorfen exposure 2.4.2.1. Long-term exposure. AEA, photosynthetic efficiency, protein content and wet weight measured in all channels at days 38 (t = 6 h), 39, 41, 59, 66 and 73 were used to study the effects of oxyfluorfen exposure on biofilms. To focus on the chemical’s effects, the effect of time was removed by carrying out a within-PCA in which the mean of the samples in a same group (i.e. collected at the same time) is substracted to each sample of a group, for each variable. Thus, the patterns of variation obtained at each sampling time can be compared among them (Dray and Dufour, 2007). To compare the variability of AEA between control and exposed biofilms, variances of AEA from all biofilms collected at days 38 (t = 6 h), 39, 41, 59, 66 and 73 were compared between them. A distance matrix was built based on Euclidean distance and a permutation-based test of multivariate homogeneity of group

63

dispersions (variance) was performed, each group corresponding to a concentration of oxyfluorfen (Anderson et al., 2006). For samples collected on day 38 (just before exposure) and 73 (5 weeks after exposure), the relative abundance of pigments between the different treatments was analysed by a PCA. In addition, DGGE profiles of biofilms from the different channels were compared for presence or absence of bands by calculating the dissimilarity index of Jaccard; matrices were then used to perform the average method of hierarchical cluster analysis (HCA). 2.4.2.2. Short-term toxicity tests. To test the influence of previous chronic exposure to oxyfluorfen on AEA of biofilms exposed to acute exposure to higher oxyfluorfen concentrations, a two-way ANOVA was performed. The AEA for which the interaction term of the twoway ANOVA was significant were selected for further analyses. For each channel, a one-way ANOVA followed by a Tukey test, as a post hoc analysis, were performed on these AEA to reveal the differences between samples after acute exposure to different concentrations of oxyfluorfen. 3. Results 3.1. Biofilm colonization Physical and chemical conditions were stable during all the experiment although small differences were observed between the two periods (Table 1). Since the water used during this experiment has previously been characterized by Serra et al. (2009b) for NO3 (1.68 ± 0.14 mg L−1 ), NO2 (0.07 ± 0.01 mg L−1 ) and NH4 (<0.1 mg L−1 ) among others (n = 20 for all), only phosphorus concentration was measured (Table 1). Total phosphorus depletion was not observed neither during colonization nor during exposure periods. 3.2. Changes in unexposed biofilms throughout time 3.2.1. Algal growth Chlorophyll-a concentration in unexposed biofilms increased during the 10 weeks of the experiment and could be successfully adjusted to a sigmoidal growth curve (Fig. 1) with K = 17.2 ± 4.2 ␮g chlorophyll-a cm−2 (p < 0.05), d0 = 43.6 ± 7.8 days (p < 0.05) and r = 0.10 ± 0.07 day−1 (p < 0.2), all parameters are presented with their corresponding standard errors and their p-value for the tstatistic. Since the transition from the exponential growth phase to the slow growth phase is observed at the inflexion point, i.e. at day d0 , biofilms collected between day 33 and 41 of colonization were in exponential growth phase while samples collected between 59 and 73 days were in slow growth phase. The loss phase was not reached during this experiment as indicated by the d0 values and the continuous increase of biomass along the experiment (Fig. 1). 3.2.2. Temporal variations in AEA Temporal variations in AEA were observed in unexposed biofilms (Fig. 2). Although CAT and GR activities seemed to increase

Table 1 Physico-chemical conditions in all channels during colonization (day 0–38) and exposure period (day 39–73). For each parameter mean values and standards errors are indicated. Flow (L min−1 )

T (◦ C)

Dissolvedoxygen (mg L−1 )

pH

Cond. (␮S cm−1 )

P concentration (␮g L−1 ) Before water changes

After water changes

Colonization

1.45 ± 0.01 n = 117

20.5 ± 0.1 n = 93

8.88 ± 0.03 n = 93

8.58 ± 0.03 n = 93

418 ± 4 n = 93

22.5 ± 2.1 n = 42

5.6 ± 0.9 n = 42

Exposure

1.42 ± 0.02 n = 84

20.2 ± 0.1 n = 102

9.26 ± 0.02 n = 102

8.64 ± 0.04 n = 102

404 ± 5 n = 102

7.5 ± 0.6 n = 54

8.9 ± 1.3 n = 54

64

C. Bonnineau et al. / Aquatic Toxicology 136–137 (2013) 60–71

Fig. 1. Chlorophyll-a concentration (␮g cm−2 ) in unexposed biofilms throughout time. Triangles correspond to the means (and standard errors) of three channels. The plain line shows the sigmoidal growth model fitted to the data.

throughout time, these differences were only slightly significant (0.05 < p < 0.1). The maximum CAT and GR activities were observed at day 59 and correspond to an increase by 142 ± 54 and 163 ± 16% of activities at day 33, respectively. APX activities from biofilms collected at days 39 and 41 were significantly higher (by 142 ± 20 and 138 ± 47%, respectively) than APX activity of biofilms from day 33 (F = 4.7, p < 0.05). 3.2.3. Temporal variations in the whole biofilm Results of the two between-PCAs showed that the factor time explained 48.5% of the total variance in the pigment matrix (relative abundance of pigments) and 56.7% of the total variance in the function-biomass matrix (AEA, photosynthetic and biomass variables) highlighting a high temporal variability for all biological variables. The inter-channel variability was also estimated and results of the between-PCAs showed that the factor channel explained 12.0% and 11.9% of the total variance in the pigment and function-biomass matrices, respectively. As a result of the co-inertia analyses on the pigment and biomass-function matrices, a significant RV coefficient of 39.0% was obtained (permutation test, p < 0.05). A high coefficient indicates simultaneous variations (either positive or negative) of the two sets of variables while a low coefficient indicates independent variations (Dray et al., 2003b). Thus, this result indicated a good degree of concordance between the two matrices. The separation of the samples was strongly driven by the axis 1, which explained 64.6% of the variance while the axis 2 explained 24.6%. Samples from day 33 to 41 were separated from those from day 59 to 73 along the first axis, whereas the second axis accounted for variability within each group of samples (Fig. 3).

Fig. 3. Temporal changes in unexposed biofilms. Ordination of the samples from the three control channels by the co-inertia analysis. Each sample is represented by one arrow on which the sampling day and the channel number (c1, c2 or c3) are indicated in a box. The black dots (beginning of the arrows) represent the samples which have been ordinated by the PCA performed on the pigment matrix while the tops of the arrows represent the samples which have been ordinated by the PCA performed on the function-biomass matrix.

Biofilms from day 33 to 41 were characterized by a higher protein concentration and a higher APX activity (Fig. 4); these biological variables were associated with a higher relative abundance of the pigments diadinochrome II, chlorophyll c and violaxanthin (Fig. 4). Various pigments (lutein, ␤,␤-carotene, chlorophyll b, diatoxanthin) were characterized by low scores (<|0.06|) on the axis 1 of the co-inertia indicating a low variability of these pigments in biofilms along this axis (Fig. 4). Biofilms from day 59 to 73 were then characterized by a higher biomass (chl-a and wet weight) and higher CAT and GR activities (Fig. 4); these biological variables were also associated with a higher relative abundance of chlorophyll-a, carotenoid P468, zeaxanthin, and antheraxanthin (Fig. 4). Variability in biofilms from day 33 to 41 was higher along the second axis than in older biofilms. In addition, samples from day 33 to 41 were distributed chronologically along the second axis, the time being negatively correlated with this axis (Fig. 3). Inside this group of samples, the oldest ones (day 39–41) were characterized by higher AEA (Fig. 4), which were associated with higher relative abundance of chlorophyll b, lutein, ␤,␤-carotene, neoxanthin (Fig. 4). The youngest biofilm samples (day 33) were characterized Table 2 List of the 18 identified pigments included in multivariate analyses of unexposed and exposed biofilms. Code and corresponding pigment name.

Fig. 2. Temporal changes in AEA of unexposed biofilms. Symbols correspond to the means (and standard errors) of CAT (, black plain line), GR (, black dashed line) and APX (, grey line) activities in biofilms from the three control channels throughout time.

Code

Pigment name

Code

Pigment name

ANT bbCAR CAR CHLa CHLb CHLc DIAD DIADcI DIADcII

Antheraxanthin ␤,␤-Carotene Carotenoid P468 Chlorophyll-a Chlorophyll b Chlorophyll c Diadinoxanthin Diadinochrome I Diadinochrome II

DIAT FUC LUT NEO PHEPa PHERa tNEO VIO ZEA

Diatoxanthin Fucoxanthin Lutein Neoxanthin Pheophytin a Pheophorbide a Trans-neoxanthin Violaxanthin Zeaxanthin

C. Bonnineau et al. / Aquatic Toxicology 136–137 (2013) 60–71

65

Fig. 4. Temporal changes in unexposed biofilms. Normed coefficients of the different variables on the axes of the co-inertia analyses shown in Fig. 3. As a result of the co-inertia analysis, the samples were ordinated by the PCA on the function-biomass matrix or by the one on the pigment matrix. The left graph shows the normed coefficients of the variables of the function-biomass matrix (Ph. eff. stands for photosynthetic efficiency). The graph on the right shows the normed coefficients of the variables of the pigment matrix, the correspondence between pigment codes and complete names is given in Table 2.

by higher wet weight and protein concentration (Fig. 4) and also by a higher relative abundance of diadinochrome I and II, fucoxanthin and chlorophyll c (Fig. 4). The scores of each variable on the co-inertia axes are available in the supporting information (Tables A1 and A2). Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.aquatox. 2013.03.009.

of AEA of biofilms control and exposed to different oxyfluorfen concentrations were not significantly different (F = 1.68, p > 0.05, n = 999 permutations), indicating that AEA’s ranges of variation in control and exposed communities throughout time were similar. Despite of these similar ranges of variation, differences between control and exposed were observed in AEA after 5 weeks of exposure, the two-way ANOVA reveals a significant effect of chronic exposure to oxyfluorfen on CAT, APX and GR activities (Table 3).

3.3. Changes in biofilms during oxyfluorfen exposure

3.4. Changes in community structure and function after 5 weeks of chronic exposure

Throughout the 5 weeks of exposure, variations were observed mainly in AEA and photosynthetic efficiency (Fig. 5). While time explained about half of the variance (intra-group variance: 49.8%), only 7.4% of the total variance could be attributed to oxyfluorfen (result of between-PCA with factor oxyfluorfen). The first two axes of the within-PCA explained 26.2% of the variance (wPCA1: 13.1%, wPCA2: 11.8%). The pattern of biological response changed throughout exposure and was never concentration-dependent, but some general trends could be observed (Fig. 5). Samples exposed to 75 and 150 ␮g L−1 were mainly distributed along axis 1 which was negatively correlated with APX activity and positively correlated with photosynthetic efficiency and CAT activity mainly (Fig. 5). Samples exposed to 3 to 15 ␮g L−1 were mainly distributed along the axis 2 which was negatively correlated with GR and APX activities mainly (Fig. 5). The difficulty to observe clear concentration-response patterns can be due to the high variability of AEA throughout time observed in all treatments. Indeed, during the 5 weeks of exposure, CAT activities fluctuated in the range of 102.2–197.1 (control) and 87.5–274.3 (exposed) ␮mol H2 O2 mg prot.−1 min−1 , APX activities in the range of 0.369–1.120 (control) and 0.272–1.136 (exposed) ␮mol ascorbate mg prot−1 min−1 and GR activities in the range of 124.4–265.9 (control) and 95.7–330.2 (exposed) ␮M NADPH mg prot−1 min−1 . Though lower and higher activities were observed in exposed biofilms than in control ones, the variances

3.4.1. Pigment composition The first two axes of the PCA explained 68.6% of the variance (PCA1: 39.7%, PCA2: 28.8%). Most of the differences between samples were due to the effect of time as observed previously. Indeed, the axis 2 clearly separated samples from day 38 (just before exposure) from those sampled on day 73, after 5 weeks of exposure. The samples exposed to oxyfluorfen were not separated from the controls, indicating that pigment composition was not influenced by oxyfluorfen exposure (Fig. 6). 3.4.2. Eukaryotic community structure Concerning the eukaryotic diversity, 36 bands were detected in the biofilm sample as a whole. In samples from the beginning of the exposure (t = 0 h) the number of bands detected was 16 or 17, whereas after 5 weeks of exposure (t = 73 days), this number ranged from 23 to 27 (average: 25). Biofilm exposed to 150 ␮g L−1 of oxyfluorfen had the lowest number of bands detected (23) within 73 days old samples. The cluster analysis, performed using the presence/absence of bands within each sample, allowed six groups to be separated (Fig. 7A). First, samples collected just before the exposure (t = 0 h) formed a group with very similar eukaryotic community and were clearly different from the samples collected after five weeks of exposure (t = 73 days). Within the chronically-exposed samples, five groups with similar eukaryotic community were

66

C. Bonnineau et al. / Aquatic Toxicology 136–137 (2013) 60–71

Fig. 5. Effects of oxyfluorfen on biofilms throughout time: results of the within-PCA performed on control and exposed biofilms collected at days 38 (t = 6 h), 39, 41, 59, 66, 73. On the left graph, the factorial maps represent the ordination of the samples along the first two axes of the within-PCA. Each exposed sample is represented by a label on which the oxyfluorfen concentration is indicated while the three control samples are represented by an ellipse the centre of which is indicated by the control label (0). Each factorial map correspond to a different sampling day and so to a different time of exposure (both are indicated on the bottom left angle of each map, wks stands for weeks). On the right graph the normed coefficients of the variables on the first two axes of the within-PCA are represented, Ph. eff. stands for photosynthetic efficiency.

Table 3 F and p-value from the two-way ANOVA analysis of biofilms’ AEA as influenced by oxyfluorfen concentration during chronic exposure (i.e. exposure to 0, 3, 7.5, 15, 75 or 150 ␮g L−1 during 5 weeks) and acute exposure (i.e. exposure to 0, 1.5, 15, 75, 150, 1000 ␮g L−1 during 6 h). Source

Chronic exposure to oxyfluorfen Acute exposure to oxyfluorfen Chronic × acute

CAT

APX

GR

F

p-value

F

p-value

F

p-value

7.37 3.05 4.39

<0.05 <0.05 <0.05

5.44 1.32 2.04

<0.05 >0.05 <0.05

7.58 0.67 0.86

<0.05 >0.05 >0.05

Values in bold indicate significant result (p < 0.05).

Fig. 6. Results of the PCA performed on the pigment composition of biofilms collected on day 38 (before exposure) and 73 (after 5 weeks of exposure). The left graph shows the ordination of the samples, the black plain dots correspond to biofilms collected on day 38, just before exposure, while the black circle corresponds to biofilms collected on day 73, after 5 weeks of exposure. Next to each dot the oxyfluorfen concentration to which the sample will be (for the black dots) or has been (for the circle) exposed is indicated. The right graph shows the normed coefficient of the variables on the two first axes of the PCA, the correspondence between pigment codes and complete names is given on Table 2.

C. Bonnineau et al. / Aquatic Toxicology 136–137 (2013) 60–71

67

Fig. 7. Cluster analysis of the biofilms eukaryotic (A) and bacterial (B) community structure before exposure (day 38) and after 5 weeks of exposure (day = 73) to different concentrations of oxyfluorfen. For each clustering, the bar plots of the node heights used to determine the number of groups are shown. t = 0 h indicates biofilms sampled on day 38 just before exposure and t = 5wks indicates biofilms sampled on day 73 after five weeks of exposure. For each sample oxyfluorfen concentration is indicated; 0a, 0b, 0c correspond to control.

observed. The first one contained samples exposed to 150 ␮g L−1 of oxyfluorfen, the second one the samples exposed to 75 ␮g L−1 , a third group was formed by the controls, a fourth one by samples exposed to 3 ␮g L−1 and a last group by samples exposed to 7.5 and 15 ␮g L−1 (Fig. 7A). 3.4.3. Bacterial community structure Concerning the bacterial diversity, 54 bands were detected in the biofilm sample as a whole. In samples from the beginning of the exposure (t = 0 h) the number of bands detected ranged from 25 to 27 (average: 26), whereas after 5 weeks of exposure (t = 73 days) this number ranged from 30 to 35 (average: 32). Biofilms exposed to 150 ␮g L−1 of oxyfluorfen had the lowest number of bands detected (30) within 73 days old samples. The cluster analysis allowed three groups to be separated (Fig. 7B). First, samples from t = 0 h were clearly separated from chronically-exposed samples; then, in this former group, biofilms exposed to 15 and 75 ␮g L−1 of oxyfluorfen were separated from the other ones (Fig. 7B). 3.4.4. Antioxidant enzymes activities After 5 weeks of chronic exposure, biofilms from all channels were exposed during 6 h to a wider range of oxyfluorfen concentrations during an acute exposure test. The two-way ANOVA showed a main effect of the chronic exposure on all the AEA tested and a main effect of the following acute exposure on CAT activity but not on APX and GR. In addition, the interaction term between acute and chronic exposure was significant for CAT and APX activities (Table 3). This last result suggests an influence of chronic exposure on the CAT and APX response of biofilms to further acute exposure to oxyfluorfen. CAT activities of biofilms from the different channels (chronically exposed and controls) not exposed to oxyfluorfen during

the short-term ecotoxicological tests were not significantly different (F = 1.33, p > 0.05) whereas the response patterns throughout the acute oxyfluorfen gradient differed between biofilms chronically exposed to different concentrations of oxyfluorfen (Fig. 8). After acute exposure, CAT activity of control biofilms (not exposed to chronic contamination by oxyfluorfen) showed an unimodal response throughout oxyfluorfen gradient with a maximum of activity reached after exposure to 15 ␮g L−1 . Indeed, CAT activity increased by 61.0 ± 5.8, 33.2 ± 4.5 and 21.5 ± 4.7% in control biofilms exposed at 15 ␮g L−1 , compared to non-exposed communities (Fig. 8). In communities chronically exposed to 3, 7.5 and 15 ␮g L−1 , no differences were found between communities exposed 6 h to oxyfluorfen and non-exposed ones (Fig. 8). In biofilms chronically exposed to 75 ␮g L−1 , CAT activity increased significantly by 27.1 ± 6.2% after acute exposure to 1000 ␮g L−1 . In communities chronically exposed to 150 ␮g L−1 , CAT activity increased significantly by 75.6 ± 4.2, 91.0 ± 17.2, 88.3 ± 8.5 and 129.6 ± 10.5% after acute exposure to 15, 75, 150 and 1000 ␮g L−1 of oxyfluorfen, respectively (Fig. 8). Consequently, after acute exposure to 1000 ␮g L−1 of oxyfluorfen, biofilms chronically exposed to 150 ␮g L−1 presented a CAT activity 2.5 times significantly higher than biofilms not chronically exposed (F = 55.47, p < 0.05). Concerning APX activity, no significant differences were observed throughout the gradient of acute exposure in control biofilms and in communities chronically exposed to 3, 7.5, 75 and 150 ␮g L−1 of oxyfluorfen. Only in biofilms chronically exposed to 15 ␮g L−1 of oxyfluorfen, a significant decrease (by 43.6 ± 7.0%) in APX activity was observed after acute exposure to 1.5 ␮g L−1 of oxyfluorfen (supporting information, Fig. A1). Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.aquatox. 2013.03.009.

68

C. Bonnineau et al. / Aquatic Toxicology 136–137 (2013) 60–71

Fig. 8. Acute toxicity test: CAT activity of biofilms after 6 h of exposure to oxyfluorfen. Each line corresponds to the response of biofilms collected from a same channel. The black square () indicated the response of the 3 channels non-exposed to oxyfluorfen during the 5 weeks while the other symbols indicates the response of biofilms chronically exposed to 3 (), 7.5 (), 15 (), 75 () or 150 () ␮g L−1 of oxyfluorfen. For each type of biofilm (control or chronically exposed), * indicates a CAT activity significantly different from the one of this biofilm non-exposed to oxyfluorfen during the acute test.

4. Discussion 4.1. Temporal variations in unexposed biofilms The use of microcosms to study ecosystems under controlled conditions is common in aquatic ecology (Taub, 1997). The system presented here (recirculating channels) allowed active and complex biofilm communities to be maintained during 10 weeks with a good reproducibility since the factor channel accounted for less than 12% of the variance in unexposed biofilms (Giddings and Eddlemon, 1979). Temporal variations of biological variables of unexposed biofilms revealed both structural and functional differences between the exponential (day 33 to 41) and slow-growth phases (day 59–73). Moreover, a higher variability was found within exponentially growing biofilms than within slow-growing biofilms. Indeed, algal richness in biofilms has been observed to increase quickly during the first days of the colonization before reaching a stable value (Hillebrand and Sommer, 2000; Szabó et al., 2008). Besemer et al. (2007) also observed a decrease in Operational Taxonomic Units turnover (i.e. the appearance of new bacterial species) throughout time during biofilm colonization in a mesocosm experiment. Sabater and Romaní (1996) reported a sharp increase in bacterial densities and ectoenzymes activities in the first 5 days of a 43 days colonization sequence in a shaded stream. Therefore, our results and those reported in the literature suggest that the changes occurring during biofilm development may be faster in exponentially growing biofilms due to the higher growth rate while communities in the slow-growth phase may have reached a steady-state and thus may be more stable. As expected, older biofilms were characterized by a higher biomass (chlorophyll-a and wet weight) than exponentially growing biofilms (Romaní and Sabater, 1999). However, the proportion of proteins in total biofilm biomass decreased throughout time, indicating temporal changes in biofilm composition. As biofilm ages, it becomes thicker and the proportion of molecules different from proteins (e.g. polysaccharides from EPS matrix) may increase (Fernandes da Silva et al., 2008). The shift from the exponential to the slow growth phase was also characterized by changes in biofilm composition. The complexity of the communities increased with time as shown by the higher bacterial and eukaryotic richness at the end of the experiment, in accordance with previous studies (Hillebrand and Sommer, 2000; Lear et al., 2008; Sabater and Romaní, 1996). Temporal variations

in pigment composition during exponential growth suggested a decrease in the proportion of diatoms (fucoxanthin, diadinochrome I, chlorophyll c) and an increase in the proportion of green algae (chlorophyll b, lutein, ␤,␤-carotene, neoxanthin) throughout time. Moreover, slow-growing biofilms were characterized by a higher proportion of cyanobacteria (zeaxanthin). These results indicate a classic succession in biofilms in agreement with previous observations (Biggs, 1996). A temporal shift in the pool of antioxidant enzymes used for oxidative stress management was also observed in biofilms. On the one hand, APX seemed to play a more important role in exponentially than in slow growing biofilms with a significant peak of activity observed at the end of the exponential growth phase. On the other hand, though CAT and GR activities seemed to be independent from time, these enzymes may have a more important role during the slow-growth phase due to the reduction in APX activity. Temporal variations of AEA may be explained by both inter-specific variations and changes in the biofilm micro-environment. In fact, these two aspects are likely to be linked as changes in microenvironment may favour the species with the most appropriate AEA pattern. For instance, communities dominated by diatoms are likely to have a low CAT activity as the occurrence of CAT in diatoms has rarely been observed (Branco et al., 2010; Wilhelm et al., 2006; Winkler and Stabenau, 1995). Besides, as biofilm ages it becomes thicker and marked gradients of light and oxygen can be observed. Therefore, the upper layers of biofilm are characterized by oxygen supersaturation resulting from a high photosynthetic activity while in the bottom layers, where light can be strongly attenuated, photosynthesis and oxygen concentrations are reduced (Carlton and Wetzel, 1987; Dodds, 1989). These changes may create microzones with different levels of oxidative stress. The balance between these zones may then determine the oxidative stress level of the whole community. Since on the first sampling of this study, colonization was already on the late exponential phase, a high oxygen concentration and pH may be expected within biofilms. Therefore the balance between the different zones may point towards a high level of oxidative stress. In this case, the stability of AEA (especially of GR and CAT) may reflect a saturation of the antioxidant defence system. 4.2. Variations due to oxyfluorfen exposure On contrary to previous observations on green algae (Geoffroy et al., 2003; Kunert and Böger, 1981; Sandmann and Böger, 1983),

C. Bonnineau et al. / Aquatic Toxicology 136–137 (2013) 60–71

oxyfluorfen did not provoke any clear decrease in chlorophyll-a concentration nor in photosynthetic efficiency in exposed biofilms at the concentrations tested (3–150 ␮g L−1 ). These lower effects observed on biofilms may be due to the lower diffusion of oxyfluorfen through the protective layer of EPS as observed for other molecules (Franz et al., 2008; Guasch et al., 2003). Further investigation including exposure to higher oxyfluorfen concentrations and determination of oxyfluorfen concentration within the EPS matrix may help elucidating this point. Chronic exposure to 3–150 ␮g L−1 of oxyfluorfen provoked slight variations in biofilms’ AEA. The activation of one or another antioxidant enzyme to cope with oxidative stress induced by oxyfluorfen seemed concentration dependent but not time dependent. Indeed, exposure to low oxyfluorfen concentration (3–15 ␮g L−1 ) led to an increase in APX, whereas exposure to higher concentration (75 and 150 ␮g L−1 ) stimulated CAT activity. This result may be related to the higher affinity of APX for hydrogen peroxide. Hence, APX is expected to be activated at lower levels of ROS than CAT (Mittler, 2002). This result may be specific to biofilm communities as Geoffroy et al. (2002, 2003) observed an activation of CAT at lower oxyfluoren concentrations (7.5–22.5 ␮g L−1 ) than APX (15–22.5 ␮g L−1 ) in Scenedesmus obliquus exposed to this toxicant. Although the set of functional and structural biofilm biomarkers measured throughout exposure was slightly affected by oxyfluorfen, the communities were strongly structured after 5 weeks of exposure, as shown by the clustering of the DGGE profiles. Oxyfluorfen was expected to affect mainly the eukaryotic community due to its direct effect on chlorophyll-a biosynthesis. Indeed, the eukaryotic community was structured in a concentrationdependent manner by oxyfluorfen. These changes could not be attributed to changes in the relative proportion of algal groups, as indicated by the lack of effects on pigments. It suggests that species resistant to oxyfluorfen were likely to be found in all algal groups. In addition, the highest concentration of oxyfluorfen had a negative impact on algae and bacteria, as indicated by the lowest bacterial and eukaryotic richness observed. This result also pointed out the potential indirect effect of oxyfluorfen on the bacterial community in biofilms. In agreement with this idea, Ricart et al. (2009) described how toxic exposure of biofilms to a photosynthesis inhibitor (the herbicide diuron) indirectly affected the bacterial community. In this study, the interaction of bacteria with the potential primary target organisms was the basis of the chain of effects that diuron caused on biofilms. Another study highlighted the importance of the bacterial-algal link in fluvial biofilms by assessing the ecotoxicological effects of the bactericide triclosan on these communities. Though triclosan directly damaged the heterotrophic compartment, the algal component was also affected, and it was attributed to the common use of space and resources within the biofilm (Ricart et al., 2010). These examples and our results emphasize the importance of indirect effects of pollutants on the aquatic system. Indeed, the toxicity of chemicals is not simply a consequence of their direct toxic effect, but it might extend to other trophic levels. The structural changes observed in biofilm communities after chronic exposure to oxyfluorfen were linked to an enhancement of the CAT capacity to answer to higher levels of oxidative stress caused by this same toxicant. Therefore chronically-exposed biofilms may better tolerate oxidative stress at intermediate oxyfluorfen concentrations and may be able to cope with higher levels of oxidative stress than control biofilms. Differences in the ability to reduce oxidative damage may, for instance, be due to various mechanisms, such as an increase in ROS excretion (Choo et al., 2004). These patterns of response of CAT activity showed that oxyfluorfen chronic exposure induced the selection of species with capacity to cope with higher oxidative stress levels, being probably

69

one of the mechanisms of adaptation to oxyfluorfen exposure. This result highlights then the importance of toxicant chronic exposure on the development of biofilm structure and function and thus on the adaptation of the community as demonstrated on several occasions (e.g. Berard et al., 2002; Dorigo et al., 2007; Schmitt-Jansen and Altenburger, 2005; Tlili et al., 2010). Nevertheless further studies are needed, in particular, to demonstrate the link between an increase in CAT capacity to answer to oxidative stress and a better tolerance to peroxidizing compounds. Previous studies reported concentrations of oxyfluorfen between 0.1 and 1 ␮g L−1 (estimated concentrations in San Joaquin River based on sediment data, US EPA/OPP, 2001) with a maximum of 541 ␮g L−1 being observed after an accidental spillage (US EPA/OPP, 2001). Then, based on our result, oxyfluorfen may represent a risk in the environment. In particular chronic exposure to environmental concentrations may provoke changes in community structure as in the present study eukaryotic diversity was affected after chronic exposure to 3 ␮g L−1 of oxyfluorfen. Chronic exposure may thus affect the community by selecting resistant species and provoking a decrease in biofilm biodiversity. Scenarios of chronic contamination by oxyfluorfen may be especially problematic in multiple stress situations. Indeed, species tolerant to other types of stresses (other herbicides, temperature increase, etc.) might disappear due to oxyfluorfen exposure and, therefore, the resistance of the whole community might be lowered (Vinebrooke et al., 2004).

4.3. AEA: temporal variation vs. ecotoxicological effects Knowing the magnitude and typology of changes occurring in non-exposed biofilms throughout the duration of the experiment facilitates the evaluation of the effects caused by toxicant exposure. In the present study, while some extreme values of AEA were detected in exposed biofilms, it has not been possible to determine activity thresholds indicative of adverse effects. This limitation was due to the high variability observed in both non-exposed and exposed communities but also to the unimodal pattern of variation of AEA. Indeed, differences between exposed and non-exposed communities were better highlighted by comparing the pattern of AEA response throughout oxyfluorfen gradient after acute exposure. Therefore, the estimation of the concentration range leading to the maximal short-term response may be especially useful to discriminate between controls and chronically exposed communities.

5. Conclusion Since the pool of AEA dedicated to oxidative stress management changed as biofilm aged, the AEA of biofilms depend strongly on biofilm age. This natural variability prevents direct interpretation of biofilm AEA, especially in the field where biofilm age is rarely known. Nevertheless, AEA brought valuable information on the oxidant effects of oxyfluorfen when the classical biomarker: chlorophyll-a was not found sensitive to this toxicant at the concentrations tested. In particular, the use of AEA in acute ecotoxicological tests revealed that the structural changes observed were linked with functional changes provoked by chronic exposure to oxyfluorfen, that is a higher CAT capacity to respond to higher levels of oxidative stress (induced by oxyfluorfen) in biofilms chronically exposed to the same compound (oxyfluorfen). Then, we proposed the use of AEA in biofilms in short-term toxicity tests to compare the oxidative stress levels undergone by different communities and determine their capacity to cope with oxidative stress.

70

C. Bonnineau et al. / Aquatic Toxicology 136–137 (2013) 60–71

Acknowledgements We thank Berta Bonet for her technical help during sampling and Wim Admiraal for his helpful and constructive comments on the manuscript. Financial support was provided by the Spanish Ministry of Science and Education (FLUVIALMULTISTRESSCTM2009-1411-CO2-01) and the EC Sixth Framework Program (KEYBIOEFFECTS MRTN-CT-2006-035695). References Aebi, H., 1984. Catalase in vitro. Methods in Enzymology 105, 121. Aizawa, H., Brown, H.M., 1999. Metabolism and Degradation of Porphyrin Biosynthesis Inhibitor Herbicides. Springer-Verlag, Berlin, Germany, pp. 347–377. Anderson, M.J., Ellingsen, K.E., McArdle, B.H., 2006. Multivariate dispersion as a measure of beta diversity. Ecology Letters 9, 683–693. Berard, A., Dorigo, U., Humbert, J.F., Leboulanger, C., Seguin, F., 2002. Application of the Pollution-Induced Community Tolerance (PICT) method to algal communities: its values as a diagnostic tool for ecotoxicological risk assessment in the aquatic environment. Annales de Limnologie – International Journal of Limnology 38, 247–261. Besemer, K., Singer, G., Limberger, R., Chlup, A.K., Hochedlinger, G., Hodl, I., Baranyi, C., Battin, T.J., 2007. Biophysical controls on community succession in stream biofilms. Applied and Environmental Microbiology 73, 4966–4974. Biggs, B., 1996. Patterns in benthic algae of streams. In: Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, pp. 31–56. Bonnineau, C., Guasch, H., Proia, L., Ricart, M., Geiszinger, A., Romaní, A.M., Sabater, S., 2010. Fluvial biofilms: a pertinent tool to assess ␤-blockers toxicity. Aquatic Toxicology 96, 225–233. Bonnineau, C., Bonet, B., Corcoll, N., Guasch, H., 2011. Catalase in fluvial biofilms: a comparison between different extraction methods and example of application in a metal-polluted river. Ecotoxicology 20, 293–303. Bonnineau, C., Sague, I., Urrea, G., Guasch, H., 2012. Light history modulates antioxidant and photosynthetic responses of biofilms to both natural (light) and chemical (herbicides) stressors. Ecotoxicology 21, 1208–1224. Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analytical Biochemistry 72, 248–254. Branco, D., Lima, A., Almeida, S.F., Figueira, E., 2010. Sensitivity of biochemical markers to evaluate cadmium stress in the freshwater diatom Nitzschia palea (Kützing) W. Smith. Aquatic Toxicology 99, 109–117. Carlton, R.G., Wetzel, R.G., 1987. Distributions and fates of oxygen in periphyton communities. Botany 65, 1031–1037. Chelikani, P., Fita, I., Loewen, P.C., 2004. Diversity of structures and properties among catalases. Cellular and Molecular Life Sciences 61, 192–208. Choo, K., Snoeijs, P., Pedersén, M., 2004. Oxidative stress tolerance in the filamentous green algae Cladophora glomerata and Enteromorpha ahlneriana. Journal of Experimental Marine Biology and Ecology 298, 111–123. Clements, W.H., Newman, M.C., 2002. Community Ecotoxicology. John Wiley and Sons, West Sussex, UK. Dewez, D., Geoffroy, L., Vernet, G., Popovic, R., 2005. Determination of photosynthetic and enzymatic biomarkers sensitivity used to evaluate toxic effects of copper and fludioxonil in alga Scenedesmus obliquus. Aquatic Toxicology 74, 150–159. Dodds, W.K., 1989. Microscale vertical profiles of N2 fixation, photosynthesis, O2 , chlorophyll a, and light in a cyanobacterial assemblage. Applied and Environmental Microbiology 55, 882. Dolédec, S., Chessel, D., 1987. Rythmes saisonniers et composantes stationnelles en milieu aquatique. I: description d‘un plan d‘observation complet par projection de variables=Seasonal successions ans spatial variables in freshwater environments. I: description of a complete two-way layout by projection of variables. Acta Oecologica – Oecologica Generalis 8, 403–426. Dolédec, S., Chessel, D., 1994. Co-inertia analysis: an alternative method for studying species–environment relationships. Freshwater Biology 31, 277–294. Dorigo, U., Leboulanger, C., Brard, A., Bouchez, A., Humbert, J., Montuelle, B., 2007. Lotic biofilm community structure and pesticide tolerance along a contamination gradient in a vineyard area. Aquatic Microbial Ecology 50, 91–102. Dray, S., Chessel, D., Thioulouse, J., 2003a. Procrustean co-inertia analysis for the linking of multivariate datasets. Ecoscience 10, 110–119. Dray, S., Chessel, D., Thioulouse, J., 2003b. Co-inertia analysis and the linking of ecological data tables. Ecology 84, 3078–3089. Dray, S., Dufour, A.B., 2007. The ade4 package: implementing the duality diagram for ecologists. Journal of Statistical Software 22, 1–20. Duke, S.O., Lydon, J., Becerril, J.M., Sherman, T.D., Lehnen Jr., L.P., Matsumoto, H., 1991. Protoporphyrinogen oxidase-inhibiting herbicides. Weed Science 39, 465–473. EFSA, 2010. Conlusion on the peer review of the pesticide risk assessment of the active substance oxyfluorfen. EFSA Journal 8, 1–78. Fernandes da Silva, C., Ballester, E., Monserrat, J., Geracitano Jr., L., Wasielesky, W., Abreu, P.C., 2008. Contribution of microorganisms to the biofilm nutritional quality: protein and lipid contents. Aquaculture Nutrition 14, 507–514. Franz, S., Altenburger, R., Heilmeier, H., Schmitt-Jansen, M., 2008. What contributes to the sensitivity of microalgae to triclosan? Aquatic Toxicology 90, 102–108.

Genty, B., Briantais, J.M., Baker, N.R., 1989. The relationship between the quantum yield of photosynthetic electron transport and quenching of chlorophyll fluorescence. Biochimica et Biophysica Acta 990, 87–92. Geoffroy, L., Teisseire, H., Couderchet, M., Vernet, G., 2002. Effect of oxyfluorfen and diuron alone and in mixture on antioxidative enzymes of Scenedesmus obliquus. Pesticide Biochemistry and Physiology 72, 178–185. Geoffroy, L., Dewez, D., Vernet, G., Popovic, R., 2003. Oxyfluorfen toxic effect on S. obliquus evaluated by different photosynthetic and enzymatic biomarkers. Archives of Environment Contamination and Toxicology 45, 445–452. Giddings, J.M., Eddlemon, G.K., 1979. Some ecological and experimental properties of complex aquatic microcosms. International Journal of Environmental Studies 13, 119. Guasch, H., Admiraal, W., Sabater, S., 2003. Contrasting effects of organic and inorganic toxicants on freshwater periphyton. Aquatic Toxicology 64, 165–175. Guasch, H., Atli, G., Bonet, B., Corcoll, N., Leira, M., Serra, A., 2010. Discharge and the response of biofilms to metal exposure in Mediterranean rivers. Hydrobiologia 657, 143–157. Heo, M., Gabriel, K.R., 1998. A permutation test of association between configurations by means of the RV coefficient. Communications in Statistics – Simulation and Computation 27, 843–856. Hillebrand, H., Sommer, U., 2000. Diversity of benthic microalgae in response to colonization time and eutrophication. Aquatic Botany 67, 221–236. Hudon, C., Bourget, E., 1981. Initial colonization of artificial substrate: community development and structure studied by scanning electron microscopy. Canadian Journal of Fisheries and Aquatic Sciences 38, 1371–1384. Ihaka, R., Gentleman, R., 1996. R: a language for data analysis and graphics. Journal of Computational and Graphical Statistics 5, 299–314. Jeffrey, S.W., Mantoura, R.F.C., Wright, S.W., 1997. Phytoplankton Pigments in Oceanography. UNESCO, Paris, France. Kunert, K.J., Böger, P., 1981. The bleaching effect of the diphenyl ether oxyfluorfen. Weed Science 29, 169–173. Kunert, K.J., Homrighausen, C., Böhme, H., Böger, P., 1985. Oxyfluorfen and lipid peroxidation: protein damage as a phytotoxic consequence. Weed Science 33, 766–770. Lear, G., Anderson, M.J., Smith, J.P., Boxen, K., Lewis, G.D., 2008. Spatial and temporal heterogeneity of the bacterial communities in stream epilithic biofilms. FEMS Microbiology Ecology 65, 463–473. Lesser, M.P., 2006. Oxidative stress in marine environments: biochemistry and physiological ecology. Annual Review of Physiology 68, 253–278. Meyer, D., Butcha, C., 2010. proxy:Distance and Similarity Measures. R package version 0.4–6. http://CRAN.R-project.org/package=proxy Mittler, R., 2002. Oxidative stress, antioxidants and stress tolerance. Trends in Plant Science 7, 405–410. Nakano, Y., Asada, K., 1981. Hydrogen peroxide is scavenged by ascorbatespecific peroxidase in spinach chloroplasts. Plant and Cell Physiology 22, 867–880. Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., O‘Hara, R.B., Simpson, G.L., Solymos, P., Stevens, M.H.H., Wagner, H., 2010. vegan: Community Ecology Package. R package version 1.17–4. http://CRAN.R-project.org/package=vegan Peterson, C.G., Stevenson, R.J., 1990. Post-spate development of epilithic algal communities in different current environments. Botany 68, 2092–2102. Pinto, E., Sigaud-Kutner, T.C., Leitao, M.A., Okamoto, O.K., Morse, D., Colepicolo, P., 2003. Heavy metal-induced oxidative stress in algae. Journal of Phycology 39, 1008–1018. R Development Core Team, 2008. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, Available at: http://www.R-project.org Ricart, M., Barceló, D., Geiszinger, A., Guasch, H., Alda, M.L., Romaní, A.M., Vidal, G., Villagrasa, M., Sabater, S., 2009. Effects of low concentrations of the phenylurea herbicide diuron on biofilm algae and bacteria. Chemosphere 76, 1392–1401. Ricart, M., Guasch, H., Alberch, M., Barceló, D., Bonnineau, C., Geiszinger, A., Farré, M., Ferrer, J., Ricciardi, F., Romaní, A.M., Morin, S., Proia, L., Sala, L., Sureda, D., Sabater, S., 2010. Triclosan persistence through wastewater treatment plants and its potential toxic effects on river biofilms. Aquatic Toxicology 100, 346–353. Robert, P., Escoufier, Y., 1976. A unifying tool for linear multivariate statistical methods: the RV-coefficient. Applied Statistics – Journal of the Royal Statistical Society Series C 25, 257–265. Romaní, A.M., Sabater, S., 1999. Effect of primary producers on the heterotrophic metabolism of a stream biofilm. Freshwater Biology 41, 729–736. Romaní, A.M., Fund, K., Artigas, J., Schwartz, T., Sabater, S., Obst, U., 2008. Relevance of polymeric matrix enzymes during biofilm formation. Microbial Ecology 56, 427–436. Romaní, A.M., 2010. In: Dürr, S., Thomason, J.C. (Eds.), Freshwater Biofilms, dans: Biofouling. , p. 137. Sabater, S., Romaní, A.M., 1996. Metabolic changes associated with biofilm formation in an undisturbed Mediterranean stream. Hydrobiologia 335, 107–113. Sabater, S., Admiraal, W., 2005. In: Azim, M.E., Verdegem, M.C.J., van A.A. Dam, Beveridge, M.C.M. (Eds.), Periphyton as Biological Indicators in Managed Aquatic Ecosystems, dans: Periphyton: Ecology, Exploitation and Management. , pp. 159–178. Sabater, S., Guasch, H., Ricart, M., Romaní, A., Vidal, G., Klünder, C., Schmitt-Jansen, M., 2007. Monitoring the effect of chemicals on biological communities. The biofilm as an interface. Analytical and Bioanalytical Chemistry 387, 1425–1434. Sandmann, G., Böger, P., 1983. Comparison of the bleaching activity of norflurazon and oxyfluorfen. Weed Science 31, 338–341.

C. Bonnineau et al. / Aquatic Toxicology 136–137 (2013) 60–71 Schaedle, M., Bassham, J.A., 1977. Chloroplast glutathione reductase. Plant Physiology 59, 1011–1012. Schmitt-Jansen, M., Altenburger, R., 2005. Predicting and observing responses of algal communities to photosystem ii-herbicide exposure using pollutioninduced community tolerance and species-sensitivity distributions. Environmental Toxicology and Chemistry 24, 304–312. Serra, A., Corcoll, N., Guasch, H., 2009a. Copper accumulation and toxicity in fluvial periphyton: the influence of exposure history. Chemosphere 74, 633–641. Serra, A., Guasch, H., Martí, E., Geiszinger, A., 2009b. Measuring in-stream retention of copper by means of constant-rate additions. Science of the Total Environment 407, 3847–3854. Sheeba Pratap Singh, V., Kumar Srivastava, P., Mohan Prasad, S., 2011. Differential physiological and biochemical responses of two cyanobacteria Nostoc muscorum and Phormidium foveolarum against oxyfluorfen and UV-B radiation. Ecotoxicology and Environment Safety 74, 1981–1993. Stevenson, R.J., Bothwell, M.L., Lowe, R.L., 1996. Algal Ecology: Freshwater Benthic Ecosystems. Academic Press. Szabó, K.É., Makk, J., Kiss, K.T., Eiler, A., Ács, É., Tóth, B., Kiss, Á.K., Bertilsson, S., 2008. Sequential colonization by river periphyton analysed by microscopy and molecular fingerprinting. Freshwater Biology 53, 1359–1371. Taub, F.B., 1997. Unique information contributed by multispecies systems: examples from the standardized aquatic microcosm. Ecological Applications 7, 1103–1110.

71

Tlili, A., Dorigo, U., Montuelle, B., Margoum, C., Carluer, N., Gouy, V., Bérard, A., Bouchez, A., 2008. Responses of chronically contaminated biofilms to short pulses of diuron An experimental study simulating flooding events in a small river. Aquatic Toxicology 87, 252–263. Tlili, A., Bérard, A., Roulier, J., Volat, B., Montuelle, B., 2010. PO4 3− dependence of the tolerance of autotrophic and heterotrophic biofilm communities to copper and diuron. Aquatic Toxicology 98, 165–177. US EPA/OPP (US Environmental Protection Agency/Office of Pesticide Programs), 2001. Environmental Fate and Effects Division, Science Chapter on Oxyfluorfen, Available at: http://www.epa.gov/pesticides/reregistration/oxyfluorfen/ Vinebrooke, R.D., Cottingham, K.L., Norberg, J., Scheffer, M., Dodson, S.I., Maberly, S.C., Sommer, U., 2004. Impacts of multiple stressors on biodiversity and ecosystem functioning: the role of species co-tolerance. Oikos 104, 451–457. Weitzel, R.L., 1979. Methods and measurements of periphyton communities: a review. ASTM International. Wilhelm, C., Büchel, C., Fisahn, J., Goss, R., Jakob, T., LaRoche, J., Lavaud, J., Lohr, M., Riebesell, U., Stehfest, K., Valentin, K., Kroth, P.G., 2006. The regulation of carbon and nutrient assimilation in diatoms is significantly different from green algae. Protist 157, 91–124. Winkler, U., Stabenau, H., 1995. Isolation and characterization of peroxisomes from diatoms. Planta, 195.