Transcriptional and Behavioral Responses of Zebrafish Larvae to Microcystin-LR Exposure

Microcystins are cyclic heptapeptides that constitute a diverse group of toxins produced by cyanobacteria. One of the most toxic variants of this family is microcystin-LR (MCLR) which is a potent inhibitor of protein phosphatase 2A (PP2A) and induces cytoskeleton alterations. In this study, zebrafish larvae exposed to 500 μg/L of MCLR for four days exhibited a 40% reduction of PP2A activity compared to the controls, indicating early effects of the toxin. Gene expression profiling of the MCLR-exposed larvae using microarray analysis revealed that keratin 96 (krt96) was the most downregulated gene, consistent with the well-documented effects of MCLR on cytoskeleton structure. In addition, our analysis revealed upregulation in all genes encoding for the enzymes of the retinal visual cycle, including rpe65a (retinal pigment epithelium-specific protein 65a), which is critical for the larval vision. Quantitative real-time PCR (qPCR) analysis confirmed the microarray data, showing that rpe65a was significantly upregulated at 50 μg/L and 500 μg/L MCLR in a dose-dependent manner. Consistent with the microarray data, MCLR-treated larvae displayed behavioral alterations such as weakening response to the sudden darkness and hypoactivity in the dark. Our work reveals new molecular targets for MCLR and provides further insights into the molecular mechanisms of MCLR toxicity during early development.


Introduction
Cyanotoxins are secondary metabolites produced by several species of cyanobacteria that exist in freshwater lakes and coastal marine ecosystems due to eutrophication. They are classified into three broad groups according to their chemical structure as cyclic peptides, alkaloids, and lipopolysaccharides [1]. Microcystins are monocyclic heptapeptides produced by several members of cyanobacterial genera including Microcystis, Anabaena, and Planktothrix [2]. To date, more than 90 variants have been identified, and among them microcystin-LR (MCLR) is the most common and toxic member [2,3]. Microcystin-LR is extremely stable under physiological conditions and resistant to chemical hydrolysis or oxidation. A slow degradation occurs either at high temperatures (40 • C) or high/low pH [4]. The toxin can also be degraded by bacterial metalloproteases, called microcystinases, but the bacterial species that produce these enzymes are often absent from surface water [5].
In aquatic environments the concentrations of MCLR vary from traces to 1800 µg/L or even higher during periods of cyanobacterial blooms [6]. Upon cyanobacteria cell death or lysis, MCLR is released into the aquatic environment and the exposure of fish may occur either by oral uptake of toxin-containing cells or via epithelium surfaces that are immersed in the polluted water [7]. Organic anion transporters (OATPs) are specifically required for active uptake of MCLR into hepatocytes and epithelial cells [8]. These transporters are highly expressed in the liver and are also found in the heart, lung, spleen, pancreas, and brain [9].
Regarding fish, MCLR toxicity develops at concentrations in the range of µg/L, depending on the aquatic species, the developmental stage, and the exposure route (intraperitoneal injection, feeding, or immersion) [7,10,11]. During early development, the toxin can disrupt embryonic hatching and growth rate, causing morphological abnormalities such as small head and curved body. It also affects the heart rate and the physiological hepatocyte structure [12]. When fish are exposed to high concentrations of MCLR (>1000 µg/L), hepatocytes die through necrosis, which often leads to tissue disruption and liver failure [12].
Previous studies have proposed that MCLR hepatotoxicity could be due to hyperphosphorylation of actin-associated proteins, which leads to cytoskeleton breakdown and ultimately disrupts the cellular architecture inducing apoptosis in liver [13,14]. Recent evidence suggests that all changes in the phosphorylation status of the cytoskeleton proteins may be attributed to the inhibition of the serine/threonine protein phosphatase 2A (PP2A) by MCLR [15,16]. Protein phosphatase 2A plays essential roles in the control of many biological processes, including cell growth, proliferation, apoptosis, and differentiation, and its activity should be very tightly regulated to maintain the cellular homeostasis [17,18]. Crystallographic experiments revealed that MCLR binds specifically to the active site of the catalytic subunit of PP2A, inhibiting its enzymatic activity with Ki value in the picomolar range [19]. Subsequent studies have shown that alterations of PP2A activity by MCLR is an early event during the toxin exposure and can be considered as one of the main mechanisms of MCLR toxicity [20].
Fish exposed to microcystins have developed detoxification mechanisms to resist the toxin risks. The glutathione pathway is an important biochemical mechanism for the formation of glutathione (GSH) conjugates that increases the water-solubility of MCLR, mediating both its metabolism and elimination [21,22]. Besides the enzymatic activity of PP2A and glutathione transferase, MCLR also affects acetylocholinesterase (AChE), an enzyme that plays key roles in neurotransmitter systems and is widely used as a bioindicator of environmental toxicity [23]. Recently, Zeng et al. have shown that MCLR exposure is accompanied by reactive oxygen species (ROS) production that consequently triggers apoptosis in developing zebrafish embryos in a caspase-dependent manner [24].
Zebrafish has emerged as a prominent organism for investigating vertebrate development and is widely used for environmental and toxicological studies [25,26]. The nervous system of larval fish is suitable for studies of fundamental neuronal pathways because zebrafish have broad homologies to mice and humans with respect to brain patterning, and the structure/function of several neural systems [27]. Recently, a number of neurobehavioral locomotor assays, including learning and memory, have been developed in zebrafish to complement genetic studies [28][29][30][31]. In the present study, we investigated the transcriptional and behavioral responses of zebrafish larvae to MCLR exposure and searched for biochemical pathways affected by the toxin at sublethal concentrations of MCLR.

Alterations of Protein Phosphatase 2A Activity in Zebrafish Larvae Following Exposure to Microcystin-LR
To understand the effects of MCLR on early zebrafish development, three days post-fertilization (dpf) zebrafish larvae were exposed to different concentrations of purified MCLR, ranging from 50-500 µg/L for a period of four days. Morphological observations of the treated larvae showed that the highest concentration of 500 µg/L of MCLR did not cause death or any obvious morphological defects in the exposed animals (not shown). Under these experimental conditions, PP2A activity was markedly reduced in the treated larvae, consistent with the established inhibitory role of MCLR to PP2A [16,19] ( Figure 1A). The reduction of PP2A activity in the MCLR-exposed larvae was calculated to be 42.4% ± 6.9% ( Figure 1B). These results suggest that exposure of three dpf larvae to 500 µg/L of MCLR within four days can be used to assess early changes of gene expression in the treated fish. Thus, it was tempting to speculate if this dose, effective for a metabolic change without overt toxic consequences, can be used to detect early changes of gene expression in treated larvae. (A) PP2A activity was measured in control and larvae exposed to 500 µg/L of MCLR, as described in the Materials and Methods section. Protein phosphatase 2A activity was expressed as nmol of free PO 4 −3 released per min using increasing amounts of total protein (µg). The data are indicated as mean ± SE obtained from three independent experiments (p < 0.05); (B) Relative PP2A specific activity of MCLR group vs. control. Control was set to 100%.

Gene Expression Analysis
To determine the global gene expression profile of larvae treated with 500 µg/L MCLR, we used DNA microarray analysis. Using a threshold of fold change ≥1.5 and p ≤ 0.05, the total number of the differentially expressed transcripts with known gene ID was 53, and among them, four genes were represented by two different transcripts: tyr (tyrosinase), cp (ceruloplasmin), ttna (titin a), lgals1l1 (lectin, galactoside-binding, soluble, 1) ( Table 1). All information concerning the differentially regulated genes and the function of the encoded proteins was obtained from the National Center for Biotechnology Information (NCBI). Our detailed analysis showed that krt96 (keratin 96) and the cytoskeleton-associated ttna and ttnb (titin b) were among the most downregulated transcripts (Table 1), consistent with the well-documented effect of MCLR on cytoskeleton structure and dynamics [13][14][15]32,33]. In addition, most genes of the melanin synthesis pathway such as tyr, dct (dopachrome tautomerase), tyrp1a (tyrosinase-related protein 1a), tyrp1b (tyrosinase-related protein 1b) pmela (premelanosome protein a) and pmelb (premelanosome protein b) were upregulated by 1.55-, 2.13-, 1.51-, 1.70-, 1.72-, and 2.06-fold, respectively (Table 1). These microarray data have been confirmed by quantitative real-time PCR (qPCR) analysis revealing 4.30-fold upregulation of tyr, 5.84-fold of dct, 4.71-fold of pmela, and 11.36-fold of pmelb at 500 µg/L of MCLR exposure (p ≤ 0.05, data not shown). Our data also demonstrated that all the enzymes of the retinoid visual cycle were differentially expressed upon the exposure of three dpf larvae to 500 µg/L of MCLR ( Figure 2A). The visual cycle takes place within the outer segments of the photoreceptors and the retinal pigment epithelium (RPE) and is constituted by a series of reactions that regenerate 11-cis-retinal from all-trans-retinol (outlined in Figure 2B). The critical and apparently rate-limiting step in the retinoid visual cycle is the conversion of the all-trans isomer to its 11-cis form by the retinal pigment epithelium-specific protein 65 kDa (RPE65) [34][35][36].
Our data also demonstrated that all the enzymes of the retinoid visual cycle were differentially expressed upon the exposure of three dpf larvae to 500 μg/L of MCLR ( Figure 2A). The visual cycle takes place within the outer segments of the photoreceptors and the retinal pigment epithelium (RPE) and is constituted by a series of reactions that regenerate 11-cis-retinal from all-trans-retinol (outlined in Figure 2B). The critical and apparently rate-limiting step in the retinoid visual cycle is the conversion of the all-trans isomer to its 11-cis form by the retinal pigment epithelium-specific protein 65 kDa (RPE65) [34][35][36]. Gene expression profiling of MCLR exposed larvae (A) microarray analysis. Differentially expressed genes in zebrafish larvae treated with 500 μg/L of MCLR involved in the retinoid visual cycle, p < 0.05. (B) The retinoid visual cycle. Briefly, absorption of light by visual pigments causes isomerization of 11-cis-retinal to all-trans-retinal, resulting in phototransduction. All-trans-retinal is reduced to all-trans-retinol and transported into the retinal pigment epithelium (RPE). LRAT esterifies all-trans-retinol to all-trans-retinyl esters, which are further processed to 11-cis-retinol by RPE65. RLBP1 removes 11-cis-retinol from the reaction site to speed the isomerization. RDH5 converts 11-cisretinol to 11-cis-retinal. STRA6 receptor mediates all-trans-retinol uptake through the plasma membrane. Red check marks indicate the visual cycle-related genes detected in MCLR-treated larvae by microarray analysis. (C) Validation of microarray data by quantitative real-time PCR (qPCR) analysis. Relative mRNA expression of rpe65a, rlbp1a, rlbp1b, rgrb, lrata, stra6, and inpp5kb (inositol polyphosphate-5-phosphatase Kb) at a four-day exposure of larvae to 50 μg/L and 500 μg/L of MCLR. Three biological replicates were run in triplicate. The mRNA expression levels are expressed in relation to the control which is set to 1. * p < 0.05 and ** p < 0.01. RBP: Retinol binding protein.
To check for early effects of MCLR on gene expression, we analyzed the levels of all the above genes at 50 μg/L and 500 μg/L of MCLR using qPCR analysis, which is a more sensitive and accurate Gene expression profiling of MCLR exposed larvae (A) microarray analysis. Differentially expressed genes in zebrafish larvae treated with 500 µg/L of MCLR involved in the retinoid visual cycle, p < 0.05. (B) The retinoid visual cycle. Briefly, absorption of light by visual pigments causes isomerization of 11-cis-retinal to all-trans-retinal, resulting in phototransduction. All-trans-retinal is reduced to all-trans-retinol and transported into the retinal pigment epithelium (RPE). LRAT esterifies all-trans-retinol to all-trans-retinyl esters, which are further processed to 11-cis-retinol by RPE65. RLBP1 removes 11-cis-retinol from the reaction site to speed the isomerization. RDH5 converts 11-cis-retinol to 11-cis-retinal. STRA6 receptor mediates all-trans-retinol uptake through the plasma membrane. Red check marks indicate the visual cycle-related genes detected in MCLR-treated larvae by microarray analysis. (C) Validation of microarray data by quantitative real-time PCR (qPCR) analysis. Relative mRNA expression of rpe65a, rlbp1a, rlbp1b, rgrb, lrata, stra6, and inpp5kb (inositol polyphosphate-5-phosphatase Kb) at a four-day exposure of larvae to 50 µg/L and 500 µg/L of MCLR. Three biological replicates were run in triplicate. The mRNA expression levels are expressed in relation to the control which is set to 1. * p < 0.05 and ** p < 0.01. RBP: Retinol binding protein.
To check for early effects of MCLR on gene expression, we analyzed the levels of all the above genes at 50 µg/L and 500 µg/L of MCLR using qPCR analysis, which is a more sensitive and accurate method than microarrays ( Figure 2C). In more detail, rpe65a was up-regulated by 2.48-and 10.80-fold, rlbp1a (NM_200705.1) 1.18-fold and 0.82-fold, rlbp1b (NM_205690.2) 2.76-fold and 8.09-fold, lrata (NM_001204131) 2.44-and 2.22-fold, stra6 1.86-and 11.85-fold and inpp5kb by 2.07-fold and 5.57-fold in larvae treated with 50 µg/L and 500 µg/L MCLR, respectively. Therefore, our results show that exposure of zebrafish larvae to MCLR affects the transcription levels of cytoskeleton-associated, melanin synthesis, and visual cycle genes that may in turn affect the physiological responses of the fish.

Behavioral Alterations upon Microcystin-LR Exposure of Zebrafish Larvae
Several studies have shown that MCLR accumulates in the brain [9] and exerts toxicity in the nervous system and impacts functions, including behavioral changes [23,38,39]. Furthermore, our data have shown that MCLR affects the transcriptional levels of all visual cycle genes. Therefore, it was tempting to investigate whether the MCLR exposed fish exhibited altered behavioral responsiveness to dark/light changes [40]. In the light, zebrafish normally have less locomotor activity than in the dark. After a sudden decrease in brightness, larvae respond with a rapid increase in movement [41] and exhibit a startle response when the light is turned back on again, followed by minimal movement in the light [40]. This reaction pattern is related to stress produced by the sudden darkness. To evaluate behavioral alterations in response to MCLR, three dpf larvae were exposed to 100 µg/L and 500 µg/L of MCLR for four days. After applying 10 min dark and 10 min light conditions in order to let larvae habituate to the system, tracking was performed at two rounds of 10 min light/10 min dark phases for 40 min. As shown in Figure 3, the behavior of larvae treated with the toxin was significantly different compared to the control group. As expected, control larvae reacted strongly against light/dark changes. On the contrary, larvae exposed to MCLR displayed a weakening reaction towards light changes, showing a constant hypoactivity during dark phases. This mitigated responsiveness of zebrafish larvae to environmental changes could be related to reduced stress and possibly to motor neuron and/or skeletal muscle functional defects caused by MCLR. .57-fold in larvae treated with 50 μg/L and 500 μg/L MCLR, respectively. Therefore, our results show that exposure of zebrafish larvae to MCLR affects the transcription levels of cytoskeleton-associated, melanin synthesis, and visual cycle genes that may in turn affect the physiological responses of the fish.

Behavioral Alterations upon Microcystin-LR Exposure of Zebrafish Larvae
Several studies have shown that MCLR accumulates in the brain [9] and exerts toxicity in the nervous system and impacts functions, including behavioral changes [23,38,39]. Furthermore, our data have shown that MCLR affects the transcriptional levels of all visual cycle genes. Therefore, it was tempting to investigate whether the MCLR exposed fish exhibited altered behavioral responsiveness to dark/light changes [40]. In the light, zebrafish normally have less locomotor activity than in the dark. After a sudden decrease in brightness, larvae respond with a rapid increase in movement [41] and exhibit a startle response when the light is turned back on again, followed by minimal movement in the light [40]. This reaction pattern is related to stress produced by the sudden darkness. To evaluate behavioral alterations in response to MCLR, three dpf larvae were exposed to 100 μg/L and 500 μg/L of MCLR for four days. After applying 10 min dark and 10 min light conditions in order to let larvae habituate to the system, tracking was performed at two rounds of 10 min light/10 min dark phases for 40 min. As shown in Figure 3, the behavior of larvae treated with the toxin was significantly different compared to the control group. As expected, control larvae reacted strongly against light/dark changes. On the contrary, larvae exposed to MCLR displayed a weakening reaction towards light changes, showing a constant hypoactivity during dark phases. This mitigated responsiveness of zebrafish larvae to environmental changes could be related to reduced stress and possibly to motor neuron and/or skeletal muscle functional defects caused by MCLR.

Discussion
In the present study, we have investigated the early effects of MCLR on zebrafish larvae at the molecular, transcriptional, and behavioral level. The responses of the larvae to sublethal concentrations of the toxin were estimated by measuring the reduction of PP2A activity, in accordance with the established inhibitory role of MCLR to PP2A [16,20]. The inhibition of PP2A activity may be used as a molecular indicator of the effects of MCLR. This suggestion is based on the

Discussion
In the present study, we have investigated the early effects of MCLR on zebrafish larvae at the molecular, transcriptional, and behavioral level. The responses of the larvae to sublethal concentrations of the toxin were estimated by measuring the reduction of PP2A activity, in accordance with the established inhibitory role of MCLR to PP2A [16,20]. The inhibition of PP2A activity may be used as a molecular indicator of the effects of MCLR. This suggestion is based on the fact that, at the molecular level, the determination of PP2A enzymatic activity evaluates more accurately the effects of MCLR that depend on the synergies of different factors; toxin bioaccumulation, exposure period, developmental stage of the organisms, and environmental conditions. Moreover, due to contaminant interference and large variations in recoveries of MCLR from tissue samples, the quantitation of MCLR in tissues by enzyme-linked immunosorbent assay (ELISA)-based strategies is more complicated and less reliable than in water samples (detection limit 0.1 µg/L) [22].
Our gene expression data of larvae exposed to MCLR identified upregulation of genes involved in the melanin synthesis pathway. This effect could be a compensatory mechanism for the treated individuals to circumvent the toxin threat and maintain their physiological status [42]. In addition, MCLR affected genes encoding for the enzymes of the retinoid visual cycle, suggesting molecular targets of MCLR in the visual pathway. Interestingly, most of the genes affected by MCLR are expressed in the RPE. For example, rlbp1b, which is localized in RPE, was upregulated in MCLR-treated larvae, while, rlbp1a, which is expressed only in Müller cells [43], was not among the differentially expressed transcripts in both microarray and qPCR analysis. The preference of MCLR action on RPE is interesting, because the pigment epithelium plays a critical role in the development of the retina and in some species it participates in the regeneration of the retinal tissue [44].
The mRNA changes of the above genes may imply an increase in enzyme production and/or activity. However, it is possible that some enzymes might indeed buffer fluctuations, whereas others might accelerate responses to MCLR. These parameters have not been analyzed in our work due to the lack of commercially available antibodies against the zebrafish proteins and the difficulties to assay for the interconnected isomerization reactions of the visual cycle. Since rpe65 has been shown to be critical for the retinoid visual cycle and normal vision [34,35] and it responds transcriptionally to MCLR (our data), rpe65a transcript levels may predict a toxic outcome and can be used as an early biomarker of MCLR exposure.
A previous microarray analysis by Rogers et al. in which three dpf larvae were exposed to 100, 1000 µg/L of MCLR, and Microcystis aeruginosa extract containing 4.5 µg/L of MCLR for four days, reported MCLR changes on genes related to cell signaling and development, cytoskeleton, immune function, and oxidative stress [45]. A comparison of our microarray results with the above study revealed one gene in common; sepp1b (selenoprotein P, plasma, 1b), which was reported to be downregulated in accordance with our data (Table 1). We have validated sepp1b expression in the 500 µg/L MCLR group by qPCR and found a 0.64-fold change. The observed variations between the two studies could be attributed to differences in MCLR concentrations.
Our findings are novel and raise interesting questions concerning the mechanism of MCLR action. Protein phosphatase 2A activity, which is inhibited by MCLR, is a regulator of signal transduction pathways that coordinate the visual cycle [46]. Therefore, it is reasonable to hypothesize that upon entering the RPE cells, MCLR, through inhibition of PP2A, might affect cell signaling that orchestrates the recycling of 11-cis-retinal. In agreement with this hypothesis, a recent proteomic study by Li et al. has shown that MCLR exposure of zebrafish alters the expression of β-crystallins which are present in the lens and are involved in signal transduction [47].
Accumulated evidence demonstrates that MCLR can induce neurotoxicity in nematode, fish, birds, and mammals, and may result in behavioral changes [38,39]. However, the effects of MCLR on the behavior of fish are still largely unknown. A previous study by Baganz et al. recorded altered swimming capacity of adult zebrafish upon long-term treatment with MCLR [48]. Similar results have been obtained from adult zebrafish that were exposed to a cell culture of the microcystin-producing cyanobacterium M. aeruginosa in comparable experimental settings [49]. In our study, we have used the visual locomotor assay that tracks the locomotor responses to light onset and offset and relies on the visual pathway integrity, the nervous system development, and locomotor system development [50].
This behavioral analysis showed that MCLR-treated larvae displayed a weakening response upon rapid loss of light and reduced activity in the dark. It is known that sudden darkness produces a direct stress response to the zebrafish larvae that leads to an increased locomotor activity [40]. Based on our findings, one can speculate that the reduction of the physical activity of larvae exposed to MCLR after the sudden darkness could be due to reduced stress. Additionally, MCLR-treated larvae may display motor neuron and/or skeletal muscle functional defects. In support of this hypothesis, our microarray analysis showed that the muscle-associated ttna and ttnb were among the most downregulated transcripts (0.63-fold each). Furthermore, our experiments revealed that tubulin α-1b (tuba1b) transcript levels were significantly increased to 3.09-fold in the 50 µg/L MCLR group and to 7.78-fold in the 500 µg/L MCLR group (data not shown). Expression of tubulinα-1b is induced in the developing and regenerating central nervous system (CNS) of vertebrates and is considered to be a regulator of motor neuron synapse and axon function [51]. Future experiments using transgenic lines showing motoneuronal and/or muscular cell types could reveal in a more effective way possible effects of MCLR in these anatomical compartments.
It has been established that behavioral profiling can characterize large classes of compounds and reveal information concerning their biological targets [52]. Moreover, the whole-animal behavioral responses can be measured effectively and characterize the sublethal exposure of fish to different molecules that is not accompanied by any obvious pathological effects [53]. Recently, the visual motor response (VMR) has been used for early safety assessment of oculotoxic drugs [54]. Therefore, our behavioral assay may be used to predict MCLR effects and to assess the toxicity risks.
In summary, in the present work we have (i) quantified the effect of MCLR on PP2A activity; (ii) identified the gene expression signature of the treated individuals using DNA microarrays; (iii) validated by qPCR analysis the gene expression changes at concentrations below the toxicity threshold; and (iv) complemented our toxicogenomic analysis with a high-throughput behavioral assay. Our findings provide new insights into the molecular mechanisms of microcystin-LR toxicity during early development.

Zebrafish Maintenance
Zebrafish maintenance, spawning, and embryo collection were performed under standard conditions. Briefly, zebrafish were cultivated in a closed flow-through system with charcoal-filtered tap water at 28 ± 0.5 • C in a 14:10 h light:dark photoperiod and were fed with commercial dry food distributed manually twice a day. Zebrafish were staged according to Kimmel et al. [55] and maintained in accordance with the European Directive 2010/63 for the protection of animals used for scientific purposes and the Recommended Guidelines for Zebrafish Husbandry Conditions. Animals were anaesthetized in 0.016% Tricaine MS-222, before sacrifice. The experimental protocols described in this study were carried out with zebrafish larvae up to seven dpf and were approved by the Foundation for Research and Technology-Hellas (FORTH) Ethics Committee (FEC) (Project Code 80821/2012).

Exposure of Zebrafish Larvae to Microcystin-LR
For exposure of zebrafish larvae to MCLR, 1 mg of pure MCLR (Enzo Life Sciences (New York, NY, USA) was dissolved in 100% ethanol and the MCLR/ethanol solution was added to fish water in different concentrations (50 µg/L and 500 µg/L). The final concentration of ethanol in fish water was ≤0.05% to avoid toxic effects. A control group was exposed to 0.05% ethanol. At three dpf, 55 larvae were exposed to 50 µg/L and 500 µg/L MCLR for four days in 250-mL glass beakers containing 100 mL of MCLR solution. These developmental stages were selected because at three dpf both the hatching period and the morphogenesis of the larvae are complete and by seven dpf, the free swimming larvae have a well-developed nervous system and the ability to perceive cues from the aquatic environment. Each concentration was studied using three independent biological replicates.

Protein Phosphatase 2A Activity Assay
Seven dpf control and treated larvae (40 larvae per tube) were placed on ice and washed once with 50 mM Tris-HCl buffer (pH 7.5) and once with lysis buffer containing 50 mM Tris-HCl, 0.05% Triton X-100, 10% (v/v) glycerol, 0.5 mM phenylmethylsulfonyl fluoride (PMSF), 0.1 mM ethylenediaminetetraacetic acid (EDTA), and 0.02% (v/v) β-mercaptethanol. Then, larvae were incubated in lysis buffer supplemented with protease inhibitors cocktail (Roche) and 1 mM Na 2 VO 4 (phosphatase inhibitor) for 40 min and homogenized using a 25-gauge syringe. Homogenates were centrifuged at 10,000× g for 3 min at 4 • C and the supernatants were collected. Whole-larvae lysates were passed through Sephadex G-50 columns (Pharmacia Biotech, Piscataway, NJ, USA) to remove free phosphate. The total protein concentration (before and after the column step) was determined by the Bradford method, using bovine serum albumin (BSA) as standard.
To measure PP2A activity, different amounts of larvae lysates were incubated with the synthetic phospho-peptide substrate, RRApTVA (Enzo Life Sciences), which is specific for PP2A, for 30 min at 37 • C. The amount of released free phosphate was determined by measuring the absorbance of the molybdate-malachite green-phosphate complex (Sigma, St. Louis, MO, USA) at 630 nm. PP2A specific activity was expressed as nanomoles of free PO 4 −3 released per min per µg of total protein. Three independent experiments were performed.

RNA Preparation and cDNA Synthesis
Total RNA was extracted from control and larvae treated with 50 µg/L and 500 µg/L of MCLR (three biological replicates) using the RNeasy Mini Kit with DNase (Qiagen, Hilden, Germany) treatment to remove genomic DNA contamination. Quantification and quality analysis of RNA was performed via a Nanodrop spectrophotometer (Thermo Scientific Waltham, MA, USA) and confirmed by the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). cDNA was synthesized using the SuperScript III First-Strand Synthesis SuperMix for qPCR (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions.

Microarray Analysis
Affymetrix platform transcriptional microarray analysis was performed on RNA samples obtained from control and larvae treated with 500 µg/L of MCLR (three biological replicates from each group) using the Affymetrix GeneChip Zebrafish Gene 1.0 Arrays (Affymetrix, Santa Clara, CA, USA). RNA samples were processed and labeled for array hybridization using the Ambion WT Expression kit (Life Technologies, Carlsbad, CA, USA). Labeled, fragmented cDNA (Affymetrix GeneChip WT Terminal Labeling and Controls Kit) was hybridized to Zebrafish Gene 1.0 arrays (Affymetrix GeneChip Hybridization, Wash, and Stain Kit) for 16 h at 45 • C at 60 rpm using a GeneChip Hybridization Oven 640 (Affymetrix). Arrays were washed and stained using the Affymetrix Fluidics Station 450, and scanned using the Hewlett-Packard GeneArray Scanner 3000 7G (Palo Alto, CA, USA). Quality of array data was assessed using Expression Console (v 1.3) software (Affymetrix) prior to importing and analyzing the data in GeneSpring 13.0 (Agilent Technologies). Data was analyzed with settings for Affymetrix Expression experiment type, Exon analysis type, and normalization using the RMA16 summarization algorithm. Differentially expressed genes were determined after filtering the data sets for background level of fluorescence (as determined by the Expression Console). Differentially expressed transcripts were determined by comparing normalized average intensities (log2 GC-RMA signal intensity values) of biological triplicates of MCLR-treated larvae to control sample groups and filtering for statistically significant genes using an unpaired Student's t-test and the cut-off value of p < 0.05, and fold change of 1.5. Microarray data have been deposited to the Gene Expression Omnibus (GEO) database under the accession number (GSE73739).

Quantitative Real-Time PCR
Quantitative real time PCR was performed on a Lightcycler capillary system 2.0 (Roche, Indianapolis, IN, USA) with SYBRgreen fluorescent label (PrimerDesign, Chandler's Ford, UK). cDNA samples were synthesized as described in Section 4.4. The 10 µL qPCR reaction mix contained 5 µL of 2× Precision SYBRgreen mastermix (PrimerDesign), 0.25 µL of 10 µM of each PCR primer, 2 µL of diluted cDNA template (12-fold), and 2.5 µL of RNase-free water. The qPCR parameters consisted of initial denaturation at 94 • C for 15 min followed by 45 cycles at 94 • C for 15 s, 60 • C for 20 s, and 72 • C for 10 s. All reactions were performed in triplicates. To ensure that only one PCR product was amplified and that stock solutions were not contaminated, a non-template control was included and melting curve analysis was performed at the end of the amplification phase. Standard curves were constructed for each gene using serial dilutions of stock cDNA. The primers of the target genes were designed based on Danio rerio gene sequences found in the NCBI. Gene names, accession numbers, and forward and reverse primer sequences are shown in Table 2. Gene expression levels were quantified relative to the expression of reference genes gap43 and rpl13 (geNorm reference kit, PrimerDesign) using the qbase PLUS analysis program (Biogazelle, Gent, Belgium). This software uses a modified classic ∆∆C t (cycle threshold) method to take multiple reference genes and gene specific amplification efficiencies into account.
For the statistics of qPCR analysis, we have used the one-way ANOVA F-test (when equality of variances is assumed) or the Welch test (when unequal variances is assumed), followed by an appropriate post-hoc test (Tukey test in the case of equal variances and Dunnett's test for the case of unequal variances). The hypothesis of equality of variances was checked by Levene's test for equality of variances.

Behavior Assay
Three dpf larvae were exposed to 100 µg/L and 500 µg/L of MCLR, as well as to 0.05% ethanol as control, for four days in 250 mL-glass beakers. At seven dpf, larvae were placed in a 96-well plate, one larva per well, 30 larvae per group. The plate was introduced in the DanioVision automated tracking system powered by EthoVision (Noldus, Wageningen, The Netherlands). This system allows tracking locomotor activity of larvae under controlled temperature and lighting conditions. Temperature was set at 28.5 • C and the lighting was variable during tracking period to analyze larvae response to external stimuli. After applying 10 min dark and 10 min light conditions in order to let larvae habituate to the system, tracking was performed at two rounds of 10 min light/10 min dark phases. Total duration of the tracking was 40 min. The mean of the total distance covered by larvae in each group was measured in 2 min time bins. Statistical analysis was done by two-factor mixed-design ANOVA for comparing the mean differences between groups taking into account time ("within-subjects" factor) and MCLR concentration ("between-subjects") factor.