Daphnia magna Exudates Impact Physiological and Metabolic Changes in Microcystis aeruginosa

While the intracellular function of many toxic and bioactive cyanobacterial metabolites is not yet known, microcystins have been suggested to have a protective role in the cyanobacterial metabolism, giving advantage to toxic over nontoxic strains under stress conditions. The zooplankton grazer Daphnia reduce cyanobacterial dominance until a certain density, which may be supported by Daphnia exudates, affecting the cyanobacterial physiological state and metabolites’ production. Therefore, we hypothesized that D. magna spent medium will impact the production of cyanobacterial bioactive metabolites and affect cyanobacterial photosynthetic activity in the nontoxic, but not the toxic strain. Microcystin (MC-LR and des-MC-LR) producing M. aeruginosa PCC7806 and its non-microcystin producing mutant were exposed to spent media of different D. magna densities and culture durations. D. magna spent medium of the highest density (200/L) cultivated for the shortest time (24 h) provoked the strongest effect. D.magna spent medium negatively impacted the photosynthetic activity of M. aeruginosa PCC7806, as well as the dynamics of intracellular and extracellular cyanobacterial metabolites, while its mutant was unaffected. In the presence of Daphnia medium, microcystin does not appear to have a protective role for the strain. On the contrary, extracellular cyanopeptolin A increased in M. aeruginosa PCC7806 although the potential anti-grazing role of this compound would require further studies.


Introduction
Cyanobacterial blooms can occur in freshwater ecosystems due to eutrophication, often producing toxic and bioactive secondary metabolites that can affect the life traits of all organisms of the ecosystem [1]. At the same time, zooplankton graze on phytoplankton, including cyanobacteria, and as such it has been suggested that Daphnia are able to reduce cyanobacterial dominance in aquatic ecosystems up to a certain density and toxicity [2][3][4]. Cyanobacteria can impair planktivorous zooplankton by several mechanisms including the action of toxic metabolites or compounds inhibiting digestive enzymes, their nutritional inadequacy, or simply the mechanical interference during feeding [5,6]. Consequences are reduced growth, survival, and reproduction [7][8][9]. Microcystis is a widespread genus of freshwater cyanobacteria, that produces microcystins (MCs), the most commonly investigated cyanotoxins, with

Photosynthetic Activity
The photosynthetic activity of the strains was measured through the maximum electron transport rate ETRmax. When exposed to spent medium from 75 all age individuals of D. magna /L raised for two weeks (low density DM (2w)), MC-was not negatively affected, on the contrary, it remained in good physiological state during seven days of exposure ( Figure 1A, Table 1). However, in the MC+ strain the photosynthetic activity showed a non-statistically significant decrease after being initially significantly higher during the first three days of exposure and did not recover compared to the control ( Figure 1C, Table 1). Then, both strains were exposed to a medium in which a higher density of Daphnia (200 all age individuals of D. magna /L) was cultured for 24 h (high density DM (24 h)). This exposure had contrasting effects on the strains: The photosynthetic activity of the MC-was high during the entire experiment in the exposed cells as well as in the control ( Figure 1B, Table 1), whereas the MC+ strain suffered from a statistically significant decrease of photosynthetic activity during the whole experiment compared to the control ( Figure 1D, Table 1). This study investigates whether the presence of D. magna infochemicals (via spent medium (DM)) affects secondary metabolites production, the physiological state in the M. aeruginosa PCC7806 (MC+), and its mutant M. aeruginosa PCC7806 mcy-(MC-). We hypothesize that the presence of D. magna infochemicals will: (a) Induce the production of microcystins and other cyanobacterial bioactive compounds in both strains and (b) have negative impact on photosynthetic activity in M. aeruginosa PCC7806 mutant, but not M. aeruginosa PCC7806.

Photosynthetic Activity
The photosynthetic activity of the strains was measured through the maximum electron transport rate ETRmax. When exposed to spent medium from 75 all age individuals of D. magna /L raised for two weeks (low density DM (2w)), MC-was not negatively affected, on the contrary, it remained in good physiological state during seven days of exposure ( Figure 1A, Table 1). However, in the MC+ strain the photosynthetic activity showed a non-statistically significant decrease after being initially significantly higher during the first three days of exposure and did not recover compared to the control ( Figure 1C, Table 1). Then, both strains were exposed to a medium in which a higher density of Daphnia (200 all age individuals of D. magna /L) was cultured for 24 h (high density DM (24 h)). This exposure had contrasting effects on the strains: The photosynthetic activity of the MC-was high during the entire experiment in the exposed cells as well as in the control ( Figure 1B, Table 1), whereas the MC+ strain suffered from a statistically significant decrease of photosynthetic activity during the whole experiment compared to the control ( Figure 1D, Table 1). Photosynthetic activity was also compared between the MC+ and MC-strains in those exposures showing differences between treatment and control (low density DM (2w) and high density DM (24 h), Table 1). In the low density DM (2w) exposure, control MC+ and control MC-showed no significant difference during the experiment, except on the first day. Photosynthetic activity of the Photosynthetic activity was also compared between the MC+ and MC-strains in those exposures showing differences between treatment and control (low density DM (2w) and high density DM (24 h), Table 1). In the low density DM (2w) exposure, control MC+ and control MC-showed no significant difference during the experiment, except on the first day. Photosynthetic activity of the low density DM (2w) treatments was not significantly different between the strains during the first three days of the experiment, but it significantly decreased in MC+ on day 5 and 7 compared MC-exposed to low density DM (2w) ( Table 1). In the high density DM (24 h) exposure, there was no statistically significant difference between control MC+ and control MC-. However, photosynthetic activity in MC+ exposed to high density DM (24 h) significantly decreased during the whole experiment (except on day 5, owing to a high variability in MC-) compared to MC-exposed to high density DM (24 h) ( Table 1).   Table 2). The dynamics of extracellular MC-LR and des-MC-LR were not significantly different between the control and treated cells during the experiment in MC+ ( Figure 3C,D, Table 2). Intracellular AC D and CP A produced by MC+ followed a decreasing trend when exposed to low density DM (2w), with significant difference on day 3 in comparison with the control ( Figure 4A,B, Table 2). While the export of AC D produced by MC+ did not change significantly in the treatment compared with the control ( Figure 4D, Table 2), CP A was excreted in significantly higher amounts in the presence of low density DM (2w) in MC+ on day 7 of the experiment in comparison with the control ( Figure 4C, Table 2). Neither the intracellular nor extracellular dynamics of CP A and AC D produced in MC-were statistically different when exposed to low density DM (2w) compared to the control ( Figure 4A-D, Table 2).

Dynamics of Intracellular and Extracellular Metabolites in MC+ and MC-When Exposed to Low Density DM (2w)
Low density DM (2w) significantly decreased the concentration of intracellular MC-LR and des-MC-LR at the second day of exposure in comparison with the control ( Figure 3A,B, Table 2). The dynamics of extracellular MC-LR and des-MC-LR were not significantly different between the control and treated cells during the experiment in MC+ ( Figure 3C,D, Table 2). Intracellular AC D and CP A produced by MC+ followed a decreasing trend when exposed to low density DM (2w), with significant difference on day 3 in comparison with the control ( Figure 4A,B, Table 2). While the export of AC D produced by MC+ did not change significantly in the treatment compared with the control ( Figure 4D, Table 2), CP A was excreted in significantly higher amounts in the presence of low density DM (2w) in MC+ on day 7 of the experiment in comparison with the control ( Figure 4C, Table 2). Neither the intracellular nor extracellular dynamics of CP A and AC D produced in MC-were statistically different when exposed to low density DM (2w) compared to the control ( Figure  4A,B,C,D, Table 2).  ( Figure 4D, Table 2), CP A was excreted in significantly higher amounts in the presence of low density DM (2w) in MC+ on day 7 of the experiment in comparison with the control ( Figure 4C, Table 2). Neither the intracellular nor extracellular dynamics of CP A and AC D produced in MC-were statistically different when exposed to low density DM (2w) compared to the control ( Figure  4A,B,C,D, Table 2).  The concentration of intracellular and extracellular CP A and AC D was compared between MC+ and MC-in the low density DM (2w) exposure experiment ( Table 2). The intracellular concentration of CP A and AC D was significantly higher in the control MC-when compared with the control MC+ on day 3 and 7. The concentration of intracellular CP A was significantly higher in low density DM (2w) MC-compared to low density DM (2w) MC+ on day 3, but not on day 0 and 7. There was no significant difference in AC D intracellular concentration in low density DM (2w) MC+ compared to low density DM (2w) MC-during the whole experiment. The concentration of extracellular CP A significantly increased in control MC-compared to control MC+ on day 3 and 7. The amount of extracellular AC D was significantly higher on day 0 and 3 in control MC-compared to control MC+, while there was no difference on day 7. When MC+ and MC-were exposed to low density DM (2w), the concentration of extracellular CP A was significantly lower in low density DM (2w) MC+ compared to low density DM (2w) MC-on day 0 and 3 of the experiment, while there was no statistical significance on day 7. The concentration of extracellular AC D was significantly lower in low density DM (2w) MC+ compared with low density DM (2w) MC-on day 3, while on day 0 and 7 there was no statistically significant difference ( Table 2).

Dynamics of Intracellular and Extracellular Metabolites in MC+ When Exposed to High Density DM (24 h)
When exposed to high density DM (24 h), the concentrations of intracellular MC-LR and des-MC-LR by MC+ were reduced at three days of exposure, compared with the control, and did not recover during the experiment ( Figure 5A,B, Table 3). Reduction of intracellular metabolites was followed by a non-statistically significant lower excretion of MC-LR and significantly lower excretion of des-MC-LR in treatment compared to the control on day 7 ( Figure 5C,D, Table 3). CP A and AC D were detected in both strains in control and 24 h DM treatment, however they were below the limit of quantification. Hence, unfortunately, no results on the CP A and AC D concentrations are available for this exposure. of des-MC-LR in treatment compared to the control on day 7 ( Figure 5C,D, Table 3). CP A and AC D were detected in both strains in control and 24 h DM treatment, however they were below the limit of quantification. Hence, unfortunately, no results on the CP A and AC D concentrations are available for this exposure.

Discussion
MCs are by far the most commonly investigated toxins in cyanobacteria [57], but their role and regulation still remain unclear. While abiotic factors have been extensively studied, biotic factors including responses to grazing have received less attention. In this study, we compared the responses of both a MC-producing strain and its nontoxic mutant to Daphnia magna exudates via spent medium exposures in order to test the potential role of MC and/or other cyanopeptides as a response to the presence of grazers.
It is believed that most cyanobacterial species retained this pathway due to the potential protective role of MCs against oxidative stress, favoring the proliferation of MC-producing strains over nontoxic ones under stressful conditions [33]. Under growth limiting conditions of light, temperature, and nitrogen, MC-producing Planktothrix agardhii strains showed better fitness than nontoxic strains [30]. Similarly under growth limiting conditions of light and nitrogen, MCproducing M. aeruginosa showed better fitness than nontoxic strains [58]. A proteomic study showed

Discussion
MCs are by far the most commonly investigated toxins in cyanobacteria [57], but their role and regulation still remain unclear. While abiotic factors have been extensively studied, biotic factors including responses to grazing have received less attention. In this study, we compared the responses of both a MC-producing strain and its nontoxic mutant to Daphnia magna exudates via spent medium exposures in order to test the potential role of MC and/or other cyanopeptides as a response to the presence of grazers.
It is believed that most cyanobacterial species retained this pathway due to the potential protective role of MCs against oxidative stress, favoring the proliferation of MC-producing strains over nontoxic ones under stressful conditions [33]. Under growth limiting conditions of light, temperature, and nitrogen, MC-producing Planktothrix agardhii strains showed better fitness than nontoxic strains [30]. Similarly under growth limiting conditions of light and nitrogen, MC-producing M. aeruginosa showed better fitness than nontoxic strains [58]. A proteomic study showed that MC-producing M. aeruginosa species coped better with environmental changes that induced production of ROS than the nontoxic strain, hypothesizing possible functions of MC as a scavenger of free radicals [21]. Zilliges et al. (2011) revealed that if MC binds to proteins, it increases their stability, suggesting an important protective role of MC against stress [22]. The possibility of stress in cyanobacteria induced by grazers, such as Daphnia sp, has gained scientific interest. Most of the studies looked at the indirect effect of Daphnia on cyanobacteria used media in which Daphnia were cultivated for short periods of time, usually 24 h-48 h [25,26,38,46,59]. Effects of aged zooplankton medium on cyanobacteria are scarce in the literature, although the culturing of Daphnia for a longer duration, allowing the accumulation of D. magna metabolites over time, would be environmentally more relevant. As the spent medium of low density DM (1w) seemed to have no effect on both strains in this study, D. magna were cultivated for one more week in order to accumulate more metabolites that would potentially evoke an effect on cyanobacterial cell physiology. The photosynthetic activity (as ETRmax) is often measured as proxy of the physiological state of algal cells, as in this study. That parameter as a precursor of growth [58,60], indicates a poor physiological state from which the strain does not recover when ETRmax values are below 30, while values reaching 100 and above are associated with a healthy physiological state [58]. Contrary to our expectations, the toxic strain decreased its photosynthetic activity (ETRmax < 30), coping poorly with stress caused by low density DM (2w), while the mutant strain stayed in good physiological state during the entire experiment (seven days). These findings are supported by a recent study using the same strains, also discovering higher sensitivity of the toxic strain, when comparing responses of both strains to prolonged oxidative stress caused by hydrogen peroxide: Only the mutant was able to recover and regain initial photosynthetic yields [34]. As cyanobacterial reactions could be strain specific, different strain's reactions are important to consider: When MC-producing M. aeruginosa strains Ch10 and UTEX LB2385 were exposed to five-day old D.magna spent medium, the content of chlorophyll also decreased, similar to our results [47].
In environments that promote oxidative stress MC binds to proteins [22,61], amongst others to the same target protein as thioredoxin [62] and possibly peroxiredoxin [34]. Binding of MC to these proteins will render them inactive, and upon loosing these antioxidant activities, the MC producing cells become more sensitive to oxidative stress [34]. Furthermore, the higher expression of genes regulating anti-oxidative enzymes in the nontoxic mutant, may favor the nontoxic mutant over the toxic strain when it comes to coping with high levels of oxidative stress [34].
Initial cyanobacterial density and concentration of Daphnia infochemicals influence the interaction between Daphnia and cyanobacteria [63]. In order to address this and avoid possible bias by a bacterial degradation of Daphnia metabolites, a culture of lower cyanobacterial cell density was exposed to high density DM (24 h). This experiment confirmed tendencies observed in the previous exposures by a drastically decreased photosynthetic activity in the toxic strain (ETRmax < 30), while the mutant stayed in good physiological state during the whole experiment. This suggests that the nontoxic mutant was coping better, compared with the toxic strain, when exposed to high density D. magna spent medium.
Depending on the density of D. magna and cultivation time, concentrations of intra-and extracellular MC-LR and des-MC-LR differed in the exposures: Whereas lower D. magna densities caused increasing tendencies of the toxins, higher D. magna densities and in particular those with short time for bacterial degradation of infochemicals caused a decrease. Similarly, M. aeruginosa increased MC production when exposed to increasing D. magna spent medium concentrations, suggesting an anti-grazer role triggered by zooplankton infochemicals [23], when only 50 adult D. magna were grown for four days per L of dechlorinated water of which 10-50% were mixed with the cyanobacterial culture. In addition, the presence of large cladocerans such as D. cucullata and D. longispina in the Sulejow Reservoir (Poland) positively correlated with the increase of MCs content in Microcystis [35]. Similarly, in the presence of D. pulex, the production of MC was significantly higher in Microcystis spp in all tested cyanobacterial (0.5-4.5 × 10 6 cells/mL) and D. pulex (100-500 individuals/L) densities [27]. Different developmental stages of Daphnia can promote different production of MCs [25,35] because the same number of adults can possibly produce more, or other infochemicals than neonates, inducing higher MC production [43]. Although these studies support the hypothesis that a possible role of MC could be as anti-grazer defense [25,27,35], molecular evidence clearly shows that genes responsible for toxin production are much older than the existence of more complex organisms [40]. Furthermore, Kurmayer et al. 2016 [64] pointed out the presence and global distribution of both toxic and nontoxic cyanobacterial genotypes, suggesting their co-existence through evolution, as none of them were strongly favored by natural selection. Therefore, the potential role of cyanobacterial secondary metabolites as a defense mechanism against grazers, might be an adaptation they acquired over time [39,41]. However, not all studies support these results (e.g., [38,45]. Van Gremberghe et al. (2009) reported that Daphnia infochemicals (300 D. magna individuals/L grown for 24 h in the WC medium) generally had a weak influence on MC production in several Microcystis strains [38], while Becker (2010) demonstrated that when exposed to Daphnia medium (165 D. magna individuals/L grown in 24 h WC medium), Microcystis stopped producing MC [45]. In exposure to 2wDM our results showed similar findings: The concentration of intracellular MC-LR and des-MC-LR decreased significantly in MC+ compared with the control. Supposedly, in the high density DM (24 h) exposure, the concentration of both MC-LR and des-MC-LR decreased, owing to a lower cyanobacterial density being exposed to a higher number of D. magna producing more metabolites, which were not degraded during the short cultivation time. However, MC decrease could be due to its potential ability to bind to proteins (such as thioredoxin [62] and peroxiredoxin [34]) under stressful conditions [34], thus escaping detection. Further studies including either the expression of the genes involved in MC production or the detection of the protein-bound MC fraction are needed to clarify this.
Hence, an overall decrease in the cyanobacterial physiology and the negative effect this Daphnia medium had on the toxic strain was demonstrated.
To note, all our controls exhibited similar trends in dynamics where intracellular MC-LR increased on the third day and then decreased in the seven-day experiment. Similar results were observed by Becker 2010 [45], suggesting that this performance might not be caused by zooplankton. The transcription of mcyHIJ and mcyB genes increased in M. aeruginosa in 50% direct grazing treatments and in 90% of Daphnia infochemicals treatments after 24 h exposure, while transcription of mcyACDEFG was only expressed in 8% of the exposures [59]. The synthesis of MC, however, requires the expression of the whole MC synthetase cassette (mcyABCDEFG), therefore in that study zooplankton presence did not trigger MC production, however, it is yet to be clarified whether genes for these secondary metabolites synthesis need more time to respond than has been considered in this study [59]. mcy A was expressed in the first five days of D. magna exposure in the M. aeruginosa UTEX LB2385 strain, while in the Ch10 strain this occurred after 10 days, demonstrating a different response between strains [47]. As gene expression after stress vary in time depending on the particular gene and the strain, and moreover may or may not lead to MC production, we favored the analysis of the metabolite by chemical analysis as applied in this study over gene expression by RT qPCR. Additionally, it has been documented that Microcystis response to Daphnia medium is strongly strain specific [38], leaving the role of MC as anti-grazer compound arguable [28].
In addition to MC, Microcystis produces a range of peptides, such as cyanopeptolins, aeruginosins, and aerucyclamides that negatively influence digestive enzymes in zooplankton [56]. Metabolic pathways and triggers for production of these group of metabolites, as well as their functional role still needs more research [65]. Production of MC is energetically costly [66], allowing strains lacking MC to allocate energy in possible alternative mechanisms of defense or production of other secondary metabolites [34,58]. Our results show that the overall concentrations of CP A and AC D was higher in MC-then in MC+ in controls as well as the treatments. Interestingly, in monocultures, higher concentrations of cyanopeptolins, aerucyclamides, and aeruginosins were produced by the non-MC-producing mutant in comparison with the wild type [36]. When exposed to 1 wDM, concentrations of CP A and AC D did not change significantly in either toxic or nontoxic strain in our seven-day experiment. Similar results were observed in low density DM (2w) exposure in the nontoxic strain. The toxic strain, however, increased the export of CP A as a response to Daphnia medium, which might be a potential defensive mechanism. Cyanopeptolins are known as protease inhibitors [67] produced similarly to MC, through non-ribosomal pathways [49] with negative impact on zooplankton via inhibiting their trypsin-like digestion enzymes [6,56,68]. Thus, one of their potential roles could be grazing avoidance. If or not CP A has an anti-grazing effect or role will be proven in future studies. In order to understand fate and possible triggers for cyanopeptolins and aerucyclamides production, Sadler and von   [26] found increased production of CP B in M. aeruginosa PCC7806 against D. magna grazing, suggesting an induced defense mechanism, while relative and total amounts of CP A and C were not affected by D. magna presence. The transcription of genes encoding cyanopeptolins, microviridins, and aeruginosins in direct and indirect exposure to D. magna and D. pulex was mostly unaffected [59]. However, it is yet to be clarified whether genes for the synthesis of these secondary metabolites need more time to respond than was considered in this study (24 h) [59].
Different responses to D. magna spent medium in our experiment were strain dependent. MCused in this study is a genetically engineered strain and its physiology as well as the response to D. magna spent medium might differ from MC non-producing strains that naturally occur in aquatic environments. Therefore, more studies are needed to show how D. magna spent medium would affect other MC producing and MC non-producing cyanobacterial species.

Conclusions
To the best of our knowledge, this is the first study showing a lower physiological state of a MC+ strain when exposed to DM than its MC-mutant. D. magna spent medium of the highest D. magna density cultivated for the shortest time (high density DM (24 h)) provoked the strongest effect, which can be attributed to both high zooplankton density and a short time for bacterial degradation of the exudates.
Photosynthetic activity as a proxy of physiological state of the MC+ strain was negatively affected by low density DM (2w) and high density DM (24 h). The mutant strain remained in good physiological state in all exposure scenarios, suggesting that it coped better with the stress induced by DM. The production of intracellular MC-LR, as well as other secondary metabolites in MC+ was decreased, as a consequence of the affected cell physiology, while the dynamics of the quantified metabolites were not affected in the mutant strain (cyanopeptolin A and aeruginosin D being higher in the MC-compared with MC+).
Extracellular cyanopeptolin A concentrations, however, increased in MC+ in the presence of 2wDM, suggesting a potential anti-grazing role, which is in line with its previously evidenced inhibition of trypsin-like digestion enzymes in Daphnia [52]. Furthermore, to the best of our knowledge, this is the first study showing that the concentration of extracellular CP A increased in MC+ when exposed to DM. The MC-strain did not show this increase in the treatment, compared to the control, possibly because the concentration was already initially that high. Therefore, more research is needed to confirm this role. In the presence of the tested Daphnia media, MC-LR and des-MC-LR do not appear to have a protective role for the strain, as their intracellular concentration decreased while the extracellular remained unchanged.
This study evidenced negative impact of D. magna spent medium on the MC+ strain, that was dependent on densities of the organisms in the seven days of exposure.

Cyanobacterial Cultures
Two axenic strains were used in the experiments: MC+ and its mutant MC-, both obtained from the Pasteur Culture collection of Cyanobacteria [69]. MC-is MC+ transformed by inserting a chloramphenicol resistance cartridge into the mcyB gene involved in microcystin biosynthesis. This transformation led to the inability of derived mutant cells to produce any variant of microcystin [70], making MC-a useful tool for exploring the precise effects of microcystins on cyanobacterial physiology and metabolism, when compared to MC+ [70]. Before the start of the experiments, strains were cultivated in BG11 medium (SIGMA), with the addition of 5 µg chloramphenicol mL −1 for the MC-, to maintain the selection of the mutant cells. However, to exclude any impact of the chloramphenicol on DM, the experiments were performed without an antibiotic in the medium, as well as in the control medium. Both strains were grown under a 14 h:10 h light:dark regime using daylight white fluorescent tubes (Toshiba, 15 W, FL15D) with 20 µmol photons m −2 s −1 illumination at a constant temperature of 20 ± 1 • C (Sanyo incubator). Cultures were maintained in exponential growth phase by repeated dilution every 3 weeks in fresh culture medium, while axenicity was regularly evaluated as described in Reference [58].

D. Magna
The D. magna clone (originating from the INERIS) was provided by D. Azam and M. Coke from the PEARL INRA 1036 U3E system (The National Infrastructure in Biology and Health, France). D. magna were slowly acclimatised during three weeks to the cyanobacterial BG11 medium that was used in all the experiments. The green algae Scenedesmus communis (isolated in our laboratory, originating from the lake of Grand Lieu) was provided by Bertrand Le Rouzic (University of Rennes 1). The animals were fed every second day with Scenedesmus communis (reaching max ≈ 2 × 10 4 cells/mL in the aquarium at feeding time) grown in BG11 under the same conditions as the cyanobacteria. Some published studies showed that Scenedesmus has a negative impact on Microcystis through allopathic interactions [71,72]. In our experimental setup however, Scenedesmus have been used as Daphnia food during the medium preparations; therefore, Daphnia medium could contain Scenedesmus metabolites that could affect cyanobacteria [71,72]. However, Scenedesmus cell densities were very low when given to the Daphnia compared to those used in the above mentioned studies showing effects (strains mixed together using equal cell densities of 5 × 10 5 cells/mL (S. huji with the M. aeruginosa PCC 7005) [72] and 1 × 10 5 cells/mL (S. acuminatus with the M. panniformis BCCUSP200) [71]). In our experimental setup, Scenedesmus was given as the only food source, it was consumed by Daphnia, leaving not many, if any, Scenedesmus cells that could produce allopathic compounds affecting cyanobacteria. Noteworthy, when used in the experiment, Daphnia spent medium was mixed with cyanobacterial culture in 4:1 ration, therefore diluting even more possible Scenedesmus compounds. Hence it seems unlikely that the Microcystis strains were affected by Scenedesmus in this experimental setup. D. magna were cultivated at the constant temperature of 20 • C, light intensity of 15 µmol photons m −2 s −1 and a day/night cycle of 14 h/10 h (Sanyo incubator).

Experimental Design
Experiments were performed in the three different exposure scenarios (Table 4), where MC+ and its mutant MC-were exposed to D. magna spent medium. The preparation of the media for the experiments: In order to test environmentally relevant scenarios where Daphnia infochemicals are accumulated over time, we cultivated 75 neonatal individuals of D. magna/L for 1 week (low density DM (1w)). However, we observed no significant difference between control and treatment in this exposure in the photosynthetic activity ( Figure S1) and metabolites dynamics ( Figures S2 and S3), therefore 75 all age individuals of D. magna /L were cultivated for 2 weeks (low density DM (2w)). Moreover, we wanted to exclude potential degradation of Daphnia infochemicals, caused by bacteria coexisting in the D. magna medium during the 2-week cultivation period, as well as provoke a stronger effect in cyanobacteria. Therefore, we cultivated a high number of Daphnia (75 all age individuals of D. magna /L) for 24 h (high density DM (24 h)). One batch of spent medium was prepared per Daphnia treatment group. D. magna were cultivated in BG11 in aquariums and medium was filtered through 0.2 µm filter to remove bacteria and other particles and then stored at −20 • C. Before being used in experiments, exponentially growing cyanobacteria were centrifuged and cells were introduced in fresh BG11 medium and grown for one week to reach a cyanobacterial cell density of 1 × 10 7 cells/mL for low density DM (1w) and low density DM (2w) exposures (pre-culture). Relative densities between cyanobacteria and low density DM (1w) and low density DM (2w) exposures did not provoke strong effects, therefore, the cyanobacterial density was lowered for the 24 h DM exposure, reaching 2x10 6 cells/mL after one week. Filtered D. magna and fresh BG11 media for the treatments and for the controls, respectively, were mixed with the one week old cyanobacterial pre-cultures in 4:1 ratio in 2 L Erlenmeyer flasks. Cyanobacterial density at the beginning of the experiment was 2 × 10 6 cells/mL for low density DM (1w and 2w) exposures and 5 × 10 5 cells/mL for high density DM (24 h) exposure.
The initial nutrients (PO 4 and NO 3 ) and pH were measured for the low density DM (2w) exposure to ensure similar conditions. Dissolved nutrient concentrations were measured from filtered (GF/F 0.7 µm) water using common colorimetric methods [73]. Nitrate was measured after reduction to nitrite on a cadmium-copper column [74]. Phosphate was measured following the method of Murphy and Riley (1962) [75]. Data presented in Table S1 show no significant difference between control and treatment (p > 0.05) for PO 4 measurements in both strains, and for NO 3 in the toxic strain. NO 3 was significantly increased by 8% in the control of the MC-strain, which is still within the range of optimal growth conditions. We also measured pH that was not significantly different between control and treatment (p > 0.05) in all exposure scenarios (Table S2). Overall, there was no nutrient of pH differences that could affect the cyanobacteria physiology.
The procedure was the same for both strains, MC+ and its mutant MC-. The control and the treatment were done simultaneously. Experiments were done in triplicate, cultivated in 1 L flasks, and mixed every day for seven days.

Photosynthetic Activity
To characterize the physiological state of the cyanobacteria, the electron transport rate (ETR) was measured every two days with a pulse-amplitude-modulated fluorescence monitoring system (PhytoPAM, Walz, Germany), following Reference [76] and [60]. The PhytoPAM is equivalent to four separate PAM-fluorometers using light-emitting diodes (LED) with 10 s light pulses at four different excitation wavelengths (470, 520, 645 and 665 nm), with the 645 nm specific to cyanobacteria (due to phycocyanin and allophycocyanin absorption). The PhytoPAM was used with only one channel, corresponding to the cyanobacteria. The reference excitation spectrum measured at the factory was used, as it was not significantly different from reference excitation spectra performed on our cyanobacterial cultures. After being dark adapted for 15 min, Ft (the instantaneous steady state fluorescence at every successive step of actinic irradiance) and Fm (the maximum fluorescence) were measured during a 20 increments of actinic irradiance from 1 to 1864 µmol photons m −2 s −1 , with a 10 s time interval between successive steps. The ETR was calculated as follows: (Fm-Ft)/Fm x 0.42 x PAR (PAR: photosynthetic radiation) [60,76]. The ETRmax was inferred using the regression model of [77] to fit the ETR irradiance curves.

Cyanobacterial Cell Density
Cyanobacterial cell density was followed by measuring the optical density on a spectrophotometer (UVIKONxs SECOMAN) at a corresponding absorbance at 750 nm. Culture medium BG11 was used as reference (blank). The optical density was used to calculate the cyanobacterial cell density (cell number/mL) [58] in order to establish a relationship between the optical density at 750 nm and the cell density, calibration curves with Microcystis cell number/mL vs corresponding absorbance were established for both strains [60]. The counting of the cells was carried out using a Nageotte cell and observed under the Olympus BX50 microscope (objective 40).

Analysis of Cyanobacterial Secondary Metabolites
The dynamics of intracellular and extracellular cyanobacterial peptides of MC+ and MC-in response to different D. magna media were analyzed using a Waters Acquity Ultra-High Performance Liquid Chromatography coupled to a photo diode array detector (PDA) and Xevo quadrupole time of flight mass spectrometer in series. Samples were centrifuged to separate cells (for intracellular metabolites) from supernatant (for extracellular metabolites) and lyophilized. Freeze-dried material was extracted in 0.5 ml 50% aqueous methanol. Extracts were separated on a CORTECS C18 column (0.2 mm ID × 100 mm long), which was maintained at 40 • C. The mobile phase was Milli-Q Water plus 0.1% formic acid (A) and acetonitrile plus 0.1% formic acid (B). Separation was achieved using a gradient increasing from 20% B to 70% B over 10 min, followed by a 100% B wash step and re-equilibration. Autosampler was maintained at 6 • C at all times. UV data was acquired from 200 to 400 nm. Mass data was acquired in positive ion electrospray scanning from m/z 50 to 2000 with a scan time of 2 s and inter-scan delay of 0.1 s. Ion source parameters, i.e., a capillary and sampling cone, were 2.9 V and 25 V respectively; desolvation temperature, 300 • C; and source temperature, 80 • C. Cone gas and desolvation gas flows were 50 L h −1 and 400 L h −1 respectively. Sodium iodide (2 µg µL −1 211 in 50/50 Propan-2-ol/H2O) was used as the calibrant with leucine-enkephalin (0.5 mg mL −1 212 in 50/50 methanol/Milli-Q) as the lock spray. Instrument control, data acquisition (centroid) and processing were achieved using MassLynx v4.1 software (Waters, Milford, MA, USA). We monitored secondary metabolites produced by Microcystis that were observed to be changing due to abiotic or biotic factors [36,50,78] (Table 5). Cyanobacterial peptides were detected using extracted ion chromatograms for the respective specific masses of the different compounds (Table 5). MC-LR, des-MC-LR, cyanopeptolin A, and aerucyclamide D were quantified in our strains. Quantification was achieved using linear relationship between peak area (MC-LR and des-MC-LR at 238 nm, cyanopeptolin A at 220 nm, and aerucyclamide D at 240 nm) and known concentrations of the toxin standards. The microcystin-LR standard was purified as previously described [84]. Cyanopeptolin A standard and aerucyclamide D standard were purified using preparative HPLC (Biotage Parallex Flex, Cardiff, UK) and Flex V3 software for instrument control and data acquisition. The separation was performed on Atlantis Prep C18 column (5 µm particle size, 19 mm ID × 300 mm long; Waters, Elstree, UK) using a 30-min linear gradient from 60% to 100% methanol in MilliQ water. Absorbance was monitored at 210 nm and 280 nm. The flow rate was 20 mL/min and 4 mL fractions were collected. MassLynx v4.1 was used for both detection and quantification of the cyanobacterial peptides.

Statistics
R Core Team (2013) was used to access statistical analysis of the obtained data. All data are presented as mean ± standard deviation. Significant differences were determined at p < 0.05. We performed t-tests to determine the difference between photosynthetic activity of control and treatment for both strains (exposed to different DM (low density DM (2w) and high density DM (24 h)). T-test was also used to compare concentration of intracellular and extracellular metabolites in control and treatment for each strain. Additionally, one-way ANOVAs were used for the comparison between strains in the experiments. Prior to the ANOVAs, Shapiro-Wilk-Normality tests was performed.