Interspecific Differences between D. pulex and D. magna in Tolerance to Cyanobacteria with Protease Inhibitors

It is known that cyanobacteria negatively affect herbivores due to their production of toxins such as protease inhibitors. In the present study we investigated potential interspecific differences between two major herbivores, Daphnia magna and Daphnia pulex, in terms of their tolerance to cyanobacteria with protease inhibitors. Seven clones each of D. magna and of D. pulex were isolated from different habitats in Europe and North America. To test for interspecific differences in the daphnids’ tolerance to cyanobacteria, their somatic and population growth rates were determined for each D. magna and D. pulex clone after exposure to varying concentrations of two Microcystis aeruginosa strains. The M. aeruginosa strains NIVA and PCC− contained either chymotrypsin or trypsin inhibitors, but no microcystins. Mean somatic and population growth rates on a diet with 20% NIVA were significantly more reduced in D. pulex than in D. magna. On a diet with 10% PCC−, the population growth of D. pulex was significantly more reduced than that of D. magna. This indicates that D. magna is more tolerant to cyanobacteria with protease inhibitors than D. pulex. The reduction of growth rates was possibly caused by an interference of cyanobacterial inhibitors with proteases in the gut of Daphnia, as many other conceivable factors, which might have been able to explain the reduced growth, could be excluded as causal factors. Protease assays revealed that the sensitivities of chymotrypsins and trypsins to cyanobacterial protease inhibitors did not differ between D. magna and D. pulex. However, D. magna exhibited a 2.3-fold higher specific chymotrypsin activity than D. pulex, which explains the observed higher tolerance to cyanobacterial protease inhibitors of D. magna. The present study suggests that D. magna may control the development of cyanobacterial blooms more efficiently than D. pulex due to differences in their tolerance to cyanobacteria with protease inhibitors.


Introduction
The frequency of cyanobacterial blooms in many marine and freshwater environments has increased world wide during the last century, partly due to increasing temperatures as a consequence of global warming and partly due to the eutrophication of lakes [1]. Blooms of cyanobacteria and their toxins may sometimes be associated with harmful effects on human health and livestock [2,3]. When the temperature of the epilimnion reaches its maximum in late summer and early fall [4], the phytoplankton of many eutrophic lakes and ponds is often dominated by bloomforming cyanobacterial species of the genera Microcystis, Anabaena and/or Oscillatoria [5]. During this time cyanobacteria are often an important food source for herbivorous zooplankton in freshwater ecosystems, such as for Daphnia, which often provides an important link for the transfer from primary production, e.g. from cyanobacteria to higher trophic levels. In cases when growth of Daphnia is mainly restricted by food quantity, non-toxic cyanobacteria can act as a complementary food source for Daphnia [6,7]. However, since in eutrophic lakes growth of Daphnia is rather constrained by food quality than by food quantity, bloom-forming cyanobacteria in those habitats have been claimed to be a major factor for a constrained mass and energy transfer from primary producers to organisms of higher trophic levels [8,9].
Negative relationships between bloom-forming cyanobacteria and the abundance of Daphnia have been discussed extensively over the years, and three major quality constraints of cyanobacteria as a food source have been revealed so far: (1) The occurrence of cyanobacterial filaments and the formation of colonies hinder ingestion by interfering with the filtering apparatus of Daphnia [10]. (2) Compared to most green algae, cyanobacteria contain low levels of essential lipids such as highly unsaturated fatty acids and sterols, which leads to reduced somatic and population growth of Daphnia due to constrained carbon assimilation [11][12][13][14]. (3) Many cyanobacteria produce a variety of bioactive secondary metabolites such as hepatotoxins like microcystins [15] and/or protease inhibitors [16][17][18]. These compounds reduce the fitness of Daphnia in terms of survival, growth and reproduction [19,20]. In addition to microcystins (which are the most extensively investigated class of cyanobacterial toxins), the role of protease inhibitors in herbivore/cyanobacteria interaction has recently also become a focus of attention. More than twenty depsipeptides, which specifically inhibit the serine proteases chymotrypsin and trypsins, have been found in different genera of marine and freshwater cyanobacteria [16]. These two classes of proteases are the most important digestive enzymes in the gut of D. magna and are responsible for more than 80% of the proteolytic activity [21].
It is known that the edible size fraction of natural phytoplankton can contain compounds that inhibit Daphnia's trypsins and chymotrypsins [22]. This inhibitory potential of seston can be in the same order of magnitude as of pure cyanobacterial cultures [23]. Hence, it is reasonable to assume that an interference of cyanobacterial protease inhibitors with Daphnia's digestive proteases occurs in nature and is ecologically relevant.
However, several studies have also demonstrated that Daphnia may develop tolerances against cyanobacterial toxins at the population level [24][25][26][27]: D. magna populations that were preexposed to toxic cyanobacteria exhibited a higher tolerance to microcystin producing M. aeruginosa than populations that were not pre-exposed [25]. Furthermore, Sarnelle & Wilson [24] suggested that D. pulicaria populations, exposed to high cyanobacterial levels over long periods of time, can adapt in terms of being more tolerant to dietary toxic cyanobacteria. With regard to protease inhibitors Blom et al. [27] have shown that Daphnia sp. coexisting with Planktothrix rubescens (a cyanobacterium that contains the trypsin inhibitor oscillapeptin-J) was significantly more tolerant to oscillapeptin-J than Daphnia sp. from a lake free of this cyanobacterium. Considering the finding that almost 60% of 17 cyanobacterial blooms isolated from 14 distinct water-bodies in India contained protease inhibitors [28], it is reasonable to assume that increased tolerance to cyanobacteria in Daphnia populations may be caused by an enhanced tolerance to the cyanobacterial protease inhibitors. It has been suggested that at least two fundamental mechanisms underlie the increased tolerance to these dietary inhibitors: (1) Colbourne et al. [29] have hypothesized that the ability of Daphnia to cope with different environmental conditions is a consequence of an elevated rate of gene duplications resulting in tandem gene clusters. And indeed, a surprisingly high number of genes of digestive serine proteases have been found in the recently published genome of D. pulex [29].
(2) Von Elert et al. [18] have shown that a physiological plasticity at the protein level in Daphnia in terms of expressing different isoforms of digestive enzymes leads to increased tolerance against cyanobacterial protease inhibitors.
In the present study we tested for interspecific differences between two Daphnia species (D. magna and D. pulex) in their tolerance to cyanobacteria which produce protease inhibitors. D. pulex and D. magna are both large-bodied species and are frequently encountered in fishless ponds [30]. Due to the availability of fullgenome data (D. pulex, [29]) or EST libraries (D. magna; [31]), both Daphnia species are ideal for ecological investigations and were therefore chosen for use in the present study. To determine potential differences between D. pulex and D. magna in their tolerance to cyanobacteria containing protease inhibitors, we performed single-clone somatic and population growth experiments in which the clones were fed with various cyanobacterial mixtures containing trypsin or chymotrypsin inhibitors. Both M. aeruginosa strains used in the present study (NIVA Cya 43 and PCC7806 2 ) produce exclusively either the chemically known chymotrypsin inhibitors cyanopeptolin 954 and nostopeptin 920 (NIVA, [32]) or specific cyanopeptolins (A-D) which are known to inhibit trypsins (PCC 2 , [33]). Possible differences in tolerance to cyanobacteria with protease inhibitors might have several causes and are therefore tested in the present study: (1) We determined the specific trypsin and chymotrypsin activity of each of the investigated D. magna and D. pulex clones and hypothesized that high growth rates on cyanobacterial diets might result from high specific protease activities. (2) For each Daphnia clone, we determined the sensitivity of gut chymotrypsins and trypsins to the respective cyanobacterial protease inhibitors. We assumed that higher sensitivity values of Daphnia's gut proteases might cause reduced somatic and population growth rates for diets with cyanobacterial protease inhibitors.

Origin and Cultivation of Organisms
Two cyanobacterial strains and one green alga were used in the single-clone growth experiments: The cyanobacterium Microcystis aeruginosa NIVA Cya 43 (Culture Collection of Algae, Norwegian Institute for Water Research), subsequently labeled as 'NIVA', is known to contain the chymotrypsin inhibitors cyanopeptolin 954 and nostopeptin 920 [32]. NIVA was cultured in 2 l chemostates in sterile cyano medium [34] at a dilution rate of 0. Seven D. magna and seven D. pulex clones originating from different habitats (Table 1) were used in the somatic and population growth experiments. All clones were cultured separately in aged, membrane-filtered tap water and fed with saturating concentrations of Chlamydomonas sp. for at least three generations prior to the experiment.

Somatic and Population Growth Assays
Each of the 14 clones was assayed in single-clone experiments for somatic and population growth. Five (D. magna) to seven (D. pulex) juveniles from the same cohort of the third clutch and no older than 24 h were kept in 0.25 l aged and filtered tap water (membrane filter of 0.45 mm pore size) under constant dim light at 20uC. The animals were fed with non-limiting food concentrations (2 mg C/l) either of 100% Chlamydomonas sp. or of various mixtures of Chlamydomonas sp. and the two M. aeruginosa strains: In two treatments the animals were fed either with a mixture of 80% C. sp and 20% NIVA or with a mixture of 50% Chlamydomonas sp. and 50% NIVA. In one further treatment the animals were fed with a mixture of 90% Chlamydomonas sp. and 10% PCC 2 . Each treatment was triplicated, and animals were transferred daily into fresh water with saturating food concentrations. Somatic growth rates were calculated on day six as according to Wacker & Von Elert [36] as for which (G) is the body dry-weight of a subsample of the animals at the beginning (G 0 ) and end (G t ) of the experiment. Mean individual dry weights were mean values of two individuals. As according to Brzezinski & Von Elert [37] population growth rates (r) were calculated from daily survival and fecundity of the first clutch by using Euler's equation, 1~X l x |m x |e {rx in which l x is the survival rate and m x the size of the first clutch on day x. Population and somatic growth rates were calculated for each replicate and subsequently averaged to give the mean of the treatment. Population growth rates, calculated on the basis of fecundity at first reproduction, may be a better predictor of fitness in the presence of size-selective mortality than the calculation based on the first three clutches as used in Lampert & Trubetskova [38], since larger individuals are more vulnerable to visually orientated predators. We performed an equal variance test (Levene's test) for each treatment to ensure a comparable intraspecific variability of both Daphnia species.

Preparation of Cyanobacterial Extracts and Daphnia Homogenates
Freeze-dried NIVA or PCC 2 were thoroughly homogenized and mixed separately. 50 mg of the resulting powder was suspended in 500 ml of 60% methanol and sonicated for 15 min followed by centrifugation (3 min at 10 4 6g). The supernatant was subsequently separated from the residue and used as an extract in the protease assays. Controls with 60% methanol had no effects on protease activities.
Seven-day-old individuals of each of the seven D. magna and D. pulex clones grown on non-limiting food concentrations of Chlamydomonas sp. were homogenized with a Teflon pestle. Subsequently the homogenate was centrifuged (3 min at 1.4610 4 6g). In order to minimize protease inactivation due to autolytic degradation, the resulting supernatant was kept permanently on ice until it was used in the enzyme assays.

Protease Activity and Protease Inhibition Assays
The activity of the proteases chymotrypsin and trypsin in homogenates of all Daphnia clones was measured as according to Von Elert et al. [21]. The protein concentration of the supernatant was analyzed using a Qubit fluorometer and the appropriate Quant-iT TM Protein Assay Kit (Invitrogen, Carlsbad, USA) as according to the manufacturer's standard protocol. For the protease activity and protease inhibition assays, SuccpNA (Nsuccinyl-L-alanyl-L-alanyl-L-prolyl-Lphenylalanine 4-nitroanilide, Sigma, 125 mM in DMSO) was used as a substrate for chymotrypsins, while BapNA (N-R-benzoyl-DL-arginine 4-nitroanilide hydrochloride, Sigma, 1.8 mM in DMSO) served as a substrate for trypsins. Trypsin and chymotrypsin assays were performed in a potassium phosphate buffer (0.1 M, pH 7.5). The absorption change was measured continuously for 10 min at 30uC at 390 nm with a Cary 50 photometer (Varian). Specific proteolytic activity was determined as nmol para-nitroanilide liberated per minute and mg protein for synthetic substrates. With regard to the chymotrypsin inhibition assays, homogenates of each of the D. magna and D. pulex clones were assayed after addition of 10-15 different concentrations of the NIVA extract. The resulting chymotrypsin activities were plotted as a function of extracted NIVA biomass per ml assay volume. By fitting a sigmoidal dose response curve, the concentration of extracted NIVA biomass which resulted in a 50% inhibition of Daphnia chymotrypsin activity (IC 50 ) was calculated. The higher the IC 50 values for the analyzed D. magna and D. pulex homogenates, the more tolerant their respective chymotrypsins were to chymotrypsin inhibitors from NIVA. Trypsin inhibition of the D. magna and D. pulex clones was assayed as above, except that different concentrations of the PCC 2 extract were used.

Data Analysis
For the inhibition assays of each single clone of D. magna and D. pulex, the protease activities were plotted as a function of extracted NIVA or PCC 2 biomass. Resulting IC 50 values were calculated by fitting a sigmoid dose-response curve using the software Graph Pad Prism (GraphPad Software, Inc.). A Mann-Whitney U-test was used to determine whether one of the two Daphnia species exhibits stronger growth reductions on mixtures of M. aeruginosa, while a Student's t-test was used to determine interspecific differences in protease inhibition and specific protease activity assays between the two species. Population growth rates were calculated for each individual per replicate independently using an R-script and subsequently averaged for each replicate. Mean Population growth rate of a Daphnia clone on a diet derived from the average of three replicates. All other statistical tests were performed using SigmaPlot 11 (Systat Software, Inc.). A significance level of p = 0.05 was applied to all statistical analyses.

Results
In order to test for interspecific differences between D. pulex and D. magna with regard to their tolerance to cyanobacteria, we performed population and somatic growth experiments on four different diets. Seven clones of each species, each originating from different habitats, were fed either with pure Chlamydomonas sp. or with mixtures of Chlamydomonas sp. and either of two M. aeruginosa strains, NIVA and PCC 2 . For each treatment (20% NIVA, 50% NIVA and 10% PCC 2 ) as well as for both growth rate calculations (somatic growth and population growth) all D. magna and D. pulex clones were categorized into five groups according to their respective relative growth rate reduction (Figure 1a-f), whereas values #0 were regarded as no growth rate reduction on respective diet and were consequently grouped in the category 0-20%. Raw data underlying this classification are presented in Table 2. Equal variance tests revealed that the variability for both Daphnia species were not different for each treatment (Levene's test, p.0.05). On a mixture of 80% Chlamydomonas sp. and 20% NIVA the categorized growth rates of the D. pulex clones were significantly more reduced than those of the D. magna clones (Figure 1a, d). This applies equally for the reduction of population growth rates (U-test: U = 6, p,0.05) and of somatic growth rates (U-test: U = 6, p,0.05): While five of seven D. magna clones were grouped into the first category (0%-20%) of relative somatic and population growth rate reduction, no D. pulex clones were grouped in this category when grown on a mixture with 20% NIVA. With increasing NIVA concentration from 20% to 50%, the categorized somatic and population growth rates decreased for both Daphnia species (Figure 1b, e). On 50% NIVA, the somatic growth rates for D. pulex were significantly more reduced than for D. magna (Figure 1b, U-test: U = 8.5, p,0.05), while this was not the case for the population growth rates (Figure 1e, U-test: U = 12, p = 0.128). However, with regard to the population growth rate reduction on 50% NIVA, four of seven D. pulex clones were grouped in the category of strongest growth reduction, whereas only one clone of D. magna was grouped in this category (Figure 1e). On 10% PCC 2 , the majority of the D. magna clones (six of seven) were grouped in the category of weakest somatic-and population growth rate reduction, representing a reduction of 0%-20% compared to the control treatment on 100% Chlamydomonas sp. (Figure 1c, f). Population growth rates of the D. pulex clones exhibited a more variable classification when grown on 10% PCC 2 and were significantly more reduced than those of the D. magna clones (Figure 1f, U-test: U = 6, p = 0.017). With regard to the somatic growth rate reduction on 10% PCC 2 no statistical differences were found between D. pulex and D. magna (Figure 1c, U-test: U = 12.5, p = 0.128).
We furthermore quantified the specific activity of trypsins and chymotrypsins for each of the D. magna and D. pulex clones as possible causes for the observed differences in tolerance of the two species to cyanobacterial protease inhibitors. Mean specific chymotrypsin activity of D. magna (331.86117.5 nmol/min/mg protein standard deviation (SD)) was dramatically higher (+225%) than the activity of D. pulex (147.46106.8 nmol/min/mg protein SD; Figure 2a, t-test: t 12 = 23.073, p,0.05). Mean specific trypsin activity for D. pulex was 45.2620.6 nmol/min/mg protein SD, whereas it was 54.4617.6 nmol/min/mg protein SD for D. magna ( Figure 2b); however, these values were not different (t-test: t 12 = 20.897, p = 0.387). Microcystis aeruginosa strains, NIVA or PCC 2 . Animals were fed either with 80% Chlamydomonas sp. and 20% NIVA (a, d), with 50% Chlamydomonas sp. and 50% NIVA (b, e) or with 90% Chlamydomonas sp. and 10% PCC 2 (c, f). Daphnia clones were classified into growth reduction categories according to the relative growth rate reduction on respective treatments compared to the control treatment on 100% Chlamydomonas sp. Significant differences (Mann-Whitney U-test, p,0.05) between species are indicated by an asterisk, while no differences are labeled with ''n.s.''. doi:10.1371/journal.pone.0062658.g001 In order to test for possible additional causes of the interspecific differences in their tolerance to cyanobacterial protease inhibitors, the sensitivity of Daphnia's trypsins and chymotrypsins to cyanobacterial protease inhibitors was determined. This was achieved by specifying the concentration of extracted cyanobacterial biomass that was needed to inhibit 50% (IC 50 ) of either Daphnia's chymotrypsins or their trypsins (Fig. 2c, d).Mean IC 50 values of the effects of NIVA extracts on Daphnia's chymotrypsins were 295.6663.4 ng/ml SD for D. pulex and 239.8627.7 ng/mL SD for D. magna, although these effects were not different (Figure 2c, t-test: t 12 = 2.136, p = 0.054). When specifying the IC 50 values of effects of PCC 2 extracts on Daphnia's trypsins, no differences between D. pulex (798.76205.6 ng/ml SD) and D. magna (876.8 ng/ml 6295.2 ng/ml SD) were detected (Figure 2d, t-test: t 12 = 20.574, p = 0.577).

Discussion
Cyanobacterial blooms are harmful to many freshwater herbivorous grazers such as Daphnia. Owing to the dominance of cyanobacteria in eutrophic lakes and ponds in late summer [4], Daphnia genotypes from these habitats have to cope more frequently with the poor food quality of cyanobacterial carbon than genotypes from habitats free of cyanobacterial blooms. The causes for the poor assimilation of cyanobacterial carbon by Daphnia have been studied extensively in past decades, leading to the following observations: Firstly, colonial or filamentous cyanobacteria can mechanically interfere with Daphnia's filtering apparatus [10]. Secondly, cyanobacteria lack essential sterols [12] and sufficient amounts of polyunsaturated fatty acids [14]. Thirdly, the production of hepaptotoxins such as microcystins can significantly reduce the fitness of Daphnia [19]. In the present study, all D. pulex clones and five out of seven D. magna clones exhibited a reduction in somatic and population growth rates at a concentration of 20% NIVA. The negative effects mentioned above can be ruled out as causal factors for the observed growth reductions, since the food mixtures used here consisted of saturating concentrations of the widely used, high quality reference food Chlamydomonas sp. Furthermore, both cyanobacterial strains as well as Chlamydomonas sp. were cultured as single cells and did not contain any microcystins. However, with two D. magna clones reductions were observed only with 50% NIVA (D. magna clone P and W, Table 2). This supports data of Lürling [39], who used the same strain of M. aeruginosa (NIVA) and reported a reduction in growth of D. magna clones at a concentration of $25% NIVA. On the mixture with PCC 2 , some D. magna clones and the majority of the D. pulex clones exhibited a reduction of population and somatic growth rate already at a concentration of 10%. Higher concentrations would probably have resulted in a significant growth rate reduction in all D. magna and D. pulex clones, since several other studies [40,41] have reported a clear reduction in growth of daphnids at a concentration of 20% PCC 2 .
One possible explanation for the observed somatic and population growth rate reduction of the D. magna and D. pulex clones in response to cyanobacteria could be the result of dietary inhibition of either Daphnia's digestive chymotrypsins or trypsins. In this study somatic and population growth rates of D. magna and D. pulex served as a measure of tolerance to microcystin-free cyanobacteria and as an approach to test for interspecific differences. In several cases it has been demonstrated that coexistence of Daphnia with microcystin-producing cyanobacteria leads to local adaptation of the Daphnia populations, as was evidenced by increased tolerance to cyanobacteria [24,42]. These findings support the notion that the presence of cyanobacteria positively selects for more tolerant Daphnia genotypes.
Several studies have already demonstrated high intraspecific variability of Daphnia's tolerance to cyanobacteria [24,25]. In the present study we also found intraspecific variability of D. magna and D. pulex in the tolerance to cyanobacteria with protease Table 2. Relative reduction of somatic and population growth rate of all D. magna and D. pulex clones on mixtures containing 20% NIVA, 50% NIVA and 10% PCC 2 as well as specific activity and IC 50 values of Daphnia's chymotrypsins (CT) and trypsins (T).

Somatic growth rate reduction (%)
Population growth rate reduction (%) inhibitors. However, an equal variance test revealed that the variability for both Daphnia species were not different. With regard to interspecific comparisons we have shown for the first time that D. magna exhibits a higher mean tolerance to cyanobacteria with protease inhibitors than D. pulex. Microevolutionary adaptation of Daphnia populations to cyanobacteria has been experimentally confirmed: Exposure of a mixed population of several Daphnia clones to a microcystin-producing strain of M. aeruginosa resulted in an enhanced tolerance in subsequent generations [25]. Such a maternally transferred increase in tolerance of Daphnia's offspring generation has also been demonstrated for the microcystin-free M. aeruginosa strain PCC 2 [43] which was also used in the present study.
Currently it is not clear to which extent the observed interspecific differences between D. magna and D. pulex in tolerance to cyanobacteria with protease inhibitors are affected by physiological plasticity of the clones investigated here. Using a single clone of D. magna, Von Elert et al. [18] have shown that a physiological plasticity at the protein level in Daphnia in terms of expressing different isoforms of digestive enzymes leads to increased tolerance to cyanobacterial protease inhibitors. In the present study we could not find a correlation between less sensitive proteases and higher tolerance to the cyanobacterial strains. The protease inhibition assays revealed that the tolerance of chymotrypsins and trypsins to cyanobacterial protease inhibitors did not differ between D. magna and D. pulex. However, the mean specific chymotrypsin activity of D. magna was significantly higher than that of D. pulex. This elevated specific chymotrypsin activity of D. magna coincides with less reduced somatic and population growth rates in the presence of chymotrypsin inhibitors. Schwarzenberger et al.  50 values, which represent the concentration of extracted dry weight of cyanobacterial biomass that is required to cause a 50% inhibition of respective proteases. Low IC 50 values indicate a high sensitivity of proteases from the respective Daphnia clones. Medians are denoted by solid black lines, while the top and bottom box edges denote the first and third quartile. Whiskers denote the largest and smallest data within 1.5 times the interquartile range. Significant differences (Student's t-test, p,0.05) among species are indicated by an asterisk; no differences are labeled with ''n.s.''. doi:10.1371/journal.pone.0062658.g002 [40] have demonstrated that the same D. magna clone as used by Von Elert et al. [18] responded to dietary protease inhibitors by increased expression of trypsin and chymotrypsin genes. However, it remains to be tested, if differences in plasticity of protease expression cause the observed higher tolerance to cyanobacteria with protease inhibitors in D. magna than in D. pulex. With regard to the significantly higher population growth rates of D. magna on PCC 2 than those of D. pulex, we couldn't find differences in species-specific trypsin characteristics, neither with respect to the specific activity nor to the respective sensitivity.
Besides protease inhibitors, the observed differences in growth on cyanobacteria between D. magna and D. pulex could also result from differences in the tolerance to other toxic compounds such as lipopolysaccharides [44] or unidentified toxins. Additionally, we cannot rule out the possibility that potentially synergistic interactions between bioactive compounds may have important negative effects on the fitness of Daphnia. Such synergistic effects have been reported for the toxicity of artificial nano-particles and arsenic on the fitness of daphnids [45]. However, we are not aware of any cyanobacterial metabolites, for which synergistic effects on herbivores have been demonstrated. Furthermore, when looking at possible defense mechanisms of Daphnia against cyanobacterial toxins, more efficient detoxification mechanisms in D. magna than in D. pulex could also be a reason for the higher tolerance of D. magna to the cyanobacterial diets. One possibility of reducing toxic effects in Daphnia is to biotransform toxins by conjugation to glutathione, which has already been demonstrated for the cyanobacterial hepatotoxin microcystin-LR [46]. However, even for this conjugation of microcystins by Daphnia no intra-or interspecific comparisons are available. Although similar detoxification mechanisms of protease inhibitors in Daphnia have not been reported so far, they cannot be excluded as a possible causal factor for the higher tolerance of D. magna to diets containing cyanobacteria with protease inhibitors. Additionally, Rohrlack et al. [47] reported a reduction of the ingestion rate in D. galeata after a short time exposure (15 min.) with 100% PCC 2 , which could also result in a reduction of somatic and population growth. Here, experimental individuals were fed for at least six days a PCC 2 proportion of 10%, while 90% of the carbon resulted from the good reference food Chlamydomonas sp. With regard to NIVA there was no detectable inhibition of the ingestion process of one D. magna clone even at a proportion of 100% NIVA in the diet (Kuster, unpublished). Thus, it seems unlikely that inhibition effects of the ingestion process contributed to the observed growth reduction on cyanobacterial diets. However, since we haven't measured ingestion rates for all D. magna and D. pulex clones used here, we could not clearly rule out this possibility.
As a result of continuing global warming, negative effects of toxic cyanobacteria on human health, livestock and the whole ecosystem will become stronger in the near future [1]. Much effort is therefore invested in preventing or controlling cyanobacterial blooms. Besides reducing nutrient input into lakes and ponds, the control of cyanobacterial blooms through biological agents such as bacteria, viruses and unicellular grazers has been discussed extensively in the last decades [48,49]. Depending on initial conditions, Daphnia may also control the development of bloomforming cyanobacteria and may even suppress established cyanobacterial blooms [50]. The present study suggests that D. magna may control the development of cyanobacterial blooms more efficiently than D. pulex due to significantly higher tolerance to cyanobacteria with protease inhibitors. In light of the increasing temperatures as a consequence of global warming, our results suggest that toxic cyanobacterial blooms and coinciding harmful effects to the ecosystem will occur more frequently in lakes with D. pulex than in lakes with D. magna.