Perspective Chapter: Daphnia magna as a Potential Indicator of Reservoir Water Quality – Current Status and Perspectives Focused in Ecotoxicological Classes Regarding the Risk Prediction

Abstract

Several types of stressors come into natural water bodies, degrading their quality, and having harmful effects on aquatic biota. As a result, many attempts have been made to develop complementary techniques to those imposed by the Water Framework Directive (WFD) to improve the water quality assessment strategy in a shorter time and be more faithful to the components and contaminants of the ecosystem. Daphnia magna has been extensively used as a model organism for ecotoxicity testing, and its ecotoxicological responses to several disturbance factors have been being well characterized. From this perspective, this work aimed to evaluate the applicability of the feeding bioassays with D. magna, as well as early distress tools (biochemical biomarkers), in the assessment of water quality of natural waters of reservoirs. Samplings were performed in several sites in three Portuguese reservoirs and were conducted in the spring of 2020. Bioassays and biomarkers results showed sensitivity to different reservoirs since the sites are minimally or moderately impacted. Biological responses can be related to several environmental factors, such as surrounding areas, seston composition, and chemical analysis (WFD), among others not quantified. This set of biological responses has presented good concordance with the ecological potential of the reservoirs.

1. Introduction

Over the years, increasing demographic pressures have contributed to an exponential increase in contaminants in the water ecosystems, associated with the intensification of agricultural and industrial activities. Some of these contaminants have raised particular concern and have been classified as specific pollutants and priority substances, identified in the Directive 2013/39/EU of the European Parliament [1] and the Agência Portuguesa do Ambiente (APA) [2]. To perform the ecological assessment of aquatic ecosystems, there are two major comprehensive frameworks worldwide, using multiple lines of evidence (LoE), with special emphasis on ecological data (biological communities): the Water Framework Directive (WFD), adopted in Europe through the Directive 2000/06/EC, and the Ecological Risk Assessment (ERA), adopted for example by the US Environmental Protection Agency. WFD and ERA approaches integrate information from different LoE, as the extent of the application of ecotoxicological evidence clarifying cause-effect relationships; the availability of expert judgment in the fine-tuning of sampling practices, strategic analysis, data interpretation, and decision procedures; on the practical meaning of the concept of LoE integration (“one-out, all-out” principle versus integrated risk quantification) [3]. The WFD is an extensive legislative framework for the protection of ground and surface waters in Europe, with defined Environmental Quality standards (EQS) for several parameters [biological, chemical, physicochemical (supporting biological elements), and hydromorphological elements] that must be complied by the different member states, and also advise additional monitoring of substances of national or regional interest [1]. However, it is estimated that a very high number of substances are present in the environment, and for Europe, it is estimated the existence of more than 100,000 compounds [4, 5]. Therefore, the definition of these lists of compounds with environmental concern, although relevant, becomes unreliable, in terms of representation in aquatic ecosystems, since complex mixtures of chemicals occur (e.g., emerging pollutants, metabolites, and transformation products). A list of priority substances, which represent a significant risk to or via the aquatic environment at the EU level, will have to be reexamined by the European Commission (EC) and should not exceed 4 years. In recent years, there has been a growing agreement among authorities and scientists that the tools currently used and proposed by the WFD for water quality assessment require a review to achieve a clearer and future-proof methodology [6, 7]. In this context, ERA is reflected as a complementary alternative for the bioassessment of the quality of freshwater, such as reservoirs. This approach considers some valuable WFD principles and metrics but, at the same time, includes complementary methods, one of which is the incorporation of effects-based tools (e.g., ecotoxicological assays; biomarkers in organisms) for a better assessment of cause-effect relationships; reflecting an effective integration of distinct LoEs (e.g., chemical, ecological and ecotoxicological) [3].

In environmental terms, pollution occurs due to a complex mixture of organic and inorganic compounds that can result in lethal and sub-lethal effects on aquatic organisms, associated with potentially significant losses of habitat and biodiversity. Currently, the assessment of water quality, using organisms as bioindicators of water quality has been widely used, since the biological responses integrate the complex influence of the stressing agents [4, 6, 8, 9, 10], in addition to the complex mixture of compounds that occur in the ecosystem under analysis. According to several studies, the effect-based water quality assessment (e.g. biological responses of organisms to natural waters) has been successful in the identification of ecotoxicological risks in surface waters and the ranking of locations based on these risks, namely for natural water bodies (e.g. rivers, transitional or coastal waters) [5, 6, 11], but also to heavily modified water bodies (e.g. reservoirs) [8, 9, 10]. Bioassays with Daphnia sp. (mostly Daphnia magna) are regularly used in ecotoxicological studies because they have high fertility values, easy to maintain in laboratory conditions, ubiquitous, and important bioindicators for aquatic environments due to their sensitivity to contaminants and position (trophic level: primary consumers) in the aquatic food webs [6, 9, 11].

To answer the research needs outlined above, the present study aimed to apply effect-based approaches (individual and biochemical responses of D. magna to natural waters) in the water quality assessment of Portuguese reservoirs, defining classes of disturbances and ranges of ecotoxicological potential. This work has been divided into two parts with specific objectives. Part 1 of this study was presented in [9], which demonstrated that biochemical parameters (metabolism, oxidative stress, and lipid peroxidation biomarkers) improved the sensitivity of the biomonitoring strategy using bioassays with the standard species D. magna, in the assessment of the ecological quality of water reservoirs, in different sampling periods (Autumn 2018 and Spring, Autumn 2019). The biochemical parameters revealed sensitivity in the evaluation of effects incited by exposure to natural waters from reservoirs, making them useful and reliable in this type of evaluation. According to the results of [9], the biomarker indicative of lipid peroxidation (levels of thiobarbituric acid reactive substances—TBARS) on D. magna represented a consistent tool for evaluation of water quality. This result reflected, in part, the prooxidative state of organisms, food performance, and possible stress scenarios, mainly due to the components of seston (e.g., quantity and quality of phytoplankton) and chemical contamination.

Part 2, the focus of the here-presented study, is intended to evaluate whether effect-based methods can be applied in natural waters quality assessment of reservoirs, by the definition of disturbance classes and ecotoxicological potential values. The integration of physical and chemical and effects-based monitoring approaches can complement and improve the water quality assessment strategies in the future, with the main objective of a nontoxic environment. For this, several ecotoxicological tools were applied, to gain insight into the ecotoxicological potential of the reservoirs understudy and compare the ecological potential with tools proposed by the WFD, as well as with previous studies, in the same areas under study. To this end, the present approach combined the evaluation of the individual (feeding rates) and cellular/molecular responses in D. magna after acute exposure to natural waters [e.g., biochemical biomarkers of oxidative response as activities of catalase (CAT) and glutathione S-transferases (GSTs), the latter also involved in the biotransformation process, lipid peroxidation (LPO measured as TBARS levels), and acetylcholinesterase (AChE) activity, involved in neurotransmission process].

2. Material and methods

2.1 Study areas, water sampling, and physicochemical parameters quantifications

Three reservoirs were selected for conducting this study (Figure 1): Miranda (M) and Pocinho (P) are main course reservoirs and belong to the hydrographic basin of the Douro river; and Aguieira (Ag) which is a northern reservoir that belongs to the Mondego hydrographic basin.

Figure 1.

Map of sampling areas (Miranda, Pocinho and Aguieira reservoirs) with the location of the sampling sites. The different colors represent the 1st level of detail of land occupation according to the land use report (2018).

The purpose for defining these reservoirs and respective sampling sites was based on previous studies of our research group [8, 9, 10, 12, 13, 14], where the water quality was assessed, in the last years, using different indicators and methodologies complementary to the WFD.

Water samples were collected during the spring of 2020, in six sites of the three reservoirs (Figure 1): one site in Miranda, one site in Pocinho, and four sites in Aguieira. The sites were well-defined based on previous works conducted in these reservoirs [8, 9, 10, 12, 13, 14].

In situ, the abiotic parameters pH, dissolved oxygen (mg/L and %), conductivity (μS/cm), and temperature (°C) were measured with a multiparameter probe (Multi 3630 IDS SET F). For conducting chemical analysis [e.g., nutrients, specific pollutants, and priority substances] and bioassays, 5 L of water were collected at each sampling site and transported to the laboratory at 4°C and in the dark. Chemical analyses were carried out within a maximum period of 48 hours after collection. D. magna assays were started within a maximum period of 24 hours, after sample collection.

2.2 Chemical analyses

A set of specific pollutants and priority substances were measured, according to the recommendations defined in the Agência Portuguesa do Ambiente [2] and the Directive 2013/39/EU [1]. Nitrites (NO₂⁻) and nitrates (NO₃⁻) were quantified by liquid chromatography of ions, as dissolved anions [15]. Total Kjeldahl nitrogen (N_Kj) determination was performed by the Kjeldahl nitrogen method after mineralization with selenium [16]. Calcium and magnesium determinations were effectuated by ion chromatography, as dissolved cations [17]. For the elements, total phosphorus (P_total), arsenic, cadmium, copper, mercury, nickel, lead, and zinc, the analysis was performed by the application of inductively coupled plasma mass spectrometry (ICP-MS) [18]. Pesticides and polycyclic aromatic hydrocarbons were not quantified in our study, because, according to previous studies [8, 9, 10, 12, 13, 14] the values of these specific pollutants and priority substances (quantified in autumn of 2018) were below the detection limits of the analytical method, in addition to, that no significant changes in the areas adjacent to the reservoirs were documented during the last years (2019, 2020).

2.3 Biological parameters by WFD—Ecological quality ratio (EQR) for phytoplankton

The phytoplankton community characterization was performed according to the Instituto da água I.P. [19] and Agência Portuguesa do Ambiente (APA) [2] guidelines and briefly described in [9]. For the determination of the ecological potential (EP), the results were expressed in an EQR, determined according to the WFD approach. According to APA [2], the EQS used in the classification of the biological quality (EQR) for Miranda and Pocinho reservoirs was carried out based on the typology “main course”. For main course typology, taking into account the biological elements proposed in the WFD to Portuguese reservoirs, the EP is only classified into two classes: moderate or less, and good or more (Table 1). Aguieira is a northern type of reservoir, and the EP is classified into four classes: Good or more, Moderate, Poor, or Bad (Table 1).

				Main course		North
		EQS	M	P	Ag1	Ag2	Ag3	Ag4
Physical and chemical elements	Temp (°C)	—	19.0	22.8	21.9	20.5	20.7	22.4
	Cond (μS/cm)	—	438	268	74	85	90	73
	pH	6–9 [2]	8.6	9.2	9.6	9.7	9.0	9.4
	O2 (mg/L)	≥5 [2]	11.0	15.9	12.9	14.2	12.4	13.3
	O2 (%)	60–120 [2]	124.1	185.0	150.2	160.1	141.0	156.5
	NO2^− (mg NO2^−/L)		0.16	0.05	0.04	0.04	0.05	0.01
	NO3^− (mg NO3^−/L)	≤ 25 [2]	7.4	3.5	2.7	2.2	3.3	0.6
	NKj (mg N/L)	—	<0.5	0.7	<0.5	0.7	<0.5	<0.5
	Ptotal (mg P/L)	≤ 0.05 [2]	0.13	0.09	0.02	0.03	0.08	0.03
Specific Pollutants	As (μg/L)	50 [2]	2.13	3.30	1.41	1.31	1.62	2.83
	Zn (μg/L)	7.8 [2]	26.6	80.7	32.2	27.7	28.9	24.7
	Cu (μg/L)	7.8 [2]	1.84	1.59	1.86	2.09	2.92	1.70
Priority substances	Cd (μg/L)	0.45 [1]	0.01	0.02	0.02	0.02	0.02	0.01
	Hg (μg/L)	0.07 [1]	1.02	1.06	1.85	1.57	1.33	0.63
	Ni (μg/L)	34 [1]	2.1	2.4	2.3	1.7	1.6	1.5
	Pb (μg/L)	14 [1]	0.5	0.5	0.4	0.3	0.4	0.3
	Ca (mg/L)	—	63	40	3	4	4	3
	Mg (mg/L)	—	10	6	1	1	2	1
Ecological potential (chemical and physico-chemical elements)			Moderate	Moderate	Moderate	Moderate	Moderate	Moderate
Biological (Phytoplankton EQR)	North [2]; [1.0–0.60] – Good or more; [0.6–0.4] – Moderate; [0.4–0.2] – Poor; [0.2–0] – Bad; Main course [2]; ≥0.17 – Good or more; <0.17 – Moderate or less		0.16	0.12	0.77	0.37	0.33	0.61
Ecological potential (biological)			Moderate or less	Moderate or less	Good or more	Poor	Poor	Good or more

Table 1.

Results of the physical and chemical parameters, and specific pollutants and priority substances concentrations of Portuguese reservoirs. The bold values represent the values outside of the environmental quality standards (EQS). The biological parameter phytoplankton and respective EQS for the main course and north reservoirs are also presented.

Sampling sites: Miranda – M, Pocinho – P and Aguieira – Ag1 to Ag4. Temperature (Temp), Conductivity (Cond), pH, Dissolved oxygen (O₂), nitrites (NO₂⁻), nitrates (NO₃⁻), Ammonium (NH₄⁺), Total Kjeldahl nitrogen (N_Kj), Total phosphorus (P_total), Arsenic (As), Zinc (Zn), Copper (Cu), Cadmium (Cd), Mercury (Hg), Nickel (Ni), Lead (Pb), Calcium (Ca), Magnesium (Mg), Iron (Fe), Manganese (Mn), and Ecological Quality Ratio (EQR).

2.4 Water treatments

The water collected in each sampling site of each reservoir was processed in 3 treatments, namely: NF (Non-Filtered water with all components present in the sample); F1 (water filtered through a Whatman GF/C filter with 1.2 μm porosity); and F2 (water filtered through a sterile filter system with a porosity of 0.22 μm) as already defined in previous studies of our group [10, 12].

2.5 Test organisms

2.5.1 Culture maintenance of D. magna

Successive generations of monoclonal cultures of D. magna were continuously kept in controlled laboratory conditions of 16^h light and 8^h dark photoperiod and temperature of 20 ± 2°C. Cultures were renewed every 2 days and were maintained in synthetic water medium “ASTM hard water” [20], supplemented with a standard organic additive, Ascophyllum nodosum extract [21], to provide essential microelements to Daphnia. D. magna were fed with Raphidocelis subcapitata at a rate of 3.0x10⁵ cells/mL/day. For conducting the bioassays, daphnids with 4 or 5 days, born between the 3rd and 5th broods were used.

2.5.2 D. magna feeding rate assays

D. magna feeding rate assays were conducted according to [22] with some adaptations described in [9]. For each water sample, bioassays were performed on 6-well plates, where each plate corresponded to specific water treatment (NF, F1, or F2). For each water treatment and control (ASTM hard water medium), 5 replicate wells with 5 D. magna individuals, and a blank well (water sample with Raphidocelis subcapitata without daphnids) were performed. The blank well is performed to account for the potential algal growth during the assay period. Mortality was also considered in this study. Feeding rate results were expressed according to [23]. The percent inhibition in feeding rate (% I_FR), relatively to the control, was calculated for each water treatment (NF, F1, and F2) as follows:

$$\%I_{\mathit{FR}}=\left[\left(\mathit{FRc}-\mathit{FRt}\right)/\mathit{FRc}\right]\times 100$$E1

where:

% I_FR: percent inhibition of feeding rate;

FRc: mean value for feeding rate in the control group;

FRt: value for feeding rate for the water treatment.

At the end of the feeding rate assays, pools of organisms from each treatment were preserved for posterior biochemical determinations (oxidative stress, lipid peroxidation, and neurotransmission biomarkers) and stored in microtubes at—80°C until analyses were performed.

2.5.3 Biochemical determinations

For determination of biomarkers of oxidative stress [catalase (CAT) and isoenzymes glutathione S-transferases (GSTs) activities] and levels of lipid peroxidation (LPO) [levels of thiobarbituric acid reactive substances (TBARS)], samples were thawed on ice, and a 1 mL of ice-cold phosphate buffer (50 mM, pH = 7.0 with 0.1% of Triton X-100) was added to each biological sample. Samples were sonicated for 20 s and centrifuged at 14,000 rpm, for 10 min, at 4°C, in a refrigerated centrifuge (Eppendorf 5810R). The supernatant fraction was divided into aliquots and used to perform the biochemical analyses. For the quantification of acetylcholinesterase (AChE) activity, the samples were homogenized with a sonicator, in a volume of 750 μL of ice-cold phosphate buffer (0.1 M phosphate buffer, pH = 7.2), and centrifuged at 6000 rpm for 3 min. The supernatants after centrifugation were collected and used for AChE activity determinations.

All biochemical analyses were adapted to 96-well microplates [9, 24], and spectrophotometric readings were performed in a microplate reader Thermo Scientific, model Multiskan GO (version 1.00.40), with SkanIt Software 3.2.

The total soluble protein concentration of samples was performed according to the Bradford method [25], using a standard of γ-globulin 1 mg/mL. This method is based on the binding of a dye (Bradford’s reagent) to the total soluble proteins, forming a stable complex [24]. This determination permits expressing the enzymatic activities and TBARS levels, as a function of the total soluble protein content of the samples.

CAT is an antioxidant enzyme with peroxidic activity and is responsible for the decomposition of hydrogen peroxide (H₂O₂) in H₂O + O₂, where H₂O₂ consumption occurs with the oxidation of hydrogen donors (phenols, formic acid, and methanol) [24]. The method consists of the monitoring of this decomposition and was performed according to [26]. The enzymatic activity was expressed in nmoles of H₂O₂ consumed, per minute, per milligram of total soluble protein.

The GSTs activity was quantified according to [27]. GSTs catalyze the conjugation of glutathione in its reduced form (GSH) with the substrate 1-chloro-2,4-dinitrobenzene (CDNB), forming a thioether, whose formation was observed by measuring the increase of the absorbance [24]. Enzyme activities were expressed in mmol of thioether produced, per minute, per milligram of total soluble protein.

LPO was measured through the determination of the levels of TBARS, according to [28], which measures the absorbance of the complex resulting from products of oxidative free radical attack to membrane lipids, with thiobarbituric acid. Results were expressed as millimoles of malondialdehyde (MDA) and MDA-like compounds equivalents, per mg of total soluble protein.

The quantification of the AChE activity was performed by the Ellman method [29]. This enzyme is responsible for the degradation of the synthetic substrate acetylthiocholine into acetate + thiocoline [24], which occurs with the increase in the yellow color produced when thiocoline is complexed with dithiobis nitrobenzoate (DTNB). The enzymatic activity was expressed as nmol of the complex formed, per minute, per milligrams of total soluble protein.

2.6 Water ecotoxicological assessment

At the end of all quantifications of enzyme activities and LPO levels, the percent inhibition in each biochemical determination (% I_X), comparatively to the control, was calculated for each water treatment (NF, F1, and F2) as follows:

$$\%I_X=\left[\left(\mathit{Xc}-\mathit{Xt}\right)/\mathit{Xc}\right]\times 100$$E2

where:

% I_X: percent inhibition of: CAT activity (% I_CAT), GSTs activity (% I_GSTs), TBARS levels (% I_TBARS) or AChE activity (% I_AchE);

Xc: mean value for CAT activity, GSTs activity, TBARS levels, or AChE activity in the control group;

Xt: value for CAT activity, GSTs activity, TBARS levels, or AChE activity for the water treatment.

2.7 Statistical analyses

The data from all test variables (the percent inhibition of feeding rate, CAT and GSTs activities, TBARS levels, and AChE activity) were previously analyzed to assure normality and uniformity of variance (Shapiro–Wilk and Levene tests, respectively). All parameters were analyzed by analysis of variance (one-way ANOVA), followed, when significant differences were detected (p < 0.05), by a Tukey test to discriminate differences between treatments (NF, F1, and F2). The data are presented as mean and respective standard errors. The analyses were performed using software SPSS Statistics (version 26) and Sigmaplot (version 11.0).

3. Results and discussion

3.1 General physicochemical characteristics and trace elements concentrations (chemical analysis) in the water samples

Table 1 summarizes the results of physicochemical parameters and chemical analyses including the quantifications of the concentrations of the specific pollutants and priority substances, measured for each site over the sampling period, as recommended by APA [2] and European Parliament and the Council [1]. According to the physicochemical parameters used in the WFD, for Portuguese heavily modified and artificial water bodies, only the pH, O₂, NO_3^- and P_total have an environmental quality standard (EQS) values established for a good ecological potential (GEP).

In general, water samples from the three reservoirs were characterized by a basic pH (values range from 8.6 and 9.7, Table 1), with almost values above the EQS (> 9.0, except M and Ag3). The water pH is an important parameter as it can determine the solubility and biological availability of nutrients, but also metals [30]. Dissolved oxygen (%) showed values above the maximum of 120%, in all sites studied. Considering the electrical conductivity, the values ranged between 73 and 438 μS/cm. Higher values of this parameter were registered in the reservoirs belonging to the Douro river basin, Miranda and Pocinho (> 260 μS/cm). In different Aguieira sites, the values showed a low variation between the locations of 73–90 μS/cm. The sites M, P, and Ag3 showed higher contents of nutrients (mainly total phosphorus) when compared with the values of EQS (Table 1). Two types of reservoirs are referred to in the work of [31]: Type 1—lowland “run-of-river” reservoirs located in the main rivers (e.g., Douro), at lower altitudes, had larger catchments, lower residence time, and were higher in mineral content (hardness and conductivity), than Type 2, which are deeper high altitude reservoirs (e.g., Mondego). Considering this distinction, Miranda and Pocinho are reservoirs of Type 1 and Aguieira is Type 2 (for more information see [31]), and in fact, our physicochemical results are supported by these assumptions (Table 1). Higher nutrient concentrations (P_total) were observed at Miranda and Pocinho sites than at the Aguieira sites (Table 1). According to [31], Type 1 reservoirs are more nutrient-rich (total phosphate and nitrates due to more extensive agriculture and intensive) than Type 2, corroborating our results. If we consider land occupation (Figure 1) in the area surrounding the sampling site of Miranda, the water pressures are associated with the artificialized territories and forests. For Pocinho, agriculture is highly representative in terms of land occupation. The land occupation in the Mondego river basin area, where the Aguieira reservoir is located, has the surrounding areas mainly represented by forests, agricultural areas, and artificialized territories (Figure 1). Kroll et al. [32] show a solid association between land occupation (urban, agriculture, and forest areas) and nutrients indices in nearby aquatic ecosystems. These findings demystify and support some of the results (e.g., nutrient levels) presented here, as well as work previously developed in these same locations [9].

Concerning the metals, only mercury (Hg) and zinc (Zn) exceeded the EQS (Table 1) defined by the Directive 2013/39/EU of the European Parliament [1] and APA [2], respectively. Hg was present in concentrations above 0.07 μg/L (EQS) at all sites (> 0.63 μg/L). For Zn, concentrations above 7.8 μg/L (EQS) were quantified in all locations of the three reservoirs (> 24.7 μg/L). Several metals such as mercury and zinc (among others) can be highly toxic even in residual quantities [33]. Hg is an important pollutant of water throughout the world, and several human activities are linked to Hg pollution (silver and gold mining, coal combustion, and dental amalgams), and is known to be an inhibitor of enzymes’ activities [9, 30, 34, 35]. The speciation of Zn in water is modulated by pH and dissolved organic matter, which normally binds most of the aqueous zinc [30]. Zn concentrations in natural waters span six orders of magnitude and are strongly influenced by human activities [30]. There are a comprehensive set of proteins that function as transporters, chelators, and molecular sensors for Zn, and are involved in the regulation of Zn uptake by homeostatic processes that are partially understood. However, several studies have proposed theories to explain how zinc compounds affect aquatic animals [30]. However, inter- and intra-specific differences cannot be disregarded, as well as doses and exposure times.

Anthropogenic activities have been found to contribute more to environmental contamination (e.g., water eutrophication which was recognized in the middle and late stages of the twentieth century) due to the everyday manufacturing of materials to meet the demands of the population [36, 37], in its various aspects that include, agriculture, industry, and urban areas. As mentioned by [37] human interference is to a greater extent caused by social and economic pressures, which are associated with the largest changes that occurred in agricultural and forest areas as a result of the extensification of agriculture, deforestation, afforestation, and urbanization. In Europe, these are the trends observed over the last years [38, 39].

3.2 Ecological quality ratio (EQR) for phytoplankton

In general, the phytoplankton EQR (Table 1) shows that the Miranda and Pocinho reservoirs had the worst water quality (moderate or lower), taking into account the defined classes for the main water course typology. The Aguieira reservoir tended to have low water quality, with Ag2 and Ag3 being the most problematic sites with the lowest EQR values recorded. All reservoirs were characterized as eutrophic, especially due to high concentrations of P_total recorded over the last few years [8, 9, 10, 12, 13, 14, 31, 35, 40], a condition also observed in the present study (Table 1). The bioavailability of nutrients such as phosphorus favors the overgrowth of phytoplanktonic communities [8, 9, 10, 12, 13, 14], namely, cyanobacteria organisms, that were already associated with poor water quality and recurrently reported blooms was been in these reservoirs [8, 9, 10, 12, 13, 14]. The most prevalent and main group of cyanobacteria detected in all reservoirs was Microcystis. However, Anabaena, Woronichinia, and Pseudanabaena (and others) were also detected, but in a much smaller percentage. The percentage of cyanobacteria detected in Miranda and Pocinho was 54.05, and 57.32%, respectively. Lower percentages of cyanobacteria were observed in the Aguieira sites, namely, 1.07% (Ag1), 4.75% (Ag2), 10.75% (Ag3), and 27.30% (Ag4). However, other phytoplankton indicators are included in the assessment of this reservoir typology (e.g., Algae Group Index (AGI) [8, 10], which is strongly interfering with the final classification of the Ag4 site.

3.3 Water ecotoxicological assessment

The proposal of classes of disturbances (defined by colors) and ecotoxicity results for D. magna, after exposure to treatments of the natural waters of Miranda, Pocinho, and Aguieira reservoirs, are presented in Figures 2–4, respectively. For each figure, the values represent the percentage of inhibition of: A – Feeding rate (% I_FR); B – CAT activity (% I_CAT); C – GSTs activity (% I_GSTs); D – TBARS levels (% I_TBARS); E – AChE activity (% I_AChE), comparatively to the control. Based on the biological responses under study (percent inhibition of different parameters, previously mentioned), ecotoxicity classes were proposed (Figures 2–4) to achieve an approach to the ecotoxicological potential for each sampling site (Figure 5). Based on the criteria to define the equivalent quality potential, to those presented in the WFD, an estimation of the ecotoxicological potential has been suggested.

Figure 2.

Proposal of classes of disturbances (defined by colors) and ecotoxicity results for D. magna, after exposure to treatments of the natural waters of Pocinho reservoir (NF—Non-filtered water; F1 and F2—Filtered with 1.2 μm and 0.22 μm respectively). The values represent the percentage of inhibition of: A – Feeding rate (% I_FR); B – CAT activity (% I_CAT); C – GSTs activity (% I_GSTs); D – TBARS levels (% I_TBARS); E – AChE activity (% I_AChE), comparatively to the control. Different letters (a, b, and c) stand for significant differences between treatments, detected by the Tukey test (p < 0.05).

Figure 3.

Proposal of classes of disturbances (defined by colors) and ecotoxicity results for D. magna, after exposure to treatments of the natural waters of Miranda reservoir (NF—Non-filtered water; F1 and F2—Filtered with 1.2 μm and 0.22 μm respectively). The values represent the percentage of inhibition of: A – Feeding rate (% IFR); B – CAT activity (% ICAT); C – GSTs activity (% IGSTs); D – TBARS levels (% ITBARS); E – AChE activity (%IAChE), comparatively to the control. Different letters (a and b) stand for significant differences between treatments, detected by the Tukey test (p < 0.05).

Figure 4.

Proposal of classes of disturbances (defined by colors) and ecotoxicity results for D. magna, after exposure to treatments of the natural waters of Aguieira reservoir (NF—Non-filtered water; F1 and F2—Filtered with 1.2 μm and 0.22 μm respectively). The values represent the percentage of inhibition of: A – Feeding rate (% I_FR); B – CAT activity (% I_CAT); C – GSTs activity (% I_GSTs); D – TBARS levels (% I_TBARS); E – AChE activity (% I_AChE), comparatively to the control. Different letters (a, b, and c) stand for significant differences between treatments, detected by the Tukey test (p < 0.05).

Figure 5.

Ecotoxicological potential of the sampling sites, according to natural water treatments (NF, F1, and F2), ecotoxicity results (defined in previous Figures 2–4), and ecological potential according to WFD parameters (Table 1).

For the parameters, feeding rate (FR), and TBARS levels (A and D for each Figures 2–4) only two classes of ecotoxicity were defined: non-disturbed (green) and disturbed (yellow). Different aspects of Daphnia biology, as feeding rate is affected by quality (i.e., the number of organic compounds and carbon/nitrogen/phosphate ratio) and quantity of available food [41]. In all reservoirs and sites studied, regarding the parameter of feeding rates (A of Figures 2–4), all reservoirs are characterized as not disturbed, since the feeding rates were positive, especially after filtration treatments. However, we draw particular attention to the Ag3 site, where a significant decrease in the percentage of inhibition of the feeding rate, between NF and F2 was observed. This means that, after the filtrations, seston components were removed, including suspended particles, phyto- and zooplanktonic elements, and bacteria, which could be interfering with the feeding capacity of Daphnia magna. Lari et al. [41] hypothesized that in addition to physical cues (e.g., concentration and physical properties of seston composition in the water), Daphnia detect and uses chemical cues, using their chemosensory system, to locate the most nutritious patches of food in the surrounding environment.

According to [9], lipid peroxidation (LPO) measured as TBARS levels were the most responsive biomarker, in the evaluation of the water quality of the reservoirs under study. If we consider that LPO corresponds to the chain of reactions of oxidative degradation of lipids, resulting in cell damage (e.g., tissue damage), in which free radicals “steal” electrons from the lipids in cell membranes, the distinction of only two classes of ecotoxicities seemed to us to be the most correct and coherent form (D of Figures 2–4). We considered that this biological response, that is, the occurrence or not of oxidative damage, results in the generality of the ability of antioxidant defenses to act to prevent, avoid, or neutralize this oxidative damage by free radicals. Organisms can adapt to increasing free radicals (as reactive oxygen or nitrogen species) production by upregulating antioxidant defenses, such as the activities of antioxidant enzymes (e.g., CAT, GSTs, among others) [42]. Failure of antioxidant defenses to detoxify excess free radicals production can also lead to significant enzyme inactivation, protein degradation, DNA damage, and lipid peroxidation [42]. In particular, LPO is considered to be a major mechanism, leading to impaired cellular function and alterations in physicochemical properties of cell membranes, which in turn disrupt vital functions of D. magna, such as growth, longevity, and reproduction but also feeding behavior. Therefore, an increase in LPO was considered a negative consequence, representing oxidative damage; and a significant decrease will be a positive consequence, that is, the nonoccurrence of oxidative damage, which may be associated with several pathways that avoided, prevented, or neutralized it.

Relatively to the other parameters analyzed (enzymatic activities: B, C, and E of Figures 2–4), based on the biological responses under study (percentage of inhibition), five ecotoxicity classes were proposed, as represented in all figures. Based on the criteria to define the equivalent quality potential, to those presented in the WFD, an estimation of the ecotoxicological potential has been suggested. Classes of ecotoxicity have been defined and to facilitate the analysis of global results, different colors were assigned to each class, according to the ecotoxicity degree of the percent inhibition of the parameter under evaluation. For the present work, we consider the following ranges of values (%) and respective ecotoxicity classes: ≤ − 5 to ≥5 (non disturbed—blue); ≤ − 5 to −30 and ≥ 5 to 30 (slightly disturbed—green); ≤ − 30 to −60 and ≥ 30 to 60 (marginally disturbed—yellow); ≤ − 60 to −90 and ≥ 60 to 90 (moderately disturbed—orange); ≤ − 90 and ≥ 90 (highly disturbed—red). The definition of this range of ecotoxicity classes had as main influences the percentage of effect of 10, 50, and 90% (values with high significance in ecotoxicology), as previously reported by [10]. The range of purposed ranges of ecotoxicity was adjusted, whereby equivalent variations were defined with five ecotoxicity classes, as suggested in the works by [10, 11]. Roig et al. [11] considered an approach to evaluate the ecotoxicological status of rivers (Ebro River watershed, NE Spain), in which the ecotoxicity of pore water has been evaluated in several models organisms, including D. magna. Rodrigues et al. [10] and Roig et al. [11] also proposed five classes of ecotoxicity, based on different endpoints, since they evaluated the effects in several aquatic organisms. Roig et al. [11] for D. magna, this range was demarcated according to the EC50 values and was expressed as % dilution, for pore water assays, from nontoxic (>100) and highly toxic (<10). Rodrigues et al. [10] defined five ecotoxicity classes for R. subcapitata, and this range was defined according to the percent inhibition of yield, from non perturbed (≥ − 10) and highly perturbed (<−90).

Enzymes (e.g., CAT, GSTs, AChE) are proteins that catalyze non-spontaneous chemical reactions in different metabolic pathways, with different physiological functions. CAT is an antioxidant enzymatic defense, GSTs have a dual role in detoxification but also antioxidant defense, and AChE is involved in the neurotransmission process. Enzyme and substrate concentrations influence the reaction rate, altering their activities, which can be significantly inhibited or stimulated [43] after different compound exposure. Antioxidant enzymes (e.g., CAT and GSTs) can be induced by increasing the production of reactive oxygen species (ROS) as a protection mechanism against oxidative stress (adaptation to stress resulting from directly or indirectly generating ROS). In contrast, they can be inhibited when deficiency of the system occurs, inducing a precarious state, making organisms more susceptible to toxic agents (e.g., significant enzyme inactivation or protein degradation by toxicants, or due to the potential attack by excessive concentrations of free radicals) [42, 43]. For example, Hg concentrations (0.08, 0.4, and 2 μg/L) promote perturbations in antioxidant enzymes (e.g. superoxide dismutase; glutathione peroxidase; glutathione reductase; and GSTs) and generate oxidative stress/damage indirectly by binding to antioxidant enzymes containing the thiol group and resulting in depletion of nonenzymatic antioxidant GSH, a scavenger of ROS, for 24 h and 48 h, in neonates and juveniles of D. magna [34]. This study corroborates our work, as the quantified Hg concentrations varied between 0.63 and 1.85 μg/L. However, we cannot neglect the mixture of potential compounds present, as well as their interactions and other features of water. Several factors can alter the catalytic activity of enzymes. Altogether, they reflect the current metabolic situations and trigger changes in the inherent characteristics of the enzyme and its interaction to promote or impede enzymatic reactions. Factors such as pH, temperature, effectors, and inhibitors (e.g., chemical compounds dissolved in water) can modify the enzyme concentration and/or conformation but also the substrate concentrations, influencing the reaction rate, and altering its catalytic activity.

AChE is an enzyme involved in the physiological hydrolytic degradation of the neurotransmitter acetylcholine (ACh), in cholinergic synapses and neuromuscular junctions of most organisms [29], and as such, it is indispensable for the normal functioning of the nervous system (neuromuscular transmission) [44]. This biomarker is used as an indicator of neurotoxicity since it results in severe neurotransmission impairment, which leads to ACh accumulation at synaptic clefts, causing nervous overstimulation and eventually death [45]. Changes in normal neurotransmission may have adverse impacts on key functions, such as food consumption, energy metabolism, growth, and reproduction; ultimately, the impairment of neuronal transmission may result in the death of exposed organisms [46]. The US EPA [47] suggests that a significant AChE activity alteration by 20% or more can be considered a clear toxicological effect of stress exposure. However, we agree that the greater the effect on this biomarker, the worse the final consequence, in terms of the aforementioned sub-individual effect (e.g., food consumption, growth, reproduction, and escape from predators). Another factor affecting AChE activity is allosteric control, which can involve stimulation of enzyme action as well as inhibition. Allosteric stimulation and inhibition allow the production of energy and materials by the cell when they are needed and inhibit production when the supply is adequate [48]. The rate of an enzymatic reaction increases with increased substrate concentration, reaching maximum velocity when all active sites of the enzyme molecules are engaged. The cholinergic system plays a major role in the neurotransmission process, and the simultaneous stimulation of nicotinic and muscarinic receptors by ACh may be necessary to synchronize and balance ionic and metabolic events within cells, which are perturbed [49]. Thus, an increase in AChE activity can be associated with perturbations in several metabolic pathways, which can be mediated by ACh. External factors such as food supply, ambient temperature, and water quality (e.g. contaminants mixture) can also alter the activity of cholinesterases [50]. These factors impair the determination of the “normal” activity of ChE and thus hinder the identification of “abnormal” activity, including that caused by anticholinesterases [50].

3.3.1 Ecotoxicity results from the reservoir and its relationship with WFD parameters (physical and chemical elements, chemical analysis, and biological element)

3.3.1.1 Miranda reservoir

Figure 2 corresponds to the proposal of classes of disturbances (defined by colors) and ecotoxicity results for different parameters quantified in D. magna, after exposure to treatments of the natural waters of Miranda reservoir. Based on the WFD parameters previously discussed, the feeding rates of D. magna exposed to natural waters from Miranda (NF) were not affected (Table 2) by the high presence of cyanobacteria, nor by the levels of total phosphorus, or the concentrations of zinc and mercury, which were recorded above the EQS. Even after the filtration treatments (F1 and F2), these rates were not altered and D. magna did not show disturbances in terms of feeding rates (Figure 2A; Table 2). However, in sub-individual terms (biomarkers), regarding CAT activity, we found that with the application of treatments with filtrations (F1 and F2) there was an increase in oxidative stress (Table 2), which resulted in a worse classification class in F2 treatment (moderately disturbed; Figure 2B). These results are supported by previous work [8, 9], since higher concentrations of phosphorus can promote an overgrowth of phytoplanktonic organisms [9] and, consequently, the blooms of cyanobacteria as Microcystis, which fact confirms our findings, in this work. In turn, these cyanobacterial blooms can result in dangerous levels of toxins such as microcystin-LR toxic to D. magna [9, 51, 52, 53]. Furthermore, it is important to note that the observed variations in CAT activity may still be associated with mercury and zinc levels above the Eq. [9], which may be more bioavailable for D. magna after F2 treatment since only the seston components were removed.

	FR			CAT			GSTs			TBARS			AChE
	d.f.	F	p	d.f.	F	p	d.f.	F	p	d.f.	F	p	d.f.	F	p
M	2, 13	0.020	0.981	2, 8	10.13	0.012	2, 8	5.554	0.043	2, 8	7.186	0.026	2, 8	7.282	0.025
P	2, 14	3.163	0.079	2, 8	5.803	0.040	2, 8	33.136	0.001	2, 8	47.753	<0.001	2, 8	6.777	0.029
Ag1	2, 14	1.522	0.258	2, 8	4.756	0.048	2, 8	5.659	0.154	2, 8	0.870	0.466	2, 8	1.347	0.329
Ag2	2, 14	0.065	0.938	2, 8	21.056	0.002	2, 8	8.066	0.020	2, 8	87.577	<0.001	2, 8	2.425	0.169
Ag3	2, 14	4.791	0.030	2, 8	2.740	0.143	2, 8	6.310	0.033	2, 8	99.326	<0.001	2, 8	1.054	0.405
Ag4	2, 14	2.889	0.095	2, 8	4.756	0.048	2, 8	2.596	0.154	2, 8	408.549	<0.001	2, 8	3.646	0.092

Table 2.

ANOVA summary table (test differences between natural water treatments – NF, F1, and F2) for the D. magna feeding rate (FR), CAT and GSTs activities, TBARS levels, and AChE activity, for Miranda, Pocinho and Aguieira. For each one, the degrees of freedom (d.f.), F statistics, and associated p-value was shown. Bold values stand for statistically significant differences. Significant values, (after Tukey test, p < 0.05), were represented in the figures with different letters (a, b, c).

Regarding the activity of GSTs, there was an improvement in the ecotoxicological classification, with the application of the F2 treatment (slightly disturbed; Figure 2C). As mentioned in previous studies, changes in the activities of antioxidant enzymes, such as CAT and GSTs, may be associated with the physiological responses of organisms to environmental adaptations, through the influence of phyto and zooplanktonic communities and suspended particles [9]. Chemical analyzes showed very low levels of most quantified contaminants, except for mercury and zinc. On the other hand, as suggested by [54] nutrients seem to be very important in controlling the performance of D. magna, and in fact, this corroborates the results observed in the F2 treatment, in the case of GSTs activity.

The results of the TBARS levels showed significant differences between NF and F2 treatments, although some degree of oxidative damage is still observed (Figure 2D; Table 2). This may indicate that the samples contained some type of disturbing and oxidizing agent, and potentially triggered oxidative stress (previously discussed), with a consequent increase in peroxidative damage (LPO). As mentioned earlier, high amounts of cyanobacteria present in this site, concomitantly with high concentrations of phosphorus and in addition to the high levels of mercury, can be associated with the results of TBARS levels. The accumulation of nutrients (e.g., phosphorus total) in Miranda reservoir can lead to eutrophication causing abnormal growth of the primary producers, which can compromise the quality and balance of the aquatic ecosystem, including the balance between biochemical pathways and physiological functions of organisms, as mentioned by [9, 53].

The results of the AChE activity showed significant differences between NF and F2 in the water treatments, although some degree of neurotoxic alteration was still observed between NF and F2 treatments, an improvement was observed considering the associated ecotoxicological class (Figure 2E; Table 2). AChE activity stimulation and inhibition allow the production of energy and materials by the cell when they are needed and inhibit production when the supply is adequate [48], and in this study seems to have been affected by the seston components, as can be seen from the changes between NF and F2. A direct relationship between the degree of AChE inhibition and toxicity might not always be expected. The reason for such variability can mainly be attributed to biological differences between species that include AChE sequence differences as well as differences in molecule affinities for the AChE-active site.

3.3.1.2 Pocinho reservoir

Figure 3 corresponds to the proposal of classes of disturbances and ecotoxicity results for different parameters quantified in D. magna, after exposure to treatments of the natural waters of Pocinho reservoir. Similar to what was observed for Miranda the feeding rates of D. magna exposed to natural waters from Pocinho (NF) were not significantly affected (Table 2) by the high presence of cyanobacteria (57.32%), nor by the levels of total phosphorus, or the concentrations of zinc and mercury, which were recorded above the EQS (Table 1). Even after the filtration treatments (F1 and F2), these rates were not significantly altered, which shows that in terms of feeding behavior, D. magna did not show significant disturbance in terms of feeding rates, despite apparent differences between treatments (Figure 3A). However, in sub-individual terms (biomarkers), regarding CAT activity, we found that with the application of treatments with filtrations (F1 and F2), an improvement in the ecotoxicological category was observed between the NF and F1 treatments, but the classification from F1 to F2 worsened again (Figure 3B). However, in general, there was an improvement between NF and F2, although not significant (Table 2), but sufficient to decrease the ecotoxicological category from moderately disturbed to marginally disturbed. Thus, at this location, we can see a potential negative interference of the various components of seston in CAT activity, which were removed by the F1 treatment, essentially highlight by the percentage of cyanobacteria. In F2, we may have a potential influence of the high concentration of zinc (80.7 μg/L), compared to the other sites and reservoirs (Table 1). In the absence of other biological communities, zinc bioavailability may be greater and result in increased toxicity to D. magna. Zn toxicity thresholds of D. magna can alter by a factor > 10 as a result of ecological interactions and are highly dependent on P_total and pH value, with the lowest Zn thresholds found in higher-P and higher-pH waters [55]. However, only a few cases corroborate this finding in this work, and Pocinho does not fit into this perspective. Furthermore, Fettweis et al. [55] evaluate the effects of 25 to 310 μg/L of Zn and pH 7.3 and 7.8 on 21-d daphnid population size and they concluded that the indirect effects of Zn via producer-consumer relationships can outweigh the direct toxic effects. According to the mentioned work, a higher phytoplankton Zn sensitivity at higher pH, affecting food supply to D. magna, and an increased algal P content at higher Zn, offering a nutritional benefit to daphnids that counteracts direct Zn toxicity under P limitation [55]. These explanations can help to understand what happened between NF for F1 and F2, not only in sub-individual but also individual responses (feeding rate).

Regarding GSTs activity (Figure 3C), a worsening was observed between NF and filtration treatments (F1 and F2). In this case, we can refer to a potential greater bioavailability of metal levels (Zn and Hg or others), in the absence of all seston components, essentially the various biological communities in F2 treatment. These results (Table 2) may indicate a potential interference of GSTs, in the antioxidant defense or detoxification of eventually dissolved compounds, which increased along with the water treatments, from NF to F2.

The levels of TBARS (Figure 3D; Table 2), all previous findings, both in terms of analysis of our results (potential greater bioavailability of dissolved compounds in F2 treatment for Daphnia, which is a filtering organism) and by comparison with other studies [9, 55], are reflected in this parameter, since it obtained a worse classification (disturbed), after F2 treatment, comparatively with NF treatment. In fact, Rodrigues et al. [9] had already selected TBARS levels as a relevant parameter in ecotoxicological assessment studies of water quality with D. magna. Then, since oxidative damage (TBARS levels) is an indicator of LPO as a potential consequence of oxidative stress, we can infer that it may be associated with the functional inefficiency of antioxidant enzymes. However, the changes observed in the enzymatic activities of CAT and GSTs support this fact.

The results regarding the AChE activity show an improvement between NF and F2, wherein in this last treatment we achieved the classification of not disturbed. These results indicate that the seston components (present in NF) could be causing some degree of neurotoxic stress, but when organisms were exposed to F2 treatment, these effects were mitigated. In this case, based on EQS values, the altered WFD parameters (pH, P_total, Zn, and Hg), did not show potential toxicity in terms of neurotransmission.

3.3.1.3 Aguieira reservoir

Figure 4 corresponds to the proposal of classes of disturbances (defined by colors) and ecotoxicity results for different parameters quantified in D. magna, after exposure to treatments of the natural waters of Aguieira reservoir. Similar to what was observed for the studied Douro River reservoirs (Miranda and Pocinho), the feeding rate parameter, in all treatments demonstrated a classification of not disturbed for all studied locations (Figure 4A). Apart from the Ag3 site, no significant differences in this parameter, among the treatments were recorded (Table 2). This site presented characteristics that differentiated them from the other Aguieira sites, namely the higher levels of total phosphorus recorded (0.08 mg P/L). Furthermore, higher percentages in terms of cyanobacteria biovolume (10.75%) were observed at this site. In fact, this site was the one that showed the greatest concern in previous studies [8, 10, 12] due to the high nutrient levels, and other metal elements (Zn, Hg), conditions also observed in the present study. In this sense, eutrophication in this reservoir, reported by other authors [9, 10, 12, 31, 40], has been a concern in terms of water quality, and in fact, supports our results. Indeed Ag3 site, with the application of filtration treatments (mainly F2), showed an improvement in feed rates, which reinforces the idea that the presence of several seston components is potential stress inductors for D. magna that can be reflected in food behavior.

Contrary to expectable, and about the antioxidant defense biomarker, catalase activity (Figure 4B; Table 2), a worse classification was observed in the locations Ag1 and Ag4 (in general moderately disturbed), compared to Ag2 and Ag3. In Ag2 and Ag3, a classification of slightly disturbed was observed after filtrations (e.g., F2 treatment). This classification is taking into account the seston components; on the other hand, Ag1 and Ag4 report different scenarios. These sites presented high pH values, and levels of Zn and Hg higher than defined EQS (Table 1). In addition, an ecological potential, taking into account biological EQR, the classification was good or more (Table 1). For this site, we observe that from NF to F2 there was an increase in oxidative stress, possibly associated with greater availability of dissolved compounds (besides quantified) to D. magna, in F2 treatment. In fact, the results of GSTs activity support this potential finding (Figure 4C), but only for Ag1. According to Rostern [30], the pH and concentration of inorganic metal ions (e.g., Zn and Hg) are key factors for toxicity in the physiology and behavior of aquatic organisms. In previous studies, metals such as Zn and mainly Hg are known to be inhibitors of enzymes’ activities and can disrupt antioxidant defenses [9, 30, 34, 35]. However, some care must be considered when comparing sites in the same reservoir, as we only analyze a tiny part of potential dissolved compounds. Moreover, the biological responses observed in the biomarkers, represent the results of an integrated response to complex mixtures. Furthermore, the land occupation also represents an important source of variation of contaminants input, and consequently biomarkers’ response. In addition, land occupation is different in each site and also between reservoirs (Figure 1).

Regarding the TBARS levels (Figure 4D), despite significant variations between treatments (Table 2), in the various sites, the results allowed classifying all treatments as not disturbed. Based on this evidence, we can observe that the antioxidant defenses, despite being altered and indicating potential oxidative stress associated with natural water treatments, in terms of lipid peroxidation, were able to prevent the occurrence of oxidative damage in exposed organisms. In fact, we cannot consider only the antioxidant defenses involved in this work. Other unquantified defenses and metabolic pathways may have acted to neutralize and prevent the occurrence of lipid peroxidation, which could indicate disturbances in cell membranes, which were not observed.

The activity of acetylcholinesterase, despite showing high variations in terms of percentage of inhibition (Figure 4E), was not significant between treatments (Table 2), for organisms exposed to treatments, with water samples from Aguieira site. The nonoccurrence of significant differences may be associated with high differences between replicates for the same treatment. The different toxicity for organisms exposed to the same conditions, but belonging to different replicates, can be considered if we considered the intraspecific variations [56]. In this sequence, the authors refer that the origin of population, animal body size, and pre-exposure history [e.g., organisms from different cultures, different broods (although meeting the criteria of the assay guidelines)] are realistic variables for zooplankton populations that cause different acute toxicities in D. magna [56]. If we consider, for example, the results of the mercury concentrations detected (Table 1), potential neurotoxicity would be expected. Mercury is known to be a neurotoxin that causes structural damage to the brain and inhibits enzymes’ activities needed for normal neurotransmission [30]. Tsui and Wang [56] suggested that acute Hg toxicities were not simply caused by the different Hg body burdens, and several other mechanisms may operate to result in such a varied Hg toxicity (e.g., reduction of Hg uptake, enhancement of intrinsic tolerance, and increase of antioxidant/detoxification activity). The same authors also measured the metal concentrations in water and living D. magna and the results provided useful information to explain whether the apparent tolerance modification was due to a change in metal accumulation and/or to a change in other subtle parameters (e.g., intrinsic tolerance and detoxification activity), which can also alter the neurotoxicity results.

3.3.1.4 Physicochemical and ecological potential vs. ecotoxicological approach

One of the fundamental using biomarkers in ecological risk assessment is based on their potential ability to anticipate effects at higher levels of biological organization. Due to the different sensitivity between ecotoxicological tools evaluated in this work, and the presence of some confounding factors that could play an important role in the final ecotoxicity evaluation, the result of the final ecotoxicological potential has been calculated as the worst classification of all biochemical parameters quantified in D. magna (Figure 5). According to this methodology, CAT and AChE activities parameters were the ones that most contributed to the final ecotoxicological potential. Except for Miranda, for the final NF treatment, in which the yellow color resulted from the parameters GSTs activity and TBARS levels. If we consider the totality of the data, we do not observe an improvement in the ecotoxicological potential, with the filtration treatments, with a few exceptions. The only site where this improvement was evident was Ag2, which is corroborated by previous studies conducted in the same site but using the microalga R. subcapitata [10]. As was done in the previous study, and due to the different sensitivity between WFD parameters (physicochemical, chemical, and biological elements) individually considered, the final ecological potential according to WFD parameters has been calculated. This approach is more in line with the results of the ecotoxicological potential, since considering the contamination classes of ecotoxicity and respective colors, they present greater similarities (Figure 5) with NF treatment (with all components), with the site Ag3 representing an exception. Several reasons were pointed out throughout the manuscript, evidenced by this site, compared to the remaining Aguieira sites reservoir, or even with Miranda and Pocinho. In general, when we considered the sampling points where the ecological potential qualification was bad or poor the similarities between ecologic and ecotoxicological potential were fully agreed upon. The results of the present work allow us to confirm that, when chemical stressors or seston components affect the organism’s homeostasis, an ecotoxicological approach, provided by suitable ecotoxicological tools, could detect these changes with accurate sensitivity. In fact, Rodrigues et al. [9] already demonstrated that feeding bioassays and biomarkers (e.g., antioxidant defense and TBARS levels) proved to be useful and reliable tools in the assessment of water quality. Notice that an accurate battery of ecotoxicological tools is a direct measure of organism functional responses, and they could have more impact on the decision-making process than criteria based on concentrations of chemicals or other physical and chemical parameters, as previously demonstrated by [10, 11].

4. Conclusions

The results of the current case study corroborate that cost-effective and rapid screening short-term ecotoxicological tools, performed with natural waters, using the model organism D. magna, could be useful to complement the determination of the water ecological potential of reservoirs. In the case study of Portuguese reservoirs, ecotoxicological tools evaluated (feeding bioassay and biomarkers) have been performed obtaining good sensitiveness and complementarity between methodologies, in most situations (intra- and inter-reservoirs). Moreover, high coincidences with the ecological potential, recognized following the WFD parameters evaluation, have been found especially when ecosystems’ disturbance due to several stressors was observed (e.g., seston components as some phytoplanktonic organisms as cyanobacteria, presence of organic pollutants, and metals). For future comparative studies, we also suggest the evaluation of the water treatments like those performed in this work, mainly the F2 (water filtered through a sterile filter system with a porosity of 0.22 μm), to evaluate seston quality, as it proved to be an important source of stress for D. magna. This set of biological responses has presented good concordance with the ecological potential of the reservoirs. These results encourage working further on the applicability of cost-effective ecotoxicity tests and early warning tools for the evaluation of water quality and their integration into the current monitoring programs.

Acknowledgments

This work was supported by National Funds (through the FCT—Foundation for Science and Technology) and by the European Regional Development Fund (through COMPETE2020 and PT2020) through the research project ReDEFine (POCI-01-0145-FEDER-029368) and the strategic program UIDB/04423/2020 and UIDP/04423/2020. Sara Rodrigues and Sara Antunes are hired through the Regulamento do Emprego Científico e Tecnológico – RJEC from the Portuguese Foundation for Science and Technology (FCT) program (2020.00464.CEECIND and CEECIND/01756/2017, respectively).

Conflict of interest

The authors declare no conflict of interest.

Ethical approval

All applicable international, national, and/or institutional guidelines for the care and use of animals were followed.