Sensitivity of freshwater species under single and multigenerational exposure to seawater intrusion

Abstract

Salinization of coastal freshwater ecosystems is already occurring in some regions of the world. This phenomenon raises serious concerns on the protection of coastal freshwater ecosystems, since many of them support and shelter a large number of species and are considered hotspots of biodiversity. This work intended to assess the adverse effects that salinization, caused by the intrusion of seawater (SW), may pose to freshwater organisms. In this study, three specific goals were addressed: (i) to assess if sodium chloride (NaCl) may be used as a surrogate of natural SW at early-stages of risk assessment; (ii) to identify the most sensitive freshwater species to salinity NaCl; and (iii) to determine if increased tolerance to salinity may be acquired after multigenerational exposure to low levels of salinization (induced with NaCl). A total of 12 standard monospecific bioassays were carried out by exposing organisms from different taxonomic groups (Cyanobacteria: one species, Tracheophyta: two species, Rotifera: one species, Arthropoda: two species and Mollusca: one species) to a series of concentrations of NaCl (ranging from 0.95 to 22.8 mS cm^–1) or dilutions of SW (ranging from 1.70 to 52.3 mS cm⁻¹). In general, NaCl exerted similar or higher toxicity than SW, both at lethal and sublethal levels, suggesting that it may be proposed as a protective surrogate of SW for first tiers of salinization risk assessment. Among all tested species, the cyanobacterium Cylindrospermopsis raciborskii, the daphnid Daphnia longispina and the rotifer Brachionus plicatilis were the most sensitive taxa to salinization (EC₅₀ ≤ 4.38 mS cm⁻¹). Given their position at the basis of the food web, it is suggested that small increments of salinity may be enough to induce structural changes in freshwater communities or induce changes in trophic relations. No clear evidences of increased tolerance after multigenerational exposure to low levels of salinity were found.

This article is part of the theme issue ‘Salt in freshwaters: causes, ecological consequences and future prospects’.

1. Introduction

Seawater (SW) intrusion in coastal freshwater ecosystems is already occurring in several regions of the globe (e.g. [1–4]). Following the projections made by the Intergovernmental Panel on Climate Change (IPCC) on climate change, it is expected that other areas will probably be affected in the near future [5]. SW intrusions have been causing an increase in the salinity of freshwater coastal ecosystems [6] and it is expected for that phenomenon to occur more frequently in the coming years [7]. Salinization of freshwater systems may occur through both surface flooding (which may be the most probable scenario under extreme weather events) and/or groundwater intrusion (predicted to occur more gradually owing to the depletion of groundwater supplies and the consequent retreat of the SW/freshwater interface) [5].

Depending on its intensity and duration, increased salinity events may cause several adverse effects in freshwater organisms, including mortality [8,9] or alterations in life cycle traits that impair fitness, like reduced growth or reproduction rates [10,11], or metabolic costs [12]. These effects, observed at the individual level, may translate into adverse consequences at higher levels of biological organization, including the disruption of ecosystem functions and, subsequently, the services it provides [13,14].

Following the need to set protective environmental levels regarding SW intrusion, it is then necessary to identify the most sensitive group of organisms to salinity. However, the toxicity studies on sea salts, particularly sodium chloride (NaCl), have reported highly heterogeneous results, with effects ranging from very low levels—e.g. 50% reduction in growth rates of Pseudokirchneriella subcapitata and Cylindrospermopsis raciborskii at 0.87 and 1.46 g l⁻¹ of NaCl (1.72 and 2.89 mS cm⁻¹ at 20°C) [15]—to very high levels—e.g.: increased mortality of Australian endemic freshwater fishes [8] at 8.2 or 14.6 g l⁻¹ of NaCl (approx. 16.3 and 28.9 mS cm⁻¹ at 20°C, respectively) [8] or decrease in growth of the freshwater macrophyte Lemna gibba above 22 g l⁻¹ NaCl (approx. 43.6 mS cm⁻¹ at 20°C, respectively) [16]. This wide tolerance to salinity, exhibited by freshwater species, highlights the need to identify freshwater species at most risk.

The majority of studies addressing salinity effects on freshwater biota used mostly specific salts (NaCl, MgCl₂, among others) in monospecific standard toxicity tests [17–21]. It is still unclear whether such surrogate salts may be safely used for the preliminary risk assessment framework of SW intrusion because available literature comparing toxicity data from surrogate salts and toxicity data of SW is still scarce [10,22–24]. Venâncio et al. [24] found that NaCl elicits effects at lower or similar concentrations than natural SW for six clonal lineages of Daphnia longispina. In the same line, Ghazy et al. [10] showed marked differences in the median lethal concentration (48 h LC₅₀ for Daphnia magna) across NaCl (3.21 g l⁻¹, 5.92 mS cm⁻¹), synthetic SW (approx. 4.29 g l⁻¹ of NaCl; 7.76 mS cm⁻¹) and filtered natural SW (9.54 mS cm⁻¹). However, this profile was altered after acclimation to salinity, with synthetic SW becoming the most toxic form [10]. This study demonstrates there is still some uncertainty in the risk assessment of salinization, since benchmark values can vary substantially by choosing to use artificial or natural SW.

Adding to the above, most studies involved exposure to salinity in a single generation, but, in the field, exposure occurs over various generations. In fact, both exposure to pulsed SW intrusions (e.g. extreme climate phenomena) and continued exposure (e.g. through groundwater intrusion) are of relevance, especially for short-lived and multivoltine biota. Therefore, it is important to understand if organisms are capable to acclimate to increased salinity, a phenomenon already reported for algae [23] and daphnids [18,25]. To better cope with salinization episodes, acclimation may occur via short-term compensatory physiological/osmoregulatory changes within their life cycle, like vacuolar compartmentalization or ion sequestration in producers [26], or the triggering of an osmoregulatory reaction in cladocerans [27]. More recently, the role of epigenetic phenomena (i.e. phenotypic characteristics that are passed on to organisms of subsequent generations without genetic alteration; e.g. DNA methylation) has also been discussed as a potential mechanism to acquire increased tolerance through multigenerational exposure [28].

This work aimed at addressing the effects of salinization on coastal freshwater ecosystems, by: (i) evaluating the suitability of NaCl as a surrogate of natural SW to be used at early-stages of ecological risk assessment; (ii) identifying the most sensitive freshwater species to increased salinity scenarios; and (iii) determining if increased tolerance to salinity may be acquired after multigenerational exposure to low levels of salinization.

2. Material and methods

(a). Test solutions

NaCl was supplied by Merck (St Louis, MO, USA), and SW was collected at the northwest Atlantic Ocean (40°38'33″ N 8°44'55″ W, Aveiro, Portugal). The collection site was located in front of a Natural Reserve (São Jacinto dunes, created in 1979) and its waters are subjected to periodic monitoring programmes according to the European Union Directive 2006/7/CE. The waters from this location are considered good to excellent quality and, for the past 28 years, this site has received the prestigious Blue Flag award granted by the Foundation for Environmental Education (please see [23]).

The NaCl concentrations were prepared by dissolving the salt in the control medium used for each tested species (please see §2b(ii) of Material and Methods). Only fresh solutions were used to perform the toxicity assays. Collected SW was always filtered through cellulose nitrate membranes of 0.20 µm (ALBET-Hannemuehle S.L., Barcelona, Spain) before being used in toxicity assays, to remove particles in suspension and organisms. The SW dilutions were made by mixing SW with the control medium used for each tested species (please see §2b of Material and Methods).

(b). Test species

The species used in this study were selected on the basis of several characteristics, among which are their sensitivity to chemicals, their easy maintenance under laboratorial conditions and well-known physiology, biology and ecology. Furthermore, they have been extensively used in ecotoxicological studies as experimental models and are representative of different taxonomic and functional groups.

Laboratory cultures of the microalgae Raphidocelis subcapitata (Chlorophyta, Sphaeropleales), Chlorella vulgaris (Chlorophyta, Chlorellales) and Cy. raciborskii (Cyanophyceae, Nostocales), were maintained in Woods Hole Marine Biological Laboratory medium (Woods Hole MBL; [29]) in 250 ml Erlenmeyer flasks, under controlled conditions of light (100 µE m⁻² s⁻¹) and temperature (23°C), according to the Organization for Economic Cooperation and Development (OECD) guideline 201 [30]. The cyanobacterium Cy. raciborskii was maintained at a lower light intensity (≈ 40 µE m⁻² s⁻¹) and all cultures were renewed once per week.

Cultures of the macrophyte Lemna minor and L. gibba (Tracheophyta, Arales) were maintained in 250 ml glass vessels with Steinberg medium [31], at 23°C with a light intensity of 100 µE m⁻² s⁻¹. Cultures were renewed once per week.

Neonates of the freshwater rotifer Brachionus calyciflorus (Rotifera, Ploimida) and the ostracod Heterocypris incongruens (Arthropoda, Podocopida) were obtained after the hatching of cysts available in commercial kits (MicroBioTests, Ghent, Belgium). For B. calyciflorus, the cysts were hatched at 23°C, for 24 h, at a constant light intensity of 3000–4000 lux. For the ostracod, the cysts were left to hatch at 25°C, for 52 h, also at a constant light intensity of 3000–4000 lux. For both species (B. calyciflorus and H. incongruens), hatching was performed in American Society for Testing and Materials (ASTM) moderately hard synthetic freshwater medium [32].

Six clonal lineages of D. longispina (Arthropoda, Cladocera) were maintained in laboratory conditions under room temperature of 20 ± 2°C and 16 L : 8 D photoperiod, in ASTM hard water medium [32] supplemented with an organic additive extract [33]. Cultures were renewed every other day and fed with R. subcapitata (1.5 × 10⁵ cells⁻¹ ml⁻¹ day⁻¹). Organisms used for experiments were born between the 3rd and the 5th broods and were less than 24 h old.

The culture of Chironomus riparius (Arthropoda, Diptera) was maintained with an inorganic fine sediment layer and ASTM hard water medium [32] (proportion 1 : 4), at a room temperature of 20 ± 2°C and a 16 L : 8 D photoperiod cycle. Organisms were fed three times a week with a suspension of commercial fish food TetraMin^® (Tetrawerke, Melle, Germany) [34].

The snail Theodoxus fluviatilis (Mollusca, Cycloneritimorpha) was collected at the Anços River spring (for further details on this location see [35,36]), where they are abundant in stream-bed stones. The organisms were transported to the laboratory in refrigerated chambers with local water. Prior to use in ecotoxicity assays, organisms were acclimated to laboratorial conditions (room temperature of 20 ± 2°C and 16 L : 8 D photoperiod cycle) and to the test medium (ASTM; [32]), at least for one week (please see [37]).

(c). Toxicity assays

Lethal and sublethal toxicity assays were carried out with eight freshwater species, representing different taxonomic and functional groups. A summary of the procedures employed to perform each of the bioassays described within this section is given in table 1.

Table 1.

Summary of procedures, range of concentrations and additional information relatively to the test species. (Light (L) : dark (D) cycle, natural seawater (SW), sodium chloride (NaCl), after multigenerational exposure to NaCl (NaCl_GE), conductivity causing 50% of mortality (LC₅₀) or another effect (EC₅₀). n.a., not applicable.)

test species	endpoint	acclimation concentration	dilution water	test conditions	test water conductivity (mS cm−1)		geometric factor	replicates (rep) organisms/rep
Chlorophyta and cyanobacteria
Chlorella vulgarisa	growth EC50 growth	EC50 growth/4	MBL	24 L : 0 D 23°C 72 h	SW: NaCl: NaClGE:	5.15–19.0 7.10–21.8 4.10–21.8	1.15	3 n.a.
Raphidocelis subcapitataa	growth EC50 growth	EC50 growth/4	MBL	24 L : 0 D 23°C 72 h	SW: NaCl: NaClGE:	5.15–19.0 7.10–21.8 4.10–21.8	1.15	3 n.a.
Cylindrospermopsis raciborskii	growthc EC50 growth	EC50 growth/3	MBL	24 L : 0 D 23°C 10 d	SW: NaCl: NaClGE:	2.72–14.1 1.90–9.81 2.72–14.1	1.2	3 n.a.
Tracheophyta
Lemna minor	growthd EC50 growth	EC50 growth/6	Steinberg medium	24 L : 0 D 23°C 7 d	SW: NaCl: NaClGE:	4.93–52.3 2.80–22.8 2.80–22.8	1.3	3 12 fronds/rep
Lemna gibba	growthd EC50 growth	—	Steinberg medium	24 L : 0 D 23°C 7 d	SW: NaCl: NaClGE:	14.1–52.3 6.15–22.8 —	1.3	3 12 fronds/rep
Rotifera
Brachionus calyciflorus	mortalitye LC50	LC50 /4	ASTM	0 L : 24 D 25°C 24 h	SW: NaCl: NaClGEi:	3.50–13.4 2.74–10.5 2.74–14.7	1.4	6 5 organisms/rep
	reproductione EC50 reproduction	LC50/2	ASTM	0 L : 24 D 25°C 48 h	SW: NaCl: NaClGE:	1.70–7.26 0.95–4.9 2.64–7.26	1.2	8 1 organisms/rep
Arthropoda
Heterocypris incongruens	mortalityf LC50		ASTM	0 L : 24 D 25°C 48 h	SW: NaCl: NaClGE:	1.78–18.2 1.40–14.7 —	1.4	3 10 organisms/rep
	somatic growthf EC50 growth	—	ASTM	0 L : 24 D 25°C 6 d	SW: NaCl: NaClGE:	6.0–10.5 2.43–4.87 —	1.4	3 10 organisms/rep
Daphnia longispinab	mortality LC50	≈ LC50/8	ASTM	16 L : 8 D 20°C 48 h	SW: NaCl: NaClGE:	2.47–13.3 1.67–14.6 2.71–14.6	1.4	4 5 organisms/rep
	feeding EC50 feeding	≈ LC50/8	ASTM	0 L : 24 D 20°C 24 h	SW: NaCl: NaClGE:	1.53–8.54 1.32–5.42 1.76–6.41	1.2	4 5 organisms/rep
	somatic growth EC50 growth	≈ LC50/8	ASTM	16 L : 8 D 20°C 72 h	SW: NaCl: NaClGE:	2.6–6.27 1.96–4.76 1.87–4.88	1.1	10 1 organisms/rep
Chironomus riparius	mortalityg LC50	EC50emergence/6	ASTM	16 L : 8 D 20°C 48 h	SW: NaCl: NaClGE:	10.1–21.0 8.44–17.5 8.44–17.5	1.2	4 5 organisms/rep
	growthh EC50 growth	EC50emergence/6	ASTM	16 L : 8 D 20°C 10 d	SW: NaCl: NaClGE:	7.23–18.0 4.89–12.2 4.89–12.2	1.2	4 10 organisms/rep
	emergenceh EC50 emergence	EC50emergence/6	ASTM	16 L : 8 D 20°C 28 d	SW: NaCl: NaClGE:	6.02–12.5 4.08–10.1 4.08–10.1	1.2	4 10 organisms/rep
Mollusca
Theodoxus fluviatilis	mortalityi LC50	—	ASTM	16 L : 8 D 20°C 48 h	SW: NaCl: NaClSTE:	16.5–44.0 6.04–14.0 —	1.15	4 5 organisms/rep
	feeding post-exposurei EC50 feeding	—	ASTM	0 L : 24 D 20°C 3 h	SW: NaCl: NaClSTE:	6.0–10.5 2.43–4.87 —	1.15	4 5 organisms/rep

^aData generated from Venâncio et al. [23].

^bData pooled from Venâncio et al. [24].

^cOECD [30].

^dOECD [31].

^eRotoxkit F acute (MicroBioTests Inc., Belgium).

^fOstracodToxKit F chronic (MicroBioTests Inc., Belgium).

^gOECD 235 [38].

^hOECD 219 [39].

ⁱCorreia et al. [37].

During the toxicity assays, the following parameters were measured to warrant quality control criteria: salinity/conductivity, pH and dissolved oxygen with WTW (Weilheim, Germany) portable probes 440i, pH330i and OXI 330i, respectively.

(i). Toxicity assays with Chlorophyta and cyanobacteria

Regarding the Chlorophyta species R. subcapitata and Chl. vulgaris, the data generated in Venâncio et al. [23] were used to compute the EC₅₀ values used in this study.

The growth inhibition assay with Cy. raciborskii was conducted following OECD guideline 201 [30] with adaptations in light conditions and time of exposure (table 1; please see [40]). Microalgae exposure was done in 100 ml sterilized Erlenmeyer flasks filled with 50 ml of test solution or control medium (MBL medium; [29]), under 40 µE m⁻² s⁻¹ of continuous light intensity (approx. 3000 lux) and 23°C. Three replicates were carried out for each treatment. Experimental treatments consisted of a control, 10 NaCl concentrations and 10 SW dilutions (table 1). The initial cell concentration in each replicate was 10⁴ cells ml⁻¹. At the end of the assay, cell density (D, cell ml⁻¹) was calculated by counting sample aliquots using a Neubauer Improved Counting Chamber, and average specific growth rate (µ, day⁻¹) was calculated as for green algae (OECD guideline 201; [30]), with equation (2.1). The assay duration was 10 days:

2.1

where D_b is the cell density at the end of the assay, D_a is the initial cell density and t_b − t_a is the exposure time interval (10 days).

(ii). Toxicity assays with Tracheophyta

A 7 day growth assay was run with the macrophytes L. minor and L. gibba according to OECD guideline 221 [31]. For both species, at the beginning of the assay three colonies of similar size, totalling 12 fronds, were introduced per replicate flask. This consisted of a 150 ml sterilized Erlenmeyer flask filled with 100 ml of test solution or Steinberg medium (control) (table 1). Experimental treatments consisted of a control, nine NaCl concentrations and 10 SW dilutions for L. minor (table 1) or six NaCl concentrations and six SW dilutions for L. gibba (table 1). Three replicates were performed per treatment. Exposures were carried out at 23°C and continuous light (100 µE m⁻² s⁻¹) (table 1). At the end of the assay, the total number of fronds was counted in each vessel and the dry weight of the macrophytes was estimated after drying at 60°C for 24 h. Average growth rates for both species (µ, day⁻¹) were calculated according to equation (2.1).

(iii). Toxicity assays with Rotifera

Mortality (24 h) and reproduction (48 h) assays were performed with the rotifer B. calyciflorus. Both assays were performed according to the standard procedure for the Rotoxkit F (MicroBioTests, Ghent, Belgium) in multiwell plates, in total darkness at 23°C. For the mortality assay, five replicates with five newly hatched rotifers were assigned to each treatment: control (ASTM medium), five NaCl concentrations and five SW dilutions (table 1). An organism was considered dead if it did not exhibit any movement within 5 s of observation after gentle agitation of the medium. Reproduction assays were carried out under the same conditions as the mortality assays, with eight replicates being performed per treatment: control (ASTM medium), 10 NaCl concentrations and nine SW dilutions (table 1). At the beginning of the assay, one rotifer was introduced per well. After 48 h of incubation of the test plates, the total number of swimming organisms was counted.

(iv). Toxicity assays with Arthropoda

The 6 day mortality and somatic growth assay with the ostracod H. incongruens followed the standard operation procedure for the Ostracodtoxkit F (MicroBioTests). Assays were run in total darkness at 25°C. Four replicates with 10 ostracods each were set per treatment: control, eight NaCl concentrations and eight SW dilutions (table 1). After the 6th day of incubation of the test plates, all swimming organisms were retrieved and measured. Somatic growth rate of the ostracods was determined through measurement of the length of the organisms at the beginning and end of the test. Measurements were performed under the stereomicroscope (MS5; Leica Microsystems, Houston, TX, USA), using a calibrated eyepiece micrometre.

Each endpoint assessed for D. longispina was obtained by pooling data from Venâncio et al. [24]. Lethal and sublethal values presented here are the mean lethal and sublethal values obtained in Venâncio et al. [24] for six clonal lineages of D. longispina. The intention was to increase the number of species represented on the identification of the most sensitive freshwater species. The range of conductivities of SW and NaCl to which the organisms were exposed is presented in table 1.

To conduct mortality, growth and emergence assays with Chi. riparius, 2-day-old larvae (1st stage) were used. The lethal assays followed the OECD guideline 235 [38], using neither food nor sediment. Four replicates, with five organisms, were established per treatment or control (ASTM medium; [32]) (table 1). For growth and emergence, another test was prepared consisting of a set of six concentrations replicated eight times (table 1). Each replicate contained 10 organisms. Vessels contained a 2 cm high layer of sediment and 300 ml of respective test solutions ([39], Guideline 219). At day 10, half of the replicates were sacrificed to evaluate growth. At this stage, all larvae removed from the sediment were preserved in 70% ethanol and, afterwards, the body and head length (millimetre) of each larva was determined using a stereomicroscope (Leica MS5, Leica Microsystems, Houston, TX, USA) fitted with a calibrated eyepiece micrometre. Growth (mm) was evaluated by subtracting the average initial length of a subsample (n = 30) of the batch of organisms to the final length of each organism. The remaining replicates were maintained until the 28th day and emergence of adults was recorded every day.

(v). Toxicity assays with Mollusca

Ecotoxicity assays with T. fluviatilis were carried out with organisms of a shell height and length of approximately 4 and 5 mm, respectively. A 48 h lethal toxicity assay was followed by a 3 h feeding test. To assess mortality, organisms were exposed, at 20°C and in a 16 L : 8 D photoperiod cycle, to a range of NaCl concentrations and SW dilutions plus a control (only ASTM medium) (table 1). Four replicates with five organisms were assembled and each replicate consisted of a 50 ml vessel filled with the respective solution, covered with a net to avoid organisms from escaping. Mortality was verified at 24 and 48 h. At the end of the mortality test, surviving organisms were transferred individually to 24-well plates. Each well was previously filled with 2 ml of clean medium and 150 brine shrimp nauplii (Artemia salina) to serve as food source. After 3 h, snails were immediately removed, and the remaining food items counted; this feeding assay was run at 20°C in total darkness. The feeding procedure was previously developed and validated by Correia et al. [37].

(d). Multigenerational exposure to low levels of salinity

To understand if multigenerational exposure to low levels of salinity could lead to increased tolerance of organisms to this abiotic stressor, the studied species were exposed for two generations to low levels of NaCl (table 1). The salt NaCl was used instead of SW because it proved to be a safe surrogate of SW in the assays described in the previous section, as it was of similar or higher toxicity to the majority of the tested species (see Results). These generations were designated as F₀ (no previous exposure to low levels of salinity) and F₁ and F₂ (previous exposure to low levels of salinity). After exposure for the two generations at low levels of salinity, the above mentioned lethal and sublethal endpoints were reassessed by performing the previously described assays (see §2c Toxicity Assays). This allowed comparing the sensitivity of organisms before (F₀) and after multigenerational exposure (F₂) to low levels of salinity.

Conductivities set up for multigenerational exposure are described in table 1 for each species. The conductivity values were selected based on the LC₅₀ for each tested species divided by a factor of 2. This approach intended to expose the organisms to sublethal salinity levels, to simulate high-stress levels and promote responses that would allow later generations to cope better with increased salinity levels. Whenever high mortality was observed during multigenerational exposure, this factor was adjusted in order to lower the conductivity (LC₅₀/3, LC₅₀/4, LC₅₀/5 and so on) until no significant mortality was observed during the multigenerational exposure (table 1).

For Chlorophyta and Cyanophycae, each new generation was obtained from the previous one when cultures were at their exponential growth phase (in case of R. subcapitata and Chl. vulgaris this period was 4/5 days, for Cy. raciborskii was 10 days).

Regarding the Tracheophyta, only L. minor was tested to assess increased tolerance after multigenerational exposure to low levels of salinity, because we were unsuccessful in rearing healthy L. gibba up to EC_50,growth/6. A generation period was considered whenever the macrophytes fronds covered the whole medium surface of a 250 ml Erlenmeyer, filled with 100 ml medium. Each generation lasted for 7/8 days.

For B. calyciflorus, this period corresponded to two generations of asexual reproduction and for chironomids corresponded to two generations of sexual reproduction. Regarding chironomids, cultures with medium spiked with NaCl (table 1) were maintained as previously described in §2c Toxicity Assays. For B. calyciflorus, a culture was implemented from rotifers hatched from cysts. The rotifers were collected within 6 h after hatching (to avoid much discrepancy in the age of organisms) and individualized in 24-well plates. Each well was filled with 1 ml of the solution of NaCl at the conductivity level desired (table 1). Organisms were fed and kept under control conditions of darkness and temperature as described in the procedure Rotoxkit F (MicroBioTests, Ghent, Belgium). After 24 h, organisms were checked for the release of first brood, and then every 12 h. This allowed us to follow the organisms until the release of their third or fourth broods, which were used either to carry on the next generation or to reassess the toxicity of NaCl at the end of the second generation.

For H. incongruens and T. fluviatilis, it was not possible to assess their capacity to acquire an increased tolerance to salinity after multigenerational exposure because the maintenance of cultures under standard laboratorial conditions at low levels of salinity was unsuccessful and organisms did not reproduce.

For D. longispina, data regarding multigenerational exposure to low levels of salinity, corresponded to the mean values obtained by pooling data previously presented for six clonal lineages of D. longispina in Venâncio et al. [24].

3. Data analysis

Lethal conductivities (as a measure of salinity, comparable across NaCl and SW) provoking 50% of mortality (LC₅₀) were computed with the PriProbit software [41]. Estimation of median effective conductivities (EC₅₀) for other endpoints was made by using nonlinear regressions, fitting the datasets to a three-parametric logistic or sigmoid curve (choice of curve depending on the best fit), using the program Statistica for Windows 4.3 (StatSoft, Aurora, CO, USA).

To evaluate the hypothesis of using NaCl as a possible surrogate for SW, lethal and sublethal effective conductivities between NaCl and SW were compared through a generalized likelihood test. Lethal/sublethal effective conductivities between NaCl and SW were considered statistically different whenever p = 0.05 [42].

In order to evaluate the hypothesis that freshwater organisms were able to increase their tolerance to salinity after a two-generation period of exposure to low levels of salinity, the same method described previously was used (a generalized likelihood test). Effective conductivities before and after generational exposure to low levels of salinity were considered statistically different whenever p = 0.05 [42]. This approach was followed because for the majority of the species the range of conductivities tested before and after multigenerational exposure to low levels of salinity was not the same.

4. Results

(a). Sodium chloride versus seawater toxicity

Estimations on the goodness of fit of the curves from which LC₅₀ and EC₅₀ were derived are depicted in electronic supplementary material, table S1 and table S2, respectively. The conductivities of diluted SW inducing 50% of mortality (LC₅₀) were always higher than the ones of NaCl (figure 1a). Nevertheless, NaCl was only significantly more toxic than SW for the cladoceran D. longispina ( p = 0.05) and the snail T. fluviatilis ( p = 0.05). Among the tested species, the decomposer Chi. riparius and the freshwater snail T. fluviatilis were the most tolerant species to SW, with LC₅₀ (95% confidence limits (CL)) values of 17.86 mS cm⁻¹ (17.1–18.8) and 26.4 mS cm⁻¹ (25.8–27.2), respectively. The rotifer B. calyciflorus and the cladoceran D. longispina were the most sensitive species both to SW and NaCl (LC₅₀ of 5.09 and 4.01 mS cm⁻¹ for B. calyciflorus and 5.99 and 2.49 mS cm⁻¹ for D. longispina, respectively).

Figure 1.

(a) Median lethal (LC₅₀) and (b) median effective conductivities (EC₅₀) of sodium chloride (NaCl, dark grey bars) and natural seawater (SW, light grey bars), for all the studied species. Vertical bars correspond to the 95% confidence limits. Different letters (a,b) indicate significant differences (generalized likelihood test: p = 0.05). *Data from Venâncio et al. [23]. ^#Data from Venâncio et al. [24].

Regarding sublethal endpoints (figure 1b), a similar tendency was observed, with SW being less toxic than NaCl for almost of the tested endpoints and species. In 6 out of 12 cases, NaCl showed to be significantly more toxic than SW ( p = 0.05; figure 1b). Exceptions, where SW was more toxic than NaCl, were recorded for Chl. vulgaris (growth EC₅₀, 21.8 and 12.2 mS cm⁻¹ for SW and NaCl, respectively), B. calyciflorus reproduction (reproduction EC₅₀, 1.96 and 3.88 mS cm⁻¹ for SW and NaCl, respectively) and Chi. riparius emergence (emergence EC₅₀, 6.99 and 8.17 mS cm⁻¹ for SW and NaCl, respectively). At sublethal levels (figure 1b), the most sensitive organisms were the producer Cy. raciborskii and primary consumers (D. longispina and B. calyciflorus), with the reproductive output of B. calyciflorus being the most sensitive endpoint (figure 1b). The macrophytes L. minor and L. gibba were the most tolerant species, namely to SW (EC₅₀ of 51.9 mS cm⁻¹ L. minor and EC₅₀ of 39.5 mS cm⁻¹ for L. gibba). The macrophyte L. minor, the ostracod H. incongruens and the freshwater snail T. fluviatilis were the species presenting the highest difference in their sensitivity to SW and NaCl (2.7, 2.6 and 2.14-fold difference, respectively), with EC₅₀ (95% CL) of: 51.9 (42.8–61.1) and 19.5 (16.7–22.2) mS cm⁻¹ for growth; 12.4 (10.6–14.3) and 4.86 (4.38–5.34) mS cm⁻¹ for somatic growth, and 24.6 (21.6–27.6) and 11.5 (9.79–13.2) mS cm⁻¹ for feeding rate, for SW and NaCl, respectively.

(b). Multigenerational exposure to increased salinity

Estimations on the goodness of fit of the curves from which LC₅₀ and EC₅₀ were derived are depicted in the electronic supplementary material, tables S1 and S2, respectively.

Multigenerational exposure to low levels of salinity did not significantly change the lethal sensitivity of B. calyciflorus and Chi. riparius but increased the lethal tolerance of D. longispina to salinization (figure 2a). For the latter species, the LC₅₀s were, before and after multigenerational exposure, 2.49 and 6.28 mS cm⁻¹, respectively (figure 2a). Analysis of the sublethal effects showed that in 5 out of the 7 tested species, organisms were able to withstand low levels of salinity, although no clear increase of tolerance occurred ( p > 0.05) (figure 2b). The only significant increases were registered for the producer Cy. raciborskii (cyanobacteria) ( p = 0.05) and the primary consumer D. longispina ( p = 0.05 for feeding and p = 0.05 for reproduction). However, their tolerance only increased by a factor of 1.6-fold (for Cy. raciborskii growth) and 2.5 or 1.6-fold (for D. longispina feeding and reproduction endpoints, respectively), (figure 2b). The most evident decreases in tolerance to salinity after multigenerational exposure to low salinity levels occurred for the growth of R. subcapitata and L. minor and for the growth and emergence of Chi. riparius. For R. subcapitata, the EC₅₀s for growth rate were of 8.92 and 6.88 mS cm⁻¹ before and after multigenerational exposure, respectively ( p = 0.05) and for L. minor, the EC₅₀s for dry weight was 19.5 and 10.9 mS cm⁻¹ before and after multigenerational exposure, respectively ( p = 0.05). For Chi. riparius, the growth EC₅₀s were 11.0 and 7.90 mS cm⁻¹ and the emergence EC₅₀s were 7.60 and 6.99 mS cm⁻¹, before and after multigenerational exposure, respectively, but no significant differences were observed in this case ( p > 0.05) (figure 2b).

Figure 2.

(a) Median lethal (LC₅₀) and (b) median effective conductivities (EC_x) of NaCl before (dark grey bars) and after multigenerational exposure to low levels of NaCl (light grey bars) for the studied species. Vertical bars correspond to the 95% confidence limits. Different letters (a,b) indicate significant differences (generalized likelihood test: p = 0.05). *Data from Venâncio et al. [23]. ^#Data from Venâncio et al. [24].

5. Discussion

(a). Sodium chloride as a surrogate of natural seawater

The results obtained with the lethal and sublethal toxicity assays revealed that NaCl exhibited a higher or similar toxicity than SW to the tested species, which may be related to the fact that SW has a more complex chemical composition. Previous works indicated that natural or artificial SW exerts less harmful effects than NaCl, for instance, for aquatic macrophytes [43], daphnids [10] and macroinvertebrates [22]. SW is composed of two major ions (Na⁺ and Cl⁻) but also includes ions in minor concentrations, such as calcium (Ca²⁺), magnesium (Mg²⁺) and potassium (K⁺) [44]. These ions have important roles in cellular functions. For example, calcium (Ca²⁺) is an important mediator in cell signalling [45], boron (B³⁺) plays a role in cell signalling and in cell wall stabilization [46], while magnesium (Mg²⁺) mediates the control of ion channels, enzymes and metabolic pathways [47]. So, the presence of these ions may constitute an advantage and help to explain the lower toxicity of SW when compared to NaCl.

In this work, the highest differences in lethal tolerance to NaCl and SW were registered for L. minor, H. incongruens and T. fluviatilis. The presence of extra ions (e.g. Ca²⁺, Mg²⁺, K⁺) in SW may constitute an advantage for these species. For instance, Mg²⁺ may intervene in chlorophyll synthesis in the macrophyte L. minor, helping the plant to continue performing photosynthesis under stress conditions; also, the presence of K⁺ ions may help in specific cellular processes in the cytosol and chloroplasts, as this particular ion cannot be replaced by other cations [48]. The ostracods, more specifically, hold a heavily calcified exoskeleton, so the presence of an external source of Ca²⁺ may favour the moulting process and therefore their growth [49]. In the case of the snail species, they not only also possess an external structure rich in calcium, but this species is known to have preference for calcium-rich waters and is considered euryhaline [50,51]. The enrichment of the medium with extra ions (e.g. K⁺ and Mg²⁺) may have a buffer effect against the toxicity exerted by Na⁺ or Cl⁻ (which is not present in the NaCl exposure). Mount et al. [52] showed that the interaction between ions could diminish their toxicity, namely for Cl⁻, SO₄²⁻ and K⁺. In particular, the NaCl LC_{50,48 h} in D. magna was 4.77 mg l⁻¹ while in a mixture of NaCl/Na₂SO₄ was 5.7 mg l⁻¹ [52]. These complex interactions among ionic constituents of SW may be particularly relevant at low levels of salinity, in sublethal responses (such as growth or reproduction). In fact, the difference in the response of NaCl versus SW was higher under sublethal exposure scenarios, in the case of the macrophyte L. minor, the ostracod H. incongruens and the snail T. fluviatilis.

In very few cases (and without statistical difference; e.g. Chl. vulgaris and B. calyciflorus reproduction) SW was slightly more toxic than NaCl. This result may evidence the role of other SW minor constituents (e.g. sulfate and potassium) that might be highly toxic to aquatic biota. For instance, Mount et al. [52] verified that sodium sulfate (Na₂SO₄) was more toxic than sodium chloride (NaCl) to Ceriodaphnia dubia, D. magna and Pimephales promelas if considered the conversion into molar units. In line with this finding, Goetsch & Palmer [53] reported that mortality of the freshwater mayfly Tricorythus sp. was induced at much lower concentrations of Na₂SO₄ comparing to NaCl. The reported LC₅₀ values were 660 mg l⁻¹ for Na₂SO₄ and within the range of 2200–4500 mg l⁻¹ for NaCl. Also, there is a high potential for potassium (K⁺) to cause toxicity towards freshwater biota. For instance, zebra mussel adults (Dreissena polymorpha) died eight times faster when exposed to a potassium chloride (KCl) concentration of 10 g l⁻¹ than when exposed to a similar NaCl concentration [54].

Though the present work focused on single-species tests, it seems reasonable to state that NaCl can be used as a protective surrogate for early stages of preliminary risk assessment of SW intrusion on freshwater ecosystems, since SW is mostly composed of Na⁺ and Cl⁻ ions and, in most cases, NaCl induced similar or higher toxicity than natural SW. Considering this last statement, information gathered within this study with NaCl might prove to be of added value since it can also be used in risk assessment frameworks regarding the use of surrogate salts in road deicing or salinization scenarios owing to industrial activity. Nevertheless, it must be taken into account that in some cases, the use of NaCl may lead to a risk overestimation. Still, even if overestimated, NaCl-based derived safety values would be protective for the most sensitive species.

(b). Most sensitive freshwater species

Daphnia longispina, B. calyciflorus and Cy. raciborskii were the most sensitive freshwater species to increased salinity. The EC_50,NaCl for feeding and somatic growth of D. longispina were 3.21 and 3.48 mS cm⁻¹, respectively; the EC_50,NaCl for reproduction of B. calyciflorus was 1.7 mS cm⁻¹ and the EC_50,NaCl for growth rate of Cy. raciborskii was 2.65 mS cm⁻¹. These values are above the salinity threshold for freshwaters (considered to be within the range 0.15–0.5 mS cm⁻¹, [55,56]), which suggests that even the most sensitive freshwater species may be able to cope with very low levels of salinization. However, these EC₅₀ values are more than 15 (D. longispina sublethal endpoints) and 30-fold (B. calyciflorus reproduction) smaller than the SW conductivity (approx. 52 mS cm⁻¹ in the Atlantic Ocean; e.g. [57]) and are likely to be easily achieved or surpassed in the case of SW intrusion. Jeppesen et al. [58] refer 2‰ (approx. 3.8 mS cm⁻¹) as the threshold salinity for freshwater zooplankton, which is in line with our results.

The sensitivity of freshwater organisms to SW and NaCl is linked to their osmotic tolerance and osmoregulation abilities. Most freshwater organisms are hyper-regulators and need to actively uptake ions from the external medium. In rotifers, this is carried out by a rudimentary system, the protonephridia system [59], while in daphnids it is assigned to thin-walled and lamellar modifications in the gills—the epipodites [60,61]. In daphnids, this process seems to change throughout ontogenesis, being mediated by an Na⁺/K⁺-ATPase and an Na⁺/Cl⁻ exchanger in neonates, and by an Na⁺/K⁺/2Cl⁻ co-transporter in adults [62,63]. Such specialized structures facilitate the entrance of ions and other substances, which may contribute to their low tolerance to salinity and other pollutants. Actually, the sensitivity of Daphnia and Brachionus genera made them relevant and widely used ecotoxicological models [64].

Since ionic regulation is an energetically costly process, the shift on energy investment to osmoregulation may impair other life traits, such as reproduction (as in B. calyciflorus), growth and feeding rates (as in D. longispina) [65]. Despite being among the most sensitive group within this study, one must consider that cladocerans and rotifers are potentially able to respond to stressful conditions via other strategies, such as producing dormant or resistance eggs. These egg banks are an important strategy of survival and resilience of these species after stressful conditions, including increased salinity. However, the hatching success of these egg banks may be compromised at high salinity levels. Delayed or impaired hatching may compromise the following generation to colonize salinity-impacted systems, therefore, compromising the resilience of the populations at these locations. For instance, Santangelo et al. [66] verified that dormant eggs did not hatch at 16 and 32 g l⁻¹ of NaCl (approx. 31.7 and 63 mS cm⁻¹); when returned to freshwater, some eggs were still viable and were able to hatch. In accordance, Bailey et al. [67] verified that, although B. calyciflorus time to hatching was delayed at a salinity of 8 (approx. 13.9 mS cm⁻¹), emergence ability was not compromised. However, at this same salinity level, the development of Daphnia and Bosmina embryos always stopped before hatching, even when transferred to freshwater [67]. Therefore, increased salinity may delay hatching processes in some species while hampering it in others.

Together with the daphnid D. longispina and the rotifer B. calyciflorus, the cyanobacterium Cy. raciborskii was also one of the most sensitive species to increased salinity levels. Other studies performed with this species have already reached similar evidences [68]. Its growth may be reduced at salinity levels higher than 2 g l⁻¹ of NaCl (approx. 3.96 mS cm⁻¹). In fact, Moisander et al. [68] verified that at an NaCl concentration of 10 g l⁻¹ (approx. 19.8 mS cm⁻¹) CO₂ fixation by the cyanobacterial cells totally ceased.

(c). Multigenerational exposure to low levels of salinity

After the multigenerational exposure to low levels of salinity, most of the studied species did not exhibit an increased tolerance to salinization. Previous studies have shown that continuous exposure (e.g. six generations in D. pulex) to a stressor could trigger mechanisms that allow the maintenance of fitness along generations [69]. When D. magna was exposed to zinc, the reproductive success in the F₀ generation (no previous exposure of the mothers to the contaminant) decreased, but in the subsequent generations (F₁ and F₂), this decline was not observed when organisms were returned to the control medium [70]. Moreover, when D. magna was continuously exposed to zinc, a reduction in reproduction was observed not only in F₀ but also in F₁ (exposed to zinc). However, the absence of a decrease in reproduction in F₂ during this multigenerational exposure suggests that D. magna was able to deal with stress by possibly triggering physiological mechanisms in response to the chemical stress from F₁ to F₂. Nonetheless, it is important to note that this probable increase of tolerance was associated with fitness costs, such as reduced metabolic rates and decreased lethal tolerance to salt throughout the experiment [70].

Many studies suggest that the development of stress mechanisms may be triggered quickly in rapid growth species. For instance, algae can increase their tolerance to salinity within three generations when exposed to sublethal levels of salinity [23]. According to Hart et al. [71], the period provided for organisms to acclimate is essential: the longer period organisms have been exposed to a particular salinity level, the higher their ability to later cope with salt stress. In this work, the multigenerational exposure period seemed not enough to trigger such response. Actually, in a few cases (R. subcapitata, Chi. riparius and L. minor), we observed that organisms became more sensitive after multigenerational exposure to low levels of salinity. In studies with macroinvertebrates, Paradise [72] showed that salinity values causing minimal effects on short-term survival provoked severe impacts on reproduction, affecting the recruitment potential of the second generation. The increased sensitivity of those species after multigenerational exposure to low levels of salinity might scale to other changes at the freshwater community structure. For instance, there might be expansion of more salt tolerant species that come to occupy the niches left by those more sensitive, but effects may also be felt in other species that depend on the most sensitive ones for survival (e.g. prey-predator, algae-grazer) or in competitive relationships (e.g. [23,73]).

In a few cases, an increase in tolerance was observed after multigenerational exposure to low levels of salinity (for instance, Cy. raciborskii growth and D. longispina mortality, feeding and reproduction). Previous studies have already reported the ability of Daphnia spp. to increase salinity tolerance over a few generations, namely after 5–10 generations (D. pulex; [25]) and after a 3 generation period (D. longispina clonal lineages; [24]) and, more, that tolerance can be maintained over a 30 generation period (e.g. [74]). Recent epigenetic data are also in line with the evidence that Daphnia spp. salinity-induced tolerance may be achieved in very short periods of time (one generation; [75]). Such epigenetic changes were closely related with genes involved in DNA repair, protein synthesis and cytoskeleton organization [75]).

The higher values of conductivity effects for D. longispina reported in this work after multigenerational exposure to low levels of salinity may be related to energy expenditure in osmoregulation functions. Furthermore, we observed a similar tolerance over multigenerational exposure regarding somatic growth but an increase in tolerance in the reproductive output, in D. longispina. This may be indicative that, under salt stress, organisms may invest in different strategies to assure the population maintenance at impacted sites, for instance, by reducing investment in growth and diverting the energy towards reproduction. This strategy has already been reported for other chemicals (e.g. exposure to metals such as cadmium and copper or to the cationic surfactant, cetyltrimethylammonium bromide; [76]). Regarding the cyanobacterium (Cy. raciborskii), we observed that this species was one of the most sensitive freshwater species to increased salinity; however, after multigenerational exposure to low levels of salinity, the effective conductivities more than doubled. Cyanobacteria can adapt to nearly all environments and, as reported above for daphnids, epigenetic traits might also be underlying that success [77]. For instance, 72 h of nutrient depletion in the model cyanobacterium Synechocystis were enough to induce DNA methylation in a partially inheritable fashion [77]. Though salinity proved to be a limiting factor on the population growth of Cy. raciborskii [68], it should not be discarded that, in more complex scenarios and in conjunction with other environmental factors, its fast ability to increased tolerance towards salinity may interfere with the resilience of other species, for instance, by interfering in prey-predator/grazer-algae balances or competition relationships; i.e. increased salinity may interfere with food quality leading to changes in trophic interactions (e.g. [78]). Moreover, there might be a risk also of toxin production, with consequences for other organisms inhabiting freshwaters [79].

6. Conclusion

Within a scenario of climate change, sea-level rise, increase in storms and intensification of drought are threats to many coastal freshwater ecosystems owing to the consequent saltwater intrusion. This work is based solely on single-species standardized toxicity tests, thus not accounting for interaction between species (e.g. changes in predatory or competitive relationships) or interaction of salinity with other abiotic factors. Still, data provided by this work are valuable to early-risk assessment frameworks on saline intrusion in low-lying coastal areas. First, NaCl proved to be a safe substitute of SW for preliminary risk assessment of salinization, as it induced similar or higher toxicity than SW for most of the studied species. Since salinity (and SW composition) is variable according to the location around the globe, the use of NaCl as a surrogate proved to facilitate the comparison of data gathered on tolerance to NaCl with other datasets, and further contributing to regulatory strategies (e.g. [80]). Second, from the battery of standard bioassays comprising several taxonomic groups, it was possible to identify Cy. raciborskii, D. longispina and B. calyciflorus as the most sensitive freshwater species (LC₅₀ ≤ 4.98 mS cm⁻¹ and EC₅₀ ≤ 4.38 mS cm⁻¹, both for NaCl), with median effective conductivities being much lower than natural SW conductivity (≈ 52 mS cm⁻¹). Among the direct effects of increased salinity on these species, disturbances in other trophic levels depending on them might also be expected (e.g. imbalances in predator-prey or competition relations). Third, no evidence of increased tolerance was found after multigenerational exposures to low levels of salinity. Most of the studied species presented comparable tolerance before and after this exposure, but it should be pointed that the producers R. subcapitata and L. minor presented an increased sensitivity. Despite being small, such an increase in sensitivity may be intensified in future generations. Therefore, their resilience in salt-disturbed freshwater environments may be compromised, subsequently inducing changes at the community level.

Data accessibility

Data available as part of the electronic supplementary material.

Competing interests

We declare we have no competing interests.

Funding

This study was funded by National Funds (OE) through FCT/MEC and co-funded by FEDER, through COMPETE (POFC), by FSE and POPH (Programme Ciência 2007-8) and the research project SALTFREE (PTDC/AAC-CLI/111706/2009). C.V. is a grant holder from FCT (ref. SFRH/BD/81717/2011).