Introduction

Ecotoxicological Risk Assessment (ERA) at the ecosystem level lies upon the study of the fate and effects of toxicants on living organisms. It can be carried out through laboratory microcosms that allow to reconstitute with a simplified experimental scheme a physical medium and organisms that interact (Cairns and Cherry 1993). In such bioassays, which can last several weeks to several months, all variables depend on time, including the concentration of the toxicant in the different abiotic (water, sediment) and biotic (organisms) compartments. Linking observed effects and toxicant presence, which is the final aim of ERA, raises the question of the reference medium for the quantification of toxicant (water, sediment, or biota?) and the determination of the dose-effect relationship. Recently, we started to modelize the functioning of a laboratory microcosm (Clément et al. 2013, 2014) and the effects of cadmium, a model toxicant, on this functioning. The laboratory microcosm is composed of a synthetic medium (volume 2 L), a sediment (natural or formulated, 200 g fresh weight), and five species (the micro-alga Pseudokirchneriella subcapitata, the duckweed Lemna minor, the cladoceran Daphnia magna, the amphipod Hyalella azteca, the chironomid Chironomus riparius). In a recent work (Lamonica et al. 2016) dealing with modeling of interactions between duckweeds and micro-algae in an aqueous medium (no sediment) in the presence of cadmium, the cadmium concentration of water was supposed constant, and the effect models were based on a log-logistic function which did not involve the internal concentration. However, effects are linked to the internal concentration which depends on the bioavailable fraction of the toxicant which is not the total concentration in water due to chemical speciation and selective permeability of biological membranes to only some chemical forms. For cadmium, the free Cd2+ ionic form can be considered as a good proxy of the bioavailable fraction (Sunda and Huntsman 1998; Errecalde et al. 1998; Campbell et al. 2002; Verheyen et al. 2012).

Cadmium is a non-essential element recognized as one of the most toxic metals (Bertin and Averbeck 2006). As regards microalgae and macrophytes, at concentrations ranging between a few µg/L and 1 mg/L found in contaminated environments, cadmium significantly impairs cell growth, induces cell death, increases cell volume and Chl a content per cell, impairs the photosystem II maximum quantum yield (Echeveste et al. 2017), reduces absorption of nutrients (Andresen et al. 2016), increases oxidative stress, changes the activity of antioxidant enzymes (Razinger et al. 2008), promotes the accumulation of reactive oxygen species (Schützendübel et al. 2001). Microalgae also respond to cadmium by forming low molecular weight sulfhydryl compounds (e.g. phytochelatins) to prevent its toxicity (Folgar et al. 2009).

In the present study, we assessed the effects of cadmium on the growth of microalgae (P. subcapitata) and duckweeds (L. minor) exposed separately, and tried to link them to the concentration of free cadmium in water determined by chemical modelling using the PHREEQC software (Parkhurst and Appelo 1999) and to internal cadmium concentration. The results will be used to increase our knowledge of the exposure conditions of both organisms in dynamic conditions (growth along 2 weeks) corresponding to the conditions of the microcosm assay of Lamonica et al. (2016), and finally to allow us to improve the bases of our model.

Materials and methods

Algal bioassays

Algal and duckweed bioassays were carried out in the conditions of batch microcosm assays (Clément et al. 2013, 2014). Algae (Pseudokirchneriella subcapitata, CCAP 278/4) were grown as recommended by standard methods (ISO 2012). Axenic cultures were maintained by weekly transferring a few mL of an exponentially growing culture in fresh medium sterilized by autoclaving. The exposure medium was not sterilized and bioassays were not carried out in sterile conditions. Algal bioassays were carried out in cylindrical glass beakers (diameter 12 cm, height 24 cm) containing 2 L of synthetic water used in microcosm bioassays (pH 7.7, hardness 60 mg CaCO3/L, alkalinity 120 mg CaCO3/L, conductivity 290 µS/cm, phosphorus 100 µg/L, nitrogen 1308 µg/L, FeEDTANa2 109.4 µg/L, oligo-elements and vitamins of M4 medium; Elendt and Bias 1990) and prepared from deionized water. For the assays, microalgae were obtained from an exponentially growing culture after two centrifugation steps (2057 g, 15 min) to eliminate the residual culture medium, wash the cells and concentrate (10×) the inoculum so as to minimize the volume introduced into the beakers. On Day 0, prior to introducing algae, cadmium was added to the synthetic water as a 10 mg/L solution of Cd(NO3)2 in the following concentrations: 0–50 µg/L (assay 1), 0–20–30–40 µg/L (assay 2), 0–30–40–45–50 µg/L (assay 3). The first assay was aimed at developing the procedure of cadmium measurement in algae, the second one at testing several concentrations, and the third one at testing a final range of concentrations on the basis of the previous ones. After introduction of cadmium and microalgae (4 × 107 cells/beaker), the water column was gently homogenized with a stainless-steel spatula. Continuous but moderate aeration of the water column was started immediately, using Pasteur glass pipettes connected to a Rena 100 aquarium pump. The temperature was kept at 20 ± 2 °C and the beakers were illuminated for 16 h d−1 (2000 lux). The beakers were sealed with a plastic film to reduce evaporation. The sedimentation of algae was allowed in order to reproduce the processes that take place in microcosm assays. The duration of assays was variable: 21 days (assays 1 and 2), 14 days (assay 3). For the assay 1, there were 18 replicates of each treatment on Day 0 since 3 to 6 beakers/treatment were sacrificed at time intervals (Day 2, Day 7, Day 14, Day 21). For the assay 2, there were 9 replicates of each treatment on Day 0, with 3 replicates sacrificed on Day 7, Day 14 and Day 21. For the assay 3, there were 6 replicates on Day 0, with 3 replicates on Day 7 and Day 14. Algal density was monitored with a particle counter (Beckman Coulter®, Z1 model). On sacrifices, the water column was thoroughly mixed so as to re-suspend the settled cells, a volume of 20 mL was sampled and a sub-sample of 100 µL was mixed to 10 mL Isoton® for counting. The Coulter® threshold was set to 3.6 µm, enabling to count at least 99% of algal particles. There were three measurements per sample, the mean value was calculated. To monitor dissolved cadmium concentrations, 10 mL out of the 20 mL of water sampled for algal counting were kept and filtered through a 0.45 µm Whatman™ filter. These samples were acidified to pH 2.0 with ultra-pure nitric acid, and stored at 4 °C until they were analyzed by graphite furnace atomic absorption spectrometry (AAS) at 1500 °C (Perkin Elmer, model PinAAcle 90 T, detection limit: 0.1 µg/L). For the cadmium accumulated by microalgae, 400 to 800 mL (depending on algal density) of the mixed water column of sacrificed beakers were centrifuged for 15 min at 2057 g in 50 mL HDPE tubes, the supernatant was discarded and the cells were re-suspended in 20 mL of a 2 × 10−4 M EDTA solution so as to remove cadmium sorbed to the surface of algal cells (Campbell et al. 2002; Maloney 2007), and again centrifuged 15 min at 2057 g. After discarding the supernatant, the pellet was collected, re-suspended in a small volume of deionized water and measured for cell density in the same way as the water column, after dilution when necessary. A known volume (4 or 5 mL) of this concentrated suspension was then transferred into pre-weighed cups, dried in an oven at 70 °C and weighed on an ultra-precise balance (Sartorius Cubis MSE6.6S, 1 µg) so as to determine the dry weight of algae. Following the weighing step, the dried algae were collected and, based on an internal protocol, submitted to a digestion step in a microwave oven (CEM/MARS Xpress, 2 mL ultra-pure HNO3 + 6 mL ultra-pure HCl, 180 °C during 30 min, and dilution of the residual in 10 or 25 mL ultra-pure water). Cadmium concentration of the digested samples was then analysed following the same protocol as for water and expressed as µg cadmium per g of dried algae or µg per cell.

Duckweed bioassay

The duckweed Lemna minor was axenically cultured in Erlenmeyer flasks containing 250 mL of a nutrient-rich synthetic medium (AFNOR 1996; pH 5.1, hardness: 985 mg CaCO3/L, alkalinity 2.5 mg CaCO3/L, conductivity 1930 µS cm−1, phosphorus 15.5 mg/L, nitrogen 140 mg/L, EDTA 2.9 mg/L). Cultures were maintained by transferring some colonies in fresh medium sterilized by autoclaving. The exposure medium (same medium as for algal assays) was not sterilized and bioassays were not carried out in sterile conditions. Duckweeds were exposed 14 days to the following cadmium concentrations: 0, 10, 20 and 30 µg/L. There were 6 replicates per treatment at the start of test. The abiotic conditions (water composition and volume, light intensity and photoperiod, temperature) were the same as for algal bioassays, but air bubbling was not necessary since duckweeds use CO2 of the gaseous phase. On Day 0, 6 colonies of two fronds were introduced into the beakers after selection from the duckweed culture so as to start with fronds of homogenous size. Fronds and colonies were enumerated every other day. On Day 7 and Day 14, three replicates per treatment were sacrificed for the measurement of duckweed biomass and cadmium internal concentration. The cadmium concentration in water was measured on Day 7 and Day 14 following the same methods as for algae. For the measurement of cadmium internal concentration, duckweeds were immersed 5 min into 100 mL of a 2 × 10−4 M EDTA solution, rinsed in ultra-pure water, dried at 70 °C, weighed on an ultra-precise balance (Sartorius Cubis MSE6.6S, 1 µg) and digested, based on an internal protocol, in a microwave oven (CEM/MARS Xpress, 8 mL ultra-pure HNO3 + 2 mL ultra-pure H2O2, 180 °C reached in 15 min and maintained for 10 min, and dilution of the residual in 25 mL ultra-pure water). Hydrogen peroxide (H2O2) is known to improve the digestion process by its oxidizing capacity (http://www.berghof.com/…/MW_Theorie_Probenvorbereitung_PT_en.pdf). Cadmium concentration of digested samples was analyzed following the same protocol as for water and algae and expressed as µg cadmium per g of dried duckweed.

Chemical speciation

Chemical speciation of cadmium in the water was carried out using PHREEQC (version 2), a computer program freely available on the web. We used the database MINTEQ+ which allows to take into account organic species such as EDTA present in the medium. The modelling was carried out with a solution composition based on the synthetic medium used, with known concentrations of salts and known pH. A second step consisted in taking into account the sorption of cadmium to algal cell surfaces. This sorption is allowed by the deprotonation of sites of this surface:

$$Surf_{sOH} = Surf_{sO - } + H^ +$$

Followed by the complexation of the same sites with Cd2+:

$$Surf_{sO - } + Cd^{2 + } = Surf_{sOCd + }$$

The respective equilibrium constants KH+ and KCdBL (with BL = biotic ligand, i.e the sites) of these reactions were taken from Kaulbach et al. (2005): log KH+ = −3.9, log KCdBL = 4.1, and the number of sites = 2.77 × 10−5 moles per gram of dried mass. The total number of sites in a beaker was calculated by multiplying this figure by the total dry mass of algae.

Data treatment and analysis

Growth inhibition at time t for microalgae was expressed as (ISO 2012):

$$\% inhib = \left[ {\frac{{\left( {\left( {mean\,total\,cells\,at\,time\,t\,in\,the\,control} \right)-\left( {mean\,total\,cells\,at\,time\,t\,in\,the\,treatment} \right)} \right)}}{{\left( {mean\,total\,cells\,at\,time\,t\,in\,the\,control} \right)}}} \right]*100$$

Growth inhibition at time t for duckweeds was expressed as (AFNOR 1996):

$$\% inhib = \left[ {\frac{{\left( {\left( {mean\,frond\,number\,at\,time\,t\,in\,the\,control} \right)-\left( {mean\,frond\,number\,at\,time\,t\,in\,the\,treatment} \right)} \right)}}{{\left( {mean\,frond\,number\,at\,time\,t\,in\,the\,control} \right)}}} \right]*100$$

Relative growth for microalgae was defined for each cadmium concentration as 100−%inhib.

To detect an effect (i.e. significant differences among treatments), an ANOVA at each time of exposure was first performed for continuous data following a normal distribution. When the ANOVA showed an effect and the residues were normally distributed, pairwise comparisons between group levels with corrections for multiple testing were carried out in order to determine which treatments were different from the control. In addition, we used a method based on 95% confidence intervals for ratios of each mean to the mean of the control using the plug-in method provided within the R package mratios (Dilba et al. 2004; Delignette-Muller et al. 2011).

Results

Algal bioassays

Analyses of dissolved cadmium in the water phase during the algal bioassays are displayed in Table 1.

Table 1 Concentration of cadmium dissolved in water ([Cd]diss, µg/L) and growth inhibition (% inhib) of microalgae (total cells and dry biomass) during the three algal bioassays
Full size table

Measurements at t = 0 were in agreement with expected nominal values. The variability between replicates was low (mean coefficient of variation (CV) of 7 ± 9%, median of 4%). A significant decrease of dissolved cadmium concentration during the assays could be noticed. This decrease was higher for assay 2 (from 34 to 50% depending on concentration) than for assays 1 and 3 (from 22 to 27% for the same concentration range). For the three assays, the growth in the controls over 14 days was homogenous, with a mean of 2.99 ± 0.22 × 109 cells (total number of cells). After 21 days, the mean numbers for assays 1 and 2 were close (3.23 ± 0.045 × 109 and 2.68 ± 0.25 × 109 cells). The variability was moderate, as shown by a mean coefficient of variation for all data of the three assays of 9.88 ± 6.83% and a median value of 7.38%. In the presence of cadmium (Table 1), the responses of algae in assays 1 and 3 were similar for the concentration 50 µg/L tested in both assays, whereas cadmium in assay 2 was found to be 2 to 3 times less toxic than in assay 3 for the concentrations 20 and 30 µg/L. Note that effects on biomass were not always correlated with effects on cell number, more or less pronounced depending on cadmium concentration and assay.

The free cadmium (as Cd2+) concentrations calculated by chemical modelling based on the composition of synthetic water at t = 0 and expressed against measured concentrations of dissolved cadmium for the three bioassays are displayed in Fig. 1. The model calculated also the cadmium fraction sorbed onto algal cells; for all assays, all concentrations and all times of calculation, taking into account the increase of algal cells during the assays, this fraction was maximum 0.13% of total cadmium. The free cadmium fraction increased with dissolved cadmium concentration. For dissolved cadmium concentrations <25 µg/L, it ranged between 0.7 and 1.7%. For the range 25–50 µg/L it increased rapidly from 2 to 30%. The modelling did not take into account the decrease of nutrient concentrations (due to consumption by primary producers) such as NO3 −, NH4 + and PO4 3−, which play a minor role in cadmium speciation; as a matter of fact, cadmium was mainly linked to EDTA, at the start of assays, [CdNO3 +] was 10,000 × lower than [CdEDTA] and [Cd2+].

Fig. 1
figure 1

Concentrations of free cadmium modelized with PHREEQC vs. concentrations of cadmium dissolved in water measured during the course of algal bioassay

Full size image

On the range of tested concentrations, cadmium was bioaccumulated by microalgae at internal concentrations increasing with the cadmium concentration of water (Table 2). For assays 1 and 3, the internal concentration increased with time and started to stabilize around Day 21 for assay 1, whereas, though increasing until Day 7 in assay 2, it decreased afterwards. Bioconcentration was very rapid, as shown by an internal concentration of 1167 µg/g as soon as Day 2 vs. 1437 µg/g on Day 7 in assay 1. Variability was acceptable, with a mean CV of 12% (median value: 9%). For the controls, a background internal concentration of a few µg/g was measured, far below measured internal concentrations for contaminated algae. The fraction of cadmium internalized in microalgae for the 3 assays, calculated from a mass equilibrium based on the cadmium internal concentration per cell and the total number of cells in beakers, is displayed in Table 2. Internalization was high (up to 38%) for cadmium concentrations above 20 µg/L. This is likely due to the fact that, as previously shown, the proportion of free cadmium was lower at 20 µg/L. These results are in accordance with that of dissolved cadmium in water displayed in Table 1, the difference between total cadmium and dissolved cadmium being attributable to internalization of cadmium by algal cells. From these assays it was possible to estimate a bioconcentration factor of ca 111,000 based on free cadmium.

Table 2 Cadmium internal concentration (µg/g) of microalgae and fraction (%) of total cadmium internalized in the three algal bioassays
Full size table

As a matter of fact, the relationship between internal concentration and free cadmium concentration, expressed as a mean of the 3 assays combining all test durations, could be modeled by linear regression as [internal cadmium concentration] = 111.15 * [free cadmium concentration] (R² = 0.96) (Fig. 2). Note that when expressed from dissolved cadmium concentrations, the relationship is not as good (R² = 0.86 and intercept of regression line >0) and it is more difficult to estimate a bioconcentration factor. The log-logistic relationship between relative growth and cadmium internal concentration of microalgae, based on the results of the 3 assays and the 3 exposure durations (7, 14 and 21 days), is displayed in Fig. 3. This concentration-effect relationship is acceptable and, together with raw results, allows to roughly estimate the No Effect Concentration (NEC, i.e. highest internal concentration with no effect) around 400 µg/g, keeping in mind that inhibition values less than 20% are not significant, and an IC50 (internal concentration leading to growth inhibition of 50%) around 850 µg/g. Note that this IC50 corresponds to an EC50 around 40 µg/L for an exposure duration of 7 or 14 days, equivalent to a free cadmium of ca 6 µg Cd2+/L.

Fig. 2
figure 2

Linear relationship fitting of mean concentration of internal cadmium in algae vs. mean concentration of free cadmium in water in the three algal assays

Full size image
Fig. 3
figure 3

Log-logistic relationship fitting of mean cadmium concentration in algae (observed) vs. mean relative growth in the three algal bioassays

Full size image

Duckweed bioassay

The variability was acceptable with CV not higher than 10% for frond number and dry weight on Day 7, and respectively 20 and 25% on Day 14 (mean CV on Day 14: 11 and 14%). Cadmium inhibited moderately but significantly duckweed growth for concentrations ≥20 µg/L. The effects were higher for frond number than for biomass (Table 3). The concentration of dissolved cadmium varied only slightly during the assay. The corresponding calculated concentrations of free cadmium were <2 µg Cd2+/L. Cadmium uptake by the duckweeds increased with cadmium concentration of water and time for the first concentration but stabilized quickly for 20 and 30 µg/L (Table 3). Owing to this stabilization, it seems possible to estimate a mean bioconcentration factor of 17812 ± 1120 based on total dissolved cadmium concentration. As a matter of fact, since the fraction accumulated by duckweeds was not higher than 9%, cadmium concentration of water could be considered as stable and the equilibrium between water and duckweed concentrations could be considered as reached. This was corroborated by a good linear relationship between mean dissolved cadmium concentration and mean cadmium internal concentration whatever test duration (R² = 0.97). Due to limitation of inhibition to 47% (final frond number) or 21% (final dry weight), it was not possible to find a clear relationship between internal concentration and effect. However, it seems that, based on our results (Table 3), no effects were found on frond number for concentrations between 140 and 352 µg/g, which leads to propose a NEC around 200–300 µg/g.

Table 3 The concentration of dissolved cadmium in water ([Cd] µg/L), cadmium concentration in duckweed ([Cd] µg/g) and growth inhibition (frond number and dry biomass) in the duckweed test
Full size table

Discussion

Effect of cadmium to Pseudokirchneriella subcapitata

A summary of literature data and our own data regarding cadmium toxicity to Pseudokirchneriella subcapitata is displayed in Table 4. Comparison of results from various authors is difficult since several parameters may vary: test duration, composition and pH of the medium, toxicity endpoint (cellular density, photosynthetic activity), unit of toxicity (total or free cadmium). The chelator EDTA present in the synthetic medium used in several studies to keep iron in solution is known to complex metals and reduce their toxicity (Koukal et al. 2003; Guéguen et al. 2003), since this one is generally assigned to their free ionic form. The calculation of the free cadmium fraction (Cd2+) showed that the concentration of Cd2+ in our algal assays was low (<2 µg/L) for the range of total cadmium concentrations 10–30 µg/L, whereas it increased rapidly from 2 to 11 µg/L for total cadmium concentration in the range 30–50 µg/L. This nonlinear evolution can be explained by the fact that EDTA is present at 0.28 µmol/L in the test medium. For cadmium concentrations <20 µg/L, i.e. 0.27 µmol/L, EDTA can complex most part of cadmium, which explains low values of Cd2+ concentration, whereas for concentrations >20 µg/L the fraction of free cadmium rapidly increases to reach about 30% at 50 µg/L. Koukal et al. (2003) chose to remove EDTA from their AAP medium, but they worked with a short exposure duration (1 h) which allows to bypass iron deficiency (Koukal et al. 2003), whereas our tests lasted up to 21 days. In their medium, the free cadmium fraction was 10% with EDTA vs. 80% without. As expected, they obtained a higher toxicity of cadmium to Pseudokirchneriella subcapitata in the absence of EDTA (EC50-1 h = 200 µg/L vs. 270 µg/L). Pardos et al. (1998) used the same method but with a 4 h exposure and natural water (enriched Lake Léman water), and a more classical protocol based on 48-hr growth inhibition. They obtained an EC50-4 h of 600 µg/L and an EC50-48 h of 118 µg/L. Rodgher et al. (2012) cultured Pseudokirchneriella subcapitata in oligo-LC medium without EDTA. They exposed it to cadmium for 96 h and obtained an EC50-96 h of 67 µg Cd2+/L, which, compared to our EC50-7d of 7 µg Cd2+/L, seems to demonstrate a lower toxicity which might be due to other differences between media. Errecalde et al. (1998) exposed for 72 h P. subcapitata in media with or without NTA, another chelator, at free cadmium concentrations ranging between 0 and 101 µg Cd2+/L. They obtained a slightly higher toxicity without NTA. Apart from humic substances and chelators such as EDTA, NTA or citric acid, algal exudates may reduce metal toxicity (Koukal et al. 2007). It is likely that in our long-term algal tests the production of exudates was significant and might have contributed to reduce cadmium toxicity, since the concentration of DOC increases from 0 to 1.3 mg C/L in an algal culture of 7 days, when algal density varies from 104 to 4.5 × 106 cells/mL (Koukal et al. 2007).

Table 4 Cadmium toxicity to Pseudokirchneriella subcapitata according to various studies
Full size table

Cadmium uptake by Pseudokirchneriella subcapitata

Cadmium is more easily bioaccumulated in chlorophyllian organisms (INERIS 2011). There are very few results in the literature on cadmium uptake by microalgae. Bioconcentration factors (BCF) ranging between 1636 and 23143 have been recorded for microalgae (INERIS 2011). Rodgher et al. (2012) found, for P. subcapitata, internal concentrations of 560 to 5600 µg/g for a test duration of 96 h and a range of free cadmium concentrations of 33 to 134 µg Cd2+/L. Errecalde et al. (1998) also measured cadmium uptake by P. subcapitata at 28 µg Cd2+/L but on a very short time (6 min) and by determining a flux of free cadmium through cell membranes which cannot be easily converted into internal concentration. Our results showed that cadmium accumulated into algal cells was linearly linked to free cadmium concentration in water and this allowed to estimate a bioconcentration factor of ca 111,000 (on the basis of the free cadmium concentration in water). Since this result was obtained by considering all cadmium water concentrations and all test durations, we suggest that it can be assimilated to a true BCF, though this one must be obtained in specific conditions, namely when internal and external concentrations no longer vary with exposure duration (OECD 1996; USEPA 1996). However, in algal assays where the conditions of exposure are far from static conditions where organisms are exposed to a constant concentration, and the rapid multiplication of algal cells contributes to increase influx of metal through biological membranes but also dilution of metal, it is difficult to evaluate the gap between experimental conditions of assays and the ones required for BCF estimation. Noticeably, the relationship between cadmium concentration of water and internal cadmium concentration was better when cadmium water concentration was expressed as free cadmium rather than as dissolved cadmium. This is likely due to the fact that free cadmium is a better proxy of bioavailable cadmium than total dissolved cadmium.

Relationship between toxicity and bioconcentration in Pseudokirchneriella subcapitata

In our study, a clear relationship between cadmium internal concentration of algae and growth inhibition was found, based on measurements carried out along the course of assays of 21 days exposure. The variable time is a hidden variable in this relationship which directly links internalized cadmium and toxicity, which means that the same effect may be obtained for a high cadmium concentration of water on a short duration and a lower concentration on a longer duration. Such a relationship allowed us to define a NEC around 400 µg/g. Unfortunately, we could not find previous works in the literature for comparison with our own results.

Duration and repeatability of algal bioassays

Our results clearly show that a duration of 21 days for the algal assays 1 and 2 did not provide much added value, due to the fact that algal growth was stabilized after Day 14 and that growth inhibition and bioconcentration reached a plateau on Day 14, as shown in Tables 1 and 2 for all Cd concentrations. For this reason, the assay 3 lasted only 14 days.

The comparison of algal assays 2 and 3, limited to the concentrations 30 and 40 µg/L and the duration of 14 days, shows a higher decrease of cadmium concentration in water for the assay 2. It can also be noticed that growth inhibition and bioconcentration were lower in assay 2, which seems logical as regards the lower cadmium pressure after 7 days. These findings suggest that cadmium in assay 2 was partly captured by some compartment and rendered unavailable. Though microalgae were axenically cultured, the algal assays were not carried out in sterile conditions, hence leading to a possible development of other microorganisms such as bacteria. These ones are known for their ability to biosorb or accumulate metals which can then be converted to more innocuous forms by binding or precipitation (Gadd 1990; Oves et al. 2013).

Effect of cadmium to Lemna minor

As for microalgae, literature results are difficult to compare with our own results (Table 5), due to different conditions (mainly concentration range, medium composition, pH, temperature and test duration). EDTA is a current component of duckweed media such as Steinberg’s and Hoagland’s medium, for the same reason as for microalgae media (keeping iron in solution), and it is likely that this chelator reduces cadmium bioavailability, as demonstrated for instance by Saygıdeğer and Doğan (2004), or Kwan and Smith (1991). In the Steinberg medium, EDTA concentration is 4 µM, pretty higher than in our medium, but final free cadmium concentration will depend upon the ratio between EDTA concentration and total cadmium concentration. Some authors also used low concentrations of EDTA, for example Khellaf and Zerdaoui (2009) with 0.75 µmol/L.The medium we used is not optimal for duckweed growth, since it was defined for a microcosm community, with plants, microalgae and invertebrates. As a consequence, duckweed growth is lower than in specific media, and this might have consequences on sensitivity to toxicants. Among published results, Naumann et al. (2007) found an EC10-7d of 22 µg/L which is close to the magnitude of effects found in our study. Note that they also obtained 50% inhibition at 323 µg/L, which shows that the slope of the internal concentration-effect relationship was not steep.

Table 5 Cadmium toxicity to Lemna minor according to various studies
Full size table

Cadmium uptake by Lemna minor

In reported studies, cadmium uptake in Lemna minor was studied at rather high concentrations ranging between 0.1 and 5 mg/L. Saygıdeğer and Doğan (2004) studied the effect of EDTA on cadmium accumulation by L. minor (0.5 mg/L + 0.5 mM EDTA) for 7 days. They obtained an internal concentration of 1136 µg/g dw without EDTA and 115 µg/g dw with EDTA, equivalent to a concentration factor (CF) of respectively 2272 and 230. Balen et al. (2011) exposed L. minor in Steinberg medium for 7 days at a cadmium concentration of 562 µg/L. They found an internal concentration of 834 µg/g after 4 days and 1056 µg/g after 7 days, which corresponds to a concentration factor of respectively 1484 and 1879, much lower than the value we found in this study (17812 ± 1120). This might be due to differences in media or/and in cadmium concentration ranges. Charlier et al. (2005) exposed L. minor to 4 mg Cd/L and observed after 3 days an internal concentration of 10.4 µg/g fw, i.e a concentration factor of 2 on a fresh weight basis, which is very low even when expressed in a dry weight basis (concentration factor of 32 to 42 when considering a fresh weight:dry weight ratio between 15 and 20 as in our results). Drost et al. (2007) exposed L. minor to cadmium for 7 days and found an EC50-3d of 392 µg/L and a corresponding internal concentration of 538 µg/g equivalent to a BCF of 1371 calculated as the ratio between internal concentration and external EC50day3 concentration. As for Balen et al. (2011) the BCF value was much lower than the one we found, but the exposure duration was only 3 days. Kwan and Smith (1991) exposed L. minor for 300 h (ca 14 days) to 168 µg/L in Steinberg medium continuously renewed to keep cadmium concentration constant, and obtained an internal concentration of 1960 µg/g dw, corresponding to a BCF of 12600, a value closer to our results. We should nevertheless be aware that the range of cadmium concentrations are different. In addition, the Steinberg medium used by Kwan and Smith and the microcosm medium used here are also very different (especially regarding EDTA concentration, much higher in Steinberg medium) as well as duckweed biomasses concerned, leading to different bioavailabilities of cadmium and hindering sound comparisons of data. As a matter of fact, the dynamics of cadmium uptake seem to be quite different in both studies: after almost 14 days at 168 µg/L bioconcentration did not reach saturation in Kwan and Smith study, whereas in our study at 10, 20 and 30 µg/L a plateau was reached as soon as 7 days. Bianconi et al. (2013) studied the response of L. minor after 72 h exposure to CdSO4 (0 to 5.15 mg Cd/L) in Hoagland medium. They found a high inhibition at 5.15 mg/L, and calculated a BCF of 1275 and 1075 respectively at 1.46 and 2.46 mg/L for 72 h exposure. Singh et al. (1994) exposed L. minor to 100 µg Cd/L for 48 h and measured cadmium uptake. They found an internal concentration of 44 µg/g fw at 48 h. Transposed to dry weight, assuming a ratio FW: DW between 15 and 20, this would be equivalent to 660 to 880 µg/g dw. As shown by the large range of BCF values obtained by various authors, it is again difficult to draw conclusions regarding cadmium uptake by L. minor. It seems that test conditions, namely, cadmium concentration magnitude, medium composition (pH and presence of chelators) and test duration influence the final results even expressed as BCF. Furthermore, concentration factor values were obtained by the previously referred authors in varied conditions, often far from the required conditions for estimating a BCF. This brings uncertainty in comparison of data.

Relationship between bioconcentration and toxicity in Lemna minor

As for P. subcapitata, we tried to explore a possible relationship between toxicity and cadmium internal concentration but since effects were limited to 47% inhibition, no clear relationship was found. It was nevertheless possible to roughly estimate a NEC around 200–300 µg/g biomass in our experimental conditions (10–30 µg Cd/L, 0.28 µM EDTA). Unfortunately, we could not find much data to discuss our findings. Saygıdeğer and Doğan (2004) found that growth of L. minor, expressed through dry weight to fresh weight ratio and chlorophyll concentration of fronds, was affected after 1 week by cadmium at 0.5 mg/L with EDTA corresponding to 115 µg/g. Though they do not provide precise data of inhibition, it is visible from their figures that effects were low, from which it can be inferred that they obtained the first detectable effects at these internal concentrations. Given that the EDTA concentration of their medium was high (4 µM) but comparable to the cadmium concentration (0.5 mg/L equivalent to 4.5 µmol/L), it is likely that the free cadmium concentration was low (fraction of µg/L to maximum a few µg/L), as suggested by our own results, but it is not possible to know it with precision in the absence of details on the composition of their medium. However, 115 µg/g corresponding to a free cadmium concentration of a few µg/L and first effects seems roughly comparable (same order) to the NEC of 200–300 µg/g inferred from our own data.

Interest for the analysis of microcosm assay results

In order to assess the interest of describing the fate of contaminant in bioassays, we re-analysed previous results on duckweed-microalgae interactions in the presence of cadmium (Lamonica et al. 2016) in the light of our own results. In Lamonica et al. (2016), using protocols very similar to ours, we assessed the effect of cadmium on duckweeds alone, on microalgae alone, and on both species together. Briefly, for cadmium concentrations ranging between 11.5 and 51.1 µg/L, the effects found for algae were slightly more severe than those found in the present study (i.e. 76 vs. 52–59% at 36–40 µg/L, 92 vs. 65–73% at 50 µg/L). For duckweeds the effects were also more severe (68% for a large range of concentrations 11–50 µg/L vs. 39–47% at respectively 20 and 30 µg/L). In the presence of algae, Lamonica et al. observed an inhibition of duckweed growth due to competition for nutrients (68%), whereas the presence of duckweeds had no effect on algal growth, and a lower effect of cadmium on duckweed growth (68% without algae vs. 35% with algae). If we apply the results of the present study regarding cadmium partitioning between water, algae and duckweeds in the three conditions of Lamonica et al. (algae alone, duckweeds alone and algae + duckweeds), we show that the internalization of cadmium by algae, as extrapolated from the results of the present study, is not high enough to significantly reduce duckweed exposure and to explain the lower effects on duckweeds. Although the absence of cadmium measurements in the algae and duckweeds of Lamonica et al. assays renders difficult any sound interpretation of observed effects, it can be hypothesised that the presence of microalgae, by mainly inhibiting duckweed growth, also seems to modify the sensitivity of this one to cadmium, which could also be due to modifications of the water chemistry and hence the chemical speciation of cadmium.

Conclusion

In this study, we focused on the fate of cadmium in growth assays of the microalga Pseudokirchneriella subcapitata and the duckweed Lemna minor under abiotic conditions of a laboratory microcosm (specific medium with a realistic composition, duration of 2 or 3 weeks, photoperiodic lighting, settling of microalgae allowed). The objective was to obtain first results in a perspective of integrating such data in a model under development (Lamonica et al. 2016). Our results bring additional information on cadmium toxicity and bioconcentration regarding two primary producer species, Pseudokirchneriella subcapitata and Lemna minor, particularly on the criteria of exposure duration, which in the literature is limited to a few days. In addition, this work considers in the same time two major factors that govern the fate of cadmium, chemical speciation and bioconcentration, and brings light on its toxicity to P. subcapitata and L. minor. Free cadmium concentration was linked to total cadmium and governed by the concentration of EDTA. For the microalga P. subcapitata, it was possible to infer a bioconcentration factor, despite complex conditions of assays (duration of assays, modification of water composition, decrease of cadmium water concentration in the algal assays, increase of biomass). For this species, a clear relationship was found between effects and cadmium internal concentration, leading to propose a NEC of 400 µg/g not available in the literature. For the duckweed Lemna minor, the relationship was not as precise, and we could only propose an approximate NEC of 200–300 µg/g. However, the test conditions (stability of cadmium water concentrations and duration of tests) enabled to propose a BCF of 17812 for this macrophyte. Such results are useful for the future integration of a fate model in the modelling of microcosm functioning under chemical pressure, but more assays are needed to build this fate model which should be based on chemical speciation (as a proxy of the bioavailable form of cadmium), uptake and depuration (kinetic model).

As a summary, this study showed that cadmium is bioconcentrated by the green microalga Pseudokirchneriella subcapitata (BCF = 111,000 on the basis of free cadmium concentration in water) and the duckweed Lemna minor (BCF = 17,812 on the basis of total dissolved concentration), and that No Effect Concentrations were estimated at 400 µg/g for Pseudokirchneriella subcapitata and 200–300 µg/g for Lemna minor.