The experimental organism D. quadricauda (Turpin) Brébisson (previously known as Scenedesmus quadricauda), strain Greifswald/15, was obtained from the Culture Collection of Autotrophic Organisms, Institute of Botany (CCALA, Czech Acad. Sci., Třeboň). The microalga was cultivated under controlled experimental conditions of temperature (30°C) and continuous light in liquid mineral medium (ŠS-medium, Zachleder and Šetlik, 1982, Table 2), in laboratory scale tubular photobioreactors, as described previously in Vítová et al. (2011), (hereafter referred as “standard medium”). Cultures were aerated with air containing 2% carbon dioxide (v/v). The photobioreactor was illuminated from one side by fluorescent lamps (Osram DULUX L, 55W/840, Italy) at a surface incident irradiance of 500 μmol m⁻² s⁻¹. The starting pH of the suspension was 7.03 varying during 3 days between 6.9 and 7.2; final value of pH at the end of the third day was 6.9.

The compositions of metal-deficient nutrient solutions (deficient controls) were similar to the standard medium (Table 2) with the following equivalent modifications: (a) Mn²⁺-deficient medium was prepared by replacement of MnSO₄ with 8.44 μmol L⁻¹ (1.20 mg L⁻¹) Na₂SO₄; (b) Ca²⁺-deficient medium was prepared by replacement of CaCl₂ with 256.6 μmol L⁻¹ (15 mg L⁻¹) NaCl (Table 2).

To prepare cultures limited for Mn or Ca, the initial inoculum (from the agar plate) contained approximately 0.5 μg of algal biomass resuspended in 100 mL Mn- or Ca-free medium to a known final cell concentration (absorbance = 0.1). The element-limited cultures were cultivated for 2 weeks under the same conditions as control cultures. Finally, before the start of the experiment, the medium was again removed and replaced with Ca- or Mn-free medium. Changes in the concentrations of Ca and Mn during preparation of starved cultures were measured by ICP-MS, see Section Determination of Element Content (ICP-MS).

We tested different Ln (lanthanum, cerium, neodymium, gadolinium, and europium) that belong to the group of light rare earth elements (LREE). In the case of Ca²⁺-deprivation experiments, we used a reduced concentration of salt (2.5 mg L⁻¹), which was equivalent to 5.65 μmol L⁻¹ of Eu, 5.72 μmol L⁻¹ of La, 5.76 μmol L⁻¹ of Nd, 5.74 μmol L⁻¹ of Ce, and 5.61 μmol L⁻¹ of Gd. For the Mn²⁺-deprivation experiments, we used a manganese-equivalent concentration of salt (0.5 mg L⁻¹), which was equivalent to 1.11 μmol L⁻¹ of Eu, 1.13 μmol L⁻¹ of La, 1.09 μmol L⁻¹ of Nd, 1.14 μmol L⁻¹ of Ce, and 1.07 μmol L⁻¹ of Gd. In the second case, the concentration was equivalent to the amount of Mn added to the standard medium (Table 2). This was not the case for the Ca²⁺ concentration due to a potential risk of Ln toxicity (determined from a prior test, see below Concentration Range Finding Experiment). For that purpose we used analytical grade (99%, Sigma-Aldrich) chloride compounds (LaCl₃, CeCl₃, EuCl₃, NdCl₃, and GdCl₃).

We performed several prior tests (standard medium + Ln) to observe the reaction of D. quadricauda to different amounts of Ln, in order to determine the range of concentrations to use in our photobioreactor experiments. Therefore, the experiments were carried out in polystyrene 96-well microplates (NuncMicroWell 96-Well Microplates, Thermo Fisher Scientific Inc., Germany) with flat bottom wells of 300 μL. Each well-received 200 μL of test solution, a nutrient spike (10 μL) and an algal inoculum (10 μL), with a final volume per well of 220 μL (Blaise and Vasseur, 2005). Peripheral wells were filled with distilled water to reduce evaporation. Eight different metal concentrations were tested, and five replicates per test solution were performed, from the highest to the lowest concentration. The negative control was the respective medium + algae + water (no metal), and we incorporated a blank well (medium + elements + water, no algae) for each metal concentration tested. The microplates were incubated under controlled experimental conditions of temperature (30°C) and continuous light (100 μmol m⁻² s⁻¹) in an incubation cabinet (CLF Plant Climatics CU-22L, Germany). In this case, microalgal growth was measured by absorbance at 750 nm at 0, 24, 48, and 72 h, using an automated microplate reader (Infinite^® F200, Tecan, US) controlled by Tecani-control software (Tecan Group Ltd.). Microplates were shaken for 10 s in the microplate reader before measuring the absorbance. A calibration curve to establish the relationship between absorbance on the microplate and cell density of D. quadricauda was previously performed for various stages of the culture.

For the determination of Ca, Mn, and Nd in the nutrient medium, the ICP-MS analytical method was used. ICP-MS measurements were performed using an Elan DRC-e (Perkin Elmer, Concord, Canada) equipped with a concentric PTFE nebuliser, a cyclonic spray chamber, a high-efficiency quartz torch and a dynamic reaction cell (DRC) for the elimination of spectral interference. The IS solution for the total metals concentration contained Ca and Mn (2.5 mg/L) in dilute (1:100) HNO₃ (Suprapur, Merck), and 0.25 mg/L of Nd. Distilled and demineralised water (Millipore, Bedford, MA, USA) was used to prepare all solutions. Samples were passed through a 0.45 μm nylon syringe filter and diluted 1:10 using water.

Growth was observed by counting cells using a Bürker chamber under a light microscope. Observations under transmitted light were carried out using a BX51 microscope (Olympus, Japan). Values were expressed in number cells mL⁻¹. Dry weight was determined gravimetrically. Aliquots (2 mL) of samples were harvested from the photobioreactors in pre-weighed tubes by centrifugation. The supernatant was discarded and pellets were dried at 105°C until they reached a constant weight (Giovanardi et al., 2014).

For the current experiment, samples were taken every 24 h for 3 days; although PAM samples were taken twice a day (every 24 and 30 h) (see Section Photosynthesis as Chlorophyll Fluorescence Measurements). Every treatment was independently replicated three times. To check that the microalgae were not seriously damaged by the deficit conditions, we carried out an experiment where we separately replaced the deficient medium with the original standard nutrient medium (hereafter as “recovery condition” or Rec).

Specific growth rates in the preliminary experiment (Table 3) were determined from the slopes of linear regressions of the natural log of cell concentration vs. time for the data plotted in Figure 1.

In vivo chlorophyll a fluorescence was determined using a Junior-PAM fluorimeter (Walz GmbH, Effeltrich, Germany), provided with blue light emitting diodes for measuring excitation and actinic light. Rapid light curves (RLCs) were carried out twice per day by sampling 5 mL of culture and transferring this to light-protected chambers for dark adaptation (15 min) in order to measure F_o (basal fluorescence in dark adapted samples). After that, a saturating flash (600 ms ~ 9000 μmol m⁻² s⁻¹) was applied in order to obtain maximal fluorescence (F_m). Maximal quantum yield (F_v/F_m) was calculated according to Schreiber et al. (1986) and the effective quantum yield (ΔF/F'_m) was calculated as ΔF/F'_m = (F'_m – F_t)/F'_m (Schreiber et al., 1995b), where F'_m represents the maximal fluorescence and F_t the current steady-state fluorescence in light adapted algae. Samples were exposed for 20 s to twelve increasing E_PAR levels between 0 and 1500 μmol photons m⁻² s⁻¹ to conduct RLCs according to Schreiber et al. (1995a). Relative electron transport rates (rETR, μmol electrons m⁻² s⁻¹) were computed by multiplication of ΔF/F_m' and the incident irradiance (E, μmol photons m⁻² s⁻¹) as given by the Junior-PAM. rETR values were fitted according to Eilers and Peeters (1988) using a least squares error calculation and the Solver function of Excel (Microsoft, Redmond, U.S.A.) in order to obtain photosynthetic parameters i.e., photon-capturing efficiency of PSII in the light limited range (α), maximum rETR (rETR_max), and the light saturation irradiance (E_k).

All experiments were repeated at least twice. Two different Two-Way ANOVA analyses were performed: (1) to determine significant differences (p < 0.05) among Ca²⁺ and Mn²⁺ treatments and (2) to determine significant differences (p < 0.05) between treatments and controls. In the case of significant effects, the Student–Newman–Keuls post-hoc test was applied (Underwood, 1997). Three replicates (n = 3) of each nutrient treatment and controls were used for each sampling time. The software Statistica for Windows (version 7.0, Statsoft, Inc., 1984–2004) was used for analyses. Data were presented as means ± SD.

In prior rapid tests to establish the correct concentration range, we confirmed that non-essential Ln produce biological effects on algae and have demonstrated stimulatory and toxic effects on growth at lower (<25 μmol L⁻¹) and higher concentrations (>25 μmol L⁻¹), respectively (Figure 1).

In one graphic example of exposure to Ln, we showed that concentrations of cerium of 3, 6, and 12 μmol L⁻¹ produced an increase of the specific growth rate of D. quadricauda of 13, 16 and 7%, respectively; which started to decrease after the exposure to 23 μmol L⁻¹ (Figure 1, Table 3).

We measured the concentrations of Ca and Mn in each medium using ICP-MS. Replete medium contained 48.6 μmol L⁻¹ Ca and 7.2 μmol L⁻¹ Mn, respectively. Our depletion techniques greatly reduced their concentrations to 3.7 and 0.067 μmol L⁻¹ for Ca and Mn, corresponding to 8 and a 1% of the original medium concentration, respectively (see Table 4). In the case of treatments with lanthanides, we tested each medium by exposure to 9.9 or 1.9 μmol L⁻¹ of neodymium, as an example of Ln, for Ca and Mn, respectively. There was a serious reduction in the concentration of Nd in each medium after 3 days of algal growth (Table 4).

The growth of D. quadricauda under complete mineral medium as a control (Ctrl), or under calcium- or manganese-deficient mineral medium is shown in Figure 2. The control and deprived conditions are graphically represented with the red and blue lines, respectively. Our results showed that a deficiency in either metal (independently), but especially Mn²⁺, significantly decreased cellular growth of the microalga (p < 0.05) (as “Def” in Figure 2). Re-establishment of the standard medium resulted in recovery of growth in the metal-deprived strain (as “Rec” in Figure 2).

The effect of metal ion-deprivation (Ca²⁺ and Mn²⁺) on the growth of D. quadricauda exposed to different lanthanides (at low concentrations) is also shown. In the case of the Ca²⁺-deficient experiment, all Ln treatments increased cellular growth in comparison with the ion-deprived nutrient medium, to reach levels close to the standard conditions (Figure 2). By contrast, none of the Ln treatments alleviated the deleterious effects of Mn deficiency on cellular grow. Instead, the addition of Ce, Eu, and Gd led to a further decrease in cellular growth in the Mn deficient cultures (Figure 2). Significant differences in growth (expressed as dry weight) for the treatments, compared to controls, are shown in Supplementary information Figure S1, Table S1.

The effects of treatment on in vivo chlorophyll fluorescence of D. quadricauda are shown in Figures 3, 4, and Table 5. Under complete mineral medium, the maximum quantum yield of PSII (F_v/F_m) showed no significant differences (p < 0.05) between the first and second days. Under these standard (replete) conditions, the F_v/F_m mean value was 0.66 ± 0.00 (Table 5). However, omission of either Ca²⁺ or Mn²⁺ from the culture medium significantly decreased the maximum quantum yield. These nutrient stresses (metal-limited conditions) in microalgae were detected by a 21% ± 0.05 decrease (Ca²⁺) or 88% ± 2.50 decrease (Mn²⁺) compared with the controls (mean ± SD, n = 3; Figures 3, 4, respectively). In the first 6 h, stresses produced by sampling, centrifugation and nutrient medium replacement was observable in all tests (control and treatments), with recovery within the first 24 h. Only in the case of Ca²⁺ (Figure 3), but not Mn²⁺-deficiency (Figure 4), did the addition of low concentrations of Ln³⁺ produce a recovery in F_v/F_m and apparently alleviated the symptoms of an ion-deficit.

rETR_max values, determined from the RLCs, were significantly higher (p < 0.05) in the first hours under standard than under Ca²⁺-deficient conditions (“Def” see Supplementary information Figure S3). After 30 h, the rETR_max increased in all Ln treatments under Ca²⁺-deficiency, reaching values significantly higher (p < 0.05) than in the controls (Figure 3). After that, values started to decrease, and although after 48 h, they were significantly higher than values under deficiency conditions (Def), no significant differences were found compared to the control.

On the other hand, treatments of the algae with Ln, under Mn²⁺-deficiency, showed lower rETR_max values compared to those of the deficiency conditions during the entire experiment (Figure 4, Table 5).

Photosynthetic efficiency (α), obtained from the RLCs, was significantly higher in cultures grown in standard nutrient medium than under either ion-deficiency conditions, or in Mn²⁺-deficient + Ln³⁺ treatments. Almost no differences were observed between the Mn²⁺-deficient control and those after exposures to Ln³⁺ (Figure 4). However, under Ca²⁺-deficient treatments exposed to Ln³⁺, significant differences were observed. While in both controls, the photosynthetic efficiency (α) tended to decrease, cultures in Ca²⁺-deficient treatments exposed to Ce, Eu, and Gd achieved photosynthetic efficiency values significantly higher than those observed in the controls (Figure 3).

The saturated irradiance (E_k), obtained from the RLCs, showed significant differences between standard conditions and both deficiency treatments (Supplementary information Figure S3). E_k tended to increase within the first 24–30 h and after that a decrease was observed in all treatments under both nutrient deficiency conditions. In general, no significant differences were found between Ca²⁺-deficiency treatments and controls, although at certain times, when maximum values were achieved in some treatments, values were higher than those in the controls (see Figure 3). For Mn²⁺-deficiency treatments, values were similar to those in the controls in the first hours although in some of the treatments i.e., Ce, Eu, La and Nd, they were lower than those in the control at the end of the experiment (Figure 4).

Replacement of the deficient medium (either Ca²⁺ or Mn²⁺) with standard conditions produced a significant recovery, observable in all fluorescence (F_v/F_m, rETR_max, α, and E_k) parameters (Figures 3, 4).

Ln are non-essential elements that have been shown to produce diverse physiological effects (Jin et al., 2009; Tai et al., 2010). In the last two decades, they have been suggested to play possible roles in terrestrial organisms as replacements for essential elements like Ca²⁺, Mg²⁺, and Mn²⁺ (Brown et al., 1990; Squier et al., 1990). This was also observed for microalgae in our Ca²⁺ treatment experiments (Figure 3), but further studies are needed to identify specific mechanisms of action. In this sense, Wang et al. (2003) stated that the biological behavior of non-essential metal ions like Ln³⁺ originates from the principle of analogy to an essential metal ion. Their properties are close to but not the same as the original. Therefore, displacement (by Ln) can have different consequences depending on the role of the native metal, possibly explaining why, in some cases, functionality can be retained (Yocum, 2008). This could explain our present (and divergent) results when exposing Ca²⁺-deficient and Mn²⁺-deficient algae to lanthanides (Figures 2–4).

Of all the metals, Ca²⁺ may exert the widest range of effects on biological systems, including functions related to structure, regulation of enzymatic activity and intercellular and intracellular signaling (Brand and Becker, 1984; Yocum, 2008). In previous studies with microalgae under Ca²⁺ deficient conditions, photosynthetic oxygen evolution and respiratory oxygen uptake were severely affected, in parallel with growth rate and chlorophyll content (Adam and Issa, 2000). These authors suggested that many enzymes (e.g., arginine kinase, adenosine triphosphatase, adenylate kinase) that are regulated by this metal are directly involved in vital processes such as photosynthesis and respiration. Dvořáková-Hladká (1976) also associated Ca²⁺ with energy, nitrogen and phosphorus metabolism of microalgae such as Scenedesmus obliquus. This could explain the observed decrease in physiological state of the deprived green algae (Figures 2–4).

Manganese (and Ca²⁺) exist in the oxygen-evolving complex of plant and algal PSII, and participate in the water-splitting reaction; moreover, they could be involved in maintenance of the chloroplast structure. Photosynthetic water oxidation takes place at a catalytic Mn(4)-Ca site within the oxygen-evolving complex of PSII, which is embedded in the thylakoid membranes of green plants, cyanobacteria and algae (Yachandra and Yano, 2011; Hou and Hou, 2013). Mn²⁺ can be also a redox cofactor or activator at metal-binding sites of many enzymes and coenzymes (e.g., manganese superoxide dismutase), so the redox balance and PSII are expected to be the prime targets of Mn²⁺-deficiency in photosynthetic organisms (Cao et al., 2011). The important role of these two metals in algae, and the effects of deficiencies on algal physiology (e.g., slow growth, structural alterations, accumulation of less chlorophyll, loss of PSII and enzyme activity), have been previously reported for a few marine phytoplanktonic species and freshwater green microalgae (Constantopoulos, 1970; Dvořáková-Hladká, 1976; Adam and Issa, 2000; Allen et al., 2007; Cao et al., 2011; Hsieh et al., 2013, and references therein).

Physiological stress by nutrient limitation was specifically measured not only by declines in cellular growth and reproduction, but also in terms of photosynthetic parameters. PAM fluorometry has been shown to be an effective method to study stress (see Komenda, 1998; Mallick and Mohn, 2003; White et al., 2011; Figueroa et al., 2013; Giovanardi et al., 2014), and in this work, demonstrated a significant decrease in F_v/F_m, rETR_max, E_k, and α in comparison with standard conditions (Table 5, Figures 3, 4). For comparison, values of F_v/F_m expressed for different species of the genus Desmodesmus under different standard nutrient media ranged from 0.60 to 0.74 (see Komenda, 1998; Koblížek et al., 2001; Karsten et al., 2007; Hu et al., 2013; Samorí et al., 2013). Interestingly, by adding individual Ln to the Ca²⁺-deficient treatment we observed different, although irregular, increases in F_v/F_m, rETR_max, E_k, and α, toward the reestablishment of standard conditions in the control (Ctrl). This is surprising because previously, it was demonstrated that although Ln are successful competitors with Ca²⁺ for binding sites in PSII, none of them retained functionality and treatment did not result in reactivation of oxygen evolution activity (Ghanotakis et al., 1985; Bakou et al., 1992; Bakou and Ghanotakis, 1993; Ono, 2000; Yachandra and Yano, 2011). There is some support for our observed Ln³⁺ stimulation of the growth of the control cultures (Figure 1) and Ca-limited treatments (Figure 2). Kruk et al. (2003) described a possible stimulation of oxygen evolution in PSII membranes by low concentrations of Eu³⁺ and Dy³⁺ ions. Under other nutrient-deficient conditions, Huang et al. (2008a,b) and Liu et al. (2008) demonstrated alleviation effects of Ce³⁺ on the photosynthetic rate (electron transport rate) of spinach chloroplasts under Ca²⁺-deficiency and the same metal improved ETR and yield values of Mg²⁺-deficient maize (Zhao et al., 2012). The stimulatory effect (or reduced stress) on D. quadricauda could be derived from secondary processes affecting algal physiology, but, because Ca²⁺ has a multifaceted role in photosynthesis (see Brand and Becker, 1984), it will be necessary to undertake a detailed molecular study. It has been suggested that Ln could modulate plant photosynthesis by interactions with K⁺, Na⁺, or Ca²⁺, ribulose-1,5-bisphosphate carboxylase/oxygenase, oxidative damage and redox systems, and indolyl-acetic acid (Chen et al., 2000; Wang et al., 2011, and references therein).

In the case of the Mn²⁺-deficiency treatments of D. quadricauda, there was no such stimulation by adding Ln to the deficient medium. Furthermore, in many cases the algae were more stressed than under deficient conditions, as was reflected by even lower values for the photosynthetic parameters (Figure T4). There is only one previous report of Mn²⁺-deficient maize treated with Ce³⁺ (Qu et al., 2012). The authors demonstrated that the Ln could significantly relieve reductions in F_v/F_m, Y(I and II), ETR(I and II) and the photochemical quenching coefficient (qP), among other parameters, as compared to those of the control. They suggested that Ce³⁺ may improve the function of PSI and PSII under Mn²⁺-deprived stress, although the mechanisms are unknown. In our work, we did not observe any improvement. Furthermore, although certain chemical similarities with Ln, this metal serves a key redox role in water oxidation in the Mn4 clusters of PSII (Hou and Hou, 2013), a highly specific function that other metals are probably not able to duplicate.

We want to make clear that these are simulation experiments where most of the conditions are controlled. We used a monoculture, a standard enriched medium, bubbling with a high CO₂ level, and concentrations of Ln that were picked subjectively (related to algal behavior and the original concentration of the omitted metal). In this way, although it gave us valuable hints to study the effects of Ln on algae, it is by no means an ecological or environmental study, requiring more natural conditions.

Lanthanides represent a potential environmental threat, particularly in high metal-exposure areas such as sites for mining, refining and recycling of REEs (Tyler, 2004). As these elements have become indispensable for a number of critical technologies, the demand for REEs in the next few years is expected to increase (Loell et al., 2011). Similarly to trace elements, Ln exhibit both positive and negative effects on algal growth and development, at low concentrations and high concentrations, respectively (Chen et al., 2000). Nevertheless, as we have demonstrated, their behavior is not simple, and in certain cases, even at low concentrations, Ln can be toxic to microorganisms. These and further studies are essential to understand the physiological and ecological effects that Ln produce in nature.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.