Introduction

Water pollution is one of the major problems of our time. Domestic, agricultural, and industrial water use are the main causes of water pollution, that can, if not properly treated, contaminate receiving waterbodies having deleterious effects on both human health and entire aquatic ecosystems (Masotti and Verlicchi 2005; Schwarzenbach et al. 2010). Among a large range of potential pollutants, nutrients are generally the most common, with domestic and agricultural wastewater often containing high levels of nutrients (mainly nitrogen and phosphorus) that are often responsible for the eutrophication of receiving waters (Welch and Lindell 2004; Yang et al. 2008).

Nature-based solutions (NBS) in the management of water resources, i.e., solutions that are inspired and supported by nature to face socio-economic and environmental challenges, such as eutrophication (Unesco 2018), are becoming an increasingly attractive option. Phytoremediation falls under the NBS umbrella and generally represents an environmental-friendly and cost-effective technique for water re-sanitation by the removal of diverse pollutants by using aquatic plants.

Lemna species (commonly known as duckweeds) are small aquatic vascular plants, consisting of floating leaflets (fronds) and submerged roots, which generally grow in stagnant or slow-flowing nutrient-enriched waters throughout tropical and temperate zones (Landolt and Kandeler 1987; Landolt 1992). Most Lemna species have been shown to be efficient in wastewater phytoremediation, as they are perennial, have a rapid vegetative growth, can produce high biomass per unit area, and can bioaccumulate various contaminants, such as organic pollutants, nutrients, heavy metals, and phenols (Mkandawire and Dudel 2007 and reference therein; Pietrini et al. 2016). The absence of a vegetative rest in some Lemna species (e.g., L. minor L., L. gibba L., L. minuta Kunth) in temperate and tropical areas (Landolt 1992; Ceschin et al. 2016b) signifies that these species can effectively be used for wastewater treatment the whole year-round (Mkandawire et al. 2004). Various studies highlight the high vegetative growth rates of some Lemna species (Lemon et al. 2001; Ziegler et al. 2014) and their rapid colonization of wide areas, forming dense free-floating mats (Dussart et al. 1993; Ceschin et al. 2016b). To sustain this high growth, nutrient uptake from water is high and rapid (Körner et al. 2003) and this renders duckweeds very effective in phytoremediation of nutrient-enriched waters (Cheng et al. 2002; Raju et al. 2010; Cui and Cheng 2015).

Although the phytoremediation potential of Lemna species is high, some problems related to their tendency to form thick and dense floating mats have been identified (Ceschin et al. 2019b). The formation, over time, of overly thick Lemna mats, results in the death of Lemna fronds in the undersurface of the mat. An increase in microbial degradation of the dead fronds results in a release of nutrients to the water (Laube and Whole 1973) and essentially stalling any phytoremediation activity. However, these problems can be avoided by active periodic harvesting of fronds to prevent the formation of necrotic layers and to encourage continued exponential growth (Xu and Shen 2011; Ceschin et al. 2019b).

Although several studies have focused on the performance of the common European duckweed L. minor to remove nutrients from wastewater (Cheng et al. 2002; Ozengin and Elmaci 2007; Patel and Kanungo 2010; Raju et al. 2010), on this topic, there is very little available information on the American duckweed L. minuta. One study on the Functional Response (FR) and Relative Growth Rate (RGR) of L. minor and L. minuta at different nutrient concentrations showed no differences between the two species (Van Echelpoel et al. 2016), even though other studies have highlighted that L. minuta has a much higher growth rate compared with L. minor (Njambuya et al. 2011; Ziegler et al. 2014; Paolacci et al. 2016). In a constructed wetland system for domestic wastewater treatment, where a mixed Lemna community that included L. minuta was utilized (Ceschin et al. 2019b), little phytoremediation activity was shown because of a mat over-growth. In other phytoremediation studies that did not include nutrient removal, L. minuta was tested as bioaccumulator of other contaminants, such as phenols and heavy metals (Chiudioni et al. 2017; Paisio et al. 2017). Therefore, it was decided to investigate through water physical and chemical measurements and physiological and biochemical analysis the ability of the American L. minuta and the European L. minor to remove nitrate and phosphate from simulated domestic wastewater. In addition, it was verified which of the two species of Lemna showed the better performance in removing and bioaccumulating these nutrients. The inclusion of L. minuta, that is an invasive species in many European countries (DAISIE 2008; Ceschin et al. 2018a), is justified because its more rapid growth rate than L. minor could suggest that it has a larger phytoremediation potential.

Materials and methods

Experimental design

The ability of L. minuta and L. minor to remove nitrate and phosphate from simulated wastewater was tested through two synchronous outdoor experiments (L. minuta_Test, L. minor_Test). Samples of both Lemna species were collected from two wetlands located in the Appia Antica Regional Park (Rome, Italy). For the certain identification of the two species, which are very similar morphologically (Landolt 1986; Iamonico et al. 2010), the collected Lemna samples were observed under a stereoscope (Olympus SZX16) and identified according to Ceschin et al. (2016c). In particular, the discrimination between the two species was performed considering the combination of some morphological characters, such as vein number (L. minuta: 1 vs L. minor: 3), frond length (mm) (0.8–2.0 vs 1.0–8.0), width (mm) (0.9–1.6 vs 1.3–3.3), and frond area (mm2) (0.8–2.8 vs 2.2–10.4).

Samples of each species were maintained separately and grown in cylindrical PVC tanks (diameter 74 × height 32 cm), filled with tap water (70 l) enriched with 3 g of KNO3 and 4 g of KH2PO4, obtaining solutions with 6.15 mg/l (6.10 × 10−2 mMol/l) of NO3− and 9.0 mg/l (6.60 × 10−2 mMol/l) of H2PO4−, to simulate concentrations of these nutrients usually found in domestic wastewater (Borin 2003; Whitton et al. 2018). The tanks were then inoculated with Lemna fronds (around 700 g) of both species, arranged to form a uniform free-floating mat on the water surface with an initial thickness of 0.5 cm. An experimental control consisted of a third identical PVC tank containing the same medium but with no added Lemna fronds (Control_Test). The experiments were run synchronously and replicated three times for both species and control.

In each Lemna_Test, water physical and chemical factors (also in Control_Test), and frond physiological and biochemical parameters, were analyzed at the start of the experiment and then every 7 days for 4 consecutive weeks (0 days, 7, 14, 21, 28). From each tank, one representative water and Lemna sample was taken and used to carry out the diverse measures (see following sections, n = 3).

Physico-chemical analysis

On each sampling occasion, water temperature (T, °C), pH, conductivity (Cond, μs/cm), and dissolved oxygen (DO, mg/l) were measured using an immersion multi-probe (Hach Lange 40 HQD). Water samples were collected in sterile 50-ml tubes at a depth of 15 cm from the water surface, and nitrate (NO3−, mg/l) and phosphate (H2PO4−, mg/l) concentrations were analyzed by colorimetrically using standard protocols (standard Hach Lange Method 8039 and 10209/10210, respectively). In addition, for a characterization of the microclimatic conditions under which the outdoor experiment was run, air temperature (Tair,°C) and relative humidity (RH, %) were measured using a thermo-hygrometer (Lafayet DT8820). The values of these last parameters are reported in Table S1.

Plant growth analysis

Variations in Lemna mat thickness (mm) and weight (g) were measured for evaluating biomass production during the experiment. Mean thickness (mm) of the Lemna mat was calculated by measuring in situ the thickness of five random undisturbed sections of mat using a precision digital caliper. Mean weight was measured on three samples taken with a simple instrument of known section, and these were then dried for 24 h at 70 °C. The final measurements were given as grams of dry weight (g, DW).

For each Lemna species, on the whole experimental time (28 days), Relative Growth Rate (RGR) (g/g day) was calculated using the following equation (Radford 1967):

$$ \mathrm{RGR}=\ln {\mathrm{DW}}_f-\ln {\mathrm{DW}}_i/{T}_f-{T}_i $$

where DWf = final dry weight (g), DWi = initial dry weight (g), Tf = overall incubation period (day), and Ti = day zero.

Determination of nitrate and phosphate content in Lemna fronds and bioconcentration factor

Ten-gram samples of each species were collected weekly for the determination of nitrate and phosphate content in the fronds, making sure to take the full depth profile of the mat. Fresh weights (FW, g) were determined after rinsing and blotting on tissue paper. Samples were then dried at 65 °C for 3 days. The inorganic anions nitrate (NO3−) and phosphate (H2PO4−) extractions broadly followed D’Imperio et al. (2018). In particular, recoveries of NO3− and H2PO4− were previously determined from 0.5-g dry plant samples with 50 ml solution of carbonate (3.5 mM) and sodium bicarbonate (1 mM), which was then stirred on an orbital shaker (OS-10) for 20 min at a speed of 140–150 rpm. The solutions were then filtered through a 0.45-μm nylon membrane filter and nitrate and phosphate determinations were performed using ion exchange chromatography (ICS-1000 Ion Chromatograph - Dionex).

To evaluate Lemna ability to bioaccumulate nitrate and phosphate in tissues, the bioconcentration factor (BCF) was determined on the Lemna samples according to Pietrini et al. (2016):

$$ \mathrm{BCF}=\mathrm{nutrientconcentrationinplantbiomass}\left(\mathrm{mg}/\mathrm{kg}\right)/\mathrm{nutrientconcentrationinsolution}\left(\mathrm{mg}/\mathrm{l}\right) $$

A higher BCF value indicates greater bioaccumulation ability of a species. Plants with BCF values ranging from 1 to 1000 are accumulators, and those with values greater than 1000 are considered as hyperaccumulators and would be more suitable for use in phytoremediation (Baker 1981; Zayed et al. 1998; Yoon et al. 2006).

Determination of chlorophyll and malondialdehyde content in Lemna fronds

Total chlorophyll content was determined according to Huang et al. (2007) where fresh Lemna fronds (0.2 g) were soaked in 10 ml of 95% (v/v) ethanol for 3 days in a stoppered tube at room temperature in the dark. The samples were then centrifuged at 3000×g for 10 min and the absorbances at 663 and 645 nm of the supernatant were determined. The concentrations of chlorophyll were calculated according to the following equations and the results expressed in mg of total chlorophyll per gram of fresh weight plant tissue (mg/g FW):

$$ \mathrm{Chl}a=12.72{A}_{663}-2.69{A}_{645} $$
$$ \mathrm{Chl}b=22.90{A}_{645}-4.68{A}_{663} $$
$$ \mathrm{Chl}\ \mathrm{Tot}=\mathrm{Chl}\ a+\mathrm{Chl}\ b $$

where Chl a, Chl b, and Chl Tot represent the content of chlorophyll a, chlorophyll b, and total chlorophyll; A663 and A645 are the absorbance at 663 and 645 nm.

Lipid peroxidation was determined by estimation of the frond malondialdehyde (MDA) content following the protocol of Heath and Packer (1968), considering MDA as biomarker of oxidative damage of plant tissue (Bailly et al. 1996). Frozen samples were homogenized in a pre-chilled mortar and pestle with two volumes of ice-cold 0.1% (w/v) trichloroacetic acid (TCA) and centrifuged for 15 min at 16.000×g. An assay mixture containing 1 ml aliquot of supernatant and 2 ml of 0.5% (w/v) thiobarbituric acid (TBA) in 20% (w/v) TCA was heated to 95 °C for 30 min and then rapidly cooled in an ice bath. After centrifugation (16.000×g for 10 min at 4 °C), the supernatant absorbance was read at 532 nm, and values corresponding to non-specific absorption at 600 nm were subtracted. MDA concentration was calculated using the extinction coefficient (ε = 155 mM/cm).

Statistical analysis

One-way analysis of variance (ANOVA) was performed to check differences in plant physiological and biochemical parameters (biomass, nutrient and chlorophyll content, peroxidative damage, BCF) in response to nutrient-enriched water. The differences were compared by employing the Tukey’s test with a significance of P < 0.05, using R statistical package vers. 3.5.0 (R Core Team 2018).

Results

L. minuta_Test

In L. minuta_Test, water pH and conductivity remained constant, while dissolved oxygen, temperature, and nutrient concentrations, changed over the time (Fig. 1). Final dissolved oxygen levels (2.88 mg/l) were over 65% lower than initial values (8.39 mg/l), while water temperature increased by about 30% ranging from 15.0 to 19.8 °C. Aqueous phosphate levels decreased from 9.2 to 4.43 mg/l (50%) over the full duration of the experiment, but minimum values were reached by day 21 (3.50 mg/l) representing a 62% reduction after which there was a little increase. Nitrate concentrations decreased from 6.1 to 0.94 mg/l (85%).

Fig. 1
figure 1

Absolute values of water temperature (T), dissolved oxygen (DO), and nutrients (nitrate and phosphate) in L. minuta_Test, L. minor_Test, and Control_Test (mean ± SD, n = 3)

Full size image

Frond phosphate content increased significantly (165%) from about 11680 to about 31000 mg/kg DW after 21 days and then remained constant for the last 7 days (Fig. 2a). Over the first 7 days, frond nitrate content significantly increased by over 10% from 8500 to 9320 mg/kg DW during the first week, after which it decreased slowly until the end of the incubation period (Fig. 2b). BCF values greater than 1000 were recorded both for phosphates (mean value = 6830) and nitrates (2310) (Fig. 2c).

Fig. 2
figure 2

Phosphate (a, d) and nitrate (b, e) content in L. minuta and L. minor fronds and relative BCF (c, f) (mean ± SD, n = 3). Means with different letters represent significant difference at P < 0.05

Full size image

L. minuta biomass slowly increased for the first 21 days then rapidly increased from day 21 to 28. There was overall 10-fold increase from 29.8 at day 0 to 306 g DW at day 28 (Fig. 3). In parallel, the Lemna mat doubled in thickness, growing from 0.5 cm to over 1 cm. The RGR calculated for L. minuta was 0.083 ± 0.001 g/g day.

Fig. 3
figure 3

L. minuta biomass recorded during the experiment (mean ± SD, n = 3). Means with different letters represent significant difference at P < 0.05

Full size image

Total chlorophyll content almost linearly decreased by 43% from 0.96 to 0.55 mg/g FW (Fig. 4) after 28 days. A significant increase in MDA content (80%) occurred across the first 14 days from 2.35 to 4.23 nmol MDA/g FW, after which there was a plateau (Fig. 4).

Fig. 4
figure 4

Total chlorophyll (Chl tot) and Malondialdehyde (MDA) content in L. minuta fronds (mean ± SD, n = 3). Means with different letters represent significant difference at P < 0.05

Full size image

L. minor_Test

In L. minor_Test, water temperature and dissolved oxygen levels changed through experiment, with temperature increased ranging from 15.2 to 20.1 °C (about 30%) and oxygen decreased from 8.87 to 6.12 mg/l (30%). Medium nutrient levels decreased significantly, with maximum reductions in phosphate from 8.43 to 2.23 mg/l (over 70%) by day 21 and nitrate from 6.18 to 3.06 mg/l (about 50%) by day 14, after which there was an increase in phosphate and nitrate by day 28 (Fig. 1). Frond phosphate content increased significantly from about 11400 to about 26400 mg/kg DW (130%) by day 14 and decreased thereafter (Fig. 2d). Frond nitrate content decreased rapidly from 720 to 280 mg/kg DW (60%) by day 7, and after that, levels remained almost constant (Fig. 2e). BCF values for phosphate were always higher than 1000 during the whole experiment (mean value = 3118), while for nitrate were lower than 1000 (mean value = 220) (Fig. 2f).

L. minor biomass decreased significantly during the experiment, resulting a threefold reduction from 86.6 to 29 g DW by day 28 (Fig. 5). The mean thickness of the Lemna mat mainly decreased by day 7 (from 0.54 to 0.42 cm) and then remained fairly constant for the remainder of the incubation period. At the end of the experiment, a negative value of RGR was calculated (RGR = − 0.039 ± 0.004 g/g day). Total chlorophyll content decreased (over 60%) throughout the experiment, dropping from an initial value of 0.47 down to 0.18 mg/g FW by the final day (Fig. 6). MDA increased linearly (over 150%) from 4.95 at day 0 to 12.81 nmol MDA/g FW at day 28 (Fig. 6) .

Fig. 5
figure 5

L. minor biomass (g) recorded during the experiment (mean ± SD, n = 3). Means with different letters represent significant difference at P < 0.05

Full size image
Fig. 6
figure 6

Total chlorophyll (Chl tot) and Malondialdehyde (MDA) content in L. minor fronds (mean ± SD, n = 3). Means with different letters represent significant difference at P < 0.05

Full size image

Control_Test

In Control_Tests, without floating Lemna mats, no significant variation of the water physical and chemical factors occurred, except for water temperature that changed in line with the changing air temperature.

Discussion

BCF values greater than 1000 were recorded for phosphate in the two Lemna species, identifying them both as hyperaccumulators for this nutrient. Lemna species have also been shown to be more effective than a range of other aquatic macrophytes in removing phosphate from treated eutrophic wastewater (Sudiarto et al. 2019). Additionally, it was found here that the alien duckweed L. minuta was more efficient in the phosphate bioaccumulation than the native L. minor, likely to support higher growth rates and biomass production. Some empirical studies have also demonstrated L. minuta grow faster and produce a greater biomass than L. minor (Ziegler et al. 2014; Ceschin et al. 2016a) and this difference become more evident when ambient phosphorus concentrations are higher (Njambuya et al. 2011; Paolacci et al. 2016).

Although both species were shown to be highly efficient in removing phosphate, only L. minuta can be considered a hyperaccumulator of nitrate (BCF > 1000). It should be noted that although the field-collected and utilized L. minor fronds had a much lower initial nitrate content than L. minuta (720 vs 8500 mg/kg DW), making to suppose a greater need by L. minor to uptake this nutrient from the enriched water, L. minuta was better for bioaccumulating nitrates than L. minor. This could suggest the two species have different kinetics of nitrate uptake and/or nitrate accumulation under the prevailing experiment conditions. As it stands here, it certainly seems as though L. minuta is more efficient than L. minor in the bioaccumulation of nitrates and that this also results in higher overall growth. Furthermore, L. minuta has already been shown to further increase its relative growth rate over that of L. minor with increasing levels of nitrogen availability (Njambuya et al. 2011; Paolacci et al. 2016; Ceschin et al. 2018b).

In both Lemna_Tests, phosphate and nitrate removal by the two Lemna species was not constant throughout the whole experiment (Figs. 2c, f). The bioaccumulation of phosphate by both species increased until day 21, then values decreased significantly by day 28, although this drop was much more evident in L. minor. The bioaccumulation values of nitrates in L. minuta were generally higher than those of L. minor and remained relatively constant with a small but significant decrease between days 7 and 14. Differently, in L. minor, the nitrate BCF values remained relatively steady throughout the full incubation period. Differences in frond nitrate contents in the two species, in spite of being prepared for the experiment under the same conditions, may partially explain their nitrate-BCFdifferences by possibly demonstrating metabolic differences in nitrate requirement (Glass et al. 2002). In L. minuta_Test, the combined recording of the higher BCF values, the more rapid increase of nitrate and phosphate in the fronds, and the concomitant reduction of these nutrients in the medium, seems to reveal in L. minuta a better performance of the regulation of high-affinity transporter system (Glass et al. 2002; Smith et al. 2003) which control nitrate and phosphate accumulation and resource uptake rate. These results might contribute to explain how the invasive success of L. minuta is further facilitated by water eutrophication in nature (Dukes and Mooney 1999; Paolacci et al. 2016).

The observed decrease in BCF for both species could be linked to the progressive reduction of available nutrients for Lemna fronds and physical stress due to the overcrowding (high plant density), phenomenon known to cause growth limitation by inhibition contact (Driever et al. 2005). In this study, as the mats developed an increase in peroxidative stress (increase in MDA content), a loss in vitality of fronds in the underlying layers of the mat, and a decrease in overall chlorophyll content could all be implicated in reducing the capacity of the Lemna species to bioaccumulate both nitrate and phosphate. The same stress responses were observed in both Lemna species, but they were more evident in L. minor (Figs. 4 and 6). A reduction in nutrient removal by Lemna species after about 2-week incubation has also been shown by Sudiarto et al. (2019) and was linked to fronds losing vitality and dying off. As a mat thickens so do, the underlying necrotic layers form a nutrient diffusion-slowing barrier between the water and the most vital superficial layers of Lemna and doing so, the processes of bioabsorption, uptake, and immobilization of nutrients are slowed (Landolt and Kandeler 1987; Boniardi et al. 1994).

Therefore, it is suggested that a periodical and partial harvesting of Lemna mats will maintain a higher phytoremediation performance in the removal of nutrients from eutrophic waters (Xu and Shen 2011). The periodic harvesting of Lemna biomass is useful, not only to eliminate non-vital Lemna fronds, but also to reduce the possibility that nitrate and phosphate from the decomposition of dead fronds are released back to the surrounding water environment. This phenomenon was observed in a recent case study of a constructed wetland system for domestic wastewater treatment in which the decomposition of dead fronds in a dense mixed Lemna mat acted as a direct source of nutrients resulting in a large decrease of the system performance (Ceschin et al. 2019b). Taking into consideration previously published evidence (Mkandawire and Dudel 2007; Xu and Shen 2011; Ceschin et al. 2019b) and the findings from the present study, a periodic harvesting plan of partial removal every 3 weeks is proposed. However, there is still a need to perform studies using real wastewater to determine the effect of other contaminants and environmental factors on the phytoremediation processes.

Conclusions

It is evident from the results of the present study that the alien L. minuta is more efficient than the native L. minor in nutrient removal from treated water and it should, therefore, be the obvious choice between the two species for being used in constructed wetland systems. However, it should be specified that the use of L. minuta in a phytoremediation plant be carried out under strictly controlled conditions, since it is recognized as an invasive alien species in many European countries (Ceschin et al. 2018a). In fact, an eventual spread of L. minuta from the phytoremediation system to surrounding natural aquatic habitats could imply problems of biological pollutions with serious impacts on the entire aquatic ecosystem (Pokorný and Rejmánková 1983; Dussart et al. 1993; Janes et al. 1996; Ceschin et al. 2016a, 2019a) that would be difficult to control and manage (Mariani et al. 2020). Therefore, measures should be put in place to prevent any possible spread of this alien duckweed to surrounding waterbodies. To prevent the possible diffusion of the small fronds of L. minuta across hydrographic networks, the using of fine meshed grids located at outlets and also above the tanks containing the L. minuta mats to hinder dispersal vectors such as wind or water birds and mammals, could be highly recommended.