The Effect of TiO 2 Nanoparticles on the Aquatic Ecosystem : A Comparative Ecotoxicity Study with Test Organisms of Different Trophic Levels

A comprehensive ecotoxicological assessment was carried out with Degussa VP nano TiO2 suspension applying a bioluminescent bacterium (Aliivibrio fischeri), algae (Pseudokirchneriella subcapitata, Scenedesmus subspicatus and Chlorella vulgaris), a protozoon (Tetrahymena pyriformis), the water flea (Daphnia magna) and an aquatic macrophyte, Lemna minor. TiO2 nanoparticles were toxic in the set of the conducted tests, but the toxicity level varied with the organisms and endpoints. According to our results the concentrations, the duration and the mechanisms of exposure are contributing factors to the toxicity of nanoparticles. The Tetrahymena phagocytic activity, the Daphnia heartbeat rate and the Lemna total chlorophyll content as ecotoxicity endpoints showed outstanding sensitivity. These organisms showed significant behavioural and physiological changes when exposed to low TiO2 nanoparticle concentrations (0.1 and 0.05 μg/L) considered to be lower than the predicted environmental concentration in surface waters. These results reveal the importance of behavioural and physiological assays in assessing the impact of nanoparticles and indicate that nanosized TiO2 may pose risks to the aquatic ecosystem.


Introduction
The risk assessment and the investigation of the adverse effects of emerging pollutants are common topics in the current literature.These potentially harmful chemical substances can be grouped as pharmaceuticals, detergents, cosmetics, industrial additives and agents, pesticides or nanomaterials [1,2,3,4].The growth of the environmental concern of nanomaterials is inevitable due to their increased manufacture and widespread use, which results in a need to focus on and explore their impact on the environment [5,6,7].In the near past scientists started to assess the risk of engineered nanoparticles (NPs) in different ecosystems.These investigations are still at an early stage.
Significant contradictions can be found between the results of different authors concerning the toxicity of nanoparticles.The effect of NPs depends on many physico-chemical factors and processes, which all affect the fate and behaviour of NPs [8].The effect of the particle shape, size, surface area, surface charge on the aggregation properties, nanosize range and ecotoxicity must be considered [9,10].An important factor is the adsorption capacity of the NPs onto different surfaces, including the cell boundary surfaces of microorganisms.Finally, we must consider the effects of other abiotic factors such as pH, ionic strength, water hardness, and the naturally occurring organic materials in the test medium [11,12].The standard test medium, which is essential for the test organism, may affect the dispersability, aggregation and sedimentation of the tested NPs [13].
The toxic properties of NPs have been revealed by determining the EC 50 , LC 50 , NOEC and LOEC values for different aquatic testorganisms [9,14,15,16,17,18,19,20].Kahru and Dubourguier [10] analyzed 77 effect values for nTiO 2 , C60, nZnO, nAg, SWCNT (single-walled carbon nanotube) sand, nCuO and MWCNTs (multi-walled carbon nanotubes).The applied test organisms were mainly crustaceans, bacteria, algae and fish species.They found that nAg and nZnO were the most toxic, nTiO 2 was classified as "harmful" with an L(E)C 50 value 10-100 mg L −1 .For all studied organism groups, the dispersability of the different NPs was the key factor in their aquatic toxicity. 1Department of Applied Biotechnology and Food Science, Budapest University of Technology and Economics, H-1111 Budapest, Szent Gellért sqr. 4

., Hungary
In order to gain a satisfactory picture of the environmental effects of NPs, more trials would be necessary with marine test organisms and several freshwater and terrestrial vertebrate and invertebrate species applying physiological responses.Data are also missing for terrestrial plants and other photosynthetic organisms as well [11].Therefore, there is a need for studies using environmentally relevant test organisms with feasible and sensitive measurement endpoints to reveal the unknown potential adverse effects.The ecotoxicological effects of NPs available in the literature shows high variability even within the same organism group [10].
The unicellular protozoa as test organisms are well applicable to determine the effects and toxicity of NPs on cell function [7].Protozoa play an important role in maintaining the balance of the ecosystem of microbial life, and they are important food source for larger creatures and the basis of many food chains.Therefore, the toxic effects of NPs on these unicellular organisms have great importance.Rajapakse et al. [21] exposed Tetrahymena thermophila cells to nTiO 2 for 24 h, which resulted in changes of membrane fatty acid profile, but no lipid peroxidation was detected.They interpreted these changes as acclimation to unfavourable conditions, not as toxic effects.
The algal growth inhibition tests are widely distributed in the risk assessment of waters.Pseudokirchneriella subcapitata (formerly Selenastrum capricornutum or Rhapidocelis subcapitata) and Desmodesmus subspicatus are commonly used freshwater algae testorganisms, which are proven to be very sensitive to heavy metal pollution [9,22].Ecotoxicity of nanosized TiO 2 on alga have been summarized in several papers in recent years [10,23].Extremely variable values were reported for the tested nTiO 2 , and there was no clear relationship found between the concentrations, primary particle size and toxic effect.
During experiments with Daphnia magna nTiO 2 was observed on the integument, antennas and in the digestive system of Daphnia.If greater quantities of nanoparticle aggregates accumulate on the surface of Daphnia it is likely to cause mobility problems [14].Daphnia magna behaviour (motility, feeding frequency) and heart beat rate were not affected during 60 minutes exposure to nTiO 2 [24].The standard deviations are very large between results published by different authors even in case of the same test organism and similar experimental conditions.This is probably because the applied TiO 2 NPs show a very wide range concerning the physico-chemical properties such as particle size, crystal form, coatings, surface area and purity.It seems likely that it is not the primary size of nanoparticles, but the secondary (formed by aggregation) size that affects the toxicity [25].
Kim et al. [26] investigated the toxic properties in Daphnia focusing on the activities of four antioxidant enzymes: catalase, superoxide dismutase, glutathione peroxidase, and glutathione-Stransferase.They found that mortality was significantly increased at 5 and 10 mg L −1 nTiO 2 concentration during the chronic bioassays; and the reduction of reproduction ability was not observed.
The antioxidant enzyme activities in D. magna were increased, but superoxide dismutase did not show a concentration-dependent increase.The size of the applied nTiO 2 was <40 nm.
Zhu et al. [27] tested Degussa P25 nTiO 2 with an average surface area of 50 m 2 /g and particle size of 21 nm (20% rutile; 80% anatase).Acute (72 h) and chronic (21 d) toxicity tests were carried out to investigate the accumulation of nTiO 2 in D. magna.They found minimal toxicity within 48 h exposure time, but high toxicity was registered within 72 h.In case of 21 d chronic test, D. magna displayed severe growth retardation, mortality, and reproductive defects, and a significant amount of nTiO 2 was found accumulated in D. magna.
Due to the heterogeneity of physico-chemical properties of TiO 2 NPs, it is important in all cases to document the characteristics of the used materials, because the chemical and physicochemical properties of a material essentially dictate the molecular-level interactions [25,28].
Since a daunting variety of TiO 2 NPs is produced and applied for ecotoxicity studies, it is complicated to discuss our findings in the reflection of the results of other authors.To support easier perspicuity and understanding, a summarizing table (Table 1) was prepared containing extensive data on the ecotoxic effect of different TiO 2 NPs from current literature applying the same test species as applied in our study.
This paper is a contribution to the general effort towards a better understanding of the ecotoxicity of engineered NPs in the aquatic ecosystem with a special emphasis on sublethal effects.

Materials and methods
The ecotoxicity of HCl activated Degussa VP P90 nano TiO 2 was investigated with test organisms from different trophic levels.A marine bioluminescent bacterium (Aliivibrio fischeri), algae (Pseudokirchneriella subcapitata, Scenedesmus subspicatus and Chlorella vulgaris), and a ciliate protozoon Tetrahymena pyriformis were applied as unicellular model organisms.The freshwater aquatic macrophyte Lemna minor (common duckweed) and a freshwater crustacean Daphnia magna (water flea) were used for the determination of adverse effects of nTiO 2 on higher trophic levels.

Applied nTiO 2 suspension
Degussa VP P90 nTiO 2 powder purchased from Evonik Resource Efficiency GmbH was activated by 0.1 M HCl.During the activation process 3 g of Degussa VP P90 nTiO 2 powder and 10 mL of 0.1 M HCl solution were homogenized thoroughly, then sonicated in a 65°C ultrasound water bath in a 100 cm 3 ground-neck round-bottom boiling flask for 15 min.Then the HCl solution was evaporated with a rotary vacuum evaporator merging the flask into a 50°C water bath.To this dry activated nTiO 2 powder, 23 mL of MilliQ ® ultrapure water was added in small proportions while homogenizing and the suspension was sonicated in a 65°C ultrasound water bath for 20 min.

Lemna minor
Growth inhibition-frond number (14 days, renewal) no observed effect between: 0.01-5 mg L −1 [39] -: Data not available DLS: Median values for particle size in medium determined with dynamic light scattering method *: If EC 10 is not available, other EC x values are given The Effect of TiO 2 Nanoparticles on the Aquatic Ecosystem 2016 60 4 The mass and the number size distribution were determined at 25°C by using a dynamic light scattering (DLS) device (Malvern Zetasizer ZS, Malvern Instruments, UK) that was operated with a He-Ne laser light at a wavelength of 633 nm.Light scattering was detected at an angle of 173°.Three replicate measurements were made immediately after dispersion in Mil-liQ ® ultrapure water.The characteristics of the applied nTiO 2 suspension are shown in Table 2.A dilution series was prepared from the original stock suspension by diluting with MilliQ ® ultrapure water.The preparation of this dilution series with MilliQ ® ultrapure water was crucial.Directly diluting the stock suspension with the inorganic salts containing medium should be avoided due to eliminating the possibility of any aggregation or agglomeration phenomena, hence the changing of the concentration of the nano-size particles in the medium.The members of the dilution series were added in appropriate proportion into the test systems to set the exact test concentration in the test medium.The applied nominal concentrations in case of all test organisms are collected in Table 3. ) 2 HPO 4 , 5 g peptone, 0.5 g yeast extract, 3 cm 3 glycerol per 1 L distilled water [40].

Unicellular alga species
The applied alga strains (Pseudokirchneriella subcapitata, Scenedesmus subspicatus, Chlorella vulgaris) were cultured and maintained on agar slant cultures in the laboratory.The following algal growth medium was solidified with 2% agar: Macroelement solution (100x dilution) ingredients for 1 L distilled water: NH 4 Cl: 1500 mg L −1 ; MgCl To ensure the validity of the data and sensitivity of unicellular algal species potassium dichromate as reference toxicant was measured twice a year.

Daphnia magna cultures
A colony of Daphnia magna cultured in the laboratory was used in this experiment.The test animals were cultured in a 5 L volume beaker in a 21.5±1ºC thermostatic chamber with 16:8 h light:dark cycle (illumination: Juwel Aquarium, Day-Lite, 15 W, 438 mm lamp, 560 Lumen, 6500 K).For the test adult (about 10 day old) female animals were used, which were fed every two days by an alga suspension cultivated in the laboratory containing Scenedesmus obtusiusculus.For maintaining Daphnia magna aged, dechlorinated tap water was used.Its electric conductivity value was presumably less than 500 mS cm −1 [41].
To check the sensitivity of the D. magna culture acute toxicity tests were performed with the potassium dichromate (K 2 Cr 2 O 7 ) as reference toxicant about every six months.Sensitivity of D. magna culture to K 2 Cr 2 O 7 was within the limits (EC 50 , 24 h = 0.6-2.1 mg L −1 ) as set by the guideline OECD 202.

Lemna minor cultures
A colony of Lemna minor cultured in the laboratory was used in this experiment.The test organisms were cultured in a 20x30x7 cm glass container in a 21.5±1°C thermostatic chamber with 16:8 h light:dark cycle (illumination: Juwel Aquarium, Day-Lite, 15 W, 438 mm lamp, 560 Lumen, 6500 K).Twoleaf L. minor individuals, cultivated in Hoagland's nutrient medium, were used for the test [42].3,5-dichlorophenol was used as a reference chemicals, to check the sensitivity of the Lemna minor at least twice a year.

Test methods and evaluation 2.3.1 Aliivibrio fischeri bioluminescence inhibition test
The Aliivibrio fischeri bioluminescence inhibition test was carried out by modification of the protocol described by Leitgib et al. [40].The luminescence intensity was measured with Fluostar Optima microplate reader after 30, 60 and 120 min of contact time.The evaluations of the results were carried out as described by Ujaczki et al. [43].Copper sulphate was used as standard toxicant in all experiments, to check the sensitivity of the Aliivibrio fischeri culture.

Algae growth inhibition test
Alga species applied for the experiment were in logarithmic growth phase.Alga cells were washed off from the agar slant with fresh, sterile alga growth medium.The cell number of the alga inoculum was determined using a Burker chamber and set to 3×10 6 cells/mL.100 µL nTiO 2 suspension was pipetted in the appropriate concentration to wells of a microtiter plate, then 100 µL alga suspension to each well.As a control, distilled water and alga growth medium were used.
The optical density of the members of the nTiO 2 suspension dilution was also tested for further correction.During the experiment the plates were continuously illuminated (Juwel Aquarium, Day-Lite, 15W, 438 mm lamp, 560 Lumen, 6500 K; 21.5±1°C).
The optical density of the alga suspension in the wells was measured over a period of five days every 24 h at four different wavelengths (405, 450, 490, 630 nm) by Dialab EL800 spectrophotometer.

Tetrahymena pyriformis phagocytic activity test
24 h age T. pyriformis culture was prepared for the test by pipetting 5 mL of the culture into 20 mL of sterile medium and 600 µL of antibiotic mix solution, then Tetrahymena cultures were incubated in a flat bottom flask at 21±0.5°C in a shaking incubator at 150 rpm for 24 h.The expansion cultures were checked under microscope for healthy morphology and motility.For the phagocytic activity bioassay, 500 µL of T. pyriformis cell suspension, 100 µL test solution and 400 µL Chinese ink solution were pipetted into an Eppendorf micro test tube and incubated for 30 min at 21±0.5°C under dark circumstances.The Chinese ink solution was prepared with Losina-Losonsky solution [44] and sterile filtered with 0.02 µm sterile filter.After 30 min of contact time the samples were fixed with 20 µL 1.5% formaldehyde solution.The Chinese ink particle granules formed by phagocytosis (test particles) were counted in 80 cells oculometrically with light microscope (Nikon CH20) under 400x magnification.Each measurement was repeated three times.

Daphnia magna heartbeat rate test
The experiment was carried out based on the assay of Villegas-Navarro et al. [45] and Dzialowski et al. [46] with modifications.For the test, 10 day old female non-pregnant animals were used, which did not derive from the first brood and were not fed during the test.Test vessels were kept under the same circumstances as described previously in case of D. magna laboratory culture.10 Daphnia magna individuals were placed into 50 mL of the test solutions in a 150 cm 3 volume beaker with the help of a special fabric spoon.As a control, 10 test animals were placed into 50 mL of the culturing medium in a 150 cm 3 volume beaker.After 24 and 48 h contact time the heartbeat rate of the Daphnia individuals was measured.The measurement was carried out applying NIKON SMZ800 stereomicroscope.The test animals were placed onto a single cavity microscope slide into a 50-100 µL droplet of the test solution, where the heartbeat rate of each test animal was measured three times for 10 seconds.The electric conductivity value of each sample was less than 500 mS cm −1 as recommended by Hebert et al. [41], the dissolved O 2 concentration of each sample was more than 3 mg L −1 as recommended by the OECD 202 Guideline [47].

Lemna minor growth inhibition test
The experiment was carried out with three parallels in 150 cm 3 beakers as described by Fekete-Kertész et al. [42].On the first day 10 healthy and two-leaf L. minor individuals were placed into 50 mL of each dilution member of the test suspension.Hoagland's nutrient medium was applied as control.The beakers were covered with a translucent plastic film to avoid evaporation and concentration of the test solutions during the experiment.The assembled test systems (beakers) were incubated in a 21.5±1°C thermostatic chamber for 7 days under the following light conditions: 16:8 h light:dark cycle (illumination: Juwel Aquarium, Day-Lite, 15W, 438 mm lamp, 560 Lumen, 6500 K).
On the seventh day L. minor individuals were removed from the test solutions, then surface-dried on filter paper to constant weight.The dried biomass was placed into ground-necked test tubes containing 5 mL of 96% ethanol.After 24 hours the optical density of the samples was determined spectrophotometrically (Sanyo SP55 UV/VIS spectrophotometer) at 470, 649 and 664 nm wavelength values.The total chlorophyll content The Effect of TiO 2 Nanoparticles on the Aquatic Ecosystem 2016 60 4 was determined based on the calculation described by Fekete-Kertész et al. [42].
All toxicity tests used a negative control there no test substance was present.These negative controls served as a quality control in an experiment as well as a reference point.Test samples were compared to these negative controls.

Statistical analysis
One-way analysis of variance (ANOVA) was performed by STATISTICA 12 ® software identifying significant effects (p < 0.05).Univariate Tests of Significance were performed and the homogeneity of variances was examined.In case of significance, the lowest observed effects concentration value (LOEC) was determined using Dunnett's test (α = 0.05).

Aliivibrio fischeri bioluminescence inhibition test
The effect of Degussa VP P90 nTiO 2 suspension was tested on Aliivibrio fischeri bioluminescence intensity.The luminescence intensity was measured with Fluostar Optima microplate reader after 30, 60 and 120 minutes of contact time.Table 4 summarises data on inhibition percentage values.After 30 and 60 minutes contact time only the most concentrated sample showed significant inhibitory effect on the luminescence intensity.After 120 minutes of contact time significant inhibition was found in the concentration range 100-100,000 µg L −1 .The lowest observed effect concentration was 100 µg L −1 with H%=18.8.

Alga growth inhibition test
The effect of Degussa VP P90 nTiO 2 suspension on alga growth rate was tested applying three different freshwater alga species (Pseudokirchneriella subcapitata, Scenedesmus subspicatus and Chlorella vulgaris).Each of the three applied alga species showed high sensitivity to nTiO 2 .In case of C. vulgaris 12,500; 25,000 and 50,000 µg L −1 concentration resulted in 33, 36 and 49% inhibition, respectively.In case of P. subcapitata the applied nTiO 2 caused 32-50% inhibition in the 3125-25,000 µg L −1 concentration range, while it was slightly toxic in the highest applied concentration (H%=10 in 50,000 µg L −1 ).In case of S. subspicatus, the applied nTiO 2 resulted in 20-45 H% in the 3125-50,000 µg L −1 concentration range (Fig. 1).Monotonic concentration-dependent adverse effect was experienced only in case of the C. vulgaris species.Significant inhibition is marked by asterisk (*).

Tetrahymena pyriformis phagocytic activity test
The modulation of phagocytic activity was investigated as a novel sublethal ecotoxicity measurement endpoint.Figure 2 represents the distribution of the dataset by Gaussian Kernel fitting.The number of test particles formed by phagocytosis is plotted based on the number of observations in 80 cells.As the supplement of Fig. 2, Table 5 collects data on the geometric mean of test particles in the samples.The applied concentration range was 0.1-10,000 µg L −1 in which significant inhibition effect was found within the 0.1-1,000 µg L −1 concentration range.We found an inverse relationship between the geometric mean values and concentrations.At the lowest applied concentration Tetrahymena could form half the amount of test particles as compared to the control sample.This ratio was nearly the same in case of the 1 and 10 µg L −1 samples (Table 5).At the highest concentration of nTiO 2 no significant inhibition was experienced.

Daphnia magna heartbeat rate test
The adverse effect of Degussa VP P90 nTiO 2 suspension was tested on D. magna heartbeat rate as an innovative sublethal ecotoxicity endpoint.The applied concentration range was 0.5-500 µg L −1 in which significant inhibitory effect was observed in all of the tested concentrations after 24 h of contact time (Fig. 3).Table 6 collects inhibition percentage (H%) data on the effect of Degussa VP P90 nTiO 2 on D. magna heartbeat rate.

Lemna minor growth inhibition test
The adverse effect of Degussa VP P90 nTiO 2 suspension was investigated on the total chlorophyll content of L. minor (Fig. 4).1-1000 µg L −1 concentration nTiO 2 resulted in 17-32% inhibition, while in case of higher concentration (10,000 µg L −1 ) visible aggregation was experienced and there was no significant inhibitory effect on L. minor total chlorophyll content (Table 7).
Fig. 4 The effect of nTiO 2 on Lemna minor total chlorophyll content.Significant decrease of the chlorophyll content is marked by asterisk (*).According to the results, the cladocerans D. magna and ciliated protozoon T. pyriformis demonstrated high acute sensitivity compared to the other test organisms.Compared to the results obtained with D. magna and T. pyriformis, the alga growth inhibition test applying three different freshwater alga species (Pseudokirchneriella subcapitata, Scenedesmus subspicatus and Chlorella vulgaris) was less sensitive to Degussa VP P90 nTiO 2 suspension.

Discussion
In our study the ecotoxicity of Degussa VP P90 nano TiO 2 was investigated with test organisms of different trophic levels.A marine bioluminescent bacterium (Aliivibrio fischeri), freshwater algae (Pseudokirchneriella subcapitata, Scenedesmus subspicatus and Chlorella vulgaris), and a ciliate protozoon Tetrahymena pyriformis were applied as unicellular model organisms.The freshwater aquatic macrophyte Lemna minor (common duckweed) and a freshwater crustacean Daphnia magna (water flea) were used to determine the adverse effects of nTiO 2 on higher trophic levels.
Supplementary information about the literature data discussed here, can be found in Table 1 regarding the properties of the applied nTiO 2 and the ecotoxicity measurement endpoints.
During our research the Aliivibrio fischeri bioluminescence inhibition test showed significant sensitivity (H% = 18.8-73.8)after 120 minutes of contact time in the 100-100,000 µg L −1 concentration range.However a few researchers reported that the ~6 nm and 25-70 nm particle size nTiO 2 had no inhibitory effect on A. fischeri neither at 20,000 mg L −1 or 0.4395-112.5mg L −1 concentration range [17,29].According to our results the toxic effect depends on duration of exposure.While 20 µg L −1 nTiO 2 did not show any difference in bioluminescence of Aliivibrio fischeri after 30 min and 60 min contact time, significant decrease (~20%) was observed after 120 min exposure.
In case of testing with algae, growth inhibition is a very common ecotoxicity endpoint.Most of the current literature data is acquired from growth inhibition studies.Ji et al. [30] found EC 30 (6 day)=30 mg L −1 , while Sadiq et al. [31] reported EC 50 (72 h)=16.12mg L −1 based on growth inhibition of Chlorella vulgaris.In our study the EC 50 value was 50 mg L −1 .
In our research the 73 nm particle size nTiO 2 caused 50% inhibition at 3.125 and 12.5 mg L −1 , but the effect was not concentration-dependent.Based on these results, we can observe that very small and very large nTiO 2 particle sizes exert lower inhibition than particles from an intermediate size range (25-70 nm).Scenedesmus subspicatus was nearly as sensitive as C. vulgaris to the 25 nm size nTiO 2 (EC 50 (72 h) = 21.2 and 16.12 mg L −1 ), respectively.In our research the 73 nm nTiO 2 particle size caused 45% inhibition at 50 mg L −1 , but the effect was not concentration-dependent.Authors reported that NPs can be bound on the cell surface of the algal species e.g.Pseudokirchneriella subcapitata and Chlorella sp.[31,48].Large nTiO 2 aggregates can entrap algal cells, which may contribute to the inhibition of algae growth [9].The mode of action of nTiO 2 particles to algal cells is still unknown, but they were shown to induce the production of reactive oxygen species, causing cell membrane damage, protein oxidation and possible DNA damage [26,49,50].
As seen in Table 1, the particle size, the specific surface area, the crystalline phase composition, and the presence of surface coating are variable factors, which may have strong influence on the toxicity even within the same particle size range.
The modulation of phagocytic activity by nTiO 2 was tested with Tetrahymena pyriformis.Currently only few data can be found in literature about the effect of nTiO 2 on T. pyriformis.According to our results, the phagocytic activity bioassay showed outstanding sensitivity to TiO 2 nanoparticles: administration of 0.1 µg L −1 nTiO 2 resulted in 50% inhibition.Ud-Daula [35] did not experience any inhibition effect of the <10 nm size nTiO 2 at up to 400 mg L −1 concentration in a 28 h contact time experiment.But there are some reports on decrease in membrane fluidity after exposure of Tetryhymena species to titanium dioxid [21].So our findings can be in associated with alterations in membrane structure.
Our microscopic studies provided exciting information about the fate of nTiO 2 particles inside the protozoan cell.Although the protozoan cells are able to ingest nTiO 2 particles, they are unable to digest them.The nTiO 2 particles in the test medium are ingested immediately and food vacuoles filled with nTiO 2 particles are formed.This filling of vacuoles was also detected by Rajapakse research group [21].In these vacuoles the cell stores the agglomerated nTiO 2 particles until they are exocytosed into the test medium as a larger aggregate (Fig. 5).Therefore, Tetrahymena can influence the bio-aggregation of nanoparticles in the aquatic environment [35].Throughout these processes of endocytosis and exocytosis, TiO 2 particles interfere with cell growth and consequently can induce acute toxicity.Because Tetrahymena can internalize and accumulate TiO 2 nanoparticles this accumulation may exert harmful effects via a food web transfer by increasing the risks of transfer of these NPs to higher trophic levels in the ecosystem.
At this phase of the research we aimed to study the nTiO 2mediated alterations in feeding behaviour and to assess potential impact on it.Then further studies are necessary to explore the mechanisms of inhibition of feeding behaviour and demonstrate how the nTiO 2 concentration and contact time influence the phagocytotic activity.
An innovative sublethal ecotoxicity endpoint, the heartbeat rate, was applied in a 24 h contact time Daphnia magna acute ecotoxicity test.The Daphnia magna heartbeat rate test presented outstanding sensitivity with significant inhibition percentage values of 15-20% at the applied concentration range 0.5−500 µg L −1 .Heinlaan et al. [17] did not experience inhibition effect of the applied 25-70 nm particle size nTiO 2 in case of 48 h contact time mortality test even at 20,000 mg L −1 concentration.Strigul et al. [29] reported no immobilization effect of a ~6 nm nTiO 2 particle size after 48 h in the 2.5-250 mg L −1 concentration range, while Clément et al. [38] measured 1.3, 3.15 and 3.44 mg L −1 EC 50 (72 h) values for immobilization when testing with 15, 25 and 32 nm size TiO 2 particles, respectively.Chronic reproduction tests could detect the adverse effect of nTiO 2 to D. magna in smaller concentration ranges.Seitz et al. [37] determined 0.06 and 0.2 mg L −1 as LOEC values in case of cumulative offspring and body length measurement endpoints in 21 days reproduction tests, respectively for the 6 nm particle size nTiO 2 .However, they did not find any inhibition effect when testing with a 21 nm particle size nTiO 2 at the 0.02-2 mg L −1 concentration range.Since D. magna is a filter feeding organism, it can accumulate a significant amount of TiO 2 NPs from the test environment, which may cause abnormal food intake or defecation, ultimately affecting physiological parameters, growth and reproduction [27].
Both Tetrahymena sp. and Daphnia sp. are important members of food chain in aquatic environments.Alterations in behaviour and physiology of these organisms can be indicator of larger ecosystem effects.
Our study demonstrated that TiO 2 nanoparticles were accumulated in protozoon and in water flea, consequently they can be transferred to higher trophic level animals with the occurrence of biomagnification.
The adverse effect of nano-scale TiO 2 suspension on the total chlorophyll content of L. minor was investigated.We experienced the aggregation and sedimentation of the nTiO 2 particles in the 10 mg L −1 concentration sample while the total chlorophyll content was not reduced in this sample compared to the control.In our study 0.001-1 mg L −1 concentration nTiO 2 resulted in 17-32% inhibition, sample compared to the control.
In our study 0.001-1 mg L −1 concentration nTiO 2 resulted in 17-32% inhibition, while Li et al. [39] reported no inhibition effect in the 0.01-5 mg L −1 concentration range, when testing a 21 nm particle size nTiO 2 .Li et al. [39] found that L. minor accumulated nano-TiO 2 by surface accumulation.The accumulation was caused by surface attachment of the particles onto plant cell walls with strong adhesion.Although TiO 2 NPs are not able to penetrate into the cells of plants and algae, the reactive groups of the plant cell walls give an opportunity to interact with nanoparticles in the test medium [31].
Currently there are only a limited number of data concerning observed environmental concentrations of TiO 2 nanoparticles.In most cases the levels of its removal and the release from waste water treatment plants were measured [25,49].The TiO 2 nanoparticle concentrations in sewage treatment plant effluents ranged from 1.37 µg L −1 to 220 µg L −1 [25,51].
A few studies modelled the quantities of nanosized TiO 2 released into the surface waters.
In our study the lowest observed effect concentrations determined by Tetrahymena pyriformis phagocytotic activity and Dahnia magna heartbeat rate tests were 0.1 µg L −1 and 0.05 µg L −1 , respectively.These quantified results of behavioural and physiological assays reflect the realistic concentrations that may occur in the ecosystem.Consequences of our study clearly indicate that nanosized TiO 2 may impact the aquatic ecosystem.

Conclusion
In the reflection of our results and the current literature data, the ecological consequences of TiO 2 NPs may have a greater importance than previously considered.Since bacteria, protozoa, unicellular algae, daphnids and duckweed are important members of the aquatic food chain and can serve as food to other aquatic organisms increased bioaccumulation can be expected even if results show that the applied TiO 2 NP has no adverse effect on the particular test organism.
This study applying sensitive endpoints clearly indicated the adverse effects of TiO 2 nanoparticles to aquatic life.In addition the study confirms the benefit of behavioural and physiological assays in assessing the impact of nanoparticles and demonstrates the uses of these sublethal endpoints to better understand exposure to nanoparticles at sublethal levels.As the production and the use of nanoparticles increases the exposure and the impact becomes more likely.
There are conflicting opinions about the testing of NPs, whether they should be tested in their stabile form by continuous medium renewal during the test or rather testing the environmentally more relevant aggregated form.Due to the tunable properties of NPs determined by their minute particle size both recommendations should be considered.The mode of action of nTiO 2 can vary from test organism to test organism and in most cases the mechanism underlying the interaction is still unknown.
Due to the high variety of TiO 2 NPs, data from different authors can be contradictory even if applying the same test organism and the same test protocol.Therefore extensive data is required to get a comprehensive perspective of nTiO 2 ecotoxicity.

Fig. 1
Fig. 1 Effect of the tested nTiO 2 suspension on freshwater alga species, given in inhibition percentages (H%).

Fig. 2
Fig. 2 Distribution of the number of test particles formed by phagocytosis in the dataset by Gaussian Kernel fitting.The results are plotted based on the number of observations in 80 cells.

Fig. 3
Fig. 3 Box Plot diagram of the effect of the tested nTiO 2 suspension on D. magna heartbeat rate after 24 h of contact time.Statistically significant decrease in heartbeat rate compared to the control group is marked by asterisks (*).In the diagrams the midpoint represents the mean, the upper and the lower line of the box represent the 25th and 75th percentiles of the distribution, respectively; the whiskers represent the mean ± SD.

Table 1
The Effect of TiO 2 Nanoparticles on the Aquatic Ecosystem 2016 60 4 Toxicity data on different types of nTiO

Characteristics of the tested nano-TiO Test organism Toxicology endpoint NOEC LOEC EC EC Reference
Primary particle diameter: ≤ 10 nm

Table 2
Characteristics of the applied Degussa VP P90 nTiO 2 suspension a Mean particle diameter by mass : the summarized volume of 50% of particles found in the dispersion is above this value, 50% is under this value.Determined by DLS method.b Mean particle diameter by number: 50% of the particles found in the dispersion is above this size, 50% is under this size.Determined by DLS method.c Mass percentage; 100 g of suspension contains 4 g of nTiO 2 by dry weight

Table 5
Geometric mean of the test particles formed by phagocytosis in 80 cells.

Table 6
Inhibition percentage values (H%) of D. magna heartbeat rate test after 24 h of contact time.

Table 7
Inhibition percentage values (H%) of L. minor growth inhibition test after 7 days of contact time.Lowest observed effect values determined by the applied test organisms at different trophic levels are summarized in Table8.