Comparison of the Level and Mechanisms of Toxicity of Carbon Nanotubes, Carbon Nanofibers, and Silicon Nanotubes in Bioassay with Four Marine Microalgae

Nanoparticles (NPs) have various applications in medicine, cosmetics, optics, catalysis, environmental purification, and other areas nowadays. With an increasing annual production of NPs, the risks of their harmful influence to the environment and human health is rising. Currently, our knowledge about the mechanisms of interaction between NPs and living organisms is limited. Additionally, poor understanding of how physical and chemical characteristic and different conditions influence the toxicity of NPs restrict our attempts to develop the standards and regulations which might allow us to maintain safe living conditions. The marine species and their habitat environment are under continuous stress due to anthropogenic activities which result in the appearance of NPs in the aquatic environment. Our study aimed to evaluate and compare biochemical effects caused by the influence of different types of carbon nanotubes, carbon nanofibers, and silica nanotubes on four marine microalgae species. We evaluated the changes in growth-rate, esterase activity, membrane polarization, and size changes of microalgae cells using flow cytometry method. Our results demonstrated that toxic effects caused by the carbon nanotubes strongly correlated with the content of heavy metal impurities in the NPs. More hydrophobic carbon NPs with less ordered structure had a higher impact on the red microalgae P. purpureum because of higher adherence between the particles and mucous covering of the algae. Silica NPs caused significant inhibition of microalgae growth-rate predominantly produced by mechanical influence.


Introduction
The growing industry of nanotechnology inevitably results in an increase of risks associated with the nanomaterials (NMs) and nanoparticles (NPs) to affect living organisms including humans. Carbon-
Carbon nanotubes and nanofibers were synthesized and characterized in the Boreskov Institute of Catalysis (Novosibirsk, Russia) [47]. The structural features of carbon NPs ware assessed by Raman spectroscopy in our earlier report [41]. The length of carbon nanotubes was hundreds of times larger than the diameter and in water suspension; the particles could cohere into the spheres up to tens of micrometers in diameter. The toxicity of these NP samples was previously evaluated on mice [49], rats [50,51], human cell lines [52], and microalgae Heterosigma akashiwo [41].
Silicon nanotubes were kindly provided by the Department of Chemistry, Inha University Republic of Korea [48]. The samples had a significantly lower ratio of length to diameter compared to carbon nanotubes.

Microalgae Cultures
Microalgal cultures were provided by The Resource Collection Marine biobank of the National Scientific Center of Marine Biology, Far Eastern Branch of the Russian Academy of Sciences (NSCMB FEB RAS). The toxicity bioassay of NPs was carried out on four marine microalgae: two types of diatom species Attheya ussuriensis [53] and Chaetoceros muelleri [54], a raphidophyte Heterosigma akashiwo [55], and a red algae Porphyridium purpureum [56]. Culturing of microalgae and toxicity test conditions were maintained in accordance with the OECD Guidelines for the Testing of Chemicals, STest No.201 [57] with minor modifications as previously described [58,59].
The microalgae species were selected based on its abundance among the microalgae of the Sea of Japan [60], and its suitability as test objects in ecotoxicology [59,[61][62][63]. Earlier, we confirmed the sensitivity of used microalgae species as test organisms with common reference toxicant potassium dichromate [58].

Bioassay
The samples of NPs were added to filtered seawater to obtain the working suspensions with a concentration of 1000 mg/L. Before each series of bioassays, the working suspensions of NPs were sonicated with ultrasound homogenizer Bandelin Sonopuls GM 3100 (Bandelin Electronic GmbH & Co. KG, Berlin, Germany) using maximal intensity for 30 min.
The exposition of microalgae cells to the suspensions of NPs was carried out in 24-well plates. Each well was filled with 2 mL of microalgae cell aliquot and the corresponding volume of the working suspension to obtain the final concentrations 1, 10, and 100 mg/L. The filtered seawater without NPs was added to the control group. The exposure of each used concentration and control group was performed in four biological replicates.

Flow Cytometry
Microalgae cell counting and registration of morphological and biochemical changes during the experiment were carried out with flow cytometer CytoFLEX (Beckman Coulter, Indianapolis, IN, USA) with the software package CytExpert v.2.0. The changes of microalgae cells after exposure to NPs were evaluated using specific fluorescent dyes. Microalgae growth-rate inhibition was determined by staining with propidium iodide (PI) according to the standard bioassay protocol [64]. Esterase activity of microalgae exposed to the NPs was evaluated using non-fluorescent lipophilic dye fluorescein diacetate (FDA) [65,66]. Membrane potential of microalgae cells was assessed by a lipophilic, positively charged fluorescent dye 3,3 -dihexyloxacarbocyanine iodide (DiOC 6 ) [67,68]. To determine the size of microalgae cells, a size calibration kit, batch F13838 (Molecular probes, Eugene, OR, USA) with the certified size distribution of 1, 2, 4, 6, 10, 15 µm was used for the forward scatter emission channel. The emission channels were selected according to the maximum emission of the dyes, provided by the manufacturer (Molecular Probes, Eugene, OR, USA). The blue laser (488 nm) of the CytoFLEX flow cytometer was chosen as a source of excitation light. The endpoints of toxicity used in this work and the parameters of their registration are listed in Table 2. Each sample was measured at a flow rate of 100 µL/min for 30 s. Prior to the assessment of growth-inhibition, esterase activity, and membrane potential of each microalgae species, we made a series of preliminary measurements to determine the optimal concentration of fluorescent dyes and the optimal duration of staining as described in our previous report [59]. The registration time for used endpoints was selected according to the standard methods commonly used to assess toxicity of a test substance in an aqueous system with microalgae model organisms [69][70][71][72].
Growth-rate inhibition and changes in the size of microalgae cells should be estimated as a benchmark of direct cytotoxic effects or as an indicator of mortality. The 96 h and 7 days half maximal effective concentration (EC 50 ) are one of the most common values used for evaluation cytotoxic effects in macroalgae bioassay [63,73].
Esterase activity and membrane potential changes can indicate either the preliminary stage of toxic action or display the adaptational ability of organisms to influence toxic substances [74,75]. Changes of microalgae esterase activity are mostly caused by the deficiency of enzyme action or by disruption of membrane integrity, and it can be evaluated as a sensitive endpoint of algal sublethal toxicity [65,66]. The 3 h and 24 h registration points were chosen to detect possible early metabolic response of the algae over short exposure periods and dynamic change of that response, respectfully [71]. Reduction of membrane potential (depolarization) can be accompanied by changes of membrane elasticity, loss of lipid microdomains, and changes of ion permeability [76]. Integrity and normal operation of membranes are vital parameters for organisms as they provide barriers and transportation functions. The 6 h and 24 h registration points were chosen to detect changes in membrane potential of microalgae cells [72].

Microscopy
Morphological changes of microalgae cells were observed and captured by optical microscope Axio Observer A1 (Carl Zeiss, Oberkochen, Germany).

Statistical Analysis
Statistical analyses were performed using the software package GraphPad Prism 7.04 (GraphPad Software, San Diego, CA, USA). The one-way ANOVA test was used for analysis. A value of p ≤ 0.05 was considered statistically significant.

Results
For all the samples of NPs, we calculated EC 50 concentrations of microalgae growth-rate, FDA fluorescence intensity (esterase activity), and DiOC 6 (membrane potential) fluorescence intensity compared to control. The calculated mean EC 50 concentrations and 95% confidence limits are given in Table 3. 95% confidence limits presented in the parentheses; n/a, measured effect was not observed even at the highest concentrations of the sample; *In the cases when influence of NPs caused stimulation of the registered endpoint, the data were represented for concentration of 100 mg/L compared to control.
For visualization of calculated data ( Table 3) and analysis of dynamic changes of microalgae cells, we created a heatmap ( Figure 1). Nanomaterials 2020, 10, x FOR PEER REVIEW 6 of 14 For visualization of calculated data ( Table 3) and analysis of dynamic changes of microalgae cells, we created a heatmap ( Figure 1). (c) C. muelleri; (d) P. purpureum; white square, the calculated EC50 was higher than 250 mg/L; crossed square, measured endpoint was not observed; green square, the influence of the NPs caused the stimulation of the measured endpoint.
Silica nanotubes SNT-1 and SNT-2 demonstrated the most pronounced influence on the microalgae growth rate. The level of toxicity increased for all microalgae species after seven days of exposure except for the samples SNT-1 and SNT-2 after 96 h of the treatment affected growth-rate only for C. muelleri and P. purpureum.
Carbon nanotubes CNT-1 and CNT-2 had almost no influence on the growth rate of A. ussuriensis, C. muelleri, and H. akashiwo both after 96 h (acute toxicity) and seven days (chronic toxicity) of the treatment. Carbon nanofibers CNF-1 and CNF-2 also did not reveal any significant influence on the growth rate of these three species in acute toxicity assessment but chronic toxicity and growth-rate inhibition was detected for A. ussuriensis and H. akashiwo. Moreover, all the carbon NPs caused significant inhibition of esterase activity and depolarization of membranes for A. ussuriensis, H. akashiwo, and C. muelleri. However, the nanofiber sample CNF-2 did not affect the esterase activity of diatomic algae A. ussuriensis and stimulated esterase activity of diatomic algae C. muelleri.
The toxicological profile of the red algae P. purpureum significantly differed from three other microalgae species. P. purpureum was the only species responded by growth-rate inhibition to the influence of carbon NPs. At the same time, the red algae had the lowest changes in esterase activity and membrane potential.
The data of flow cytometry analysis indicated the changes in the size of microalgae cells after the treatment of NPs as demonstrated in Figure 2. (c) C. muelleri; (d) P. purpureum; white square, the calculated EC 50 was higher than 250 mg/L; crossed square, measured endpoint was not observed; green square, the influence of the NPs caused the stimulation of the measured endpoint.
Silica nanotubes SNT-1 and SNT-2 demonstrated the most pronounced influence on the microalgae growth rate. The level of toxicity increased for all microalgae species after seven days of exposure except for the samples SNT-1 and SNT-2 after 96 h of the treatment affected growth-rate only for C. muelleri and P. purpureum.
Carbon nanotubes CNT-1 and CNT-2 had almost no influence on the growth rate of A. ussuriensis, C. muelleri, and H. akashiwo both after 96 h (acute toxicity) and seven days (chronic toxicity) of the treatment. Carbon nanofibers CNF-1 and CNF-2 also did not reveal any significant influence on the growth rate of these three species in acute toxicity assessment but chronic toxicity and growth-rate inhibition was detected for A. ussuriensis and H. akashiwo. Moreover, all the carbon NPs caused significant inhibition of esterase activity and depolarization of membranes for A. ussuriensis, H. akashiwo, and C. muelleri. However, the nanofiber sample CNF-2 did not affect the esterase activity of diatomic algae A. ussuriensis and stimulated esterase activity of diatomic algae C. muelleri.
The toxicological profile of the red algae P. purpureum significantly differed from three other microalgae species. P. purpureum was the only species responded by growth-rate inhibition to the influence of carbon NPs. At the same time, the red algae had the lowest changes in esterase activity and membrane potential.  (Figure 2b). The cell sizes of C. muelleri and P. purpureum slightly decreased after the treatment with the sample CNT-1 (Figure 2c,d).
The visual observation of P. purpureum after seven days of exposure to the NPs is presented in Figure 3 and Figure 4.
It should be noted that despite a visible dissimilarity in the sensitivity and responses of different microalgae cells to the treatment of the NPs, we highlighted general mechanisms of action for assessed samples. In the next section, we discuss the principal findings of this investigation.   (Figure 2b). The cell sizes of C. muelleri and P. purpureum slightly decreased after the treatment with the sample CNT-1 (Figure 2c,d).
The visual observation of P. purpureum after seven days of exposure to the NPs is presented in Figures 3 and 4.  (Figure 2b). The cell sizes of C. muelleri and P. purpureum slightly decreased after the treatment with the sample CNT-1 (Figure 2c,d).
The visual observation of P. purpureum after seven days of exposure to the NPs is presented in Figure 3 and Figure 4.
It should be noted that despite a visible dissimilarity in the sensitivity and responses of different microalgae cells to the treatment of the NPs, we highlighted general mechanisms of action for assessed samples. In the next section, we discuss the principal findings of this investigation.

Carbon Nanotubes and Nanofibers
One of the most peculiar observations that we can highlight from experimental data is a relatively high sensitivity of the red microalgae P. purpureum to the samples of carbon NPs and a lower sensitivity to the silica nanotubes, although the other three microalgae species had diametrically opposite responses (Table 3, Figure 1). The probable reason for such differences is a highly hydrophobic surface of the cells of P. purpureum caused by the presence of mucous covering around the cells of the red algae [77]. It was shown that carbon NPs can bind to membranes of microorganisms by hydrophobic interaction and hydrogen bonding formed between surfaces of cells and defect areas of NPs [78]. Hence, the hydrophobic surface of cells might have facilitated adhesion of more hydrophobic carbon NPs to microalgae (Figure 3b,c). Therefore, it could be stated that the red algae P. purpureum received higher influence from the samples of carbon NPs.
The current statement correlated with the results of the structural and surface analysis of the samples of carbon NPs (Table 1). It was shown (Figure 1) that the carbon nanotubes sample CNT-2 having the most ordered structure had a lower influence on the growth-rate of P. purpureum as compared to the other carbon NPs having unordered structure and a relatively higher hydrophobicity. Therefore, based on inhibition of growth-rate of P. purpureum and low influence of carbon NPs (except CNT-2) on esterase activity and membrane potential, we can conclude that the main mechanism of toxic action of these samples on the red microalgae was a physical damage caused by adhesion of clusters of NPs with extracellular mucopolysaccharides of P. purpureum [79]. Almost no effect of carbon NPs on growth-rate of A. ussuriensis, H. akashiwo, and C. muelleri was probably caused by the lower mechanical interaction between microalgae cells and the NPs. However, observed esterase activity inhibition, membrane depolarization, and size changes of microalgae cells might be produced by the influence of metal impurities containing in carbon NPs (Table 1). Prior research demonstrated similar effects of metal ions on esterase activity, membranes, and size of microalgae cells [80][81][82].
In general, nanofibers reveal a lower influence on esterase activity and membrane polarization compared to carbon nanotubes. Such a difference is probably caused by the unequal physical accessibility of toxic impurities from different types of NPs to microalgae cells [83]. It should be noted that despite a visible dissimilarity in the sensitivity and responses of different microalgae cells to the treatment of the NPs, we highlighted general mechanisms of action for assessed samples. In the next section, we discuss the principal findings of this investigation.

Carbon Nanotubes and Nanofibers
One of the most peculiar observations that we can highlight from experimental data is a relatively high sensitivity of the red microalgae P. purpureum to the samples of carbon NPs and a lower sensitivity to the silica nanotubes, although the other three microalgae species had diametrically opposite responses (Table 3, Figure 1). The probable reason for such differences is a highly hydrophobic surface of the cells of P. purpureum caused by the presence of mucous covering around the cells of the red algae [77]. It was shown that carbon NPs can bind to membranes of microorganisms by hydrophobic interaction and hydrogen bonding formed between surfaces of cells and defect areas of NPs [78]. Hence, the hydrophobic surface of cells might have facilitated adhesion of more hydrophobic carbon NPs to microalgae (Figure 3b,c). Therefore, it could be stated that the red algae P. purpureum received higher influence from the samples of carbon NPs.
The current statement correlated with the results of the structural and surface analysis of the samples of carbon NPs (Table 1). It was shown (Figure 1) that the carbon nanotubes sample CNT-2 having the most ordered structure had a lower influence on the growth-rate of P. purpureum as compared to the other carbon NPs having unordered structure and a relatively higher hydrophobicity. Therefore, based on inhibition of growth-rate of P. purpureum and low influence of carbon NPs (except CNT-2) on esterase activity and membrane potential, we can conclude that the main mechanism of toxic action of these samples on the red microalgae was a physical damage caused by adhesion of clusters of NPs with extracellular mucopolysaccharides of P. purpureum [79]. Almost no effect of carbon NPs on growth-rate of A. ussuriensis, H. akashiwo, and C. muelleri was probably caused by the lower mechanical interaction between microalgae cells and the NPs. However, observed esterase activity inhibition, membrane depolarization, and size changes of microalgae cells might be produced by the influence of metal impurities containing in carbon NPs (Table 1). Prior research demonstrated similar effects of metal ions on esterase activity, membranes, and size of microalgae cells [80][81][82].
In general, nanofibers reveal a lower influence on esterase activity and membrane polarization compared to carbon nanotubes. Such a difference is probably caused by the unequal physical accessibility of toxic impurities from different types of NPs to microalgae cells [83].

Silica Nanotubes
The influence of silica nanotubes caused a significant growth-rate inhibition for all species but almost did not change esterase activity, membrane polarization (Table 3, Figure 1), size (Figure 2), and shape of microalgae cells (Figure 4). According to the experimental data and the absence of impurities in the composition of SNT-1 and SNT-2 samples (Table 1), we can conclude that the main toxic mechanism for silica nanotubes was mechanical damage of microalgae cells.
Hydrophilic surface of silica NPs allows them to easily move in the water body [84]. Such properties increase the possibility of NPs to have contact with microalgae cells. Thus, a planktonic species i.e., C. muelleri, more frequently had a contact with silica nanotubes and experienced more severe mechanical damage (Figure 1c) compared to benthic laying on the bottom A. ussuriensis (Figure 1a), placed near the water surface H. akashiwo (Figure 1b), and had defense mucous covering P. purpureum (Figure 1d).
Interestingly, the toxic influence of silica nanotubes was reduced with time for all microalgae species except small planktonic algae C. muelleri. At the same time, the sample SNT-2 with a larger diameter of nanotubes had higher influence on growth-rate of two microalgae species with larger cells i.e., A. ussuriensis and H. akashiwo. Previous works showed that the toxicity from the silica nanotubes increases with the increase of microalgae cell diameter [85]. The other research described an increase in the phytotoxic action for silica NPs with an increase of particle size [86]. For C. muelleri and P. purpureum, which are the algae species having smaller cell sizes, both types of silica nanotubes demonstrated a comparable level of toxicity, though the minor predominance could be seen for the sample SNT-1 (Table 3, Figure 1) with a smaller diameter of nanotubes but a more developed surface area.
In the bioassay on common microalgae Raphidocelis subcapitata, the authors claimed the increase of silica nanotubes toxicity with increase of its surface area [84]. Hence, our results were in a good correlation with previous research, and we can conclude that the toxicity of silica NPs strongly depends on their size, surface area, and other surface properties.
However, the cell size changes observed for diatomic algae A. ussuriensis under the treatment of silica nanotubes were unexpected and it was in contradiction to the conclusion made for carbon NPs that the size of microalgae changed by the influence of metal ions from particle impurities. Nevertheless, the size of A. ussuriensis cells was reduced in the presence of silica nanotubes having no impurities (Table 1, Figure 3a).
Probably, the observed effect could be related to diatomic algae reproduction peculiarities. It is a known fact that the size of the diatomic algae population becomes smaller through the series of cell divisions, and the initial cell size can be kept only by sexual reproduction [87]. Therefore, cell size reduction for A. ussuriensis might be caused by a disorder in the reproduction processes. The reason for such disorder is a matter for further investigation and discussion. Moreover, the samples of silica nanotubes did not reveal any influence on the size of the other three microalgae species (Figure 3b-d).
Another interesting observation was the increase in microalgae membrane polarization under the influence of silica NPs (Table 3, Figure 1). Therefore, the mechanisms of toxic action for silica nanotubes cannot depend only on particle diameter and surface area, and future investigations should be focused on the searching of the parameters and their combinations which could influence the toxicity of NPs.

Conclusions
In this study, the aim was to assess the level of toxicity and the mechanisms of toxic action of carbon nanotubes, carbon nanofibers, and silica nanotubes using four microalgae species as the objects of aquatic toxicity bioassay. The results of this investigation show that (1) carbon nanotubes samples CNT-1 and CNT-2 had non-significant toxic effect on the growth rate of all four microalgae species but caused a high inhibition of esterase activity and depolarization of cell membranes, which was most probably caused by heavy metal impurities in NPs; (2) more hydrophobic carbon NPs with less ordered structure had a higher impact on the red microalgae P. purpureum because of higher adherence between the particles and mucous covering of the red algae cells; (3) silica NPs did not affect the esterase activity and membrane potential of the cells of all four microalgae species even at higher concentrations but caused significant inhibition of growth-rate, which indicated predominance of mechanical damage as a mechanism of toxicity for used samples of silica nanotubes.
The findings of this research provide insights for the formation of the principles of safe design, production, and utilization of NPs [88]. We believe that safety ensuring in nanotechnology would be provided only by international cooperation and large-scale nanotoxicology research, including the approaches of bioinformatics, system biology, and other methods of modeling, prediction, and maintaining the handling and interpretation the growing body of research data.