Carbon nanotubes affect the toxicity of CuO nanoparticles to denitrification in marine sediments by altering cellular internalization of nanoparticle

Denitrification is an important pathway for nitrate transformation in marine sediments, and this process has been observed to be negatively affected by engineered nanomaterials. However, previous studies only focused on the potential effect of a certain type of nanomaterial on microbial denitrification. Here we show that the toxicity of CuO nanoparticles (NPs) to denitrification in marine sediments is highly affected by the presence of carbon nanotubes (CNTs). It was found that the removal efficiency of total NOX−-N (NO3−-N and NO2−-N) in the presence of CuO NPs was only 62.3%, but it increased to 81.1% when CNTs appeared in this circumstance. Our data revealed that CuO NPs were more easily attached to CNTs rather than cell surface because of the lower energy barrier (3.5 versus 36.2 kT). Further studies confirmed that the presence of CNTs caused the formation of large, incompact, non-uniform dispersed, and more negatively charged CuO-CNTs heteroaggregates, and thus reduced the nanoparticle internalization by cells, leading to less toxicity to metabolism of carbon source, generation of reduction equivalent, and activities of nitrate reductase and nitrite reductase. These results indicate that assessing nanomaterial-induced risks in real circumstances needs to consider the “mixed” effects of nanomaterials.

present in the same circumstance. However, to date, it is largely unknown whether the coexistence of different nanomaterials could affect the toxicity of single nanomaterial to microbial denitrification in marine sediments.
Carbon nanotubes, a widely used class of carbon-based nanomaterials, have been shown to be able to cause negative effects on bacterial viability 22 , soil microbial communities 7 , and wastewater treatment process 23 , and their toxicities were determined by functionalization, size, and morphology 14,21 . Also, CuO NPs were reported to inhibit the denitrification efficiency of microbe 24 , mediate DNA damage 25 , or cause genotoxicity 26 . Nevertheless, once CNTs and CuO NPs co-existed in the environment, their potential environmental risks remain unclear. Here we report the potential effects of CuO NPs, CNTs, and CuO NPs + CNTs on denitrification in marine sediments. Then, the mechanisms for the different toxicities of single (CuO NPs or CNTs) and mixed (CuO NPs + CNTs) nanomaterials to microbial denitrification were explored by analyzing interacting energy, nanomaterial property (including particle polydispersity index, fractal dimension, and electrical property), cell cytoplasm density, carbon source metabolism, reduction equivalent generation, and key enzymes activities. The data from this work show the importance of the "mixed" effects of nanomaterials in real circumstances, when we assess the nanomaterial-induced environmental risks.

Results
CNTs affect the toxicity of CuO NPs to denitrification in marine sediments. Figure 1 illustrates the denitrification performances with single CuO NPs, single CNTs and CuO NPs + CNTs present in marine sediments, respectively. It can be found that the variations of nitrate in the single CNTs tests were almost the same as those in the control. Nevertheless, the concentration of nitrate was significantly higher in the CuO NPs tests than in the control (p < 0.05), which indicated that the reduction of nitrate was significantly inhibited by CuO NPs. When 50 mg/g of CNTs were present in the CuO NPs system, the final nitrate concentration, compared with single CuO NPs, was remarkably decreased from 102.6 to 72.9 mg/L, which was further declined to 52.4 mg/L with the increase of CNTs to 200 mg/g. Meanwhile, the data in Fig. 1B implied that more nitrite accumulation was caused by single CuO NPs, and the final nitrite concentration was 10.66 mg/L, which was 8 times of the control or 9 times of single CNTs. However, the presence of CNTs significantly alleviated the impacts of CuO NPs on nitrite accumulation, and the final nitrite was respectively 6.67 and 4.27 mg/L with the supplement of 50 and 200 mg/g of CNTs. Figure 1C shows that the final removal efficiencies of total NO X − -N (NO 3 − -N and NO 2 − -N) in single CuO NPs, single CNTs and control tests were 62.3%, 91.1% and 95.0%, respectively, which became to 73.5% or 81.1% after 50 or 200 mg/g CNTs were present in the CuO NPs system. Clearly, the presence of CNTs reduced the toxicity of CuO NPs to denitrification in marine sediments.
CNTs decrease the CuO NPs-induced effects on carbon source utilization, reduction equivalent generation and denitrifying enzyme activity. Denitrifying bacteria need to utilize carbon source (such as glucose) to provide energy for cell growth and produce reduction equivalent (i.e., NADH) for nitrate and nitrite bio-reductions. As shown in Fig. 2A, bacteria with CuO NPs utilized 1.33 g/L of glucose, which was significantly lower than that in the control (1.82 g/L) and that with CNTs (1.74 g/L). However, when 50 and 200 mg/g of CNTs were present in the CuO NPs system, the utilized glucose was increased to 1.57 g/L and 1.68 g/L, respectively. During the metabolism of glucose, NADH is produced, and then served as the reduction equivalent for denitrification 27 . The data in Fig. 2B revealed that CuO NPs decreased the intracellular NADH level to 68% of the control. However, the NADH content was respectively recovered to 74% and 89% of the control after the addition of 50 and 200 mg/g of CNTs.
The bio-reduction of nitrate to dinitrogen is catalyzed by a series of reductases, and the denitrification performance can be affected once the enzymatic activities are inhibited 9 . As seen in Fig. 2C, single CuO NPs caused significantly inhibitory effect on the activity of nitrate reductase (NAR) (75.1% of the control), and the presence of 50 and 200 mg/g of CNTs led to the recovery of NAR activity to 84.3% and 93.5% of the control, respectively. Likewise, the activity of nitrite reductase (NIR) was increased from 72.1% to 81.2% and 92.4% of the control with increasing the CNTs concentrations from 0 to 50 and 200 mg/g, indicating that CNTs also reduced the negative influence of CuO NPs on the activity of NIR. Figure 2D illustrates the principal component analysis (PCA) of the correlation of NO x − -N removal efficiency and activities of NAR and NIR under different exposure conditions, and the data indicated that the activity of NIR showed better association with NO x − -N removal than that of NAR.
CNTs reduce the cellular internalization of CuO NPs during denitrification. The uptake of nanoparticle by cells is closely related to the NPs-induced toxicity. Thus, the cellular internalization of CuO NPs was investigated by flow cytometry in this study 20 , and the analysis of side scatter (SSC) in flow cytometry was used to measure the uptake of nanoparticle by bacterial cells 28 . Figure 3 shows the dot plots of forward scatter (FSC) and SSC of cells under different conditions. Compared with the control, there was no significant change in the SSC of the single CNTs tests (data not shown). However, the SSC of cells exposed to CuO NPs was increased to 1.62 folds of the control according to the SSC distribution (Fig. 3B). More importantly, the presence of 50 and 200 mg/g of CNTs in the CuO NPs system decreased the SSC to 1.33-fold and 1.16-fold of the control, respectively. Therefore, although CuO NPs were taken up by cells, the presence of CNTs could reduce the cellular internalization of nanoparticle.

Influence of CNTs on the interaction between CuO NPs and cells.
Before CuO NPs were taken up by cells, they first needed to contact with the surface of microbe. As shown in Fig. 4A, CuO NPs tended to attach onto CNTs other than cells, suggesting that the presence of CNTs was likely to block the entrance of CuO NPs into cells. Previous studies reported that the entrance of nanoparticle would cause cell membrane damage, which subsequently resulted in the leakage of cytoplasm substances, such as the rise of extracellular LDH level 26 . Figure 4B demonstrates that CNTs did not cause significant LDH release, which was consistent with our previous observation 29 . However, CuO NPs led to the damage of membrane integrity for significantly elevating the released LDH (124% of control). When 50 and 200 mg/g of CNTs were present in the CuO NPs system, the levels of LDH release were decreased to 118% and 107% of the control, respectively, indicating that the presence of CNTs alleviated the membrane damage induced by CuO NPs.

Discussion
The literature reported that the cytotoxicity of CuO NPs to algae 30 , mussel 31 , or human cells 26 was attributed to nanoparticle itself, because CuO NPs showed distinct modes of action and greater magnitude influence compared with dissolved Cu 2+ . However, in other studies on murine microalgae 32 , macrophage cells 33 , and human lung cells 34 , the toxic effects of CuO NPs were resulted from the dissolved Cu 2+ . In this study, we also considered the   To investigate the underlying mechanism, a widely used model denitrifier, Paracoccus denitrificans, was used in this study. It was observed that the presence of CNTs also alleviated the negative influence of CuO NPs on the denitrification performance of P. denitrificans ( Figure S2, Supplementary Information), which was similar to that of sedimentary denitrification. As seen in Figure S3 Fig. 2 further revealed that the reason for CNTs alleviating the toxicity of CuO NPs to denitrification was attributed to the improved glucose utilization, which caused cell growth increase and more reduction equivalent generation. Meanwhile, the presence of CNTs significantly attenuated the effects of CuO NPs on denitrifying enzymes, and the influence on the activity of NIR showed the higher correlation with the performance of NO x − -N removal. Our data confirmed that the presence of CNTs reduced the cellular internalization of CuO NPs (Fig. 3). To explore the reason of CNTs decreasing the contact between CuO NPs and cells, the interacting energy (including van der Waals attraction, electrostatic double layer energy, and the total energy) between CuO NPs and cells was calculated and shown in Fig. 5. Because the measured zeta potentials of CuO NPs and cells were negative (− 14.7 mV versus − 21.5 mV), the electrostatic double layer force showed the repulsion interaction, which prevented the contact between CuO NPs and cells (Fig. 5A). The van der Waals force exhibits attractive interaction. The net energy is the sum of van der Waals and electrostatic double layer, and the maximum of net energy has been regarded as the energy barrier to overcome prior to getting close, which determined the contacting possibility of particles 13 . Figure 5B depicts that the energy barrier was 36.2, 28.7, and 3.5 kT for the pairs of CuO-Cells, CNTs-Cells, and CuO-CNTs, respectively. These results suggested that CuO NPs were most likely to contact with CNTs rather than cells, and the contacting probability of pairs was in the following order: CuO-CNTs > CNTs-Cells > CuO-Cells.
Since CNTs tended to interact with CuO NPs, it might alter the surface properties. The changes in nanomaterial properties have been reported to influence its cellular uptake and subsequent biological effect 35 . In this study, the average hydrodynamic diameters of CuO NPs and CNTs were 228 and 193 nm, respectively, but it became 268 nm in the co-existence system of CuO NPs + CNTs, indicating the formation of large particles and aggregates. Furthermore, the disperse property of nanoparticles evaluated by polydispersity index (PDI) showed that the mixed nanoparticles exhibited greater PDI (0.478) than their individuals (0.236 with CuO NPs and 0.197 with CNTs) in the reaction medium (Fig. 6A). Broader size distribution with higher PDI value indicated more polydispersed samples 21 . Therefore, the presence of CNTs in the CuO NPs system caused a more polydispersed status of nanoparticles.
The structural morphology of nanoparticles was investigated by static light scattering technique, which was used to track colloidal aggregation and aggregation state 36 . As seen in Fig. 6B, the calculated D f values of CNTs, CuO NPs, and CuO-CNTs were 2.28, 2.22, and 1.87, respectively. Generally, the D f values of colloidal aggregates fall in the range of 1 to 3, and higher values are associated with more compact morphological structures 37 . Apparently, the formed aggregation of CuO-CNTs possessed looser structure than that of CuO NPs. Figure 6C demonstrates that the heteroaggregates of CuO-CNTs had more negative surface charge for the zeta potential (− 27.5 mV) being lower than that of CuO NPs (− 14.7 mV) or CNTs (− 21.1 mV). These data revealed that the presence of CNTs led to the formation of large, non-uniformed and incompact heteroaggregates with greater negative surface charge. Because the greater size or more negatively charged nanoparticles were less taken up by cells [38][39][40] , it can be inferred that the CuO-CNTs heteroaggregates were difficult to be internalized by bacterial cells, which finally reduced the toxicity of CuO NPs to denitrification.
In summary, the data of this work indicate that the presence of CNTs highly affects the toxicity of CuO NPs to denitrification in marine sediments via altering the cellular internalization of nanoparticle and relieving the negative influences on the intracellular metabolism and the key enzyme activity of denitrifers, such as glucose utilization, reduction equivalent generation, and denitrifying enzymes. In the future, more attention should be paid to determine the potential effects of mixed nanomaterials on the way of nanoparticle entering cells and the interaction between nanoparticles and intracellular biomacromolecules. These investigations will help understand the biological responses of microbiota to "mixed" nanomaterials in real circumstances.

Methods
Preparation of nanomaterials. CuO NPs (primary particle size < 50 nm) and CNTs were purchased from Sigma-Aldrich (St. Louis, MO, USA) and Nanjing XF Nanotech Port Co. Ltd. (China), respectively. Before experiments, CNTs were treated according to the literature 21 . Firstly, CNTs were heated at 350 °C for 3 h. After being cooled down, nanotubes were placed in 12 M HCl for 8 h, bath sonicated under ambient conditions for 1 h, and then washed by copious amounts of Milli-Q water until neutral pH. Finally the powder CNTs were gotten after being dried in an oven (60 °C) overnight. The stock nanomaterial suspension was prepared by ultrasonicating the powder CuO NPs or CNTs in Milli-Q water for 1 h.

Source of denitrifiers.
Marine sediments were used as the microbiota source of sedimentary denitrification, which was collected from the Chongming Island (31° 30′ 26″ N, 121° 58′ 51″ E near the East Sea), Shanghai, China, and then sieved through 100-mesh sieve. After the determination of suspended solid concentration, the sediment was rinsed with 10 mM phosphate buffer (pH 7.4) for 3 times for the following exposure experiments. Also, Paracoccus denitrificans (ATCC 19367), purchased from American Type Culture Collection, was chosen as a model denitrifying bacteria to investigate the mechanisms due to its wide existence in sediment 41 . Prior to experiments P. denitrificans was grown in Difco nutrient broth at 30 °C overnight, harvested in midexponential phase, and then used as the inoculum for the following experiments.

Calculation of interfacial interactions between nanomaterials and cells. The interaction between
particles and model denitrifying bacteria (P. denitrificans) was evaluated by the DLVO theory, which described the total interaction as the combination of van der Waals (vdW) and electrostatic double layer (EDL) interaction 44 . The vdW interaction is attractive while EDL interaction can be repulsive in some cases, and both vary with the distance of particles surface (h). The total interaction energy (V T ) is the sum of these two interactions, which determines whether the net interaction between particles is repulsive or attractive 45,46 . According to DLVO theory, the following relationship can be derived: T v dW EDL  where A H is the particle-particle Hamaker constant, R is the center distance between particles, h is the surface distance, and a 1 and a 2 are the radii. As to A H , values of 6.95 × 10 −21 , 1.57 × 10 −20 , and 3.11 × 10 −21 were used for CuO-Cells, CuO-CNTs, and CNTs-Cells respectively, which were calculated according to literature 13 . The EDL energy is calculated as following: where ε is the permittivity of water, ψ 0 is surface potential, and κ is the inverse Debye length which is expressed as: Scientific RepoRts | 6:27748 | DOI: 10.1038/srep27748 The reaction mixture was prepared by following volumes: 50 μ L neutralized cell extract, 0.3 mL distilled water, and 0.6 mL reagent mixture. The reaction was started by adding 50 μ L of alcohol dehydrogenase (ADH, 500 U/ mL), and then the absorbance at 570 nm was recorded for 10 min at 30 °C. The concentration of NADH was calibrated with standard solutions of NADH, and the final NADH level was calculated as per milligram of protein.
Transmission electron microscopy analysis. Transmission electron microscopy was used to show the intracellular presence of nanoparticles by a Tecnai G2 spirit Biotwin (FEI, USA) with 120 kV accelerating voltage, and the samples were prepared as the literature 30 . The bacterial cells treated with or without nanomaterials for 24 h, and then fixed using 2.5% glutaraldehyde. The samples were further postfixed in 1% osmic acid for 1 h, washed with 0.1 M phosphate buffer solution (pH 7.4) for 3 times, and dehydrated in increasing concentration of acetone (30%, 50%, 70%, 90% and 100%) for 20 min each time. Then, samples were permeated and impregnated for 5 h at 60 °C, and the 60 nm ultrathin sections were made on Ni grids for subsequent imaging.
Lactate dehydrogenase release assay. Membrane integrity was determined by lactate dehydrogenase (LDH) release using cytotoxicity detection kit (Roche Applied Science). After exposed to nanomaterials, the culture supernatants were seeded in a 96-well plate, then 50 μ L of reaction mixture was added for 30 min incubation at 30 °C. The data were obtained at 490 nm absorbance on a microplate reader (BioTek, USA), and the results were expressed as a percentage of the control. A significant increase in LDH level would indicate cellular disruption.
Cell viability assay. Cell viability was measured by Cell Counting Kit-8 (Dojindo) according to the manufacturer's instructions. Briefly, after 24 h exposure, the cells were collected and placed in the 96-well plate, and then incubated with 10 μ L of CCK-8 for 30 min. The absorbance was recorded at 450 nm in a microplate reader (BioTek, USA), and the results were calculated as the relative activity based on sample without nanoparticles treatment.
Statistical analysis. In this study, all tests were performed in triplicate, and the results were expressed as mean ± standard deviation. The test of significance was carried out by analysis of variance and p < 0.05 was considered to be significantly different. Principal component analysis (PCA) was performed using XLSTAT.