ZnO nanoparticles alter redox metabolism of Limnoperna fortunei

Nanoparticles such as zinc oxide nanoparticles (ZnO-NP) that are incorporated in consumer and industrial products have caused concern about their potential ecotoxicological impact when released into the environment. Bivalve mollusks are susceptible targets for nanoparticle toxicity since nanomaterials can enter the cells by endocytosis mechanisms. The aim of this study was to evaluate the influence of ZnO-NP on the redox metabolism in Limnoperna fortunei and the DNA damage after exposure to ZnO-NP. Adult bivalves were incubated with 1-, 10-, and 50-μg mL−1 ZnO-NP for 2, 4, and 24 h. Ionic Zn release, enzymatic and non-enzymatic antioxidant activity, oxidative damage, and DNA damage were evaluated. Oxidative damage to proteins and lipids were observed after 4-h exposure and returned to baseline levels after 24 h. Superoxide dismutase levels decreased after 4-h exposure and increased after 24 h. No significant alteration was observed in the catalase activity or even DNA double-strand cleavage. The dissociation of ZnO may occur after 24 h, releasing ionic zinc (Zn2+) by hydrolysis, which was confirmed by the increase in the ionic Zn concentration following 24-h exposure. In conclusion, ZnO-NP were able to induce oxidative stress in exposed golden mussels. The golden mussel can modulate its own antioxidant defenses in response to oxidative stress and seems to be able to hydrolyze the nanoparticles and consequently, release Zn2+ into the cellular compartment.


Introduction
The increasing production and use of nanoparticles in consumer and industrial products and the large range of applications related to these nanomaterials raise some concerns about their potential ecological impact in different ecosystems and their adverse effect on human health (Simonet and Valcárcel 2009;Salieri et al. 2015;Girardello et al. 2016a;Girardello et al. 2016b). Nanoparticles have a high surface-to-volume ratio, which results in a high reactivity potential and unique physical and chemical properties that differ from those of their respective bulk materials (Mallevre et al. 2014;Girardello et al. 2016a;Girardello et al. 2016b). The nanoparticle composition, solubility, and interaction modes with biological systems are highlighted as main factors for risk assessment of metal oxide nanoparticles Wu et al. 2010).
Zinc oxide nanoparticles (ZnO-NP) have applications in many different fields, including wastewater treatment (Anjum et al. 2016), molecular biology (Navaei-Nigjeh et al.  (Khezri et al. 2018), food industry (Venkatasubbu et al. 2016), and construction materials (Hossain et al. 2014;Schaumann et al. 2015). Such widespread, expanding production and use of ZnO-NP increase the potential for their release to the environment, causing ecotoxicological problems (Jovanović and Palić 2012). The exposure of aquatic animals to nanoparticles and their aggregates is a major concern (Shang et al. 2018;Shang et al. 2019;Wu et al. 2018) due to the potential harmful effects that may occur. ZnO-NP toxicity is attributed to factors such as size, surface characteristics (specific area, charge, and defect), solubility, chemical reactivity, and exposure routes. Moreover, the nanoparticle uptake into cells and organisms may be affected by their association with naturally occurring colloids in aquatic systems, altering the nanoparticle behavior in this environmental compartment (Moore 2006).
The aquatic system has a variety of organisms, and once nanoparticles enter the organism's compartment, it may result in additional toxic effects, deregulate cell metabolism, and generate reactive oxygen species (ROS) with effects that are related to concentration and period of exposure (Azqueta and Dusinska 2015;Marisa et al. 2015). ROS can induce serious damage to biomolecules, such as lipids and proteins, and deplete enzymatic defenses, such as superoxide dismutase (SOD) and catalase (CAT). Non-enzymatic protein-bound sulfhydryl groups are also affected by ROS exposure (Katsumiti et al. 2014). Nanoparticles can cause alterations in redox metabolism, modifying markers related to biomolecule oxidation, such as CAT, SOD, carbonyl proteins, and lipids (Iummato et al. 2013;Girardello et al. 2016a;Girardello et al. 2016b). Furthermore, antioxidant defenses will respond differently according to the individual aquatic organism and depending on the intensity and frequency of exposure (Canesi et al. 2010;Faria et al. 2014;Hao et al. 2009;Zhu et al. 2011;Borase et al. 2021;Wong et al. 2020;Alkaladi 2019).
The ecotoxicity of ZnO-NP is related to their physicalchemical characteristics, such as solubilization and photoreactivity, as well as the tested species (Borase et al. 2021;Wong et al. 2020;Alkaladi 2019). ZnO-NP have photocatalytic properties and contribute to ROS generation. It is believed that solubilized Zn 2+ may contribute to nanoparticle cytotoxicity (Ma et al. 2013). Gagné and coauthor (2019) showed that exposure of the freshwater mussel Dreissena polymorpha to ZnO-NP increased the oxidative stress of these bivalves. Zn toxicity could be influenced by both solubilized and non-solubilized Zn complexes in freshwater mussels (Gagné et al. 2019). However, toxicity studies on aquatic invertebrates exposed to ZnO-NP are limited, and its molecular mechanisms related to nanomaterial exposure should be further investigated.
Bivalve mollusks are sensitive to nanoparticle toxicity and may be internalized by endocytosis mechanisms (Canesi et al. 2010;Canesi et al. 2012;Barmo et al. 2013;Canesi et al. 2014;Girardello et al. 2016a;Girardello et al. 2016b). Mussels are filter feeders and stationary organisms and for this reason, are largely used as biomonitors for environmental perturbations (Villela et al. 2007(Villela et al. , 2013 and nanoparticle toxicity assessment (Girardello et al. 2016a;Girardello et al. 2016b). Furthermore, the ability of bivalves to bioaccumulate toxic compounds in their body determines its role in the transfer of environmental pollutants to higher trophic levels (Hunt et al. 2003).
The golden mussel (Limnoperna fortunei) is an exotic organism from Asia that lives in freshwater compartments and has been used for biomonitoring environmental conditions (Mariano et al. 2006;Iummato et al. 2013;Villela et al. 2013;Girardello et al. 2016a;Girardello et al. 2016b). L. fortunei are widely distributed in Rio Grande do Sul, the southernmost state of Brazil, and can be collected during the entire year, which makes this organism an adequate tool for biomonitoring nanoparticles and an excellent candidate to be a sentinel organism (Mariano et al. 2006;Villela et al. 2006Villela et al. , 2007Villela et al. , 2013Girardello et al. 2016a;Girardello et al. 2016b). Ecotoxicological effect models using L. fortunei exposed to nanoparticles have been validated in previous studies from our group, which tested the exposure of TiO 2 -NP (Girardello et al. 2016a;Girardello et al. 2016b). These nanoparticles have photocatalytic properties similar to ZnO-NP.
Although there have been an increasing number of studies on nanoparticle toxicity, comprehensive knowledge on the impact of ZnO-NP on aquatic organisms is still not clear. Our research group has studied L. fortunei since 2006, and specifically regarding nanoparticle exposure since 2012, due to the fact that these bivalves are target organisms in the environment. In this sense, this study aims to evaluate the oxidative effects, DNA damage, and modulation of enzymatic and non-enzymatic defenses in L. fortunei.

Test organisms
The collection, handling, and maintenance of L. fortunei bivalves were based on the protocols established by Villela et al. (2006) and Girardello et al. (2016a). Briefly, L. fortunei samples were collected in the Itapuã State Park conservation unity (RS, Brazil) (S 30°21′ 25.8′′, WO 51°02′ 58.5′′), which is protected by the environmental department of the state government. The mussels and water samples were collected according to the protocol described by Villela et al. (2006). A maximum of 100 specimens (~2 cm) were acclimated in aquariums containing 2 L of water of the Itapuã, which were maintained under constant aeration and a controlled temperature (19 ± 2°C). The mussels were fed with a mixture of herbivorous fish feed and Itapuã water, which were provided every 2 days. For every 100 individuals (1 aquarium), about 340 mg of feed per month was used. After 10 days of the acclimation period, the experiments were carried out (Girardello et al. 2016a).

Zinc oxide nanoparticles (ZnO-NP)
ZnO-NP (purity > 97%) were obtained from Sigma-Aldrich. Physicochemical characterization of the ZnO-NP was carried out following the procedure described by Girardello et al. (2016a). Details of the procedure are available as Supplementary Material. Transmission electron microscopy (TEM), X-ray diffraction (XRD), Brunauer-Emmett-Teller method (BET) (Brunauer et al. 1938), dynamic light scattering (DLS), and zeta potential (ZP) were used to characterize the nanomaterial. A 1000-μg mL −1 stock solution of ZnO-NP was prepared in ultrapure water and sonicated (Sonicator USC-1400A UNIQUE) for 30 min before dilution to prepare ZnO-NP solutions containing 1, 10, or 50 μg mL −1 . Prior to exposure to the mussels, the nanoparticles were dispersed by sonication for 30 min (Girardello et al. 2016a). Nontoxic concentrations were chosen with no lethality to L. fortunei, being more suitable for tests in the range of 1 and 50 μg mL −1 . This range of concentration enabled to verify the potential effect of damages in golden mussel hemocytes. The concentrations used in the exposure assays also fell within the range of those observed in ecotoxicity tests using the same types of nanoparticles (Vijayakumar et al. 2017;Shang et al. 2019).

Exposure to zinc oxide nanoparticles
The mussels were exposed to ZnO-NP solutions (1, 10, or 50 μg mL −1 ) for periods of 2, 4, or 24 h. The cells of the golden mussel reproduce quickly, with damage and repair being observed. Thus, it is important to assess the acute and chronic impact of the nanoparticles on the organism, which justifies the choice of both early and late times. In addition, the concentration was chosen to verify the potential effect of damages in the DNA of the hemocytes of the golden mussel. Each individual exposure condition was carried out using four adult mussel specimens (2.05 ± 0.17 cm in length). The control group was exposed to the same conditions, except for the presence of ZnO-NP in the solutions. The mussel specimens were not fed during the exposure assays days. The experiments were carried out on days of feeding interval to avoid any reaction/interference with the nanoparticles.

Quantitative determination of Zn 2+ by ICP-MS
Soft tissue (body and internal organs) samples of L. fortunei exposed and not exposed (control group) to ZnO-NP were lyophilized and subsequently weighed on perfluoroalkoxy (PFA) vessels. The samples (0.0350 g) were digested in the open PFA flasks for 24 h using 1.25 mL of doubly distilled 14-mol L −1 HNO 3 (Merck, Darmstadt, Germany) and 0.125-mL 30 % H 2 O 2 (Synth, São Paulo, Brazil). The residual solution was evaporated to dryness at 70°C, and the dried residue was solubilized with 2.5 mL of doubly distilled 14mol L −1 HNO 3 followed by dilution to 50 mL using ultrapure water. The determination of zinc content was carried out using an ELAN 6000 inductively coupled plasma mass spectrometer (Perkin Elmer Sciex, Thornhill, Canada); 99.996 % argon (White Martins, São Paulo, Brazil) was used as the plasma and nebulizer gas. Rhodium (10 μg L −1 ) was used as an internal standard, and recovery tests were performed to confirm the absence of interferences (recovery values ranged from 92 to 114 %). The detection limit for Zn was 0.3 mg g −1 . The method was proven to be free from interferences (Martín-Cameán et al. 2014). The ICP-MS operating conditions are summarized in Table S1.

Enzymatic and non-enzymatic antioxidant defenses
Sample preparation for determining enzymatic and nonenzymatic antioxidant activity and evaluating oxidative damage to lipids and proteins was adapted from Iummato et al. (2013). The enzymatic activities of SOD and CAT were determined following the protocols established by Bannister and Calabrese (1987) and Aebi (1984), respectively. Absorbance measurements were obtained using a Victor-X3 multilabel counter microplate reader (Perkin Elmer, Finland). The protein sulfhydryl content assay was performed according to the protocol described by Aksenov and Markesbery (2001).
The screening method to determine the total antioxidant capacity followed the procedure described by Re et al. (1999) . Lipid peroxidation was induced following the procedure described by Girardello et al. (2006a) and Lowry and coauthors (1951). The oxidative damage to proteins was assessed following the method established by Levine et al. (1990) and modified by Girardello et al. (2016a). Detailed descriptions of all procedures mentioned in this section are available in the Supplementary Material.

DNA fragmentation assay
The DNA fragmentation assay was carried out following the procedure described by Girardello et al. (2016a). The DNA fragmentation assay was accomplished using hemolymph cells of the L. fortunei. The hemolymph was extracted from the posterior adductor muscle with a hypodermic needle using a 1-mL syringe while the mussels were still alive (Girardello et al. 2016a). Genomic DNA from the hemolymph was extracted using the Illustra tissue & cells genomicPrep Mini Spin Kit (GE Healthcare, Buckinghamshire, HP7 9NA UK) according to the manufacturer's instructions. In this protocol, proteinase K was used at the lysis step, and RNAse A was used to remove RNA from the samples. The purified genomic DNA was visualized on 1.5 % agarose gel stained with ethidium bromide. The analysis was performed using an ImageQuant LAS 500 GE® equipment, where DNA fragmentation levels were observed.

Statistical analysis
All experiments were performed with biological and technical triplicates, and the results are expressed as mean ± standard deviation (SD). Statistical analysis using the SPSS 1.0 software was carried out to establish a statistical comparison between the exposed groups and the control group. ANOVA and Dunnett's multiple comparison tests were applied to the results, as well as Tukey's post hoc test. Statistical significance was considered in cases where p ≤ 0.05.

Characterization of nanoparticles
The results from the characterization of the ZnO-NP are summarized in Table 1. ZnO-NP were analyzed by TEM, which showed an average diameter of approximately 50 nm (Fig. 1). This result is in accordance with the diameter defined by the producer. The X-ray diffraction (XRD) patterns revealed the presence of the wurtzite phase (Fig. 2) The scattering experiments using samples containing ZnO-NP revealed the presence of particle sizes larger than 500 nm in all samples, which extrapolated the maximum detection limit of the DLS technique (set as 500 nm). For all experiments using water collected in the State Park of Itapuã, the formation of aggregates with diameters larger than 500 nm was observed (Girardello et al. 2016a). The stability of the nanoparticles in aqueous solutions was evaluated by the zeta potential (ζ) parameter, which was employed to estimate the particle surface potential related to the nanoparticle aggregation rate. The zeta potential values ranged from −13.58 ± 1.44 to −17.98 ± 1.55 mV for the ZnO-NP solutions and −20.46 ± 1.98 mV for the control (Table 2).

Analysis by ICP-MS
The results from the determination of Zn by ICP-MS are shown in Fig. 3. As shown in this figure, Zn was detected in the soft body portions of L. fortunei after exposure to different ZnO-NP concentrations (Fig. 3). Zn 2+ concentrations associated to all treatments decreased after 2-h exposure when compared to the control group (Fig. 3). After 4-h exposure, the Zn concentration decreased in the specimens exposed to 1-and 50-μg mL −1 ZnO-NP. On the other hand, the Zn concentration increased significantly after exposure for 24 h to 50-μg mL −1 ZnO-NP.

Enzymatic and non-enzymatic antioxidant defenses
The results from SOD and CAT activities, protein sulfhydryl content, and the TEAC are shown in Fig. 4. With increasing exposure time of mussels to ZnO-NP, the SOD antioxidant activity decreased after 4-h exposure, and an increased activity was observed when exposed for 24 h. The exposure of L. fortunei to increasing concentrations of ZnO-NP resulted in a decreasing tendency of SOD activity after 2-h exposure, without a statistical difference. After exposure for 24 h, the SOD activity increased significantly under 10-and 50-μg mL −1 ZnO-NP treatments when compared to the control (Fig. 4A). The activity of the CAT enzyme was not statistically affected by the increase in exposure time. However, the CAT antioxidant activity increased significantly after 24-h exposure to 10-and 50-μg mL −1 ZnO-NP when compared to the control. After 4-h exposure to ZnO-NP, the CAT levels increased when compared to control (Fig. 4B), although the difference was not statistically significant.
The protein sulfhydryl content decreased significantly after 24-h exposure to ZnO-NP for the tested concentrations. With exposure of L. fortunei to increasing concentrations of ZnO-NP, the protein sulfhydryl content increased significantly after 2 h in 50-μg mL −1 ZnO-NP, and no statistically significant alterations were observed for the additional exposure conditions tested (Fig. 4C). The TEAC levels remained essentially unchanged upon changes in the exposure time and NP concentration, as shown in Fig. 4D.

Oxidative damage to lipids and proteins
The oxidative damage to lipids increased in the 4-h exposure group and decreased after 24 h, considering the exposure of mussels to ZnO-NP. No oxidative damage to lipids was observed following a 2-h exposure period to ZnO-NP (Fig. 4F). Oxidative protein damage increased after 4 h of exposure to NP, although it decreased after being exposed to the NP for 24 h. Compared to the control, no significant alterations were detected in the group that was exposed to ZnO-NP for 2 h with increasing ZnO-NP concentration. Oxidative protein damage decreased for 1-, 10-, and 50-μg mL −1 ZnO-NP after 4-h exposure when compared to the control group. In the 24-h exposure group, the oxidative protein damage decreased regardless of the ZnO-NP concentrations (Fig. 4E).

DNA fragmentation assay
The results obtained from DNA fragmentation assays show that the DNA double strand remained intact in the specimens that were exposed to ZnO-NP, as shown in Fig. S1.

Discussion
The physical and chemical properties of nanoparticles may interfere with ecosystems, since the release of NP to the environment is inevitable and originating different levels of ecotoxicity (Ma et al. 2013;Châtel and Mouneyrac 2017). The shape, size, surface area, crystalline phase, and particle surface potential are considered important factors in reactive oxidative species (ROS) generation, as demonstrated for TiO 2 nanoparticles (TiO 2 -NP) in previous studies from our group  (Girardello et al. 2016a;Girardello et al. 2016b). The results presented here indicate that ZnO-NP have larger average sizes when compared to TiO 2 -NP (Girardello et al. 2016a) and consequently, a lower surface area, which may imply a lower particle reactivity potential. These differences could be related to the lower incidence of negative effects on the biomolecules of L. fortunei cells exposed to ZnO-NP when compared to other nanoparticles (Girardello et al. 2016a;Girardello et al. 2016b). High cytotoxic and phototoxic effects are related to a larger surface area of the nanoparticles (Xiong et al. 2013). Smaller molecules present a larger fraction of the atoms on the surface, which causes high mobility and reactivity (Tarrahi et al. 2018). In addition to the physical and chemical characteristics of metal oxide nanoparticles, another factor to be considered for risk assessment is the mechanism involved in the interaction of the nanomaterial with the biological system Wu et al. 2010). The toxic effect of metal oxide nanoparticles may involve distinct mechanisms of ion release from their complexes, producing chemical radicals or interacting with biological targets (Ma et al. 2013).
The production of reactive oxygen and nitrogen species by nanoparticles may occur directly or indirectly and act as a key pathway in toxicity induction (Tarrahi et al. 2018). This reactive species generation may be related to the modulation of antioxidant defenses (Azqueta and Dusinska 2015;Girardello et al. 2016a;Girardello et al. 2016b). ZnO-NP may induce toxic effects by their ability to disturb electron transfer processes in cells, increasing ROS generation and creating alternative routes of oxidative stress that are able to cause different types of cell damage (Ma et al. 2013). Our results show that exposure of golden mussels to nanoparticles induces a redox imbalance in the mussel's cells in both enzymatic and nonenzymatic antioxidant defenses. Oxidative damage to proteins and lipids were observed after 4-h exposure, and the levels stabilized after 24 h ( Fig. 4E and F), mainly for protein oxidative damage, which is more easily repaired than lipid oxidative damage (TBARS). On the other hand, SOD decreased after 4-h exposure and increased after 24 h (Fig. 4A). To remove the produced ROS, the SOD and CAT enzymes are activated, decreasing the oxidative stress in the cell. SOD acts as a first antioxidant defensive system, reducing the superoxide ion radical (O 2•− ) to hydrogen peroxide (H 2 O 2 ), therefore suggesting that the radical O 2•− is produced upon exposure to the nanoparticles (Halliwell and Gutteridge 2015). Catalase reduces H 2 O 2 , generating H 2 O and O 2 (Halliwell and Gutteridge 2015) to reduce the deleterious effect of H 2 O 2 . The SOD enzyme is important to avoid an increase in O 2•− concentration, which may lead to oxidative damage. The O 2•− Fig. 4 Enzymatic and nonenzymatic antioxidant defenses of golden mussel's soft body in different ZnO-NP concentrations after 2-, 4-, and 24-h exposure times: A SOD; B CAT; C protein sulfhydryl; D TEAC; E carbonyl protein, and F TBARS. Different lowercase letters indicate statistically significant differences to different concentration conditions in the same exposure time. Different capital letters indicate statistically significant differences at the same concentration of exposure in different exposure time. One-way ANOVA, Dunnett's multiple comparison test, and Tukey's test, p < 0.05 is an important reducing agent that is possibly responsible for generating carbonylated proteins and oxidized lipids at higher exposure concentrations. The imbalance between ROS and cellular antioxidants, with excessive production of ROS, allows specific species to attack the cellular macromolecules, causing peroxidation of lipids in the cell and mitochondrial membranes, mitochondrial dysfunction, inhibition of the enzymes activity, and DNA damage, which ultimately results in cell death (Abdel-Daim et al. 2019).
After 24-h exposure to ZnO-NP, SOD levels increased and lipid and protein damage decreased, which is possibly related to restoration of the redox metabolism. This modulation of SOD activity in organisms exposed to ZnO-NP suggests that the mussel L. fortunei tried to stimulate its detoxification system (Fig. 4). Indeed, this observation corroborates with data obtained by Huang and coauthors, who assessed the impact of ZnO-NP and ocean acidification on the antioxidant responses of Mytilus coruscus . A study with blue mussels (Mytilus edulis) evaluated the effects of ZnO structures exposure on these bivalves and showed that ZnO-NP caused an accumulation of oxidative lesions in proteins and lipids in the intracellular compartment (Falfushynska et al. 2019).
One explanation of these changes after 24 h is that the dissociation of ZnO may occur, releasing zinc ions (Zn 2+ ) by hydrolysis (Fig. 5). After 24 h of ZnO-NP exposure, the cells had probably already encapsulated the nanoparticles in their cytosol by phagocytosis. This process may occur by the invagination of the plasmatic membrane to form vesicles that enclose the nanoparticles and transport them into the cell (Châtel and Mouneyrac 2017). After phagocytosis of the ZnO-NP, the fusion of the lysosome with the phagosome occurs to degrade the particle. The hydrolysis of ZnO-NP starts in the phagolysosomes in the presence of water, and Zn 2+ release may occur because of the reaction ZnO + H 2 O ⇌ Zn 2+ + 2OH − (Taccola et al. 2011). A pH between 4.5 and 5 favors dissociation and subsequent ion release, corroborating the results obtained by ICP-MS analysis (Fig. 3), which showed high concentrations of Zn 2+ in L. fortunei after 24 h of exposure at high assay concentrations. This phagocytosis process is probably a defense mechanism of the cell when it perceives changes in its biomolecules caused by ZnO-NP exposure at higher concentrations and longer exposure times, corroborating the model of ZnO-NP cytotoxicity on proliferating cells (Taccola et al. 2011). This autophagosome formation was also observed in previous studies from our group after exposure to L. fortunei, with the golden mussel's cells encompassing TiO 2 -NP (Girardello et al. 2016b). The proposed mechanism of action of ZnO-NP in golden mussel cells and the defense mechanisms of these cells may be observed in the illustrative diagram of Fig. 5. Fig. 5 Schematic plot highlighting a possible mechanism of phagocytosis of ZnO-NP. ZnO-NP can be internalized through a phagocytosis process, followed by the fusion of the phagosome with the lysosome, forming the phagolysosome. Inside of the phagolysosome, the hydrolysis of ZnO-NP and the release of Zn 2+ into the cytosol can occur. In the cytosol, ZnO-NP can cause oxidative stress by generating ROS, damaging cellular organelles, as protein carbonylation and lipid peroxidation. ZnO-NP can also be internalized in autophagosomes and induce cell death. High ROS production and DNA damage can lead to cell death (Girardello et al. 2016b;Taccola et al. 2011;Mihai et al. 2015;Sabella et al. 2014) Swiatek and Bednarska (2019) discussed the idea that the detoxification process may be energetically costly. The authors could not observe a clear relation between Zn exposure and Eisenia andrei energy reserves, as carbohydrates were reduced after treatment. Another ecotoxicological study showed that the clearance rate is the component of the energy budget most affected by toxic compounds (Swiatek and Bednarska 2019). As an example, in mussels, the potential toxic elements may be associated with suspended particulate matter in the water column, and the energy gain from its feed may be compromised, since there is an energy deviation for contaminant metabolism (Toro et al. 2003). Following this idea, the high energy cost to detoxicate ZnO-NP would also be related to the increase in ROS. This oxidative stress caused by the detoxification process could be increasing the amount of ROS in a synergistic way (Swiatek and Bednarska 2019).
It is interesting to note that the ZnO-NP exposure or the ROS increase did not lead to DNA fragmentation. In addition, the oxidative stress observed in the L. fortunei cells did not result in double-strand cleavage in DNA once the redox metabolism was restored after 24 h. The fragmentation assay that was employed identified solely double-strand cleavage in DNA, whereas other forms of DNA damage, such as singlestrand cleavage and alkali labile site, were not detected. As evaluated by Girardello and coauthors in 2016 using the comet assay, exposure to TiO 2 -NP may cause DNA damage to the hemocytes of L. fortunei (Girardello et al. 2016a).

Conclusions
In conclusion, ZnO-NP induce oxidative stress in exposed golden mussels by physical and chemical properties and detoxification processes. The golden mussel can modulate its own antioxidant defenses in response to oxidative stress and seems to be able to hydrolyze the nanoparticle complex, thereby releasing Zn ions. This study gives a perspective of a mechanism that relates to the internalization of ZnO-NP through phagocytosis and its oxidative effects; however, further studies are required to understand the molecular mechanism of oxidative stress induced by ZnO-NP.
Availability of data and materials All data generated or analyzed during this study are included in this published article.
Funding The authors would like to thank the Conservation Unit Division (Divisão de Unidades de Conservação) of the environmental department of the state government (SEMA-RS) for allowing us to collect samples in Itapuã State Park (RS, Brazil) and the funding agencies FAPERGS, CAPES, and CNPQ for supporting this study. The authors also thank the National Institute for Advanced Analytical Science and Technology (INCTAA, CNPq proc. 465768/2014-8) for financial support.

Consent for publication Not applicable
Competing Interests The authors declare no competing interests.