Bioaccumulation and phytotoxicity of ZnO nanoparticles in soil-grown Brassica chinensis and potential risks

A pot-culture experiment was carried out to explore the phytotoxicity and accumulation of ZnO nanoparticles (ZnO NPs) in pakchoi (Brassica chinensis), as well as its potential risk. Zn was mainly accumulated in the shoots of pakchoi and showed the order of cell wall fraction > soluble fraction > organelles fraction. Endosomes were observed in roots cell by transmission electron microscope (TEM) and it is proving that endocytosis is a possible way for NPs to enter the cells. The results showed that higher concentration of ZnO NPs treatments resulted in an increase in the content of photosynthetic pigments and malondialdehyde (MDA), Catalase (CAT) and ascorbate peroxidase (APX) activity, while the activity of superoxide dismutase (SOD) and guaiacol peroxidase (G-POD) decreased signicantly compared with the control. These changes suggest that the mechanism of ZnO NPs phytotoxicity may induce strongly oxidative stress and damage biomembrane. Generally, ZnO NPs had a similar impact on the growth and absorption of Zn in pakchoi with Zn 2+ . Simultaneously, the treatment of ZnO NPs affected the nutritional quality and food safety of pakchoi at high concentration.


Introduction
Zinc oxide nanoparticles (ZnO NPs) are one of the most used forms of metal-oxide nanoparticles (MONPs) and have been widely used in various products such as personal care products, textiles, paintings, industrial coatings, dye-sensitized solar cells, plastics, glass, ceramics, cement, nanomedicine, as well as optic and electronic materials cosmetics (Balážová et al. 2018, Liu et al. 2020, Wang et al. 2013, Zoufan et al. 2020. It has been estimated that the global annual production of ZnO NPs ranges from 550 to 35,000 tons (Liu et al. 2020, Piccinno et al. 2012, Rajput et al. 2018, and about 8-28% of them are released into the soil environment (Medina-Velo et al. 2017).
Soil is considered as a major sink of ZnO NPs compared to atmospheric and aqueous ecosystems (Liu et al. 2020, Rajput et al. 2018. In recent years, ZnO NPs have been employed as a novel macronutrient nanofertilizer in the Zn de cient soil (Song &Kim 2020, Sun et al. 2020, Yuse -Tanha et al. 2020, as an antibacterial agent for plant protection , as well as an effective agent for reducing metal uptake by plants (Rizwan et al. 2019, Sharifan et al. 2019a. ZnO NPs may enter the soil medium by direct (agricultural products) or indirect (biosolids) pathways, and then can be absorbed and accumulated by plants, causing an effect on plant growth (Zeb et al. 2021). Large scale applications of ZnO NPs are linked to a high environmental distribution, which demands a better understanding of the interaction of these NPs with plants.
Plants are essential components of various ecosystems (Balážová et al. 2018, Yuse -Tanha et al. 2020 and are prone to be directly in uenced by ZnO NPs in soils, posing possible risks to food safety and human health (Baskar et al. 2020). From an ecological perspective, understanding the phytotoxicity and uptake of ZnO NPs in plant species is of signi cant importance (Zafar et al. 2016). Phytotoxicity of ZnO NPs is largely related with nanoparticle size and concentration, surface coating, plant species and growth matrix (Liu et al. 2020). A latest study showed that lower concentrations of ZnO NPs (especially 10 and 25 mg/L) enhanced seed germination and improved seedling growth of Vicia faba, while higher concentrations (100 and 200 mg L − 1 ) resulted in phytotoxicity (Youssef &Elamawi 2020). Length and weight of Brassica nigra plants were adversely affected by ZnO NPs in culture conditions, but no signi cant effects were observed in the soil matrix (Zafar et al. 2019). Moreover, ZnO NPs were highly toxic to the leaves of Glycine max (Priester et al. 2017), while they were relatively less toxic to Medicago sativa leaves (Bandyopadhyay et al. 2015). Therefore, further studies were needed to deeply reveal the phytotoxicity mechanisms of ZnO NPs in plants.
Pakchoi (Brassica chinensis) is one of the most important leafy vegetables worldwide, which contains rich essential nutrients (Cu, Ca, Fe, and so on), abundant vitamins and minerals. Due to the strong dissolution of ZnO NPs, their phytotoxicity on pakchoi, as well as a dose-response relationship relating ZnO NPs and Zn ions, need to be investigated. Thus, the paper aims to (1) realize the absorption and accumulation of ZnO NPs in pakchoi, (2) determine the biochemical parameters of pakchoi treated with ZnO NPs, (3) compare the effects of ZnO NPs to Zn 2+ on pakchoi, determine the phytotoxicity of NPs to plant and (4) assess the potential food risks of pakchoi on safety and quality.

ZnO NPs characterization
Zinc oxide nanoparticles (ZnO NPs, 30 ± 10 nm, 99% purity) were obtained from Shanghai Macklin Biochemical Co., China. The morphology and particle size distribution of the nanomaterials used in the experiment were observed by transmission electron microscopy (TEM, JEM-2800), and the crystal form and purity were confirmed via X-ray powder diffraction (XRD, Ulitma IV). The TEM image and diffraction pattern are shown in Page 4/25 Fig. S1. From Fig. S1A, it can be clearly seen that the NPs were predominantly spherical, even though NPs with tetragonal or other irregular shapes can also be seen. The average particles size was about 30 nm. XRD diffraction peak showed that the ZnO NPs were hexagonal, with good crystallinity and high purity (Fig. S1B). ZnO NPs formed aggregation in aqueous solution definitely. The hydrodynamic size distribution of ZnO NPs suspensions with different concentrations was measured with dynamic light scattering (DLS) in this study. The relative particle size of ZnO NPs in the suspension was mainly distributed within the range of 348 -435 nm for 50 -1000 mg L -1 (Fig. S1C). The zeta potential peaked at 19.53 ± 0.55 mV when the ZnO NPs suspensions concentration was 1000 mg L -1 (Fig. S1D).

Soil treatments
Fresh and clean soil samples were collected, dried naturally, and passed through a 20mesh sieve for later use. 1 kg of soil samples were weighed and added into the pots, which is 11 cm high with an outer diameter of 18 cm. A total of nine treatment groups were prepared including a negative control (CK), with three replicates in each treatment group.
The pakchoi seedlings were treated with 50, 100, 500 and 1000 mg kg -1 of ZnO NPs, respectively. Each group was correspondingly labeled as N50, N100, N500, N1000 separately. (These concentrations were selected because they clearly affected the growth and physiological metabolism of pakchoi plants, previously). While the Zn 2+ treatment groups were labeled as I50, I100, I500 and I1000 respectively. (Zn 2+ was obtained by the guaranteed reagent ZnSO 4 , which provided an equal amount of Zn compared to ZnO NPs.) Each group of accurately weighed ZnO NPs was added to the weighed clean soil to make a contaminated soil of the aforementioned concentration through a stepwise stirring manner.
Stock solution of ZnSO 4 were prepared, and a corresponding volume of stock solution was added into the weighed soil and stirred well. The pots were placed in the dark and balanced for a month, watered to equivalent of 80% field capacity weekly (Ebbs et al. 2016).

Plant culture and harvest
Pakchoi seeds were provided by the Beijing Green gold blue Seedlings Co., China. The seeds were sterilized using 10% H 2 O 2 for about 30 min and then rinsed repeatedly with Page 5/25 ultrapure water before germination. The germination experiment was carried out in a seedling tray at room temperature and moist soil. When the seedlings grew to 3 -4 true leaves, the seedlings with relatively consistent growth were selected and transplanted to a balanced contaminated soil, with 3 plants per pot. Regular watering was done to keep the soil in the pot at 60% -80% of the field water capacity, and the plants were harvested one month after transplantation. The pot-culture experiment was carried out under a glasshouse, and the irrigation water was DI water during the whole test period.
After a month of transplanting, the plants were harvested and rinsed thoroughly with tap water to remove the surface sludge and dust attached to the plants, immersed in 20 mM Na 2 -EDTA solution for 20 min and then washed with DI water for several times. The water on the surface of the plants was dried with filter papers. Pakchoi plants were separated into roots and shoots and the fresh weight was weighed.

Transmission electron microscope (TEM) observation
About 1 mm fresh root tip of pakchoi was selected from 500 mg kg -1 treatment group using double blades, and put into a fixative composed of 2.5% glutaraldehyde solution at 4 ℃ for 5 h, and for 2 h in 1% osmic acid again. After rinsing in buffer solution, root tip was dehydrated in an ethanol series, and embedded by epoxy resin steps. Finally, uranium dyed samples were sliced by an ultrathin slicer (Leica, Germany). The detailed operation referenced to Ma et al. (2013) (Ma et al. 2013). The nickel net that was placed with the processed samples was observed using TEM (JEM-2800).

Fe, Cu, Ca, P and Zn quantitative analysis in plant
Zn, Ca, Cu, P, and Fe content in pakchoi were determined by ICP-MS (Perkin Elmer ELAN DRC-e). The fresh samples were dried in the oven at 105 ℃ for 5 min, and at 70 ℃ until constant weight as described by Liu et al. (2010) (Liu et al. 2010). Briefly, 0.10 g of dried plant tissues were transferred into a digestion tubes with a mixture of 1:1 HNO 3 and H 2 O 2 on hot block digester (Xiong et al. 2017). The experiment used a standard substance (GB/T 35871-2018, from the Standardization administration in China) to detect the recovery of heavy metal in plants. The recovery rate of standard materials was 90 -105 %.
Each sample was analyzed as triplicate.

Determination of Zn subcellular distribution of pakchoi
Zinc determination in pakchoi plants was done according to the procedure by Zhou et al (Zhou et al. 2017). Briefly, 0.5 g fresh samples of shoots and roots were weighed and added into a pre-chilled mortar, 10 mL extraction buffer was added to fully grind. The buffer was composed of 1.0 mM DTT, 500 mM sucrose, 5.0 mM ascorbic acid, 50 mM HEPES and 1% Polyclar AT PVPP (w:v, pH 7.5). The homogenate was sieved through 100 mesh nylon sieve, the residue was cell wall fraction. The filtrate was centrifuged at 10000 × g for 30 min, the deposition and supernatant were organelles fraction and soluble fraction (including vacuole, cytoplasm and cytomembrane) respectively. All operations were performed at 4 ℃. The fractions were dried to a constant weight, and then digested with HNO 3 -HClO 4 mixture (4:1, v:v). Zn content was determined by ICP-MS, according to the method used by Liu et al. (2010) (Liu et al. 2010).
2.7 Determination of biochemical parameters 2.7.1 Determination of photosynthetic pigments contents 0.2 g fresh leave samples were cut to pieces and extracted in the mixture of 10 mL ethanol and acetone (1:1, v/v). Shook until the leaves were completely discolored in the dark, and volume was adjusted to 25 mL. The absorbance of supernatant was determined at 662 nm, 645nm and 470 nm in sequence by the UV-422G spectrophotometer.

Estimated daily intake (EDI)
To assess the health risks of Zn consumed leaves of pakchoi, the EDI of Zn was calculated by the following equation (Natasha et al. 2020): Here, C FW is the concentration of Zn in the fresh tissues of pakchoi edible part, IR represents the day ingestion rate, the average value is 0.345 kg d -1 (Hu et al. 2014 The oral reference dose (RfD, mg kg -1 d -1 ) obtained from the EPA's Integrated Risk Information System online database (IRIS) is intended for use in risk assessments for health effects known or assumed to be produced through a nonlinear (possibly threshold) mode of action. The value of IRIS used was 0.3 mg kg -1 d -1 (Agency. 2005).

Data analysis
All the data were analyzed with SPSS 13.0, Excel 2010 for one-way ANOVA and Tukey's HSD test, and graphed by Origin 9.0. All treatments were replicated three times in this experiment. And the data were reported as mean ± standard error (SE). Metal concentrations of plants are expressed in terms of DW.

Zn accumulation and distribution
According to Fig. 1, Zn was mainly accumulated in the shoot tissues of pakchoi, both ZnO NPs and Zn 2+ appreciably enhanced the Zn content in pakchoi tissues. However, the zinc concentration of edible tissues with NPs treatments were not affected compared to control (CK). With the increase of the ZnO NPs dosage, the concentrations of Zn in pakchoi gradually increased. When the concentration of ZnO NPs was 1000 mg kg − 1 , the Zn concentration of shoot was 315.17 mg kg − 1 that was signi cant difference compared with CK. In addition, there was inapparent difference after Zn 2+ treatment.
There were no signi cant differences between Zn 2+ and ZnO NPs treatments, but the concentration and content of Zn in each component of pakchoi treated with ZnO NPs was higher than Zn 2+ under the same conditions. The same result was found in soybean (Glycine max), Chinese cabbage (Brassica pekinensis L.), lettuce (Lactuca sativa L.), cowpea (Vigna unguiculata) and so on (López-Moreno et al. 2010, Wang et al. 2013, Xiang et al. 2015, Xu et al. 2018a. In fact, a series of experiments suggested that changes in plant physiological characteristics exposed to ENPs or heavy metals may be due to dynamic changes in essential nutrients in plant tissues (Sharifan et al. 2019b(Sharifan et al. , 2020. Therefore, the toxicity mechanism of the two treatments on plants is similar, and further experiments are needed to investigate the action pathway of ZnO NPs in plants. The TF values were signi cantly higher than 1, Table 1. When the Zn concentration was increased to 500 mg kg − 1 , the TF values decreased as compared to CK. It has been reported that ZnO NPs can affect the performance of transporters, in uence the uptake of heavy metals (Sharifan et al. 2020). Therefore, at higher concentrations, the phytotoxicity of Zn affects the absorption and transport of heavy metals. In addition, the concentration and content of Zn in each component of pakchoi treated with ZnO NPs was higher than Zn 2+ under the same conditions. Figure 2 showed the distribution of Zn in the subcellular fraction, and the subcellular distribution of Zn in pakchoi treated showed the order of cell wall fraction > soluble fraction > organelles fraction. Zn was stored mainly in the cell wall of leaves (79.1%) and roots (85.3%) treated by ZnO NPs, which is consistent with most studies , Xin et al. 2017, Xu et al. 2018b, Zhou et al. 2017). The proportion of Zn showed an increase in the soluble fraction but decreased in the organelle of pakchoi tissues treated by ZnO NPs. Rather, exposure to Zn 2+ resulted in a decrease in the proportion of organelles and soluble fractions. The cell walls are the rst barrier for heavy metals uptake and can restrict their transport across the cytomembrane , the organelles are the main place for cell metabolic activities (Xin et al. 2017). Studies have shown that soluble components are mainly composed by vacuoles (Dou et al. 2009), and vacuolar regionalization is identi ed as one of the most important heavy metal detoxi cation mechanisms (Zhang et al. 2015b). Therefore, the subcellular result might be related to the defense mechanism of pakchoi cells, and indicated that Zn 2+ caused lower damage to normal metabolism than ZnO NPs.

TEM analysis
In order to understand the possible pathway of NPs into plant cells, the root cells were observed by TEM in this study. Figure 3 illustrated the possibility of the presence of ZnO NPs in pakchoi root cells. Compared with the TEM image of control (Fig. 3E) ZnO NPs were found in the entire cell (Fig. 3A). We could see NPs not only in the cytoplasm (Fig. 3D) but also in the intercellular spaces (Fig. 3B) of pakchoi root cells. Many of round vesicles (range from about several hundred nanometers to a few microns) were observed in Fig. 3A, these vesicles were similar in size and shape to the endosomes observed in tobacco BY-2 cells by Lam, et al (Lam et al. 2007).
As con rmation that ZnO NPs can enter cells, it is important to understand the main transmembrane pathway for uptake of ZnO NPs. We conjectured that Fig. 3C may be the early stage of the formation of vesicles. Early endosome with tubular vesicles structure morphology appearance is not observed in our study. Therefore, the vesicles observed in the TEM image may be late endosomes. The presence of endosomes suggests that ZnO NPs may enter the cells through endocytosis if our conjecture was correct, thus proving that endocytosis is a possible way for NPs to enter the cells. It was reported that NPs can also enter the plant cell through ion channels, aquaporin or carrier protein except endocytosis (Jasmina et al. 2010).

Pakchoi growth and photosynthesis
As shown in Fig. 4, the biomass of pakchoi reached the maximum value of 7.0 g at 500 mg kg − 1 of ZnO NPs. But, the biomass of pakchoi increased with the increase of Zn 2+ concentration. Under the same conditions, Zn 2+ promoted the biomass of pakchoi more than ZnO NPs,especially at 1000 mg kg − 1 . It may be related to the fact that ZnO NPs could enhance the accumulation of Zn more than Zn 2+ (Fig. 1), and caused more effect to plant growth than Zn 2+ .
The content of chlorophyll directly affects the photosynthesis e ciency of plants, which in turn affects the growth of plants. Figure 5 depicted that with increasing the concentration of ZnO NPs, the contents of chlorophyll (Fig. 5a) and carotenoid (Fig. 5b) without exception were increased. In Fig. 5, it is shown that chlorophyll and carotenoid are increased signi cantly under N1000 and I100 treatments, compared with control. There is no obvious difference in the effect of Zn 2+ and ZnO NPs treatments on the pigment content of pakchoi leaves.
ZnO NPs can signi cantly promote the growth of edible plants by increasing chlorophyll content and total biomass. It was found that certain concentration of ZnO NPs increased the growth of mung (Vigna radiata) and cotton (Gossypium spp) (Pramod et al. 2011, Venkatachalam et al. 2017). The phytotoxicity of metal NPs is considered to be due to the toxicity of NPs themselves and the ions dissolved from NPs (Rao & Shekhawat 2014). TEM analysis of pakchoi revealed that there were suspected NPs in root cells. Some studies have found that metal NPs may form small masses or break into ions inside the plant (Večeřová et al. 2016). According to the results of Zn content, biomass and photosynthetic pigments of pakchoi, ZnO NPs and Zn 2+ promoted the growth of pakchoi by increasing chlorophyll content and total biomass, but there were not statistical differences between the phytotoxicity of ZnO NPs and Zn 2+ under the same conditions.

Antioxidant enzyme activities and lipid peroxidation content
To understand the oxidative stress produced by pollutant entering plants. This study measured the activities of antioxidant enzymes (SOD, CAT, APX and G-POD) and MDA content of pakchoi leaves. SOD catalyzes the disproportionation reaction of O 2 · to H 2 O 2 , which is then reduced to H 2 O and O 2 by CAT and peroxidase. Figure 6a showed that the SOD activity increased by 6% at the 50 mg kg − 1 of ZnO NPs. As the concentration of pollutions increased, SOD activity decreased. The activity of SOD in shoots decreased by 33%, 52% in 1000 mg kg − 1 of ZnO NPs and Zn 2+ respectively. On the other hand, there was more effect of Zn 2+ on SOD activity than ZnO NPs, compared with control.
Similar to the role of CAT, APX can remove H 2 O 2 in plants, and G-POD is related to plant anti-aging. Within a certain range (less than 500 mg kg − 1 ), there was a dose-response relationship between ZnO NPs and CAT activity (Fig. 6b). But no obvious change was detected in CAT activity of pakchoi shoots when treated with Zn 2+ . CAT activity in the plants grown at 500 mg kg − 1 of ZnO NPs and 1000 mg kg − 1 of Zn 2+ increased approximately 186% and 177% separately. The difference of APX (Fig. 6c) and G-POD (Fig. 6d) activity between ZnO NPs treatments were similar to that of CAT activity. The maximum of APX activity increase to 101% at 500 mg kg − 1 and the maximum of G-POD activity was found at 100 mg kg − 1 .
Compared with ZnO NPs, Zn 2+ treatment has less effect on the activity of APX and G-POD.
The content of MDA is usually used to characterize the degree of cell membrane lipid peroxidation. To evaluate the intensity of oxidative stress under ZnO NPs treatments, the MDA content was measured. As the concentration of ZnO NPs increased to 1000 mg kg − 1 , MDA content showed an increase of approximately 187% compared with the control (Fig. 6e). MDA content was not signi cantly different from that of the control (p < 0.05) between the pakchoi treated with Zn 2+ . Similar to APX and G-POD activity, Zn 2+ treatment has less effect on MDA content compared with ZnO NPs.
The changes of antioxidant enzyme activities and lipid peroxidation content suggested that ZnO NPs within a range of concentration strongly induce oxidative stress. ZnO NPs might affect functionality of transporter. Low concentration of ZnO NPs promoted SOD removal of O 2 · . As the concentration increased, the phytotoxicity of ZnO NPs affected the performance of SOD. The activity of other antioxidant enzymes was similar to SOD. The results suggested that low concentration MONPs could alleviate the phytotoxicity by increasing the activity of enzymatic antioxidants, while higher concentration inhibited the activity of the antioxidant enzyme system. Some studies had proved this conclusion (Ajey et al. 2016, Qian et al. 2013, Zoufan et al. 2020). In the present study, Zn 2+ treatment had less effect on the activity of APX, G-POD, CAT, and the content of MDA compared with ZnO NPs.

Nutrient elements in shoots
Pakchoi is not only rich in minerals and vitamins, but also has a high content of essential nutrients (Ca, P, Fe, et al.) The concentration of several nutrients of edible tissues accumulation in pakchoi treated by ZnO NPs is shown in Fig. 7. The increase of ZnO NPs concentration, especially at 100 mg kg − 1 , promoted almost 74% of the absorption of Ca by pakchoi. The content of P is much higher than other nutrient elements. Similar to Ca, the concentration of P increased by 31.90% when treated with ZnO NPs (100 mg kg − 1 ). Fe and Cu content increased by approximately 88.3%, and 31.66% at 50 mg kg − 1 , respectively. The impact of ZnO NPs in accumulation of few essential nutrients in shoots follows the order of Fe > Ca > P > Cu. Similar results were also found in cilantro and spinach for absorbing Fe and Cu (Sharifan et al. 2020).

Health risk assessment
The hazard quotient (HQ) is widely used to assess the non-carcinogenic risk level due to heavy metal consumption (Hu et al. 2013, Xu et al. 2015. The potential health risk of Zn in the edible part of pakchoi are presented in Table 1. The HQ exceeded threshold limits (1.0) for N1000 and I100, and the HQ value was close to the limit when the concentration of ZnO NPs is higher than 100 mg kg − 1 . When the concentration of Zn 2+ is 100 mg kg − 1 , HQ reaches the highest value of 1.03 and then gradually decreases, strangely. The EDI and HQ were more than one times higher at N1000 compared with control. There was a dose-dependent relation between health risk parameters and ZnO NPs applied, but the health risk of Zn 2+ at high concentration need to more research.
Recently, some studies (Ji et al. 2018, Natasha et al. 2020, Zhong et al. 2018) have focused on the risk assessment of crops triggered by heavy metals. For the human body, zinc is an essential nutrient, but excessive zinc can cause considerable effects on protein synthesis, induce a copper de ciency in the body, and even cause cancer (Zhou et al. 2016). Xiong et al. (Xiong et al. 2017) indicates that only a part of micro-sized metal particles is bio-accessible for the plant. But the current risk assessment of NPs focuses on the effects of metal species on plants, rather than NPs on human health. Thus, the toxicity of NPs in plants and human health need further investigations. In our experiment the distribution of Zn in root cells and Zn content in various tissues of pakchoi were measured, the intake risk of ZnO NPs into human body could not be assessed. Therefore, the existence and distribution of ZnO NPs in usable tissues needs to be further explored to determine the possibility of human intake risk.

Conclusions
The effects of ZnO NPs and Zn 2+ on pakchoi were similar, but pakchoi had a greater ability to absorb and transfer Zn under the NPs treatment. ZnO NPs can promote the growth of pakchoi by increasing chlorophyll content and total biomass, yet nonetheless Zn 2+ promoted the biomass of pakchoi more than ZnO NPs. Based on the discovery of vesicles on roots cell, ZnO NPs may enter the cell through endocytosis. ZnO NPs could enhance the accumulation of Zn more than Zn 2+ , which will cause more serious impact on homeostasis of pakchoi. ZnO NPs changed the activities of antioxidant biomarkers and MDA content, suggesting oxidative stress induction in pakchoi. In particular, ZnO NPs can increase levels of lipid peroxidation and disrupt the permeability of biomembrane. Moreover, ZnO NPs affect the nutritional quality of pakchoi especially the content of trace elements (Ca, Fe, P, Cu), simultaneously, HQ values indicate that NPs would cause food safety risk.
Declarations Tables   Table 1 The values of TF, EDI (mg kg -1 day -1 ), HQ of Zn in pakchoi.  Figure 1 Zn concentrations in shoots and roots of pakchoi (DW). Different lettering indicated the signi cant difference at p < 0.05, and the error bars are standard errors.

Figure 2
Zn subcellular distribution in shoots and roots of pakchoi treated with control (CK), ZnO NPs (N) and Zn2+ (I).