Effects of Selenium–Zinc Interactions on the Bioavailability of Selenium/Zinc in Soil and Its Mechanism


 Purpose Biofortification is a core strategy in solving the malnutrition of selenium (Se) and zinc (Zn) for global human being. However, whether Se and Zn co-biofortification can manipulate soil Se/Zn bioavailability and its mechanism still remains unclear. Methods The pot experiment was conducted to investigate the co-amendment of selenate and Zn sulfate on the growth of pak choi, especially on the uptake of Se and Zn, and to elucidate the effect of soil pH and soil enzyme activity on the bioavailability of Se/Zn in soil and its mechanism. Results Results showed that plant growth inhibition caused by the application of high Se rate was significantly alleviated with Zn supplements, and the biomass in the shoots and roots of pak choi in Se2.5Zn20 and Se2.5Zn50 treatments significantly increased (67.0%–112.8%) compared with the Se2.5Zn0 treatment. Additionally, Se and Zn co-amended application significantly enhanced soil available Se/Zn content compared with correspond single Se/Zn treatments. The increase of soil available Zn content could be attributed to the significant decrease in the soil pH, while the increase of soil available Se was from the biochemical conversion caused by the activity of catalase, urease, and alkaline phosphatase. Conclusion Se–Zn co-amendment can ameliorate bioavailability of Se/Zn in soil by regulate the pH and enzyme activity of soil.


Introduction
Selenium (Se) and zinc (Zn) are essential micronutrients for humans and animals, and inadequate intake of these micronutrients could threaten human health because it weakens the immune system. Meanwhile, they have been associated with an increase in the incidence of cardiovascular diseases, asthma, epilepsy, growth retardation, decreased fertility, cognitive decline, and various cancers (Mangueze et al. 2018). Estimates pointed that ca. 30% (2.7 billion) and 15% (0.5-1 billion) of the world's population experience Zn and Se de ciency, respectively, which constitute a serious problem of public health, especially in developing countries (Li et al. 2015).
In China, approximately 51% of soil is affected by Se de ciency, and most of this area is Zn de cient as well ; Yang et al. 2007). Therefore, to meet the human nutrition's needs for Se and Zn, effective and e cient co-bioforti cation strategies of Se and Zn must be developed (Mangueze et al. 2018). Among them, agronomic bioforti cation by applying Se and Zn fertilizer in soil is an effective and reliable solution that can supplement Se and Zn in co-de cient areas to increase the Se and Zn contents in plants and then in the human body (Bañuelos et al. 2015; Xue et al. 2020). Notably, a narrow range exists among Se dietary de ciency (< 40 ug/d), suitability (~ 110 ug/d), and toxicity (> 400 ug/d) levels, which are quite di cult to regulate (Qin et al. 2016). When Zn was applied to soil, especially in calcareous soil with high pH and carbonate content, a series of physical and chemical reactions will occur with the soil components, thereby resulting in an extremely low available Zn content for plants (Lu et al. 2012). Furthermore, Se usually exists as selenite oxyanions in oxidizing conditions, whereas Zn exists largely as Zn 2+ in aqueous environments (Xue et al. 2020). Therefore, the interaction of Se and Zn in soil must be clearly understood in their co-bioforti cation.
When different elements enter the soil-plant system, their interaction will affect the bioavailability of elements in soil and the growth of crops. Previous studies found that application of L-SeMet (L-selenomethionine) and EDTA-Zn (sodium ethylenediamine tetraacetate-Zn) in soil signi cantly (p < 0.05) increased the biomass of pak choi (Dai et al. 2019). Selenate and Zn sulfate coamended treatments could reduce the binding intensity of Zn, thereby making Zn the available forms for plants and promoting the uptake and translocation of Zn by pak choi (Xue et al. 2020). Previously, Poblaciones and Rengel (2017) and Dai et al. (2019) veri ed that the co-amendment of Se and Zn could promote the uptake of Se/Zn in plants. By contrast, in a eld experiment, Ramezani et al. (2016) reported that soil co-application of selenite and Zn sulfate meliorated the uptake of Se but restrained Zn uptake of alfalfa. Fang et al. (2008) found that foliar application of sodium selenite and Zn sulfate had no signi cant in uence (p > 0.05) on the absorption of Se/Zn in rice, which may be due to the discrepancy in the metabolic pathways between Se and Zn in plants. Generally, Se enters the roots via the S transport system, P transporter (OsPT2), or a silicon transporter [Lsi1 (OsNIP2;1)] , and Zn uptake by plant roots involves an active process though Zn transporter (ZIP) (Nguyen et al. 2017). Based on the ndings, different responses on soil Se/Zn bioavailability induced by Se-Zn co-amended condition was closely related to plant species and soil properties and relied on Se/Zn species and application methods as well.
Previous studies involved in the Se-Zn interaction were mostly focused on the effect of selenite and Zn sulfate co-amendment on plant growth. Given that selenite was feasibly adsorbed onto carbonate, Fe/Mn oxides, and organic matter in soil, soil Se bioavailability was inhibited. Nevertheless, Se initially exists in the soluble fractions and binds weakly to soil particles with the addition of selenate; Se easily migrates with pore water in soil and becomes more available than other soil fractions (Wang et al. 2012). Consequently, selenate was the major Se fertilizer that is commonly used in Se bioforti cation, whereas the effect of selenate and Zn sulfate co-amendment on plant growth have not yet to be systematically conducted.
The accumulation of Se and Zn in crops not only depends on the availability of Se and Zn contents in soil (Xu et al. 2020) but also affected by the physicochemical properties soil, such as pH, cation exchange capacity, redox potential, enzyme activity, and microbe activity (Dhal et al. 2013). Soil pH plays an incomparable role in governing Se/Zn bioavailability in the soil (Antoniadis et al. 2017).
Zn availability decreases as soil pH increases. At pH values below 7.7, Zn 2+ predominates, but ZnOH + is predominant between pH 7.7 and 9.1 and Zn(OH) 2 is predominant above pH 9.1 (Zhao et al. 2018). On the contrary, the bioavailability of Se gradually decreases with pH, which determines the valence state of Se in soil, thereby, in uencing soil Se bioavailability. The de ned available Se in soil mainly appears as inorganic Se (IV) and Se(VI) ).
Soil enzymes are involved in all nutrient cycling, and its availability to plants can be used as an index of soil function (Tan et al. 2018), which could also be widely used to assess the impact of various heavy metals on soil microbial function, both in a long and short period of time (Yang et al. 2007). Studies showed that the application of Se/Zn markedly in uenced soil enzyme activity (Kunito et al. 2001;Nowak et al. 2004). Soil is a multiphase complex medium. The surfaces of soil particles, such as clay, can not only adsorb heavy metal ions but also interact with soil enzymes (Hu et al. 2014). Thus, the relationship between soil enzyme activity and the bioavailability of metals in soil needs to be further explored.
Given above, the main objectives of this study were to (1) clarify the effect of co-amendment of selenate and Zn sulfate on plant growth, especially the uptake of Se and Zn, and (2) elucidate the effect of soil pH and soil enzyme activity on the bioavailability of Se/Zn in soil and its mechanism.
Seeds of pak choi (Brassica chinensis L.) were provided by the Northwest A&F University Seeds Co. Ltd., Shaanxi, China. Pak choi seeds were immersed in 0.5% NaClO solution for 30 min and washed with deionized water before planting.

Experimental design
According to previous studies, Se (Na 2  After the soil was homogenized and equilibrated for 21 days, basal fertilizer (N, P, and K) that comprised 0.15 g/kg N (urea, AR) and 0.033 g/kg P (monopotassium phosphate, AR) were thoroughly mixed into the equilibrated soil (Qin et al. 2016). Different concentrations of Se and Zn solutions were applied to the soil using a plastic nebulizer. A total of 15 seeds were sown in each pot, thinned to 5 seedlings per pot after 10 days, and grown under natural lighting conditions. Deionized water was supplied once every 2 to 3 days to maintain 70% of the soil water-holding capacity during the entire experiment.
In 30 d after sowing, the pak choi was harvested and then were divided into shoots and roots. The shoots and roots were washed thoroughly with deionized water after being washed with tap water and then dried with absorbent tissue. Both parts of the pak choi were oven-dried at 90°C for 30 min and 60°C to a constant weight, and then dry weight (DW) was recorded. Dry samples were used to analyze Zn and Se concentration. Soil samples were collected from each pot and then completely air-dried at room temperature, homogenized, and nally ground to pass through 2, 1, and 0.149 mm nylon sieves for chemical analyses, respectively.
Plant and soil analysis Soil pH was extracted with CO 2 -free deionized water (water : soil = 2.5:1) and measured by a pH electrometer (Bao 2000).The activity of invertase (INV) in soil was determined by 3,5-dinitrosalicylic acid colorimetric method, and the amount of reducing sugar produced by enzymatic reaction with sucrose as substrate, that is, the mg number of glucoses produced 1 g soil after 24 h was expressed. The urease activity (URE) in soil was determined by phenol sodium-sodium hypochlorite colorimetric method and by enzymatic reaction of urea as substrate NH 4 + -N, expressed by the number of NH 4 + -N milligrams per 24 h per 1 g soil sample. The alkaline phosphatase (APS) activity in soil was determined by the colorimetric method of benzene disodium phosphate, and the conversion of benzene disodium phosphate to phenol was expressed by the number of micrograms P 2 O 5 in soil 1 h later. The catalase (CAT) activity in soil was titrated with standard potassium permanganate solution by adding a certain amount of hydrogen peroxide solution to the soil, and the enzyme activity was expressed as the KMNO 4 milliliter number of 0.1 mol/L per 1 g soil sample (Guan et al. 1986).

Quality control and data analysis
Different quality assurance and control measures were adopted in sample preparation and chemical analyses, including the use of certi ed reference materials for instrumental calibration, determination of the method detection limit, and analyses of reagent blanks, sample duplicates, and spiked samples. In plant standard substances, the measured Se concentrations were 0.19 ± 0.05 mg/kg The differences among different Se/Zn treatments with the same Zn/Se application concentration were evaluated using a one-way analysis of variance (one-way ANOVA, p < 0.05). Two-way ANOVA was used to further investigate the effects of Se and Zn treatments and their interaction on plant growth. Pearson correlation analysis, linear stepwise regression (SR), and path analysis were conducted to clarify the effect of soil pH, enzyme activity, and available Se/Zn on Se/Zn bioavailability. All tests were performed using Microsoft Excel 2013 and SPSS 13.0.

Plant growth
The biomass (dry weight; DW) in pak choi shoot and root was signi cantly affected by single Se/Zn and Se-Zn co-amended treatments (p < 0.01) ( Table 2). When compared with control treatment, the biomass (DW) in pak choi shoot and root showed no signi cant (p > 0.05) difference at the single Se concentration of 0.5 and 1 mg/kg but reduced by 40.4% and 27.7% at single Se concentration of 2.5 mg/kg, respectively. Meanwhile, for single Zn treatments, the biomass of pak choi shoot had no signi cant (p > 0.05) effect, and the root biomass increased by 33.4-51.0% compared with the control treatment (p < 0.05).  (Fig. 1a).
The Zn content in pak choi shoot and root was signi cantly affected by single Zn treatment (p < 0.01), whereas Zn content was not signi cantly in uenced by the single Se treatments or Se*Zn interaction (Figs. 1e and f) (p > 0.05). For single Zn treatments, Zn content in shoot and root of pak choi signi cantly increased (0.6-3.1 folds) with the increase in Zn application rates (p < 0.05) (Fig. 1b). No signi cant change in Zn contents was observed in plants exposed to Se-Zn co-amended treatments in comparison with the corresponding single Zn treatment (p > 0.05) (Fig. 1b).

Available Se/Zn in soil
The available Se/Zn contents in soil with different treatments are shown in Fig. 2. The available Se/Zn content in soil was signi cantly in uenced by Se and Zn treatments or their interaction (p < 0.01) (Figs. 2c and d).
For single Se treatments, available Se contents in soil signi cantly increased (11.7-62.6 folds) with the increasing Se rates (p < 0.05) (Fig. 2a). Compared with the corresponding single Se treatments, the available Se contents in soil in the combined application of Se and Zn treatments increased by 9.6-41.6% (only showed signi cant difference in Se1Zn20 treatment) (p < 0.05) (Fig. 2a).
For single Zn treatments, the available Zn contents of soil signi cantly increased (4.5-11.0 folds) with the Zn application rates (p < 0.05) (Fig. 2a). Compared with corresponding single Zn treatments, the available Zn contents of soil in the combined application of Se and Zn treatments increased by 0.12-70.7% (only showed signi cant difference in Se1Zn20 and Se2.5Zn50 treatments) (p < 0.05) (Fig. 2a).

Soil pH and enzyme activities
In the present study, single Se/Zn application showed no signi cant effect on soil pH. Conversely, the pH of Se-Zn co-amended soil changed with the Se/Zn content in soil (p < 0.01) ( Table 1). Compared with Se0.5Zn0 treatment, the soil pH in Se0.5Zn20 and Se0.5Zn50 treatments increased by 0.46 and 0.53 units, respectively (p < 0.05); When Se was applied at 1.0 mg/kg and 2.5 mg/kg with the addition of Zn into the soil, a slight decrease (0.10-0.44 units) in soil pH was observed as compared with the corresponding single Se treatments. When soils were treated with 1.0 and 2.5 mg/kg Se, the pH slightly decreased (0.19-0.55 units) at 20 and 50 mg/kg Zn treatments in comparison with the corresponding single Zn treatment ( Table 1).
The activities of the selected soil enzymes, including invertase (carbon cycle), urease (nitrogen cycle), alkaline phosphatase (phosphorus cycle), and catalase (hydrogen dioxide degradation), are presented in Fig. 3. Compared with control treatments, no change in invertase activity was observed in single Se/Zn treatments (Fig. 3a). Meanwhile, urease and alkaline phosphatase activities were stimulated by low level of Se/Zn (0.5-1 mg/kg Se or 20 mg/kg Zn) in soil, but were inhibited by high level (i.e., 2.5 mg/kg Se or 50 mg/kg Zn) (Figs. 3b and c). For the catalase activity, although a signi cant increase (28.3-48.8%) in single Se treatments (p < 0.05) was noted, no signi cant difference in single Zn treatments was observed compared with those in the control treatments (Fig. 3d).
Effect of soil pH, enzyme activity, and availability of Se/Zn on Se/Zn content in pak choi The content of Se/Zn in plants is the most direct indicator of the bioavailability of soil Se/Zn.

Correlation analysis
Correlation analysis estimated with the linear regression programs showed that pH, enzyme activities, and the availability of Se/Zn on soil were related to the Se/Zn content in shoots and roots of pak choi (Table 3). Se content in shoots and roots of pak choi had the best positively correlation with the available Se in soil (R 2 > 0.82, p < 0.01).
Additionally, Se content in pak choi roots was signi cantly and negatively correlated with the catalase activities (R 2 = − 0.38, p < 0.05).
The Zn content in pak choi shoots and roots was signi cantly positively correlated with the available Zn content in soil (R 2 > 0.92, p < 0.01). However, this correlation with soil urease was negative (R 2 > − 0.49, p < 0.01). The available Zn in soil also had negative correlation with soil urease activities (R 2 = − 0.54, p < 0.01). Notably, a signi cant positive correlation was found among the Zn content and invertase activities in pak choi shoots (R 2 = 0.33, p < 0.05).
Stepwise regression (SR) modeling Table 4 summarizes the results in applying the SR modeling with Se/Zn (in mg/kg) in the shoots and roots of pak choi as dependent variables, and soil pH, soil enzyme activities (e.g., invertase (in mg glucose/g/d), urease (in mg NH 4 +− N/g/d), alkaline phosphatase (in ug/g/h), catalase (in ml KmnO 4 /g)), and availability of Se/Zn in soil (in mg/kg) as independent variables. The "B" values correspond to the beta coe cients for each independent variable in the regression model. The adjusted R 2 value re ects the explanatory power of the regression model. We can determine whether the model offers a good t for the data based on the F-test and its associated signi cance level. The signi cance (Sig.) gure is 0.05 or below shows a statistically signi cant) relationship between independent and dependent variables (Xu et al. 2020).  (Table 4). Furthermore, the indirect path coe cients indicated that catalase activities affected the available Se in soil (p < 0.01) (Table S1). This nding reveals that the effect of catalase activities in soil on the Se content of root is through the available Se in soil.

Discussions
Effects of Se-Zn interaction on plant growth and Se/Zn uptake  (Table 2). Similar results also have been reported by Dai et al. (2019). Zn is an essential element in plants that is involved in the physiological processes of growth and metabolism and is regarded as a necessary co-factor of six classes of enzymes in plants, which include oxidoreductases (Mao et al. 2015). Therefore, these observations suggest that Zn may exert a Se detoxi cation effect at high Se doses.
This study found that Zn supply signi cantly increased the Se content in the root of pak choi exposed to exogenous Se treatments.
Interestingly, the shoot Se content decreased with the increase in Zn compared with the single Se treatments (Fig. 1a). Da Silva et al. . Accordingly, the expression of sulfate transporters enhances the ability of plant roots to absorb Se (Fig. 1a), thereby suggesting a positive correlation between the available Zn and Se contents of the soil in shoot of pak choi (Table S1). However, this observation was in contrast to a previous study that concluded that the soil application of Se (L-Selenomethionine) and Zn (EDTA-Zn) signi cantly (p < 0.05) increased Se/Zn accumulation in shoot and root of pak choi without affecting its formal growth (Dai et al. 2019). This discrepancy may be due to the different exogenous Se species: (1) different soil amendments, such as the addition of EDTA chelates, which have been used to increase the availability of trace elements (Cu and Zn) to plant (Jalali and Khanlari 2008); and (2) the uptake of L-SeMet by the roots was higher than that of inorganic Se (Wang et al. 2019).
The co-amendment of selenate and Zn sulfate had no signi cant (p > 0.05) effect on the Zn content in the shoot and root of pak choi (Figs. 1b, e, and f). This nding is in agreement with many previous studies on rice (Fang et al. 2008;Mangueze et al. 2018). However, Xue et al. (2020) found that the combination of exogenous selenite and Zn sulfate markedly elevated the uptake and translocation of Zn in pak choi. The cause of this phenomenon is that selenite was feasibly adsorbed onto carbonate, Fe/Mn oxides, and organic matter in soil, which left selenite remains in the soil in turn (Qin et al. 2016). Therefore, selenite can be wrapped on the surface of soil particles through electrostatic adsorption, which inhibits the contact between soil particles and Zn, thereby reducing the xation effect of soil on Zn and improving the bioavailability of Zn in the soil (Xue et al. 2020). This aspect needs to be con rmed by further investigation.

Mechanism of Se-Zn interaction that affects Se/Zn content in plants
Se-Zn co-amend treatments signi cantly (p < 0.05) increased the available Se/Zn content in the soil (Figs. 2a and b). Moreover, the available Se and Zn contents in soil were signi cantly and positively (R 2 > 0.82, p < 0.01) correlated with the Se and Zn content in the shoots and roots of pak choi, respectively ( Table 2) (Table S1). Additionally, the pH of the soil treated with low Se rate (0.5 mg/kg) and Zn signi cantly (p < 0.05) increased by 0.46-0.53 compared with Se0.5Zn0 treatment (Table 2), and the de-protonated Se 6+ was the main speciation of Se , which increased the available Se content in soil (Fig. 2a) and uptake of Se in pak choi roots (Fig. 1a). Invertase, urease, and alkaline phosphatase are involved in the C, N, and P cycles in soil, and catalase is an oxidoreductase that protects organisms from the toxicity of hydrogen peroxide (Hu et al. 2014;Li et al. 2012). In this study, the activities of these four enzymes were signi cantly (p < 0.01) affected by Se and Zn treatments in the soil (Fig. 3). Compared with single Se/Zn treatment, Se-Zn co-amended treatments signi cantly (p < 0.05) increased the activity of invertase and alkaline phosphatase in soil but inhibited the activity of urease (Fig. 3). Thus, it improved the soluble nutrient content in soil, the ability of soil to release simple sugars, the decomposition and conversion rate of organophosphorus, and reduced the transformation capability of the organic nitrogen into inorganic nitrogen (Ciarkowska et al. 2014). Monosaccharides were the main energy source of soil microorganisms (Ciarkowska et al. 2014). The increase in the activity of sucrase indicated that the microbial activity was improved, which promoted biological oxidation and released Zn related to the organic matter in rhizosphere soil (Zhao et al. 2018). Therefore, the activity of sucrase was signi cantly positively correlated with the available Zn content in soil ( Table 3). The increase in the activity of alkaline phosphatase enhanced the mineralization of soil organophosphates (Zhang et al. 2013), and the generated inorganic phosphorus compounds may form competitive adsorption with Se ). In turn, it will affect the Se content in the shoots and roots of pak choi (Table 4). Organic nitrogen in soil can be hydrolyzed to NH 3 and CO 2 by urease with the increase in soil pH (Das and Varma 2011). This nding was consistent with the conclusion of this study that the activity of urease was signi cantly positively correlated with soil pH. Thus, the signi cant (p < 0.05) negative correlation between the content of Zn in shoots and roots of pak choi, available Zn content in soil, and the activity of urease can be attributed to the relationship between urease and soil pH ( Table 3). As for catalase, only at the same Zn application rate, application of Se signi cantly (p < 0.05) inhibited the activity of catalase compared with single Zn treatment (Fig. d). This phenomenon may be due to the exogenous water-soluble Se can complicate the enzymatic substrate or mask the catalytic active group, thereby leading to the disintegration of the enzyme-substrate complex and reducing the reaction rate of the enzyme (Gianfreda et al. 2005;Nowak et al. 2002). Catalase is an intracellular enzyme involved in the metabolism of microbial oxidoreductase (Bin et al. 2013). Therefore, the reduction of the catalase activity could affect the oxidation-reduction of microorganisms, which in turn in uences the form and valence of Se. Meanwhile, the form and valence of Se in soil are closely related to the bioavailability of Se in soil ). This nding also explains that catalase can indirectly affect the Se content of roots in pak choi by negatively affecting the available Se content in the soil in this study (Table S1).

Conclusion
In the present study, although Se stress signi cantly inhibited plant growth compared with the control treatment, the inhibition of plant growth at high Se rate was completely alleviated by Zn. More importantly, Se combined with Zn amendment signi cantly increased the available soil Se/Zn content compared with the single Se/Zn treatment, and in turn increased the uptake of Se of roots in pak choi. No signi cant difference was observed in the Zn content of shoots and roots in pak choi among single Se/Zn with their added combination. Notably, Se-Zn co-amendment can regulate the pH and enzyme activity of soil, and then affected the bioavailability of Se/Zn in soil. For example, the increase in the available Zn content could be attributed to the signi cant decrease in the pH of the soil treated with Se-Zn co-amendment compared with single Zn treatments. The activity of catalase, urease, and alkaline phosphatase could induce changes in the soil available Se/Zn affecting Se/Zn uptake by pak choi. Therefore, soil amendment of Se and Zn sulfate can be used as a potential strategy for the co-bioforti cation of Se and Zn.  Available Se (a) and Zn (b) content of soil with or without Se application. Two-way ANOVA results (F value) for two groups (a: Se; b: Zn) are presented in c (available Se) and (available Zn)), * and ** represent p < 0.05 and p < 0.01, respectively. Soil enzyme activities (invertase (a), urease (b), alkaline phosphatase (c), catalase (d)) of soil with or without Se and Zn applications. Two-way ANOVA results (F value) for two groups (a: Se; b: Zn) are presented in e (invertase), f (urease), g (alkaline phosphatase) and h (catalase), * and ** represent p < 0.05 and p < 0.01, respectively.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SupplementaryMaterial.docx