Biouptake Responses of Trace Metals to Long-Term Irrigation with Diverse Wastewater in the Wheat Rhizosphere Microenvironment

Wastewater irrigation is widely practiced and may cause serious environmental problems. However, current knowledge on the effects of long-term irrigation with wastewater from different sources on the biouptake of trace metals (TMs) in the rhizosphere zone by plants in farmlands is limited. Here, we analyzed wheat rhizosphere soil and wheat roots collected from a typical wastewater irrigation area in North China to evaluate the influence of wastewater irrigation from different sources on the bioavailability of trace metals in soils. Results showed that irrigation with tanning and domestic wastewater helped enhance the bioavailability of trace metals in rhizosphere soil by increasing the active organic carbon content, soil redox potential, and catalase activity, thus enhancing the proportion of the potentially bioavailable part of trace metal speciation. Conversely, irrigation with pharmaceutical wastewater can reduce the bioavailability of trace metals in rhizosphere soil by increasing total soil antibiotics and thus decreasing the proportions of bioavailable and potentially bioavailable parts of trace metal speciation. These findings can provide insights into the migration and transformation of trace metal speciation in soil rhizosphere microenvironments under the context of wastewater irrigation.


Introduction
Wastewater irrigation is a worldwide issue, and it is particularly common in developing countries [1,2]. This situation may become widespread in the future because of fresh water scarcity, population growth, urbanization, and increasing food demands [3]. Although wastewater irrigation can provide nutrients for soil development and boost agricultural productivity [4], it leads to the accumulation of soil contaminants, such as trace metals (TMs) and toxic chemicals [5], which may have considerably negative effects on the growth of crops and even threaten the health of humans and livestock through the food chain process [6][7][8][9]. In addition, numerous soil properties, such as soil pH, organic matter (OM) content, cation exchange capacity (CEC), soil redox potential, and enzyme activity, can be potentially affected by long-term wastewater irrigation [10][11][12][13]. Altered soil properties can further influence the transport and transformation of soil contaminants [14]. Wastewater-irrigated and reference samples were collected in June 2014 (prior to harvest). Intact wheat roots were randomly sampled at five locations in each farmland. After transportation to the laboratory, rhizosphere soils were collected by using the adhering soil method. Briefly, after gently shaking the wheat roots by hand to remove bulk soils, the soil still adhering to the root surface (2 mm thick at the most) was considered rhizosphere soil. The rhizosphere soil was carefully brushed off from the roots, air-dried at room temperature, and sieved through a 2-mm mesh to remove soil fauna, fine roots, and rock fragments. After removing the rhizosphere soil, the roots were cleaned with deionized water and sealed in plastic bags. A total of 20 wheat root and 20 rhizosphere soil samples were stored at −20 °C for subsequent analysis. The effluents from the local sewage treatment plant, pharmaceutical plant, and tannery all met the integrated wastewater discharge standard of China (Table S1). Irrigation wastewaters and groundwater were collected in March, June, September, and November 2014, and their average physicochemical properties are presented in Table S2.

Soil Physicochemical Analysis
Soil pH was measured in a soil water suspension (soil-water ratio of 1:2.5). Soil redox potential (Eh) was measured with oxidation−reduction potential (ORP) depolarization automatic analyzer Wastewater-irrigated and reference samples were collected in June 2014 (prior to harvest). Intact wheat roots were randomly sampled at five locations in each farmland. After transportation to the laboratory, rhizosphere soils were collected by using the adhering soil method. Briefly, after gently shaking the wheat roots by hand to remove bulk soils, the soil still adhering to the root surface (2 mm thick at the most) was considered rhizosphere soil. The rhizosphere soil was carefully brushed off from the roots, air-dried at room temperature, and sieved through a 2-mm mesh to remove soil fauna, fine roots, and rock fragments. After removing the rhizosphere soil, the roots were cleaned with deionized water and sealed in plastic bags. A total of 20 wheat root and 20 rhizosphere soil samples were stored at −20 • C for subsequent analysis. The effluents from the local sewage treatment plant, pharmaceutical plant, and tannery all met the integrated wastewater discharge standard of China (Table S1). Irrigation wastewaters and groundwater were collected in March, June, September, and November 2014, and their average physicochemical properties are presented in Table S2.

Soil Physicochemical Analysis
Soil pH was measured in a soil water suspension (soil-water ratio of 1:2.5). Soil redox potential (Eh) was measured with oxidation−reduction potential (ORP) depolarization automatic analyzer (FJA-6, Nanjing Chuan-Di Instrument & Equipment, Nanjing, China). Soil clay content (<2 µm) was determined using a laser particle size analyzer (Malvern Mastersizer, 2000, Malvern, UK), and the cation exchange capacity (CEC) was determined using standard methods [28]. The concentrations of N and S were determined with an elemental analyzer (Vario EL cube). The total organic carbon (TOC) and soil dissolved organic carbon (DOC) extracted with 0.01 M CaCl 2 (soil: solution ratio of 1:10; 2 h) were determined with a Shimadzu 5000 TOC analyzer. Permanganate oxidizable carbon (KMnO 4 -C) was determined using the method of Vieira et al. [29]. Soil antibiotics were extracted according to the modified method of Xie et al. [30], and the target antibiotic compounds included chlortetracycline, cefazolin, cefotaxime, cefoxitin, cefaclor, cefuroxime, furazolidone, sulfadiazine, sulfamerazine, sulfamethazine, sarafloxacin, oxytetracycline, ofloxacin, trimethoprim, tetracycline, and carbadox. For the concentrations of P, K, Ca, Fe, Mn, and Mg, soil was extracted by aqua regia at 160 • C, and the concentrations were determined through inductively coupled plasma-optical emission spectrometry (ICP-OES, Thermo ICAP-6000, Waltham, MA, USA).

Soil Biomass Carbon and Enzyme Activity Determination
Aliquots of the fresh rhizosphere soil samples were used to determine microbial biomass carbon (MBC) via a modified fumigation extraction procedure [31]. The activity of soil catalase was measured using the method of Zhou et al. [32]. Briefly, 3% H 2 O 2 as the oxidizer was added to fresh soil and allowed to stand for 30 min at 3 • C. Afterward, the reaction was stopped with the addition of 1 M H 2 SO 4 . After filtration, 0.5 M H 2 SO 4 was added to the filtrate, and 20 mM KMnO 4 was used to measure the O 2 absorbed. The activities of lignin peroxidase (LiP), laccase (Lac), and manganese peroxidase (MnP) were assayed using the methods of Fujii et al. [33]. The MBC and enzyme activities of rhizosphere soils irrigated with groundwater and different wastewater types are presented in Table 1.

TM Analysis
The chemical speciation of TMs (Cu, Cr, Cd, As, Pb, and Ni) in rhizosphere soils was analyzed with a modified five-step extraction method [34]. In accordance with Tessier's method, metals were partitioned into the following five operationally defined fractions: (1) Exchangeable fraction. Samples were extracted at a solid-to-solution ratio of 1:8 with 0.5 M MgCl 2 (pH 7.0) and stirred continuously for 5 h.
(2) Fraction bound to carbonates. The residual from (1) was extracted with 1 M NaOAc (pH 5.0) at a solid-to-solution ratio of 1:8 and stirred continuously for 5 h.
To further explore the transformation of TM chemical speciation in rhizosphere soil, water-soluble fraction and humic acid fractions were added to the steps shown above. Therefore, TMs in rhizosphere soils were extracted as the following seven fractions: fraction 1 (water soluble), fraction 2 (ion exchangeable), fraction 3 (bound to carbonates), fraction 4 (bound to humic acids), fraction 5 (bound to Fe-Mn oxides), fraction 6 (bound to OM), and fraction 7 (residual).
After microwave digestion, the total concentrations of TMs contained in wheat roots and the concentration of each chemical speciation of TMs in rhizosphere soils were analyzed through ICP-OES (Thermo ICAP-6000, Waltham, MA, USA).

Statistical Analysis
The bio-accumulation factor (BAF) is the ratio of TM concentration in wheat root to that in rhizosphere soil, and was calculated in this study as follows: where C root and C soil represent the TM concentration in wheat root and rhizosphere soil, respectively. Data were expressed as arithmetic mean ± standard deviation calculated from replicates. The relationships between BAF and environmental variables, between TM chemical speciation and environmental variables (Figures S1-S3), and between environmental variables (Figures S4-S6) were assessed by correlation analysis. Analysis of variance (ANOVA) was performed to compare the differences in BAF values, physicochemical characteristics, and total concentration of TMs among farmlands irrigated with wastewaters from different sources. All data were subjected to homogeneity testing before ANOVA to ensure they were in a normal distribution. Stepwise multiple regression was used to evaluate the chemical speciation of TMs responsible for the changes in BAF values in different farmlands. Correlation analysis, stepwise multiple regression, and ANOVA were conducted with SPSS 20.0 (IBM Corporation Software Group, Somers, New York, NY, USA). The results were considered significant at the p < 0.05 level.
Structural equation models (SEMs) were constructed to determine the major pathways that affect the BAF of TMs in rhizosphere soils irrigated with different wastewater types. SEM is a modeling tool that integrates ANOVA, regression analysis, path analysis, and factor analysis to deal with multivariate complex relationships. SEM can be used to analyze the relationship between latent variables and simulate the internal logic of multiple factors [35]. The model χ 2 test (p > 0.05), normed fit index (NFI > 0.90), goodness-of-fit index (GFI > 0.90), and low root mean square errors of approximation (RMSEA < 0.05) were used to indicate the overall fitness of SEMs. SEM analyses were performed using SPSS 20.0 and AMOS 21.0 (IBM Corporation Software Group, Somers, New York, NY, USA).

Changes in Rhizosphere Soil Properties
The descriptive statistics of several physicochemical characteristics of the analyzed rhizosphere soils are summarized in Table 1. The TOC, DOC, and KMnO 4 -C contents in rhizosphere soils irrigated with wastewater increased in comparison with the control, possibly due to the high concentrations of OMs in wastewater (Table S2). The total soil antibiotic content in PWIF was significantly higher than that in other farmlands. Compared with the control, soil pH was higher in PWIF and had an average of 8.13. In contrast, lower soil pH was detected in TWIF and DWIF and had average values of 7.52 and 6.85, respectively. Soil Eh was lower with pharmaceutical wastewater irrigation (63.96 mV) but higher with domestic and tanning wastewater irrigation (85.34 and 95.98 mV, respectively) compared with the control. The mean concentrations of N, P, K, Ca, Mg, S, and Mn showed no significant differences in the farmlands irrigated with groundwater and wastewater. Compared with the control, the soil clay contents was higher in PWIF and had an average value of 39.5%, but it was lower in TWIF and DWIF and had average values of 34.47% and 30.64%, respectively.
Soil catalase activity, which is often related to the quantity and activity of aerobic microorganisms [36], increased with wastewater irrigation (Table 1). However, the soil biomass carbon decreased after pharmaceutical and tanning wastewater irrigation, which could be ascribed to the high levels of total antibiotics and toxic materials in pharmaceutical and tanning wastewater (Table S2). Similarly, the activity of LiP and Lac decreased with long-term pharmaceutical wastewater irrigation. Except for MnP, all of the analyzed enzyme activities increased with domestic wastewater irrigation. Table 2 summarizes the total concentrations of TMs in rhizosphere soils irrigated with groundwater and wastewater from different sources. China's soil environmental quality standards are recommended by the Ministry of Environmental Protection of the People's Republic of China (MEPPRC) [37]. These standards consist of three grades of threshold values, in which the first grade is the background values of the soil, the second grade is for pollution assessment in agricultural soils, and the third grade is for pollution assessment in forests or highly contaminated agricultural soils. The threshold of the second grade was used to evaluate the TMs in rhizosphere soils in this study. Compared with the control, the concentrations of Cu, Cr, Cd, Pb, and Ni increased significantly after long-term tanning wastewater irrigation. Similarly, but to a reduced extent, the concentration of Cu increased remarkably after pharmaceutical wastewater irrigation, and the concentrations of Cu and Cr increased significantly after domestic wastewater irrigation compared with the control. Furthermore, in TWIF, the mean concentrations of Cu, Cr, Cd, and Pb exceeded the reference background values of Hebei province, but remained below the second grade of environmental quality standards. In PWIF and DWIF, the mean concentrations of Cr and Cd exceeded the reference background values, but remained below the second grade of standards. The total concentration of Ni exceeded the second grade of the environmental quality standard in both wastewater and groundwater irrigated croplands.

Distribution of TM Fractions in Rhizosphere Soils
Seven types of chemical speciation of TMs in rhizosphere soils were successfully extracted by sequential extraction and summarized in three parts according to their bioavailability, namely, a bioavailable (B) part, a potentially bioavailable (PB) part, and a non-bioavailable (NB) part. The bioavailable part represents the fractions likely to be directly absorbed by plants, and consists of fractions 1-3. The potentially bioavailable part can be absorbed by plants in strong acid medium and consists of fractions 4-6 [18]. The non-bioavailable (NB) part is residual fraction (fraction 7) and cannot be absorbed by plants [38].
The proportions of TM speciation in rhizosphere soils irrigated with groundwater and different wastewater types are presented in Figure 2. The PB and B parts of the analyzed TMs were significantly higher in TWIF and DWIF and remarkably lower in PWIF than those in the control. On the contrary, the NB part of all TMs was notably lower in TWIF and DWIF, and significantly higher in PWIF than that in the control. These results show that pharmaceutical wastewater irrigation could increase the proportions of TMs in the NB part and decrease the proportions of TMs in the PB and B parts. By contrast, domestic and tanning wastewater irrigation could reduce the proportions of TMs in the NB part and increase the proportions of metals in the PB and B parts. Therefore, the bioavailability of TMs could be increased in rhizosphere soils irrigated with domestic and tanning wastewater and decreased in rhizosphere soils irrigated with pharmaceutical wastewater.

Differences in BAF among Farmlands Irrigated with Wastewater from Different Sources
BAF can be used to illustrate the strength of the enrichment of TMs by plant roots and was calculated according to the Equation (1). The bioaccumulation of TMs was observed in all tested wheat roots, as shown in Figure 3. Compared with the BAF values of the control, those of the analyzed TMs increased after domestic and tanning wastewater irrigation. Inversely, the BAF values of TMs (except for Cr) decreased with pharmaceutical wastewater irrigation. These results further confirm that irrigation with domestic and tanning wastewater could increase the bioavailability of TMs, whereas irrigation with pharmaceutical wastewater could decrease the bioavailability of TMs in the soil rhizosphere microenvironment.

Effects of Physicochemical Properties and TM Fractions on BAF in Farmlands Irrigated with Wastewater from Different Sources
After integrating the data of each wastewater-irrigated soil with the data of groundwater-irrigated soil (10 data pairs in total for each comparison), correlation analysis, stepwise regression analysis, and structural equation modeling were performed to analyze the relationships among BAF values, soil properties, and soil TM chemical speciation in different wastewater-irrigated rhizosphere soils.
The transformation and biouptake processes of TMs in soils are generally regulated by many physicochemical and biological properties. Thus, some soil physicochemical and biological indicators were selected and analyzed to investigate their influences on the BAF of TMs. The correlations between BAF values of TMs and soil properties showed evident differences among the farmlands irrigated with diverse wastewater types (Figure 4). For DWIF, the BAF values of TMs exhibited significantly positive correlations with DOC, KMnO 4 -C, P, catalase activity, and lac activity, and a remarkably negative correlation with clay content. For TWIF, the BAF values of TMs exhibited positive correlations with KMnO 4 -C content, Eh, and CEC, and a negative correlation with soil biomass carbon. For PWIF, the BAF values of TMs showed significantly negative correlations with DOC, KMnO 4 -C content, total antibiotics, clay content, N, P, and catalase, and a positive correlation with S. Seven types of chemical speciation of TMs in rhizosphere soils were successfully extracted by sequential extraction and summarized in three parts according to their bioavailability, namely, a bioavailable (B) part, a potentially bioavailable (PB) part, and a non-bioavailable (NB) part. The bioavailable part represents the fractions likely to be directly absorbed by plants, and consists of fractions 1-3. The potentially bioavailable part can be absorbed by plants in strong acid medium and consists of fractions 4-6 [18]. The non-bioavailable (NB) part is residual fraction (fraction 7) and cannot be absorbed by plants [38]. After finding out some physicochemical and biological indicators that are strongly correlated with BAF, SEM was conducted to analyze the hypothetical pathways that may explain how these physicochemical and biological indicators affect BAF under different wastewater irrigation conditions. The evaluation parameters show that the SEMs constructed in this study were well-fitted (TWIF: wastewater and decreased in rhizosphere soils irrigated with pharmaceutical wastewater.

Differences in BAF among Farmlands Irrigated with Wastewater from Different Sources
BAF can be used to illustrate the strength of the enrichment of TMs by plant roots and was calculated according to the Equation (1). The bioaccumulation of TMs was observed in all tested wheat roots, as shown in Figure 3. Compared with the BAF values of the control, those of the analyzed TMs increased after domestic and tanning wastewater irrigation. Inversely, the BAF values of TMs (except for Cr) decreased with pharmaceutical wastewater irrigation. These results further confirm that irrigation with domestic and tanning wastewater could increase the bioavailability of TMs, whereas irrigation with pharmaceutical wastewater could decrease the bioavailability of TMs in the soil rhizosphere microenvironment.

Effects of Physicochemical Properties and TM Fractions on BAF in Farmlands Irrigated with Wastewater from Different Sources
After integrating the data of each wastewater-irrigated soil with the data of groundwater-irrigated soil (10 data pairs in total for each comparison), correlation analysis, stepwise regression analysis, and structural equation modeling were performed to analyze the relationships among BAF values, soil properties, and soil TM chemical speciation in different wastewater-irrigated rhizosphere soils.   The transformation and biouptake processes of TMs in soils are generally regulated by many physicochemical and biological properties. Thus, some soil physicochemical and biological indicators were selected and analyzed to investigate their influences on the BAF of TMs. The correlations between BAF values of TMs and soil properties showed evident differences among the farmlands irrigated with diverse wastewater types (Figure 4). For DWIF, the BAF values of TMs exhibited  The distribution of TM chemical speciation in soils greatly affects the biouptake processes of TMs. Since wastewater irrigations remarkably changed the distribution of TM chemical speciation in soils (Figure 2), stepwise regression analysis was conducted to illustrate the influences of these changes on the BAF values of TMs. The results indicate that the fractions bound to OM, Fe-Mn oxides, and humic acids were the dominant chemical speciation responsible for the variations in BAF values in the farmlands irrigated with wastewater from different sources ( Figure 6). For the farmland irrigated with tanning wastewater, OM-bound fractions explained 65%, 69%, and 95% of the variations in BAF values of Cu, Cr, and Pb, respectively, while humic acid-bound fractions explained 85%, 91% and 89% of the variations in the BAF values of Cd, As, and Ni, respectively. For the farmland irrigated with domestic wastewater, the role of OM-bound and humic acid-bound fractions was replaced by Fe-Mn oxide fractions, which explained 94%, 87%, and 98% of the variations in the BAF values of Cu, Cd, and Ni, respectively. For the farmland irrigated with pharmaceutical wastewater, OM-bound fractions explained 80%, 90%, and 80% of the variations in the BAF values of Cu, As, and Ni, respectively, while humic acid-bound fractions explained 83% of the variations in the BAF values of Pb. Meanwhile, water dissolved fractions explained 78% and 40% of the variations in the BAF values of Cr and Cd, respectively. After finding out some physicochemical and biological indicators that are strongly correlated with BAF, SEM was conducted to analyze the hypothetical pathways that may explain how these physicochemical and biological indicators affect BAF under different wastewater irrigation conditions. The evaluation parameters show that the SEMs constructed in this study were well-fitted (TWIF: χ 2 = 1.42, df = 5, p = 0.92, NFI = 0.96, GFI = 0.94, RMSEA = 0.00, mean; DWIF: χ 2 = 3.02, df = 4, p = 0.61, NFI = 0.94, GFI = 0.90, RMSEA = 0.03, mean; PWIF: χ 2 = 0.013, df = 1, p = 0.91, NFI = 0.99, GFI = 0.99, RMSEA = 0.00, mean; Figure 5). On the average, our model explained 84% (Figure 5a), 85% (Figure 5b), and 86% (Figure 5c) of the variance in BAF values of TMs in rhizosphere soils irrigated with tanning, domestic, and pharmaceutical wastewaters, respectively. Among the explanatory variables, KMnO4-C content and catalase activity directly affected the BAF values of TMs in all wastewater-irrigated farmlands. For TWIF (Figure 5a), CEC, KMnO4-C content, and catalase activity had direct positive effects on the BAF values of all TMs, whereas the catalase activity was affected by soil Eh. For DWIF (Figure 5b), KMnO4-C content and catalase activity had direct positive effects on the BAF values of all TMs, soil Eh had direct positive effects on KMnO4-C content and catalase activity, and clay contents had a direct negative effect on KMnO4-C content. However, for PWIF (Figure 5c), KMnO4-C content and catalase activity had direct negative effects on the BAF values of all TMs, and total soil antibiotic content had a remarkably positive effect on KMnO4-C content and catalase activity. The distribution of TM chemical speciation in soils greatly affects the biouptake processes of TMs. Since wastewater irrigations remarkably changed the distribution of TM chemical speciation in soils (Figure 2), stepwise regression analysis was conducted to illustrate the influences of these changes on the BAF values of TMs. The results indicate that the fractions bound to OM, Fe-Mn oxides, and humic acids were the dominant chemical speciation responsible for the variations in BAF values in the farmlands irrigated with wastewater from different sources ( Figure 6). For the farmland irrigated with tanning wastewater, OM-bound fractions explained 65%, 69%, and 95% of the variations in BAF values of Cu, Cr, and Pb, respectively, while humic acid-bound fractions explained 85%, 91% and 89% of the variations in the BAF values of Cd, As, and Ni, respectively. For the farmland irrigated with domestic wastewater, the role of OM-bound and humic acid-bound fractions was replaced by Fe-Mn oxide fractions, which explained 94%, 87%, and 98% of the variations in the BAF values of Cu, Cd, and Ni, respectively. For the farmland irrigated with pharmaceutical wastewater, OM-bound fractions explained 80%, 90%, and 80% of the variations in the BAF values of Cu, As, and Ni, respectively, while humic acid-bound fractions explained 83% of

Discussion
The bioavailability of TMs in rhizosphere soil exhibited significant differences in the farmlands irrigated with wastewater from different sources. Tanning and domestic wastewater irrigation led to an increase in the bioavailability of TMs, whereas pharmaceutical wastewater irrigation led to a decrease in the bioavailability of TMs (Figure 3). Our results also showed that changes in the PB part (humic acid-and OM-bound fractions) and B part (water dissolved fractions) of TM speciation were the determinants that resulted in the decrease in BAF values in PWIF, and changes in the PB part were the determinant that led to the increase in BAF values in TWIF and DWIF ( Figure 6).
Furthermore, long-term irrigation with wastewater from different sources could lead to changes in several soil properties (Table 1). Decreased soil pH was observed in the rhizosphere soils with domestic wastewater irrigation (Table 1). We found that soil pH had an observably negative correlation with soil enzymatic activity in DWIF ( Figure S5). However, a previous study reported that soil pH was positively correlated with soil enzymatic activities after wastewater irrigation [39]. The contradiction of both studies implies that there might be no direct causal relationships between the changes of soil enzymatic activities and soil pH in DWIF. Increased soil enzymatic activities and decreased soil pH in DWIF might be attributed to the additional input of active organic matter and exchangeable cations created by the irrigation water [40]. In PWIF, soil pH was elevated with an increase in soil OM content (Table 1, Figure S6). Mancino and Pepper [41] attributed this increase in soil pH after wastewater irrigation to an increase in denitrification rate. Thus, the denitrification rate might be strengthened in rhizosphere soil after long-term irrigation with pharmaceutical wastewater. Notably, antibiotic content was the most obvious difference between PWIF and other farmlands irrigated with wastewater (Table 1). Total soil antibiotic content was a key factor that affected the BAF values of TMs in PWIF ( Figure 5). Accumulation of antibiotics could inhibit the target microorganisms, and other uninhibited microorganisms could acquire abundant resources to rapidly reproduce, thereby altering the compositions and functions of microbial communities in the soil [42]. Accordingly, the population of certain microbial species might be enhanced in PWIF due to the environmental selection of strains resistant to the antibiotics. Such changes may eventually lead to variations in some soil properties in PWIF, such as soil pH (Table 1) [41,43].
KMnO 4 -C content was one of the direct factors responsible for the variations in BAF values in the investigated rhizosphere soils ( Figure 5). The close association between BAF values and KMnO 4 -C content was possibly due to OM, especially active-soil OM (DOC, KMnO 4 -C), which could regulate BAF values by affecting dissolved TM concentrations in soil [44]. KMnO 4 -C content exerted a positive effect on BAF values in TWIF and DWIF (Figure 5a,b). One possible explanation is that the accumulation of KMnO 4 -C content could lower soil pH (Table 1) [45], thus indirectly leading to a higher solubility of TMs and root uptake of TMs. However, in PWIF, KMnO 4 -C content exerted a negative effect on the BAF values of TMs (Figures 4 and 5), suggesting that the major effect of KMnO 4 -C content on the bio-availability of TMs depends on the type of irrigation water. Generally, soil organic matter has a high affinity for metal cations due to the presence of ligands or functional groups such as carboxylic acids (-COOH), hydroxylic acids (-OH), and phenolic acids (aromatic ring-OH) [46]. Therefore, the negative effect of KMnO 4 -C content on BAF might be attributed to the fact that the accumulation of KMnO 4 -C content can directly lower the mobility and solubility of TMs in soils via adsorption and complexation processes. Moreover, the positive and negative effects of KMnO 4 -C content on the bio-availability of TMs might simultaneous exist during wastewater irrigation, and future studies need to explore the mechanisms regulating the major effects of KMnO 4 -C content on the bio-availability of TMs during wastewater irrigation.
Our results also revealed that catalase activity was the dominant factor responsible for the variations in BAF values in the rhizosphere soil (Figure 5a,b). Zeng et al. [47] indicated that the activities of soil enzymes can be stimulated by low concentrations of TMs. In this study, the relatively low concentrations of TMs in the rhizosphere soils irrigated with wastewater were in accordance with the China's soil environmental quality standards (Table 2). In addition, the consistent distribution (Tables 1 and 2) of catalase activity and TM concentrations implied that catalase activity might be stimulated by TMs in the rhizosphere soils irrigated with wastewater. Specifically, the results confirmed that soil Eh has a strong positive effect on the solubility of TMs in rhizosphere soil [48]. Therefore, increasing soil Eh could promote the solubility of TMs and further stimulate catalase activity [45], ultimately leading to a positive effect of soil Eh on catalase activity in TWIF and DWIF ( Figure 5). Catalase activity relates to the metabolism activity of aerobic organisms and can be used as an indicator of soil fertility [36,49]. The increase in catalase activity reveals that the activity of aerobic organisms might be promoted, and the flux of active organic carbon might be increased in rhizosphere soil. Newly generated active OM constantly reacted with the TMs through ion exchange, adsorption, complexation, chelation, flocculation, and redox [50,51], thereby accelerating the transformation of TMs from NB to PB and PB to B (Figures S1 and S3), and eventually increasing the concentrations of PB and B in TWIF and DWIF. In addition, high concentrations of TMs contained in tanning wastewater may be another important factor for the increase in PB and B in TWIF ( Table 2). The B part of TMs in TWIF and DWIF did not contribute to the variations in BAF values (Figure 6), possibly because its contribution to the variations in BAF values was mainly induced by increasing its flux rather than concentrations.

Conclusions
This study investigated the impact of long-term irrigation with wastewater from different sources on the uptake of TMs by plant roots in the soil rhizosphere microenvironment. Irrigation with tanning and domestic wastewater can led to an enhancement in the bioavailability of TMs in rhizosphere soils mainly by increasing active-soil organic carbon content, catalase activity and soil Eh to increase the concentrations of TMs in the PB part. Active organic carbon content and catalase activity had direct positive effects on BAF values, and soil Eh indirectly affected BAF values by influencing active organic carbon content and catalase activity in farmlands irrigated with tanning and domestic wastewater. Conversely, pharmaceutical wastewater irrigation reduced the bioavailability of TMs in rhizosphere soil by increasing soil antibiotic content to reduce the concentrations of TMs in the PB and B parts. Overall, our results indicate that the effects of wastewater irrigation on the bioavailability of TMs in rhizosphere soils were largely dependent on the source of wastewater. This work provides a new perspective on TM pollution in the soil rhizosphere microenvironment caused by diverse sources of wastewater irrigation. Further studies are required to clarify our findings and their universality in different soil types. In addition, given that tanning and domestic wastewater irrigation may enhance the risk of existing TM contamination, stringent regulations are needed for the irrigations using these wastewaters in the future, especially in developing countries.
Supplementary Materials: The following are available online at http://www.mdpi.com/1660-4601/16/17/3218/s1: Table S1. Summary of indices reflecting physicochemical characteristics and soil enzyme activities of rhizosphere soils; Table S2. Physicochemical properties of groundwater and wastewater from different sources; Figure S1. Correlations between trace metal speciation and soil physicochemical characters in rhizosphere soils irrigated with tanning wastewater; Figure S2. Correlations between trace metal speciation and soil physicochemical characters in rhizosphere soils irrigated with pharmaceutical wastewater; Figure S3. Correlations between trace metal speciation and soil physicochemical characters in rhizosphere soils irrigated with domestic wastewater; Figure S4. Correlations among soil physicochemical characters in rhizosphere soils irrigated with tanning wastewater; Figure S5. Correlations among soil physicochemical characters in rhizosphere soils irrigated with domestic wastewater; Figure S6. Correlations among soil physicochemical characters in rhizosphere soils irrigated with pharmaceutical waste water.